diff --git a/src/asset_cache/asset_cache_core.c b/src/asset_cache/asset_cache_core.c
index 687e4faa..1243312e 100644
--- a/src/asset_cache/asset_cache_core.c
+++ b/src/asset_cache/asset_cache_core.c
@@ -29,7 +29,7 @@ void AC_RefreshDebugTable(void)
 {
 #if RtcIsEnabled
     AC_SharedState *g = &AC_shared_state;
-    P_Lock lock = P_LockE(&g->dbg_table_mutex);
+    Lock lock = LockE(&g->dbg_table_mutex);
     ZeroArray(g->dbg_table);
     g->dbg_table_count = 0;
     for (u64 i = 0; i < countof(g->lookup); ++i)
@@ -40,16 +40,16 @@ void AC_RefreshDebugTable(void)
             g->dbg_table[g->dbg_table_count++] = asset;
         }
     }
-    P_Unlock(&lock);
+    Unlock(&lock);
 #endif
 }
 
 /* Returns first matching slot or first empty slot if not found.
  * Check returned slot->hash != 0 for presence. */
-AC_Asset *AC_GetAssetCacheSlotLocked(P_Lock *lock, String key, u64 hash)
+AC_Asset *AC_GetAssetCacheSlotLocked(Lock *lock, String key, u64 hash)
 {
     AC_SharedState *g = &AC_shared_state;
-    P_AssertLockedES(lock, &g->lookup_mutex);
+    AssertLockedES(lock, &g->lookup_mutex);
     LAX lock;
 
     u64 index = hash % countof(g->lookup);
@@ -96,9 +96,9 @@ AC_Asset *AC_TouchCache(String key, u64 hash, b32 *is_first_touch)
 
     /* Lookup */
     {
-        P_Lock lock = P_LockS(&g->lookup_mutex);
+        Lock lock = LockS(&g->lookup_mutex);
         asset = AC_GetAssetCacheSlotLocked(&lock, key, hash);
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
 
     /* Insert if not found */
@@ -111,7 +111,7 @@ AC_Asset *AC_TouchCache(String key, u64 hash, b32 *is_first_touch)
     }
     else
     {
-        P_Lock lock = P_LockE(&g->lookup_mutex);
+        Lock lock = LockE(&g->lookup_mutex);
 
         /* Re-check asset presence in case it was inserted since lock */
         asset = AC_GetAssetCacheSlotLocked(&lock, key, hash);
@@ -136,7 +136,7 @@ AC_Asset *AC_TouchCache(String key, u64 hash, b32 *is_first_touch)
                 .hash = hash,
                 .key = key_stored
             };
-            P_CounterAdd(&asset->counter, 1);
+            AddCounter(&asset->counter, 1);
             if (is_first_touch)
             {
                 *is_first_touch = 1;
@@ -146,7 +146,7 @@ AC_Asset *AC_TouchCache(String key, u64 hash, b32 *is_first_touch)
             AC_RefreshDebugTable();
         }
 
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
 
     return asset;
@@ -166,12 +166,12 @@ void AC_MarkReady(AC_Asset *asset, void *store_data)
 {
     asset->store_data = store_data;
     asset->status = ASSET_STATUS_READY;
-    P_CounterAdd(&asset->counter, -1);
+    AddCounter(&asset->counter, -1);
 }
 
 void AC_WaitOnAssetReady(AC_Asset *asset)
 {
-    P_WaitOnCounter(&asset->counter);
+    WaitOnCounter(&asset->counter);
 }
 
 ////////////////////////////////
@@ -193,7 +193,7 @@ void *AC_DataFromStore(AC_Asset *asset)
 AC_Store AC_OpenStore(void)
 {
     AC_SharedState *g = &AC_shared_state;
-    P_Lock lock = P_LockE(&g->store_mutex);
+    Lock lock = LockE(&g->store_mutex);
     AC_Store store = {
         .lock = lock,
         .arena = g->store_arena
@@ -203,5 +203,5 @@ AC_Store AC_OpenStore(void)
 
 void AC_CloseStore(AC_Store *store)
 {
-    P_Unlock(&store->lock);
+    Unlock(&store->lock);
 }
diff --git a/src/asset_cache/asset_cache_core.h b/src/asset_cache/asset_cache_core.h
index eb8b6a6b..f2c31109 100644
--- a/src/asset_cache/asset_cache_core.h
+++ b/src/asset_cache/asset_cache_core.h
@@ -34,7 +34,7 @@ Struct(AC_Store)
     Arena *arena;
 
     /* internal */
-    P_Lock lock;
+    Lock lock;
 };
 
 ////////////////////////////////
@@ -45,18 +45,18 @@ Struct(AC_Store)
 
 Struct(AC_SharedState)
 {
-    P_Mutex lookup_mutex;
+    Mutex lookup_mutex;
     AC_Asset lookup[AC_AssetLookupTableCapacity];
     u64 num_assets;
 
-    P_Mutex store_mutex;
+    Mutex store_mutex;
     Arena *store_arena;
 
 #if RtcIsEnabled
     /* Array of len `num_assets` pointing into populated entries of `lookup`. */
     AC_Asset *dbg_table[AC_AssetLookupTableCapacity];
     u64 dbg_table_count;
-    P_Mutex dbg_table_mutex;
+    Mutex dbg_table_mutex;
 #endif
 };
 
@@ -77,7 +77,7 @@ u64 AC_HashFromKey(String key);
 //~ Cache operations
 
 void AC_RefreshDebugTable(void);
-AC_Asset *AC_GetAssetCacheSlotLocked(P_Lock *lock, String key, u64 hash);
+AC_Asset *AC_GetAssetCacheSlotLocked(Lock *lock, String key, u64 hash);
 AC_Asset *AC_TouchCache(String key, u64 hash, b32 *is_first_touch);
 
 ////////////////////////////////
diff --git a/src/base/base.c b/src/base/base.c
index 70c7fd18..93fe0db1 100644
--- a/src/base/base.c
+++ b/src/base/base.c
@@ -10,9 +10,10 @@
 #include "base_uid.c"
 #include "base_uni.c"
 #include "base_incbin.c"
+#include "base_snc.c"
 
 #if PlatformIsWindows
 # include "win32/base_win32.c"
 #else
 # error Base layer platform backend not implemented
-#enidf
+#endif
diff --git a/src/base/base.h b/src/base/base.h
index 2fa051bc..e88a6980 100644
--- a/src/base/base.h
+++ b/src/base/base.h
@@ -8,8 +8,9 @@
 #include "../prof/prof.h"
 # include "base_intrinsics.h"
 # include "base_memory.h"
-# include "base_job.h"
 # include "base_arena.h"
+# include "base_snc.h"
+# include "base_job.h"
 # include "base_uid.h"
 # include "base_string.h"
 # include "base_uni.h"
diff --git a/src/base/base_core.h b/src/base/base_core.h
index 4cc5a5aa..71f77ab9 100644
--- a/src/base/base_core.h
+++ b/src/base/base_core.h
@@ -548,6 +548,23 @@ ForceInline void UnlockTicketMutex(TicketMutex *tm)
 }
 #endif
 
+////////////////////////////////
+//~ Fibers
+
+#define MaxFibers 4096
+
+#if !LanguageIsGpu
+ForceInline i16 FiberId(void)
+{
+    /* FIXME */
+    //return (i16)(i64)GetFiberData();
+    Assert(0);
+    return 0;
+}
+StaticAssert(MaxFibers < I16Max);  /* Fiber id type should fit max fibers */
+#endif
+
+
 ////////////////////////////////
 //~ Config
 
diff --git a/src/base/base_job.h b/src/base/base_job.h
index 42091594..fb1b1f83 100644
--- a/src/base/base_job.h
+++ b/src/base/base_job.h
@@ -33,30 +33,59 @@ typedef i32 PriorityKind; enum
 };
 
 ////////////////////////////////
-//~ @hookdecl Fiber helpers
+//~ @hookdecl Futex
 
-#define MaxFibers 4096
-
-i16 FiberId(void);
-StaticAssert(MaxFibers < I16Max);  /* Fiber id type should fit max fibers */
+/* Futex-like wait & wake */
+void FutexWait(volatile void *addr, void *cmp, u32 size, i64 timeout_ns);
+void FutexWake(void *addr, i32 count);
 
 ////////////////////////////////
 //~ @hookdecl Job helpers
 
 #define EmptySig { i32 _; }
 
+typedef void GenericJobFunc(void *, i32);
+
+Struct(GenericJobDesc)
+{
+    Arena *arena;
+    void *sig;
+    GenericJobFunc *func;
+    i32 count;
+    PoolKind pool;
+    PriorityKind priority;
+    Counter *counter;
+};
+
+Struct(JobDescParams)
+{
+    i32 count;
+    PoolKind pool;
+    PriorityKind priority;
+    Counter *counter;
+};
+
 #define JobDecl(job, sigdef) \
     typedef struct job##_Sig sigdef job##_Sig; \
-    Struct(job##_Desc) { Counter *counter; Arena *job_arena; job##_Sig *sig; }; \
-    void job(job##_Sig *)
+    Struct(job##_Desc) { Arena *arena; job##_Sig *sig; GenericJobFunc *func; i32 count; PoolKind pool; PriorityKind priority; Counter *counter;}; \
+    void job(job##_Sig *, i32)
 
-#define JobDef(job, sig_arg) void job(job##_Sig *sig_arg)
+#define PushJobDesc(job, ...) (job##_Desc *)PushJobDesc_(sizeof(job##_Sig), alignof(job##_Sig), (GenericJobFunc *)(i64)job, (JobDescParams) { .count = 1, .pool = PoolKind_Inherit, .priority = PriorityKind_Inherit, .counter = 0, __VA_ARGS__ })
+GenericJobDesc *PushJobDesc_(u64 sig_size, u64 sig_align, GenericJobFunc *func, JobDescParams params);
 
-#define RunJob(job, desc) job(&desc->sig)
+#define JobDef(job, sig_arg, id_arg) void job(job##_Sig *sig_arg, i32 id_arg)
+
+#define RunJob(_count, job, _pool, _priority, _counter, ...) do { \
+    job##_Desc *__job_desc = (job##_Desc *)PushJobDesc_(sizeof(job##_Sig), alignof(job##_Sig), (GenericJobFunc *)(i64)job, (JobDescParams) { .count = _count, .pool = _pool, .priority = _priority, .counter = _counter,}); \
+    *__job_desc->sig = (job##_Sig) { __VA_ARGS__ }; \
+} while (0)
+void RunJobEx(GenericJobDesc *desc);
 
 ////////////////////////////////
-//~ @hookdecl Wait
+//~ @hookdecl Helpers
 
-/* Futex-like wait & wake */
-void P_Wait(volatile void *addr, void *cmp, u32 size, i64 timeout_ns);
-void P_Wake(void *addr, i32 count);
+i64 TimeNs(void);
+
+u32 GetLogicalProcessorCount(void);
+u32 ThreadId(void);
+i64 GetCurrentSchedulerPeriodNs(void);
diff --git a/src/base/base_rand.c b/src/base/base_rand.c
index f554c950..9cb1e809 100644
--- a/src/base/base_rand.c
+++ b/src/base/base_rand.c
@@ -2,8 +2,10 @@
 //~ True randomness
 
 #if PlatformIsWindows
+#if 0
 # define BCRYPT_RNG_ALG_HANDLE   ((void *)0x00000081)
 u32 BCryptGenRandom(void *algorithm, u8 *buffer, u32 buffer_size, u32 flags);
+#endif
 void TrueRand(String buffer)
 {
     BCryptGenRandom(BCRYPT_RNG_ALG_HANDLE, (u8 *)buffer.text, buffer.len, 0);
diff --git a/src/platform/platform_snc.c b/src/base/base_snc.c
similarity index 76%
rename from src/platform/platform_snc.c
rename to src/base/base_snc.c
index 1013640b..6964ac9a 100644
--- a/src/platform/platform_snc.c
+++ b/src/base/base_snc.c
@@ -2,7 +2,7 @@
 //~ Mutex
 
 //- Exclusive mutex lock
-P_Lock P_LockSpinE(P_Mutex *m, i32 spin)
+Lock LockSpinE(Mutex *m, i32 spin)
 {
     b32 locked = 0;
     i32 spin_cnt = 0;
@@ -49,7 +49,7 @@ P_Lock P_LockSpinE(P_Mutex *m, i32 spin)
             }
             else
             {
-                P_Wait(&m->v, &v, 4, I64Max);
+                FutexWait(&m->v, &v, 4, I64Max);
                 spin_cnt = 0;
             }
         }
@@ -59,14 +59,14 @@ P_Lock P_LockSpinE(P_Mutex *m, i32 spin)
     Atomic32FetchSet(&m->exclusive_fiber_id, FiberId());
 #endif
 
-    P_Lock lock = ZI;
+    Lock lock = ZI;
     lock.exclusive = 1;
     lock.mutex = m;
     return lock;
 }
 
 //- Shared mutex lock
-P_Lock P_LockSpinS(P_Mutex *m, i32 spin)
+Lock LockSpinS(Mutex *m, i32 spin)
 {
     b32 locked = 0;
     i32 spin_cnt = 0;
@@ -96,30 +96,30 @@ P_Lock P_LockSpinS(P_Mutex *m, i32 spin)
             }
             else
             {
-                P_Wait(&m->v, &v, 4, I64Max);
+                FutexWait(&m->v, &v, 4, I64Max);
                 spin_cnt = 0;
             }
         }
     }
-    P_Lock lock = ZI;
+    Lock lock = ZI;
     lock.mutex = m;
     return lock;
 }
 
 //- Mutex lock wrappers
-P_Lock P_LockE(P_Mutex *m)
+Lock LockE(Mutex *m)
 {
-    return P_LockSpinE(m, P_DefaultMutexSpin);
+    return LockSpinE(m, DefaultMutexSpin);
 }
 
-P_Lock P_LockS(P_Mutex *m)
+Lock LockS(Mutex *m)
 {
-    return P_LockSpinS(m, P_DefaultMutexSpin);
+    return LockSpinS(m, DefaultMutexSpin);
 }
 
-void P_Unlock(P_Lock *l)
+void Unlock(Lock *l)
 {
-    P_Mutex *m = l->mutex;
+    Mutex *m = l->mutex;
     if (l->exclusive)
     {
 #if RtcIsEnabled
@@ -131,64 +131,64 @@ void P_Unlock(P_Lock *l)
     {
         Atomic32FetchAdd(&m->v, -1);
     }
-    P_Wake(&m->v, I32Max);
+    FutexWake(&m->v, I32Max);
     ZeroStruct(l);
 }
 
 ////////////////////////////////
 //~ Condition variable
 
-void P_WaitOnCv(P_Cv *cv, P_Lock *l)
+void WaitOnCv(Cv *cv, Lock *l)
 {
-    P_WaitOnCvTime(cv, l, I64Max);
+    WaitOnCvTime(cv, l, I64Max);
 }
 
-void P_WaitOnCvTime(P_Cv *cv, P_Lock *l, i64 timeout_ns)
+void WaitOnCvTime(Cv *cv, Lock *l, i64 timeout_ns)
 {
     u64 old_wake_gen = Atomic64Fetch(&cv->wake_gen);
-    P_Mutex *mutex = l->mutex;
+    Mutex *mutex = l->mutex;
     b32 exclusive = l->exclusive;
     {
-        P_Unlock(l);
+        Unlock(l);
         {
-            P_Wait(&cv->wake_gen, &old_wake_gen, sizeof(old_wake_gen), timeout_ns);
+            FutexWait(&cv->wake_gen, &old_wake_gen, sizeof(old_wake_gen), timeout_ns);
         }
         if (exclusive)
         {
-            *l = P_LockE(mutex);
+            *l = LockE(mutex);
         }
         else
         {
-            *l = P_LockS(mutex);
+            *l = LockS(mutex);
         }
     }
 }
 
-void P_SignalCv(P_Cv *cv, i32 count)
+void SignalCv(Cv *cv, i32 count)
 {
     Atomic64FetchAdd(&cv->wake_gen, 1);
-    P_Wake(&cv->wake_gen, count);
+    FutexWake(&cv->wake_gen, count);
 }
 
 ////////////////////////////////
 //~ Counter
 
-void P_CounterAdd(Counter *counter, i64 x)
+void AddCounter(Counter *counter, i64 x)
 {
     i64 old_v = Atomic64FetchAdd(&counter->v, x);
     i64 new_v = old_v + x;
     if (old_v > 0 && new_v <= 0)
     {
-        P_Wake(&counter->v, I32Max);
+        FutexWake(&counter->v, I32Max);
     }
 }
 
-void P_WaitOnCounter(Counter *counter)
+void WaitOnCounter(Counter *counter)
 {
     i64 v = Atomic64Fetch(&counter->v);
     while (v > 0)
     {
-        P_Wait(&counter->v, &v, sizeof(v), I64Max);
+        FutexWait(&counter->v, &v, sizeof(v), I64Max);
         v = Atomic64Fetch(&counter->v);
     }
 }
diff --git a/src/base/base_snc.h b/src/base/base_snc.h
new file mode 100644
index 00000000..d272ec11
--- /dev/null
+++ b/src/base/base_snc.h
@@ -0,0 +1,83 @@
+////////////////////////////////
+//~ Mutex types
+
+#define DefaultMutexSpin 4000
+
+AlignedStruct(Mutex, 64)
+{
+    /* Bit 31   = Exclusive lock is held
+     * Bit 30   = Exclusive lock is pending
+     * Bit 0-30 = Shared locks count
+     */
+    Atomic32 v;
+
+#if RtcIsEnabled
+    Atomic32 exclusive_fiber_id;
+    u8 _pad[56];
+#else
+    u8 _pad[60];
+#endif
+};
+StaticAssert(sizeof(Mutex) == 64);  /* Padding validation */
+StaticAssert(alignof(Mutex) == 64);  /* Prevent false sharing */
+
+Struct(Lock)
+{
+    Mutex *mutex;
+    b32 exclusive;
+};
+
+////////////////////////////////
+//~ Condition variable types
+
+AlignedStruct(Cv, 64)
+{
+    Atomic64 wake_gen;
+    u8 _pad[56];
+};
+StaticAssert(sizeof(Cv) == 64);  /* Padding validation */
+StaticAssert(alignof(Cv) == 64);  /* Prevent false sharing */
+
+////////////////////////////////
+//~ Counter types
+
+AlignedStruct(Counter, 64)
+{
+    Atomic64 v;
+    u8 _pad[56];
+};
+StaticAssert(sizeof(Counter) == 64);  /* Padding validation */
+StaticAssert(alignof(Counter) == 64);  /* Prevent false sharing */
+
+////////////////////////////////
+//~ Mutex operations
+
+Lock LockSpinE(Mutex *m, i32 spin);
+Lock LockSpinS(Mutex *m, i32 spin);
+
+Lock LockE(Mutex *m);
+Lock LockS(Mutex *m);
+
+void Unlock(Lock *lock);
+
+//- Lock assertion
+#if RtcIsEnabled
+# define AssertLockedE(l, m) Assert((l)->mutex == (m) && (l)->exclusive == 1)
+# define AssertLockedES(l, m) Assert((l)->mutex == (m))
+#else
+# define AssertLockedE(l, m) LAX l
+# define AssertLockedES(l, m) LAX l
+#endif
+
+////////////////////////////////
+//~ Condition variable operations
+
+void WaitOnCv(Cv *cv, Lock *lock);
+void WaitOnCvTime(Cv *cv, Lock *l, i64 timeout_ns);
+void SignalCv(Cv *cv, i32 count);
+
+////////////////////////////////
+//~ Counter operations
+
+void AddCounter(Counter *counter, i64 x);
+void WaitOnCounter(Counter *counter);
diff --git a/src/base/base_string.c b/src/base/base_string.c
index 94c0190f..f81bf5aa 100644
--- a/src/base/base_string.c
+++ b/src/base/base_string.c
@@ -63,12 +63,12 @@ String StringFromI64(Arena *arena, i64 n, u64 base, u64 zfill)
     u8 len = 0;
     if (n < 0)
     {
-        /* PushStruct sign */
+        /* Push sign */
         StringFromChar(arena, '-');
         len = 1;
         n = -n;
     }
-    /* PushStruct unsigned number */
+    /* Push unsigned number */
     String uint_str = StringFromU64(arena, n, base, zfill);
     return (String)
     {
diff --git a/src/base/win32/base_win32_job.c b/src/base/win32/base_win32_job.c
index 2f265e59..43add503 100644
--- a/src/base/win32/base_win32_job.c
+++ b/src/base/win32/base_win32_job.c
@@ -1,5 +1,7 @@
 W32_SharedCtx W32_shared_ctx = ZI;
 
+/* FIXME: Enable logs, panic, shutdown */
+
 ////////////////////////////////
 //~ Win32 libs
 
@@ -21,6 +23,19 @@ void StartupJobs(void)
 {
     W32_SharedCtx *g = &W32_shared_ctx;
 
+    /* Init timer */
+    {
+        LARGE_INTEGER qpf;
+        QueryPerformanceFrequency(&qpf);
+        g->ns_per_qpc = 1000000000 / qpf.QuadPart;
+    }
+    {
+        LARGE_INTEGER qpc;
+        QueryPerformanceCounter(&qpc);
+        g->timer_start_qpc = qpc.QuadPart;
+    }
+
+
     /* Init fibers */
     g->num_fibers = 1;  /* Fiber at index 0 always nil */
     g->fiber_names_arena = AllocArena(Gibi(64));
@@ -47,6 +62,7 @@ void StartupJobs(void)
     /* Start job scheduler */
     Atomic64FetchSet(&g->current_scheduler_cycle_period_ns.v, W32_DefaultSchedulerPeriodNs);
     W32_Thread *scheduler_thread = W32_AllocThread(W32_JobSchedulerEntryFunc, 0, Lit("Scheduler thread"), PROF_THREAD_GROUP_SCHEDULER);
+    LAX scheduler_thread;
 
     //- Start job workers
     /* TODO: Heuristic worker counts & affinities */
@@ -114,12 +130,12 @@ void StartupJobs(void)
         }
     }
 
-    P_OnExit(ShutdownJobs);
+    //P_OnExit(ShutdownJobs);
 }
 
 ////////////////////////////////
 //~ Shutdown
-
+#if 0
 void ShutdownJobs(void)
 {
     /* Signal shutdown */
@@ -162,7 +178,7 @@ void ShutdownJobs(void)
     /* Find any dangling threads that haven't exited gracefully by now */
     if (!Atomic32Fetch(&g->panicking))
     {
-        P_Lock lock = P_LockS(&g->threads_mutex);
+        LockTicketMutex(&g->threads_tm);
         if (g->first_thread)
         {
             TempArena scratch = BeginScratchNoConflict();
@@ -176,29 +192,13 @@ void ShutdownJobs(void)
                 ++num_dangling_threads;
             }
             threads_msg = StringFormat(scratch.arena, Lit("%F dangling thread(s):\n%F"), FmtUint(num_dangling_threads), FmtString(threads_msg));
-            P_Panic(threads_msg);
+            //P_Panic(threads_msg);
             EndScratch(scratch);
         }
-        P_Unlock(&lock);
+        UnlockTicketMutex(&g->threads_tm);
     }
 }
-
-////////////////////////////////
-//~ Win32 ticket mutex
-
-void LockTicketMutex(W32_TicketMutex *tm)
-{
-    i64 ticket = Atomic64FetchAdd(&tm->ticket.v, 1);
-    while (Atomic64Fetch(&tm->serving.v) != ticket)
-    {
-        _mm_pause();
-    }
-}
-
-void UnlockTicketMutex(W32_TicketMutex *tm)
-{
-    Atomic64FetchAdd(&tm->serving.v, 1);
-}
+#endif
 
 ////////////////////////////////
 //~ Win32 thread
@@ -219,7 +219,7 @@ DWORD WINAPI W32_Win32ThreadProc(LPVOID vt)
         SetThreadDescription(GetCurrentThread(), t->thread_name_wstr);
     }
 
-    P_LogInfoF("New thread \"%F\" created with ID %F", FmtString(StringFromCstrNoLimit(t->thread_name_cstr)), FmtUint(P_GetThreadId()));
+    //P_LogInfoF("New thread \"%F\" created with ID %F", FmtString(StringFromCstrNoLimit(t->thread_name_cstr)), FmtUint(ThreadId()));
 
     /* Enter thread entry point */
     t->entry_point(t->thread_data);
@@ -236,12 +236,12 @@ W32_Thread *W32_AllocThread(W32_ThreadFunc *entry_point, void *thread_data, Stri
     TempArena scratch = BeginScratchNoConflict();
     W32_SharedCtx *g = &W32_shared_ctx;
     Assert(entry_point != 0);
-    P_LogInfoF("Creating thread \"%F\"", FmtString(thread_name));
+    //P_LogInfoF("Creating thread \"%F\"", FmtString(thread_name));
 
     /* Allocate thread object */
     W32_Thread *t = 0;
     {
-        P_Lock lock = P_LockE(&g->threads_mutex);
+        LockTicketMutex(&g->threads_tm);
         if (g->first_free_thread)
         {
             t = g->first_free_thread;
@@ -262,7 +262,7 @@ W32_Thread *W32_AllocThread(W32_ThreadFunc *entry_point, void *thread_data, Stri
             g->first_thread = t;
         }
         g->last_thread = t;
-        P_Unlock(&lock);
+        UnlockTicketMutex(&g->threads_tm);
     }
 
     t->entry_point = entry_point;
@@ -293,7 +293,7 @@ W32_Thread *W32_AllocThread(W32_ThreadFunc *entry_point, void *thread_data, Stri
 
     if (!t->handle)
     {
-        P_Panic(Lit("Failed to create thread"));
+        //P_Panic(Lit("Failed to create thread"));
     }
 
     EndScratch(scratch);
@@ -319,7 +319,7 @@ b32 W32_TryReleaseThread(W32_Thread *thread, f32 timeout_seconds)
             success = 1;
             CloseHandle(handle);
             {
-                P_Lock lock = P_LockE(&g->threads_mutex);
+                LockTicketMutex(&g->threads_tm);
                 {
                     W32_Thread *prev = t->prev;
                     W32_Thread *next = t->next;
@@ -342,7 +342,7 @@ b32 W32_TryReleaseThread(W32_Thread *thread, f32 timeout_seconds)
                     t->next = g->first_free_thread;
                     g->first_free_thread = t;
                 }
-                P_Unlock(&lock);
+                UnlockTicketMutex(&g->threads_tm);
             }
         }
     }
@@ -710,30 +710,30 @@ W32_Fiber *W32_AllocFiber(W32_JobPool *pool)
     {
         if (pool != 0)
         {
-            LockTicketMutex(&pool->free_fibers_lock);
+            LockTicketMutex(&pool->free_fibers_tm);
             if (pool->first_free_fiber_id)
             {
                 fiber_id = pool->first_free_fiber_id;
                 fiber = &g->fibers[fiber_id];
                 pool->first_free_fiber_id = fiber->parent_id;
             }
-            UnlockTicketMutex(&pool->free_fibers_lock);
+            UnlockTicketMutex(&pool->free_fibers_tm);
         }
         if (!fiber_id)
         {
-            LockTicketMutex(&g->fibers_lock);
+            LockTicketMutex(&g->fibers_tm);
             {
                 {
                     fiber_id = g->num_fibers++;
                     if (fiber_id >= MaxFibers)
                     {
-                        P_Panic(Lit("Max fibers reached"));
+                        //P_Panic(Lit("Max fibers reached"));
                     }
                     fiber = &g->fibers[fiber_id];
                     new_name_cstr = PushStructs(g->fiber_names_arena, char, W32_FiberNameMaxSize);
                 }
             }
-            UnlockTicketMutex(&g->fibers_lock);
+            UnlockTicketMutex(&g->fibers_tm);
         }
 
     }
@@ -809,13 +809,13 @@ W32_Fiber *W32_AllocFiber(W32_JobPool *pool)
 //- Release fiber
 void W32_ReleaseFiber(W32_JobPool *pool, W32_Fiber *fiber)
 {
-    LockTicketMutex(&pool->free_fibers_lock);
+    LockTicketMutex(&pool->free_fibers_tm);
     {
         i16 fiber_id = fiber->id;
         fiber->parent_id = pool->first_free_fiber_id;
         pool->first_free_fiber_id = fiber_id;
     }
-    UnlockTicketMutex(&pool->free_fibers_lock);
+    UnlockTicketMutex(&pool->free_fibers_tm);
 }
 
 
@@ -855,6 +855,7 @@ void W32_YieldFiber(W32_Fiber *fiber, W32_Fiber *parent_fiber)
 }
 
 //- Fiber entry
+
 void W32_FiberEntryPoint(void *id_ptr)
 {
     i16 id = (i32)(i64)id_ptr;
@@ -866,12 +867,10 @@ void W32_FiberEntryPoint(void *id_ptr)
         {
             W32_YieldParam *yield_param = fiber->yield_param;
             yield_param->kind = W32_YieldKind_None;
-            P_JobData data = ZI;
-            data.id = fiber->job_id;
-            data.sig = fiber->job_sig;
             {
                 MemoryBarrier();
-                fiber->job_func(data);
+                GenericJobFunc *job_func = fiber->job_func;
+                job_func(fiber->job_sig, fiber->job_id);
                 MemoryBarrier();
             }
         }
@@ -881,7 +880,7 @@ void W32_FiberEntryPoint(void *id_ptr)
             Counter *job_counter = fiber->job_counter;
             if (job_counter)
             {
-                P_CounterAdd(job_counter, -1);
+                AddCounter(job_counter, -1);
             }
             /* Yield to worker */
             fiber->yield_param->kind = W32_YieldKind_Done;
@@ -959,7 +958,7 @@ W32_ThreadDef(W32_JobWorkerEntryFunc, worker_ctx_arg)
         PriorityKind job_priority = 0;
         i16 job_fiber_id = 0;
         i32 job_id = 0;
-        P_JobFunc *job_func = 0;
+        GenericJobFunc *job_func = 0;
         void *job_sig = 0;
         Counter *job_counter = 0;
         {
@@ -1064,7 +1063,7 @@ W32_ThreadDef(W32_JobWorkerEntryFunc, worker_ctx_arg)
                         {
                             /* Invalid yield kind */
                             TempArena scratch = BeginScratchNoConflict();
-                            P_Panic(StringFormat(scratch.arena, Lit("Invalid fiber yield kind \"%F\""), FmtSint(yield.kind)));
+                            //P_Panic(StringFormat(scratch.arena, Lit("Invalid fiber yield kind \"%F\""), FmtSint(yield.kind)));
                             EndScratch(scratch);
                         } break;
 
@@ -1133,11 +1132,11 @@ W32_ThreadDef(W32_JobWorkerEntryFunc, worker_ctx_arg)
                                                 }
                                                 else
                                                 {
-                                                    LockTicketMutex(&g->wait_lists_arena_lock);
+                                                    LockTicketMutex(&g->wait_lists_arena_tm);
                                                     {
                                                         wait_addr_list = PushStructNoZero(g->wait_lists_arena, W32_WaitList);
                                                     }
-                                                    UnlockTicketMutex(&g->wait_lists_arena_lock);
+                                                    UnlockTicketMutex(&g->wait_lists_arena_tm);
                                                 }
                                                 ZeroStruct(wait_addr_list);
                                                 wait_addr_list->value = (u64)wait_addr;
@@ -1187,11 +1186,11 @@ W32_ThreadDef(W32_JobWorkerEntryFunc, worker_ctx_arg)
                                                 }
                                                 else
                                                 {
-                                                    LockTicketMutex(&g->wait_lists_arena_lock);
+                                                    LockTicketMutex(&g->wait_lists_arena_tm);
                                                     {
                                                         wait_time_list = PushStructNoZero(g->wait_lists_arena, W32_WaitList);
                                                     }
-                                                    UnlockTicketMutex(&g->wait_lists_arena_lock);
+                                                    UnlockTicketMutex(&g->wait_lists_arena_tm);
                                                 }
                                                 ZeroStruct(wait_time_list);
                                                 wait_time_list->value = wait_time;
@@ -1291,7 +1290,7 @@ W32_ThreadDef(W32_JobSchedulerEntryFunc, UNUSED arg)
     HANDLE timer = CreateWaitableTimerExW(0, 0, CREATE_WAITABLE_TIMER_HIGH_RESOLUTION, TIMER_ALL_ACCESS);
     if (!timer)
     {
-        P_Panic(Lit("Failed to create high resolution timer"));
+        //P_Panic(Lit("Failed to create high resolution timer"));
     }
 
     /* Create rolling buffer of scheduler cycles initialized to default value */
@@ -1319,7 +1318,7 @@ W32_ThreadDef(W32_JobSchedulerEntryFunc, UNUSED arg)
         }
 
         /* Calculate mean period */
-        i64 now_ns = P_TimeNs();
+        i64 now_ns = TimeNs();
         i64 period_ns = last_cycle_ns == 0 ? W32_DefaultSchedulerPeriodNs : now_ns - last_cycle_ns;
         last_cycle_ns = now_ns;
 
@@ -1348,9 +1347,9 @@ W32_ThreadDef(W32_JobSchedulerEntryFunc, UNUSED arg)
 }
 
 ////////////////////////////////
-//~ Wait / wake
+//~ @hookdef Futex
 
-void P_Wait(volatile void *addr, void *cmp, u32 size, i64 timeout_ns)
+void FutexWait(volatile void *addr, void *cmp, u32 size, i64 timeout_ns)
 {
     W32_Fiber *fiber = W32_FiberFromId(FiberId());
     i16 parent_id = fiber->parent_id;
@@ -1390,31 +1389,46 @@ void P_Wait(volatile void *addr, void *cmp, u32 size, i64 timeout_ns)
     }
 }
 
-void P_Wake(void *addr, i32 count)
+void FutexWake(void *addr, i32 count)
 {
     W32_WakeByAddress(addr, count);
 }
 
 ////////////////////////////////
-//~ Fiber
+//~ @hoodef Job
 
-i16 FiberId(void)
+GenericJobDesc *PushJobDesc_(u64 sig_size, u64 sig_align, GenericJobFunc *func, JobDescParams params)
 {
-    return (i16)(i64)GetFiberData();
+    GenericJobDesc *result = 0;
+    /* FIXME: Pool arenas */
+    Arena *arena = AllocArena(Mebi(1));
+    result = PushStruct(arena, GenericJobDesc);
+    result->arena = arena;
+    result->sig = PushBytes(arena, sig_size, sig_align);
+    result->func = func;
+    result->count = params.count;
+    result->pool = params.pool;
+    result->priority = params.priority;
+    result->counter = params.counter;
+    return result;
 }
 
-////////////////////////////////
-//~ Job
-
-void P_Run(i32 count, P_JobFunc *func, void *sig, PoolKind pool_kind, PriorityKind priority, Counter *counter)
+#if 1
+void RunJobEx(GenericJobDesc *desc)
 {
     __prof;
     struct W32_SharedCtx *g = &W32_shared_ctx;
+    i32 count = desc->count;
+    Counter *counter = desc->counter;
+    PoolKind pool_kind = desc->pool;
+    PriorityKind priority = desc->priority;
+    GenericJobFunc *func = desc->func;
+    void *sig = desc->sig;
     if (count > 0)
     {
         if (counter)
         {
-            P_CounterAdd(counter, count);
+            AddCounter(counter, count);
         }
         W32_Fiber *fiber = W32_FiberFromId(FiberId());
         priority = ClampI32(priority, fiber->job_priority, PriorityKind_Count - 1);  /* A job cannot create a job with a higher priority than itself */
@@ -1474,3 +1488,101 @@ void P_Run(i32 count, P_JobFunc *func, void *sig, PoolKind pool_kind, PriorityKi
         }
     }
 }
+#else
+void RunJob(i32 count, GenericJobFunc *func, void *sig, PoolKind pool_kind, PriorityKind priority, Counter *counter)
+{
+    __prof;
+    struct W32_SharedCtx *g = &W32_shared_ctx;
+    if (count > 0)
+    {
+        if (counter)
+        {
+            AddCounter(counter, count);
+        }
+        W32_Fiber *fiber = W32_FiberFromId(FiberId());
+        priority = ClampI32(priority, fiber->job_priority, PriorityKind_Count - 1);  /* A job cannot create a job with a higher priority than itself */
+        if (pool_kind == PoolKind_Inherit)
+        {
+            pool_kind = fiber->job_pool;
+        }
+        W32_JobPool *pool = &g->job_pools[pool_kind];
+        W32_JobQueue *queue = &pool->job_queues[priority];
+        LockTicketMutex(&queue->lock);
+        {
+            W32_JobInfo *info = 0;
+            if (queue->first_free)
+            {
+                info = queue->first_free;
+                queue->first_free = info->next;
+            }
+            else
+            {
+                info = PushStructNoZero(queue->arena, W32_JobInfo);
+            }
+            ZeroStruct(info);
+            info->count = count;
+            info->func = func;
+            info->sig = sig;
+            info->counter = counter;
+            if (queue->last)
+            {
+                queue->last->next = info;
+            }
+            else
+            {
+                queue->first = info;
+            }
+            queue->last = info;
+        }
+        UnlockTicketMutex(&queue->lock);
+
+        /* Wake workers */
+        {
+            LockTicketMutex(&pool->workers_wake_tm);
+            {
+                Atomic64FetchAdd(&pool->num_jobs_in_queue.v, count);
+                if (count >= W32_WakeAllThreshold)
+                {
+                    WakeByAddressAll(&pool->num_jobs_in_queue);
+                }
+                else
+                {
+                    for (i32 i = 0; i < count; ++i)
+                    {
+                        WakeByAddressSingle(&pool->num_jobs_in_queue);
+                    }
+                }
+            }
+            UnlockTicketMutex(&pool->workers_wake_tm);
+        }
+    }
+}
+#endif
+
+////////////////////////////////
+//~ @hookdef Helpers
+
+i64 TimeNs(void)
+{
+    struct W32_SharedCtx *g = &W32_shared_ctx;
+    LARGE_INTEGER qpc;
+    QueryPerformanceCounter(&qpc);
+    i64 result = (qpc.QuadPart - g->timer_start_qpc) * g->ns_per_qpc;
+    return result;
+}
+
+u32 GetLogicalProcessorCount(void)
+{
+    return GetActiveProcessorCount(ALL_PROCESSOR_GROUPS);
+}
+
+u32 ThreadId(void)
+{
+    return GetCurrentThreadId();
+}
+
+i64 GetCurrentSchedulerPeriodNs(void)
+{
+    W32_SharedCtx *g = &W32_shared_ctx;
+    return Atomic64Fetch(&g->current_scheduler_cycle_period_ns.v);
+}
diff --git a/src/base/win32/base_win32_job.h b/src/base/win32/base_win32_job.h
index 8f370588..dfd9b8c2 100644
--- a/src/base/win32/base_win32_job.h
+++ b/src/base/win32/base_win32_job.h
@@ -5,7 +5,6 @@
 #  define UNICODE
 #  define WIN32_LEAN_AND_MEAN
 #  include <Windows.h>
-#if 0
 #  include <WinSock2.h>
 #  include <TlHelp32.h>
 #  include <WS2tcpip.h>
@@ -16,7 +15,6 @@
 #  include <bcrypt.h>
 #  include <avrt.h>
 #  include <shellapi.h>
-#endif
 #pragma warning(pop)
 
 ////////////////////////////////
@@ -123,7 +121,7 @@ AlignedStruct(W32_Fiber, 64)
     /* ---------------------------------------------------- */
     /* -------------------- Cache line -------------------- */
     /* ---------------------------------------------------- */
-    void *job_func;                             /* 08 bytes */
+    GenericJobFunc *job_func;                   /* 08 bytes */
     /* ---------------------------------------------------- */
     void *job_sig;                              /* 08 bytes */
     /* ---------------------------------------------------- */
@@ -158,7 +156,7 @@ Struct(W32_JobInfo)
     i32 num_dispatched;
 
     i32 count;
-    void *func;
+    GenericJobFunc *func;
     void *sig;
     struct Counter *counter;
 
@@ -217,6 +215,8 @@ AlignedStruct(W32_JobPool, 64)
 
 Struct(W32_SharedCtx)
 {
+    i64 timer_start_qpc;
+    i64 ns_per_qpc;
     Atomic32 shutdown;
 
     //- Worker thread pool
diff --git a/src/font/font_core.c b/src/font/font_core.c
index 6c647afe..dc7265bb 100644
--- a/src/font/font_core.c
+++ b/src/font/font_core.c
@@ -1,7 +1,7 @@
 ////////////////////////////////
 //~ Load job
 
-JobDef(F_LoadAssetJob, job)
+JobDef(F_LoadJob, sig, _)
 {
     __prof;
     TempArena scratch = BeginScratchNoConflict();
@@ -25,13 +25,12 @@ JobDef(F_LoadAssetJob, job)
         0xF2,0xF3,0xF4,0xF5,0xF6,0xF7,0xF8,0xF9,0xFA,0xFB,0xFC,0xFD,0xFE,0xFF
     };
 
-    F_LoadJobSig *params = job.sig;
-    String path = STRING(params->path_len, (u8 *)params->path_cstr);
-    f32 point_size = params->point_size;
-    AC_Asset *asset = params->asset;
+    String path = sig->path;
+    f32 point_size = sig->point_size;
+    AC_Asset *asset = sig->asset;
 
     P_LogInfoF("Loading font \"%F\" (point size %F)", FmtString(path), FmtFloat((f64)point_size));
-    i64 start_ns = P_TimeNs();
+    i64 start_ns = TimeNs();
 
     Assert(StringEndsWith(path, Lit(".ttf")));
     Assert(countof(font_codes) < F_LookupTableSize);
@@ -87,7 +86,7 @@ JobDef(F_LoadAssetJob, job)
         font->lookup[codepoint] = result.cache_indices[i];
     }
 
-    P_LogSuccessF("Loaded font \"%F\" (point size %F) in %F seconds", FmtString(path), FmtFloat((f64)point_size), FmtFloat(SecondsFromNs(P_TimeNs() - start_ns)));
+    P_LogSuccessF("Loaded font \"%F\" (point size %F) in %F seconds", FmtString(path), FmtFloat((f64)point_size), FmtFloat(SecondsFromNs(TimeNs() - start_ns)));
     AC_MarkReady(asset, font);
 
     EndScratch(scratch);
@@ -113,22 +112,12 @@ AC_Asset *F_LoadAsset(String path, f32 point_size, b32 wait)
 
     if (is_first_touch)
     {
-        /* Assemble task params */
-        F_LoadJobSig *params = F_AllocJobSig();
-        if (path.len > (sizeof(params->path_cstr) - 1))
-        {
-            P_Panic(StringFormat(scratch.arena,
-                                 Lit("Font path \"%F\" too long!"),
-                                 FmtString(path)));
-        }
-        CstrBuffFromStringToBuff(StringFromArray(params->path_cstr), path);
-        params->path_len = path.len;
-        params->asset = asset;
-        params->point_size = point_size;
-
-        /* PushStruct task */
         AC_MarkLoading(asset);
-        P_Run(1, F_LoadAssetJob, params, PoolKind_Background, PriorityKind_Low, 0);
+        F_LoadJob_Desc *desc = PushJobDesc(F_LoadJob, .pool = PoolKind_Background, .priority = PriorityKind_Low);
+        desc->sig->asset = asset;
+        desc->sig->path = PushString(desc->arena, path);
+        desc->sig->point_size = point_size;
+        RunJobEx((GenericJobDesc *)desc);
         if (wait)
         {
             AC_WaitOnAssetReady(asset);
diff --git a/src/font/font_core.h b/src/font/font_core.h
index 7a7e2ce6..d9bd209d 100644
--- a/src/font/font_core.h
+++ b/src/font/font_core.h
@@ -27,7 +27,7 @@ Struct(F_Font)
 ////////////////////////////////
 //~ Font load job
 
-JobDecl(F_LoadAssetJob, { AC_Asset *asset; f32 point_size; String path; });
+JobDecl(F_LoadJob, { AC_Asset *asset; f32 point_size; String path; });
 
 ////////////////////////////////
 //~ Font load operations
diff --git a/src/gpu/gpu_dx12.c b/src/gpu/gpu_dx12.c
index d08f21a4..fd0c2fb0 100644
--- a/src/gpu/gpu_dx12.c
+++ b/src/gpu/gpu_dx12.c
@@ -71,7 +71,7 @@ void GPU_Startup(void)
     P_OnExit(GPU_D12_Shutdown);
 
     /* Start evictor job */
-    P_Run(1, GPU_D12_EvictorJob, 0, PoolKind_Background, PriorityKind_Low, &g->evictor_job_counter);
+    RunJob(1, GPU_D12_EvictorJob, PoolKind_Background, PriorityKind_Low, &g->evictor_job_counter, 0);
 }
 
 P_ExitFuncDef(GPU_D12_Shutdown)
@@ -92,12 +92,12 @@ P_ExitFuncDef(GPU_D12_Shutdown)
 #endif
 
     {
-        P_Lock lock = P_LockE(&g->evictor_wake_mutex);
+        Lock lock = LockE(&g->evictor_wake_mutex);
         g->evictor_shutdown = 1;
-        P_SignalCv(&g->evictor_wake_cv, I32Max);
-        P_Unlock(&lock);
+        SignalCv(&g->evictor_wake_cv, I32Max);
+        Unlock(&lock);
     }
-    P_WaitOnCounter(&g->evictor_job_counter);
+    WaitOnCounter(&g->evictor_job_counter);
 }
 
 /* ========================== *
@@ -322,13 +322,10 @@ void GPU_D12_InitObjects(void)
             {.type = D3D12_COMMAND_LIST_TYPE_COPY, .priority = D3D12_COMMAND_QUEUE_PRIORITY_HIGH, .dbg_name = Lit("Copy queue") },
             {.type = D3D12_COMMAND_LIST_TYPE_COPY, .priority = D3D12_COMMAND_QUEUE_PRIORITY_NORMAL, .dbg_name = Lit("Background copy queue") }
         };
-        GPU_D12_AllocCommandQueueJobSig sig = ZI;
-        sig.descs_in = params;
-        sig.cqs_out = g->command_queues;
         {
             Counter counter = ZI;
-            P_Run(DX12_NUM_QUEUES, GPU_D12_AllocCommandQueueJob, &sig, PoolKind_Inherit, PriorityKind_Inherit, &counter);
-            P_WaitOnCounter(&counter);
+            RunJob(DX12_NUM_QUEUES, GPU_D12_AllocCommandQueueJob, PoolKind_Inherit, PriorityKind_Low, &counter, .descs_in = params, .cqs_out = g->command_queues);
+            WaitOnCounter(&counter);
         }
 #if ProfilingIsEnabled
         {
@@ -418,12 +415,9 @@ void GPU_InitPipelines(void)
     GPU_D12_Pipeline **pipelines = PushStructs(scratch.arena, GPU_D12_Pipeline *, num_pipelines);
     {
         __profn("Allocate pipelines");
-        GPU_D12_AllocPipelineJobSig sig = ZI;
-        sig.descs_in = descs;
-        sig.pipelines_out = pipelines;
         Counter counter = ZI;
-        P_Run(num_pipelines, GPU_D12_AllocPipelineJob, &sig, PoolKind_Inherit, PriorityKind_Inherit, &counter);
-        P_WaitOnCounter(&counter);
+        RunJob(num_pipelines, GPU_D12_AllocPipelineJob, PoolKind_Inherit, PriorityKind_Inherit, &counter, .descs_in = descs, .pipelines_out = pipelines);
+        WaitOnCounter(&counter);
     }
     for (u32 i = 0; i < num_pipelines; ++i)
     {
@@ -502,11 +496,8 @@ void GPU_D12_InitNoise(void)
                 /* Upload texture */
                 {
                     Counter counter = ZI;
-                    GPU_D12_UploadJobSig sig = ZI;
-                    sig.resource = r;
-                    sig.data = data.text;
-                    P_Run(1, GPU_D12_UploadJob, &sig, PoolKind_Inherit, PriorityKind_Inherit, &counter);
-                    P_WaitOnCounter(&counter);
+                    RunJob(1, GPU_D12_UploadJob, PoolKind_Inherit, PriorityKind_Low, &counter, .resource = r, .data = data.text);
+                    WaitOnCounter(&counter);
                 }
             }
         }
@@ -526,17 +517,16 @@ void GPU_D12_InitNoise(void)
 
 #if RESOURCE_RELOADING
 
-JobDef(GPU_D12_CompileShaderJob, job)
+JobDef(GPU_D12_CompileShaderJob, sig, id)
 {
     __prof;
-    GPU_D12_CompileShaderJobSig *sig = job.sig;
     Arena *arena = sig->arena;
-    GPU_D12_ShaderDesc *desc = &sig->descs[job.id];
-    GPU_D12_CompiledShaderResult *result = &sig->results[job.id];
+    GPU_D12_ShaderDesc *desc = &sig->descs[id];
+    GPU_D12_CompiledShaderResult *result = &sig->results[id];
 
     TempArena scratch = BeginScratch(arena);
     {
-        i64 start_ns = P_TimeNs();
+        i64 start_ns = TimeNs();
         DXC_Result dxc_result = ZI;
         {
             __profn("Compile shader");
@@ -565,7 +555,7 @@ JobDef(GPU_D12_CompileShaderJob, job)
         result->success = dxc_result.success;
         result->dxc = dxc_result.dxc;
         result->errors = dxc_result.errors;
-        result->elapsed_ns = P_TimeNs() - start_ns;
+        result->elapsed_ns = TimeNs() - start_ns;
 
     }
     EndScratch(scratch);
@@ -577,17 +567,16 @@ JobDef(GPU_D12_CompileShaderJob, job)
  * Pipeline
  * ========================== */
 
-JobDef(GPU_D12_AllocPipelineJob, job)
+JobDef(GPU_D12_AllocPipelineJob, sig, id)
 {
     __prof;
     GPU_D12_SharedState *g = &GPU_D12_shared_state;
-    GPU_D12_AllocPipelineJobSig *sig = job.sig;
-    GPU_D12_PipelineDesc *desc = &sig->descs_in[job.id];
+    GPU_D12_PipelineDesc *desc = &sig->descs_in[id];
     GPU_D12_Pipeline **pipelines_out = sig->pipelines_out;
 
     GPU_D12_Pipeline *pipeline = 0;
     {
-        P_Lock lock = P_LockE(&g->pipelines_mutex);
+        Lock lock = LockE(&g->pipelines_mutex);
         if (g->first_free_pipeline)
         {
             pipeline = g->first_free_pipeline;
@@ -597,17 +586,17 @@ JobDef(GPU_D12_AllocPipelineJob, job)
         {
             pipeline = PushStructNoZero(g->pipelines_arena, GPU_D12_Pipeline);
         }
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
     ZeroStruct(pipeline);
-    pipelines_out[job.id] = pipeline;
+    pipelines_out[id] = pipeline;
     pipeline->desc = *desc;
     pipeline->name = desc->name;
     pipeline->hash = HashFnv64(Fnv64Basis, pipeline->name);
 
     TempArena scratch = BeginScratchNoConflict();
     {
-        i64 start_ns = P_TimeNs();
+        i64 start_ns = TimeNs();
         String pipeline_name = pipeline->name;
         P_LogInfoF("Loading pipeline \"%F\"", FmtString(pipeline_name));
         b32 success = 1;
@@ -880,7 +869,7 @@ JobDef(GPU_D12_AllocPipelineJob, job)
 
         pipeline->pso = pso;
         pipeline->rootsig = rootsig;
-        pipeline->compilation_time_ns = P_TimeNs() - start_ns;
+        pipeline->compilation_time_ns = TimeNs() - start_ns;
         pipeline->success = success;
         pipeline->is_gfx = cs_dxc.len == 0;
         pipeline->error = error_str;
@@ -913,12 +902,12 @@ void GPU_D12_ReleasePipelineNow(GPU_D12_Pipeline *pipeline)
     {
         ID3D12PipelineState_Release(pipeline->pso);
     }
-    P_Lock lock = P_LockE(&g->pipelines_mutex);
+    Lock lock = LockE(&g->pipelines_mutex);
     {
         pipeline->next = g->first_free_pipeline;
         g->first_free_pipeline = pipeline;
     }
-    P_Unlock(&lock);
+    Unlock(&lock);
 }
 
 /* ========================== *
@@ -931,13 +920,13 @@ GPU_D12_PipelineScope *GPU_D12_BeginPipelineScope(void)
     GPU_D12_SharedState *g = &GPU_D12_shared_state;
     GPU_D12_PipelineScope *scope = 0;
     {
-        P_Lock lock = P_LockE(&g->pipelines_mutex);
+        Lock lock = LockE(&g->pipelines_mutex);
         if (g->first_free_pipeline_scope)
         {
             scope = g->first_free_pipeline_scope;
             g->first_free_pipeline_scope = scope->next_free;
         }
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
     Arena *arena = 0;
     if (scope)
@@ -959,7 +948,7 @@ void GPU_D12_EndPipelineScope(GPU_D12_PipelineScope *scope)
 {
     __prof;
     GPU_D12_SharedState *g = &GPU_D12_shared_state;
-    P_Lock lock = P_LockE(&g->pipelines_mutex);
+    Lock lock = LockE(&g->pipelines_mutex);
     {
         for (DictEntry *entry = scope->refs->first; entry; entry = entry->next)
         {
@@ -972,7 +961,7 @@ void GPU_D12_EndPipelineScope(GPU_D12_PipelineScope *scope)
         scope->next_free = g->first_free_pipeline_scope;
         g->first_free_pipeline_scope = scope;
     }
-    P_Unlock(&lock);
+    Unlock(&lock);
 }
 
 Readonly GPU_D12_Pipeline GPU_D12_nil_pipeline = ZI;
@@ -991,13 +980,13 @@ GPU_D12_Pipeline *GPU_D12_PipelineFromName(GPU_D12_PipelineScope *scope, String
     else
     {
         {
-            P_Lock lock = P_LockE(&g->pipelines_mutex);
+            Lock lock = LockE(&g->pipelines_mutex);
             tmp = (GPU_D12_Pipeline *)DictValueFromHash(g->top_successful_pipelines, hash);
             if (tmp)
             {
                 ++tmp->refcount;
             }
-            P_Unlock(&lock);
+            Unlock(&lock);
         }
         if (tmp)
         {
@@ -1013,7 +1002,7 @@ void GPU_D12_RegisterPipeline(u64 num_pipelines, GPU_D12_Pipeline **pipelines)
 {
     __prof;
     GPU_D12_SharedState *g = &GPU_D12_shared_state;
-    P_Lock lock = P_LockE(&g->pipelines_mutex);
+    Lock lock = LockE(&g->pipelines_mutex);
     {
         for (u64 i = 0; i < num_pipelines; ++i)
         {
@@ -1042,7 +1031,7 @@ void GPU_D12_RegisterPipeline(u64 num_pipelines, GPU_D12_Pipeline **pipelines)
             }
         }
     }
-    P_Unlock(&lock);
+    Unlock(&lock);
 }
 
 #if RESOURCE_RELOADING
@@ -1085,40 +1074,42 @@ W_CallbackFuncDef(GPU_D12_WatchPipelineCallback, name)
             pipeline_name = split.count > 1 ? split.strings[split.count - 2] : pipeline_name;
         }
         {
-            GPU_D12_CompileShaderJobSig sig = ZI;
-            sig.arena = scratch.arena;
+            GPU_D12_CompileShaderJob_Desc *job_desc = PushJobDesc(GPU_D12_CompileShaderJob, .count = 0);
+            job_desc->sig->arena = scratch.arena;
             if (is_rs)
             {
                 num_shaders = 2;
                 shader_descs = PushStructs(scratch.arena, GPU_D12_ShaderDesc, num_shaders);
                 shader_results = PushStructs(scratch.arena, GPU_D12_CompiledShaderResult, num_shaders);
-                sig.descs = shader_descs;
-                sig.results = shader_results;
-                sig.descs[0].src = data;
-                sig.descs[0].friendly_name = friendly_name;
-                sig.descs[0].entry = Lit("vs");
-                sig.descs[0].target = Lit("vs_6_6");
-                sig.descs[1].src = data;
-                sig.descs[1].friendly_name = friendly_name;
-                sig.descs[1].entry = Lit("ps");
-                sig.descs[1].target = Lit("ps_6_6");
+                job_desc->sig->descs = shader_descs;
+                job_desc->sig->results = shader_results;
+                job_desc->sig->descs[0].src = data;
+                job_desc->sig->descs[0].friendly_name = friendly_name;
+                job_desc->sig->descs[0].entry = Lit("vs");
+                job_desc->sig->descs[0].target = Lit("vs_6_6");
+                job_desc->sig->descs[1].src = data;
+                job_desc->sig->descs[1].friendly_name = friendly_name;
+                job_desc->sig->descs[1].entry = Lit("ps");
+                job_desc->sig->descs[1].target = Lit("ps_6_6");
             }
             else if (is_cs)
             {
                 num_shaders = 1;
                 shader_descs = PushStructs(scratch.arena, GPU_D12_ShaderDesc, num_shaders);
                 shader_results = PushStructs(scratch.arena, GPU_D12_CompiledShaderResult, num_shaders);
-                sig.descs = shader_descs;
-                sig.results = shader_results;
-                sig.descs[0].src = data;
-                sig.descs[0].friendly_name = friendly_name;
-                sig.descs[0].entry = Lit("cs");
-                sig.descs[0].target = Lit("cs_6_6");
+                job_desc->sig->descs = shader_descs;
+                job_desc->sig->results = shader_results;
+                job_desc->sig->descs[0].src = data;
+                job_desc->sig->descs[0].friendly_name = friendly_name;
+                job_desc->sig->descs[0].entry = Lit("cs");
+                job_desc->sig->descs[0].target = Lit("cs_6_6");
             }
             {
                 Counter counter = ZI;
-                P_Run(num_shaders, GPU_D12_CompileShaderJob, &sig, PoolKind_Inherit, PriorityKind_Inherit, &counter);
-                P_WaitOnCounter(&counter);
+                job_desc->count = num_shaders;
+                job_desc->counter = &counter;
+                RunJobEx((GenericJobDesc *)job_desc);
+                WaitOnCounter(&counter);
             }
         }
         P_CloseFIle(file);
@@ -1177,15 +1168,12 @@ W_CallbackFuncDef(GPU_D12_WatchPipelineCallback, name)
             __profn("Compile dirty pipelines");
             GPU_D12_Pipeline **pipelines = PushStructs(scratch.arena, GPU_D12_Pipeline *, num_pipelines);
             {
-                GPU_D12_AllocPipelineJobSig sig = ZI;
-                sig.descs_in = pipeline_descs;
-                sig.pipelines_out = pipelines;
                 Counter counter = ZI;
-                P_Run(num_pipelines, GPU_D12_AllocPipelineJob, &sig, PoolKind_Inherit, PriorityKind_Inherit, &counter);
-                P_WaitOnCounter(&counter);
+                RunJob(num_pipelines, GPU_D12_AllocPipelineJob, PoolKind_Inherit, PriorityKind_Low, &counter, .descs_in = pipeline_descs, .pipelines_out = pipelines);
+                WaitOnCounter(&counter);
             }
             {
-                P_Lock lock = P_LockS(&g->pipelines_mutex);
+                Lock lock = LockS(&g->pipelines_mutex);
                 for (u32 i = 0; i < num_pipelines; ++i)
                 {
                     GPU_D12_Pipeline *pipeline = pipelines[i];
@@ -1216,7 +1204,7 @@ W_CallbackFuncDef(GPU_D12_WatchPipelineCallback, name)
 
                     }
                 }
-                P_Unlock(&lock);
+                Unlock(&lock);
             }
             GPU_D12_RegisterPipeline(num_pipelines, pipelines);
         }
@@ -1237,7 +1225,7 @@ GPU_D12_Descriptor *GPU_D12_AllocDescriptor(GPU_D12_CpuDescriptorHeap *dh)
     u32 index = 0;
     D3D12_CPU_DESCRIPTOR_HANDLE handle = ZI;
     {
-        P_Lock lock = P_LockE(&dh->mutex);
+        Lock lock = LockE(&dh->mutex);
         if (dh->first_free_descriptor)
         {
             d = dh->first_free_descriptor;
@@ -1255,7 +1243,7 @@ GPU_D12_Descriptor *GPU_D12_AllocDescriptor(GPU_D12_CpuDescriptorHeap *dh)
             index = dh->num_descriptors_reserved++;
             handle.ptr = dh->handle.ptr + (index * dh->descriptor_size);
         }
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
     ZeroStruct(d);
     d->heap = dh;
@@ -1267,12 +1255,12 @@ GPU_D12_Descriptor *GPU_D12_AllocDescriptor(GPU_D12_CpuDescriptorHeap *dh)
 void GPU_D12_ReleaseDescriptor(GPU_D12_Descriptor *descriptor)
 {
     GPU_D12_CpuDescriptorHeap *dh = descriptor->heap;
-    P_Lock lock = P_LockE(&dh->mutex);
+    Lock lock = LockE(&dh->mutex);
     {
         descriptor->next_free = dh->first_free_descriptor;
         dh->first_free_descriptor = descriptor;
     }
-    P_Unlock(&lock);
+    Unlock(&lock);
 }
 
 /* ========================== *
@@ -1342,31 +1330,31 @@ void GPU_D12_ReleaseDataFenced(void *data, GPU_D12_FencedReleaseKind kind)
     for (u32 i = 0; i < countof(g->command_queues); ++i)
     {
         GPU_D12_CommandQueue *cq = g->command_queues[i];
-        P_Lock lock = P_LockS(&cq->submit_fence_mutex);
+        Lock lock = LockS(&cq->submit_fence_mutex);
         {
             fr_targets[i] = cq->submit_fence_target;
         }
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
 
-    /* PushStruct data to release queue */
+    /* Push data to release queue */
     {
-        P_Lock lock = P_LockE(&g->fenced_releases_mutex);
+        Lock lock = LockE(&g->fenced_releases_mutex);
         {
             *PushStruct(g->fenced_releases_arena, GPU_D12_FencedReleaseData) = fr;
             CopyBytes(g->fenced_release_targets, fr_targets, sizeof(fr_targets));
         }
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
 
     /* Wake evictor */
     {
-        P_Lock lock = P_LockE(&g->evictor_wake_mutex);
+        Lock lock = LockE(&g->evictor_wake_mutex);
         {
             ++g->evictor_wake_gen;
-            P_SignalCv(&g->evictor_wake_cv, I32Max);
+            SignalCv(&g->evictor_wake_cv, I32Max);
         }
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
 }
 
@@ -1380,7 +1368,7 @@ GPU_D12_Resource *GPU_D12_AllocResource(D3D12_HEAP_PROPERTIES heap_props, D3D12_
     GPU_D12_SharedState *g = &GPU_D12_shared_state;
     GPU_D12_Resource *r = 0;
     {
-        P_Lock lock = P_LockE(&g->resources_mutex);
+        Lock lock = LockE(&g->resources_mutex);
         if (g->first_free_resource)
         {
             r = g->first_free_resource;
@@ -1390,7 +1378,7 @@ GPU_D12_Resource *GPU_D12_AllocResource(D3D12_HEAP_PROPERTIES heap_props, D3D12_
         {
             r = PushStructNoZero(g->resources_arena, GPU_D12_Resource);
         }
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
     ZeroStruct(r);
 
@@ -1441,10 +1429,10 @@ void GPU_D12_ReleaseResourceNow(GPU_D12_Resource *t)
     ID3D12Resource_Release(t->resource);
 
     /* Add to free list */
-    P_Lock lock = P_LockE(&g->resources_mutex);
+    Lock lock = LockE(&g->resources_mutex);
     t->next_free = g->first_free_resource;
     g->first_free_resource = t;
-    P_Unlock(&lock);
+    Unlock(&lock);
 }
 
 void GPU_ReleaseResourceFenced(GPU_Resource *resource)
@@ -1515,12 +1503,11 @@ void GPU_D12_InsertBarrier(ID3D12GraphicsCommandList *cl, i32 num_descs, GPU_D12
 
 GPU_D12_CommandListPool *GPU_D12_AllocCommandListPool(GPU_D12_CommandQueue *cq);
 
-JobDef(GPU_D12_AllocCommandQueueJob, job)
+JobDef(GPU_D12_AllocCommandQueueJob, sig, id)
 {
     __prof;
     GPU_D12_SharedState *g = &GPU_D12_shared_state;
-    GPU_D12_AllocCommandQueueJobSig *sig = job.sig;
-    GPU_D12_CommandQueueDesc *desc = &sig->descs_in[job.id];
+    GPU_D12_CommandQueueDesc *desc = &sig->descs_in[id];
     {
         GPU_D12_CommandQueue *cq = 0;
         {
@@ -1548,7 +1535,7 @@ JobDef(GPU_D12_AllocCommandQueueJob, job)
 
         cq->cl_pool = GPU_D12_AllocCommandListPool(cq);
 
-        sig->cqs_out[job.id] = cq;
+        sig->cqs_out[id] = cq;
     }
 }
 
@@ -1587,7 +1574,7 @@ GPU_D12_CommandList *GPU_D12_BeginCommandList(GPU_D12_CommandListPool *pool)
     struct ID3D12GraphicsCommandList *old_cl = 0;
     struct ID3D12CommandAllocator *old_ca = 0;
     {
-        P_Lock lock = P_LockE(&pool->mutex);
+        Lock lock = LockE(&pool->mutex);
         /* Find first command list ready for reuse */
         for (GPU_D12_CommandList *tmp = pool->first_submitted_command_list; tmp; tmp = tmp->next_submitted)
         {
@@ -1625,12 +1612,12 @@ GPU_D12_CommandList *GPU_D12_BeginCommandList(GPU_D12_CommandListPool *pool)
         {
             cl = PushStructNoZero(pool->arena, GPU_D12_CommandList);
         }
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
     ZeroStruct(cl);
     cl->cq = cq;
     cl->pool = pool;
-    cl->global_record_lock = P_LockS(&g->global_command_list_record_mutex);
+    cl->global_record_lock = LockS(&g->global_command_list_record_mutex);
 
     HRESULT hr = 0;
     if (old_cl)
@@ -1698,20 +1685,20 @@ u64 GPU_D12_EndCommandList(GPU_D12_CommandList *cl)
     u64 submit_fence_target = 0;
     {
         __profn("Execute");
-        P_Lock submit_lock = P_LockS(&g->global_submit_mutex);
-        P_Lock fence_lock = P_LockE(&cq->submit_fence_mutex);
+        Lock submit_lock = LockS(&g->global_submit_mutex);
+        Lock fence_lock = LockE(&cq->submit_fence_mutex);
         {
             submit_fence_target = ++cq->submit_fence_target;
             ID3D12CommandQueue_ExecuteCommandLists(cq->cq, 1, (ID3D12CommandList **)&cl->cl);
             ID3D12CommandQueue_Signal(cq->cq, cq->submit_fence, submit_fence_target);
         }
-        P_Unlock(&fence_lock);
-        P_Unlock(&submit_lock);
+        Unlock(&fence_lock);
+        Unlock(&submit_lock);
     }
 
     /* Add descriptor heaps to submitted list */
     {
-        P_Lock lock = P_LockE(&g->command_descriptor_heaps_mutex);
+        Lock lock = LockE(&g->command_descriptor_heaps_mutex);
         for (GPU_D12_CommandDescriptorHeap *cdh = cl->first_command_descriptor_heap; cdh; cdh = cdh->next_in_command_list)
         {
             cdh->submitted_cq = cq;
@@ -1726,12 +1713,12 @@ u64 GPU_D12_EndCommandList(GPU_D12_CommandList *cl)
             }
             g->last_submitted_command_descriptor_heap = cdh;
         }
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
 
     /* Add command buffers to submitted list */
     {
-        P_Lock lock = P_LockE(&g->command_buffers_mutex);
+        Lock lock = LockE(&g->command_buffers_mutex);
         for (GPU_D12_CommandBuffer *cb = cl->first_command_buffer; cb; cb = cb->next_in_command_list)
         {
             GPU_D12_CommandBufferGroup *group = cb->group;
@@ -1747,14 +1734,14 @@ u64 GPU_D12_EndCommandList(GPU_D12_CommandList *cl)
             }
             group->last_submitted = cb;
         }
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
 
     /* Add command list to pool submitted list */
-    P_Unlock(&cl->global_record_lock);
+    Unlock(&cl->global_record_lock);
     cl->submitted_fence_target = submit_fence_target;
     {
-        P_Lock lock = P_LockE(&pool->mutex);
+        Lock lock = LockE(&pool->mutex);
         if (pool->last_submitted_command_list)
         {
             pool->last_submitted_command_list->next_submitted = cl;
@@ -1764,7 +1751,7 @@ u64 GPU_D12_EndCommandList(GPU_D12_CommandList *cl)
             pool->first_submitted_command_list = cl;
         }
         pool->last_submitted_command_list = cl;
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
 
     return submit_fence_target;
@@ -1786,7 +1773,7 @@ GPU_D12_CommandDescriptorHeap *GPU_D12_PushDescriptorHeap(GPU_D12_CommandList *c
     D3D12_CPU_DESCRIPTOR_HANDLE old_start_cpu_handle = ZI;
     D3D12_GPU_DESCRIPTOR_HANDLE old_start_gpu_handle = ZI;
     {
-        P_Lock lock = P_LockE(&g->command_descriptor_heaps_mutex);
+        Lock lock = LockE(&g->command_descriptor_heaps_mutex);
         /* Find first heap ready for reuse */
         for (GPU_D12_CommandDescriptorHeap *tmp = g->first_submitted_command_descriptor_heap; tmp; tmp = tmp->next_submitted)
         {
@@ -1828,7 +1815,7 @@ GPU_D12_CommandDescriptorHeap *GPU_D12_PushDescriptorHeap(GPU_D12_CommandList *c
             /* No available heap available for reuse, allocate new */
             cdh = PushStructNoZero(g->command_descriptor_heaps_arena, GPU_D12_CommandDescriptorHeap);
         }
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
     ZeroStruct(cdh);
 
@@ -1855,9 +1842,9 @@ GPU_D12_CommandDescriptorHeap *GPU_D12_PushDescriptorHeap(GPU_D12_CommandList *c
 
     /* Copy CPU heap */
     {
-        P_Lock lock = P_LockS(&dh_cpu->mutex);
+        Lock lock = LockS(&dh_cpu->mutex);
         ID3D12Device_CopyDescriptorsSimple(g->device, dh_cpu->num_descriptors_reserved, cdh->start_cpu_handle, dh_cpu->handle, D3D12_DESCRIPTOR_HEAP_TYPE_CBV_SRV_UAV);
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
 
     /* Insert into command list */
@@ -1894,7 +1881,7 @@ GPU_D12_CommandBuffer *GPU_D12__PushCommandBuffer(GPU_D12_CommandList *cl, u64 d
     GPU_D12_CommandBuffer *cb = 0;
     GPU_D12_Resource *r = 0;
     {
-        P_Lock lock = P_LockE(&g->command_buffers_mutex);
+        Lock lock = LockE(&g->command_buffers_mutex);
 
         {
             u64 group_hash = GPU_D12_CommandBufferHashFromSize(size);
@@ -1946,7 +1933,7 @@ GPU_D12_CommandBuffer *GPU_D12__PushCommandBuffer(GPU_D12_CommandList *cl, u64 d
             /* Allocate new */
             cb = PushStructNoZero(g->command_buffers_arena, GPU_D12_CommandBuffer);
         }
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
     ZeroStruct(cb);
     cb->group = cb_group;
@@ -2017,10 +2004,9 @@ GPU_D12_CommandBuffer *GPU_D12__PushCommandBuffer(GPU_D12_CommandList *cl, u64 d
  * Wait job
  * ========================== */
 
-JobDef(GPU_D12_WaitOnFenceJob, job)
+JobDef(GPU_D12_WaitOnFenceJob, sig, UNUSED id)
 {
     __prof;
-    GPU_D12_WaitOnFenceJobSig *sig = job.sig;
     ID3D12Fence *fence = sig->fence;
     u64 target = sig->target;
     if (ID3D12Fence_GetCompletedValue(fence) < target)
@@ -2100,11 +2086,8 @@ GPU_Resource *GPU_AllocTexture(GPU_TextureFormat format, u32 flags, Vec2I32 size
     {
         /* TODO: Make wait optional */
         Counter counter = ZI;
-        GPU_D12_UploadJobSig sig = ZI;
-        sig.resource = r;
-        sig.data = initial_data;
-        P_Run(1, GPU_D12_UploadJob, &sig, PoolKind_Inherit, PriorityKind_Inherit, &counter);
-        P_WaitOnCounter(&counter);
+        RunJob(1, GPU_D12_UploadJob, PoolKind_Inherit, PriorityKind_Inherit, &counter, .resource = r, .data = initial_data);
+        WaitOnCounter(&counter);
     }
 
     return (GPU_Resource *)r;
@@ -2120,10 +2103,9 @@ Vec2I32 GPU_GetTextureSize(GPU_Resource *resource)
  * Upload
  * ========================== */
 
-JobDef(GPU_D12_UploadJob, job)
+JobDef(GPU_D12_UploadJob, sig, UNUSED id)
 {
     GPU_D12_SharedState *g = &GPU_D12_shared_state;
-    GPU_D12_UploadJobSig *sig = job.sig;
     GPU_D12_Resource *r = sig->resource;
     void *data = sig->data;
 
@@ -2216,12 +2198,9 @@ JobDef(GPU_D12_UploadJob, job)
         /* Wait on fence so we know it's safe to release upload heap */
         if (ID3D12Fence_GetCompletedValue(cq->submit_fence) < fence_target)
         {
-            GPU_D12_WaitOnFenceJobSig wait_sig = ZI;
-            wait_sig.fence = cq->submit_fence;
-            wait_sig.target = fence_target;
             Counter counter = ZI;
-            P_Run(1, GPU_D12_WaitOnFenceJob, &wait_sig, PoolKind_Floating, PriorityKind_Low, &counter);
-            P_WaitOnCounter(&counter);
+            RunJob(1, GPU_D12_WaitOnFenceJob, PoolKind_Floating, PriorityKind_Inherit, &counter, .fence = cq->submit_fence, .target = fence_target);
+            WaitOnCounter(&counter);
         }
 
         /* Release upload heap now */
@@ -3040,7 +3019,7 @@ GPU_Swapchain *GPU_AllocSwapchain(P_Window *window, Vec2I32 resolution)
 
     GPU_D12_Swapchain *swapchain = 0;
     {
-        P_Lock lock = P_LockE(&g->swapchains_mutex);
+        Lock lock = LockE(&g->swapchains_mutex);
         if (g->first_free_swapchain)
         {
             swapchain = g->first_free_swapchain;
@@ -3050,7 +3029,7 @@ GPU_Swapchain *GPU_AllocSwapchain(P_Window *window, Vec2I32 resolution)
         {
             swapchain = PushStruct(g->swapchains_arena, GPU_D12_Swapchain);
         }
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
 
     /* Create swapchain1 */
@@ -3132,9 +3111,9 @@ GPU_D12_SwapchainBuffer *GPU_D12_UpdateSwapchain(GPU_D12_Swapchain *swapchain, V
         GPU_D12_CommandQueue *cq = g->command_queues[DX12_QUEUE_DIRECT];
         /* Lock direct queue submissions (in case any write to backbuffer) */
         /* TODO: Less overkill approach - Only flush GPU_D12_BlitToSwapchain since we know it's the only operation targeting backbuffer */
-        P_Lock lock = P_LockE(&cq->submit_fence_mutex);
+        Lock lock = LockE(&cq->submit_fence_mutex);
         //DEBUGBREAKABLE;
-        //P_Lock lock = P_LockE(&g->global_command_list_record_mutex);
+        //Lock lock = LockE(&g->global_command_list_record_mutex);
         {
             /* Flush direct queue */
             //ID3D12CommandQueue_Signal(cq->cq, cq->submit_fence, ++cq->submit_fence_target);
@@ -3161,7 +3140,7 @@ GPU_D12_SwapchainBuffer *GPU_D12_UpdateSwapchain(GPU_D12_Swapchain *swapchain, V
                 P_Panic(Lit("Failed to resize swapchain"));
             }
         }
-        P_Unlock(&lock);
+        Unlock(&lock);
 
         GPU_D12_InitSwapchainResources(swapchain);
 
@@ -3309,7 +3288,7 @@ void GPU_PresentSwapchain(GPU_Swapchain *gp_swapchain, Vec2I32 backbuffer_resolu
 
         __profn("Mark queue frames");
         /* Lock because frame marks shouldn't occur while command lists are recording */
-        P_Lock lock = P_LockE(&g->global_command_list_record_mutex);
+        Lock lock = LockE(&g->global_command_list_record_mutex);
         for (u32 i = 0; i < countof(g->command_queues); ++i)
         {
             {
@@ -3317,7 +3296,7 @@ void GPU_PresentSwapchain(GPU_Swapchain *gp_swapchain, Vec2I32 backbuffer_resolu
                 __prof_dx12_new_frame(cq->prof);
             }
         }
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
     {
         __profn("Collect queues");
@@ -3334,7 +3313,7 @@ void GPU_PresentSwapchain(GPU_Swapchain *gp_swapchain, Vec2I32 backbuffer_resolu
  * Evictor job
  * ========================== */
 
-JobDef(GPU_D12_EvictorJob, _)
+JobDef(GPU_D12_EvictorJob, UNUSED sig, UNUSED id)
 {
     GPU_D12_SharedState *g = &GPU_D12_shared_state;
     u64 completed_targets[DX12_NUM_QUEUES] = ZI;
@@ -3352,13 +3331,13 @@ JobDef(GPU_D12_EvictorJob, _)
             GPU_D12_FencedReleaseData *fenced_releases = 0;
             {
                 __profn("Copy queued releases");
-                P_Lock lock = P_LockE(&g->fenced_releases_mutex);
+                Lock lock = LockE(&g->fenced_releases_mutex);
                 num_fenced_releases = g->fenced_releases_arena->pos / sizeof(GPU_D12_FencedReleaseData);
                 fenced_releases = PushStructsNoZero(scratch.arena, GPU_D12_FencedReleaseData, num_fenced_releases);
                 CopyBytes(fenced_releases, ArenaBase(g->fenced_releases_arena), g->fenced_releases_arena->pos);
                 ResetArena(g->fenced_releases_arena);
                 CopyBytes(targets, g->fenced_release_targets, sizeof(targets));
-                P_Unlock(&lock);
+                Unlock(&lock);
             }
 
             /* Wait until fences reach target */
@@ -3374,14 +3353,9 @@ JobDef(GPU_D12_EvictorJob, _)
                         {
                             __profn("Wait on fence");
                             {
-                                GPU_D12_WaitOnFenceJobSig sig = ZI;
-                                sig.fence = cq->submit_fence;
-                                sig.target = targets[i];
-                                {
-                                    Counter counter = ZI;
-                                    P_Run(1, GPU_D12_WaitOnFenceJob, &sig, PoolKind_Floating, PriorityKind_Low, &counter);
-                                    P_WaitOnCounter(&counter);
-                                }
+                                Counter counter = ZI;
+                                RunJob(1, GPU_D12_WaitOnFenceJob, PoolKind_Floating, PriorityKind_Inherit, &counter, .fence = cq->submit_fence, .target = targets[i]);
+                                WaitOnCounter(&counter);
                             }
                         }
                     }
@@ -3415,15 +3389,15 @@ JobDef(GPU_D12_EvictorJob, _)
             }
             EndScratch(scratch);
         }
-        P_Lock lock = P_LockE(&g->evictor_wake_mutex);
+        Lock lock = LockE(&g->evictor_wake_mutex);
         {
             while (!g->evictor_shutdown && g->evictor_wake_gen == 0)
             {
-                P_WaitOnCv(&g->evictor_wake_cv, &lock);
+                WaitOnCv(&g->evictor_wake_cv, &lock);
             }
             shutdown = g->evictor_shutdown;
             g->evictor_wake_gen = 0;
         }
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
 }
diff --git a/src/gpu/gpu_dx12.h b/src/gpu/gpu_dx12.h
index e5002003..100478ea 100644
--- a/src/gpu/gpu_dx12.h
+++ b/src/gpu/gpu_dx12.h
@@ -133,7 +133,7 @@ Struct(GPU_D12_CommandQueue)
     ID3D12CommandQueue *cq;
     Arena *arena;
 
-    P_Mutex submit_fence_mutex;
+    Mutex submit_fence_mutex;
     u64 submit_fence_target;
     ID3D12Fence *submit_fence;
 
@@ -148,7 +148,7 @@ Struct(GPU_D12_CommandListPool)
 {
     GPU_D12_CommandQueue *cq;
     Arena *arena;
-    P_Mutex mutex;
+    Mutex mutex;
     struct GPU_D12_CommandList *first_submitted_command_list;
     struct GPU_D12_CommandList *last_submitted_command_list;
 };
@@ -197,7 +197,7 @@ Struct(GPU_D12_CommandList)
     GPU_D12_CommandListPool *pool;
     struct ID3D12CommandAllocator *ca;
     struct ID3D12GraphicsCommandList *cl;
-    P_Lock global_record_lock;
+    Lock global_record_lock;
 
     GPU_D12_Pipeline *cur_pipeline;
 
@@ -232,7 +232,7 @@ Struct(GPU_D12_CpuDescriptorHeap)
 {
     enum D3D12_DESCRIPTOR_HEAP_TYPE type;
     Arena *arena;
-    P_Mutex mutex;
+    Mutex mutex;
 
     u32 descriptor_size;
     u32 num_descriptors_reserved;
@@ -352,23 +352,23 @@ Struct(GPU_D12_SharedState)
     Atomic32 initialized;
 
     /* Descriptor heaps pool */
-    P_Mutex command_descriptor_heaps_mutex;
+    Mutex command_descriptor_heaps_mutex;
     Arena *command_descriptor_heaps_arena;
     GPU_D12_CommandDescriptorHeap *first_submitted_command_descriptor_heap;
     GPU_D12_CommandDescriptorHeap *last_submitted_command_descriptor_heap;
 
     /* Command buffers pool */
-    P_Mutex command_buffers_mutex;
+    Mutex command_buffers_mutex;
     Arena *command_buffers_arena;
     Dict *command_buffers_dict;
 
     /* Resources pool */
-    P_Mutex resources_mutex;
+    Mutex resources_mutex;
     Arena *resources_arena;
     GPU_D12_Resource *first_free_resource;
 
     /* Swapchains pool */
-    P_Mutex swapchains_mutex;
+    Mutex swapchains_mutex;
     Arena *swapchains_arena;
     GPU_D12_Swapchain *first_free_swapchain;
 
@@ -376,7 +376,7 @@ Struct(GPU_D12_SharedState)
     TAR_Archive dxc_archive;
 
     /* Pipeline cache */
-    P_Mutex pipelines_mutex;
+    Mutex pipelines_mutex;
     Arena *pipelines_arena;
     GPU_D12_Pipeline *first_free_pipeline;
     Dict *pipeline_descs;
@@ -385,7 +385,7 @@ Struct(GPU_D12_SharedState)
     GPU_D12_PipelineScope *first_free_pipeline_scope;
 
     /* Fenced release queue */
-    P_Mutex fenced_releases_mutex;
+    Mutex fenced_releases_mutex;
     Arena *fenced_releases_arena;
     u64 fenced_release_targets[DX12_NUM_QUEUES];
 
@@ -407,14 +407,14 @@ Struct(GPU_D12_SharedState)
     GPU_D12_CpuDescriptorHeap *rtv_heap;
 
     /* Command queues */
-    P_Mutex global_command_list_record_mutex;
-    P_Mutex global_submit_mutex;
+    Mutex global_command_list_record_mutex;
+    Mutex global_submit_mutex;
     GPU_D12_CommandQueue *command_queues[DX12_NUM_QUEUES];
 
     /* Evictor job */
     Counter evictor_job_counter;
-    P_Cv evictor_wake_cv;
-    P_Mutex evictor_wake_mutex;
+    Cv evictor_wake_cv;
+    Mutex evictor_wake_mutex;
     i64 evictor_wake_gen;
     b32 evictor_shutdown;
 };
@@ -457,7 +457,7 @@ void GPU_D12_InitNoise(void);
  * Shader compilation
  * ========================== */
 
-JobDecl(GPU_D12_CompileShaderJob, { Arena *arena; GPU_D12_ShaderDesc *decs; GPU_D12_CompiledShaderResult results; });
+JobDecl(GPU_D12_CompileShaderJob, { Arena *arena; GPU_D12_ShaderDesc *descs; GPU_D12_CompiledShaderResult *results; });
 
 /* ========================== *
  * Pipeline
diff --git a/src/json/json_core.c b/src/json/json_core.c
index 81ceebd4..3ae5bfad 100644
--- a/src/json/json_core.c
+++ b/src/json/json_core.c
@@ -873,7 +873,7 @@ void JSON_Parse(Arena *arena, JSON_Parser *p)
 
         if (is_new_parent)
         {
-            /* PushStruct self back to stack to re-check for closing brace later */
+            /* Push self back to stack to re-check for closing brace later */
             *PushStructNoZero(scratch.arena, JSON_Blob *) = json;
             ++stack_count;
 
diff --git a/src/mixer/mixer_core.c b/src/mixer/mixer_core.c
index 074bfe2f..26468130 100644
--- a/src/mixer/mixer_core.c
+++ b/src/mixer/mixer_core.c
@@ -53,10 +53,10 @@ MIX_Track *MIX_TrackFromHandle(MIX_Handle handle)
     }
 }
 
-MIX_Track *MIX_AllocTrackLocked(P_Lock *lock, SND_Sound *sound)
+MIX_Track *MIX_AllocTrackLocked(Lock *lock, SND_Sound *sound)
 {
     MIX_SharedState *g = &M_shared_state;
-    P_AssertLockedE(lock, &g->mutex);
+    AssertLockedE(lock, &g->mutex);
     LAX lock;
 
     MIX_Track *track = 0;
@@ -100,10 +100,10 @@ MIX_Track *MIX_AllocTrackLocked(P_Lock *lock, SND_Sound *sound)
     return track;
 }
 
-void MIX_ReleaseTrackLocked(P_Lock *lock, MIX_Track *track)
+void MIX_ReleaseTrackLocked(Lock *lock, MIX_Track *track)
 {
     MIX_SharedState *g = &M_shared_state;
-    P_AssertLockedE(lock, &g->mutex);
+    AssertLockedE(lock, &g->mutex);
     LAX lock;
 
     /* Remove from playing list */
@@ -155,12 +155,12 @@ MIX_Handle MIX_PlaySoundEx(SND_Sound *sound, MIX_TrackDesc desc)
     MIX_SharedState *g = &M_shared_state;
     MIX_Track *track;
     {
-        P_Lock lock = P_LockE(&g->mutex);
+        Lock lock = LockE(&g->mutex);
         {
             track = MIX_AllocTrackLocked(&lock, sound);
             track->desc = desc;
         }
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
     return MIX_HandleFromTrack(track);
 }
@@ -175,7 +175,7 @@ MIX_TrackDesc MIX_TrackDescFromHandle(MIX_Handle handle)
     if (track)
     {
         /* TODO: Only lock mutex on track itself or something */
-        P_Lock lock = P_LockE(&g->mutex);
+        Lock lock = LockE(&g->mutex);
         {
             /* Confirm handle is still valid now that we're locked */
             track = MIX_TrackFromHandle(handle);
@@ -184,7 +184,7 @@ MIX_TrackDesc MIX_TrackDescFromHandle(MIX_Handle handle)
                 result = track->desc;
             }
         }
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
 
     return result;
@@ -198,7 +198,7 @@ void MIX_UpdateTrack(MIX_Handle handle, MIX_TrackDesc desc)
     if (track)
     {
         /* TODO: Only lock mutex on track itself or something */
-        P_Lock lock = P_LockE(&g->mutex);
+        Lock lock = LockE(&g->mutex);
         {
             /* Confirm handle is still valid now that we're locked */
             track = MIX_TrackFromHandle(handle);
@@ -207,19 +207,19 @@ void MIX_UpdateTrack(MIX_Handle handle, MIX_TrackDesc desc)
                 track->desc = desc;
             }
         }
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
 }
 
 void MIX_UpdateListener(Vec2 pos, Vec2 dir)
 {
     MIX_SharedState *g = &M_shared_state;
-    P_Lock lock = P_LockE(&g->mutex);
+    Lock lock = LockE(&g->mutex);
     {
         g->listener_pos = pos;
         g->listener_dir = NormVec2(dir);
     }
-    P_Unlock(&lock);
+    Unlock(&lock);
 }
 
 ////////////////////////////////
@@ -260,7 +260,7 @@ MIX_PcmF32 MIX_MixAllTracks(Arena *arena, u64 frame_count)
     MIX_MixData **mixes = 0;
     u64 mixes_count = 0;
     {
-        P_Lock lock = P_LockE(&g->mutex);
+        Lock lock = LockE(&g->mutex);
 
         /* Read listener info */
         listener_pos = g->listener_pos;
@@ -276,7 +276,7 @@ MIX_PcmF32 MIX_MixAllTracks(Arena *arena, u64 frame_count)
             mixes[mixes_count++] = mix;
         }
 
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
 
     //- Process mix data
@@ -461,7 +461,7 @@ MIX_PcmF32 MIX_MixAllTracks(Arena *arena, u64 frame_count)
     //- Update track effect data
     {
         __profn("Update track effect data");
-        P_Lock lock = P_LockE(&g->mutex);
+        Lock lock = LockE(&g->mutex);
         for (u64 i = 0; i < mixes_count; ++i)
         {
             MIX_MixData *mix = mixes[i];
@@ -475,7 +475,7 @@ MIX_PcmF32 MIX_MixAllTracks(Arena *arena, u64 frame_count)
                 }
             }
         }
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
 
     EndScratch(scratch);
diff --git a/src/mixer/mixer_core.h b/src/mixer/mixer_core.h
index 98a05823..1f480f8e 100644
--- a/src/mixer/mixer_core.h
+++ b/src/mixer/mixer_core.h
@@ -79,7 +79,7 @@ Struct(MIX_Track){
 
 Struct(MIX_SharedState)
 {
-    P_Mutex mutex;
+    Mutex mutex;
 
     /* Listener */
     Vec2 listener_pos;
@@ -106,8 +106,8 @@ MIX_StartupReceipt MIX_Startup(void);
 
 MIX_Handle MIX_HandleFromTrack(MIX_Track *track);
 MIX_Track *MIX_TrackFromHandle(MIX_Handle handle);
-MIX_Track *MIX_AllocTrackLocked(P_Lock *lock, SND_Sound *sound);
-void MIX_ReleaseTrackLocked(P_Lock *lock, MIX_Track *track);
+MIX_Track *MIX_AllocTrackLocked(Lock *lock, SND_Sound *sound);
+void MIX_ReleaseTrackLocked(Lock *lock, MIX_Track *track);
 
 ////////////////////////////////
 //~ Mixer state operations
diff --git a/src/mp3/mp3_mmf.c b/src/mp3/mp3_mmf.c
index f35dfc91..ca2ddd2a 100644
--- a/src/mp3/mp3_mmf.c
+++ b/src/mp3/mp3_mmf.c
@@ -11,6 +11,7 @@
 # include <mfidl.h>
 # include <mfreadwrite.h>
 # include <Shlwapi.h>
+# include <objidl.h>
 #pragma warning(pop)
 
 #pragma comment(lib, "mfplat")
@@ -121,7 +122,8 @@ MP3_Result MP3_Decode(Arena *arena, String encoded, u32 sample_rate, MP3_DecodeF
 
     IMFSourceReader_Release(reader);
     IMFByteStream_Close(byte_stream);
-    IStream_Release(i_stream);
+    /* FIXME: Enable this */
+    //IStream_Release(i_stream);
     MFShutdown();
 
     return result;
diff --git a/src/net/net_core.c b/src/net/net_core.c
index 7fcb4329..a57d5764 100644
--- a/src/net/net_core.c
+++ b/src/net/net_core.c
@@ -542,7 +542,7 @@ N_EventList N_BeginUpdate(Arena *arena, N_Host *host)
     TempArena scratch = BeginScratch(arena);
 
     N_EventList events = ZI;
-    i64 now_ns = P_TimeNs();
+    i64 now_ns = TimeNs();
     {
         __profn("Read packets");
 
diff --git a/src/net/net_core.h b/src/net/net_core.h
index 6e0273eb..434e73d4 100644
--- a/src/net/net_core.h
+++ b/src/net/net_core.h
@@ -246,7 +246,7 @@ Struct(N_Host)
     u64 num_msg_assembler_lookup_bins;
 
     /* Double buffer for incoming data */
-    P_Mutex rcv_buffer_write_mutex;
+    Mutex rcv_buffer_write_mutex;
     N_RcvBuffer *rcv_buffer_read;
     N_RcvBuffer *rcv_buffer_write;
 
diff --git a/src/platform/platform.c b/src/platform/platform.c
index cad458c5..096ed58d 100644
--- a/src/platform/platform.c
+++ b/src/platform/platform.c
@@ -1,6 +1,5 @@
 #include "platform.h"
 
-#include "platform_snc.c"
 #include "platform_log.c"
 
 #if PlatformIsWindows
diff --git a/src/platform/platform.h b/src/platform/platform.h
index 053f2d51..57722ec8 100644
--- a/src/platform/platform.h
+++ b/src/platform/platform.h
@@ -3,7 +3,6 @@
 
 #include "../base/base.h"
 
-#include "platform_snc.h"
 #include "platform_core.h"
 #include "platform_log.h"
 
diff --git a/src/platform/platform_core.h b/src/platform/platform_core.h
index 60efb84c..1b7c8a01 100644
--- a/src/platform/platform_core.h
+++ b/src/platform/platform_core.h
@@ -311,7 +311,6 @@ typedef P_ExitFuncDef(P_ExitFunc);
 //~ @hookdecl Time helper operations
 
 P_DateTime P_LocalTime(void);
-i64 P_TimeNs(void);
 
 ////////////////////////////////
 //~ @hookdecl File system operations
@@ -403,9 +402,6 @@ void P_WriteSock(P_Sock *sock, P_Address address, String data);
 void P_MessageBox(P_MessageBoxKind kind, String message);
 void P_SetClipboardText(String str);
 String P_GetClipboardText(Arena *arena);
-u32 P_GetLogicalProcessorCount(void);
-u32 P_GetThreadId(void);
-i64 P_GetCurrentSchedulerPeriodNs(void);
 
 ////////////////////////////////
 //~ @hookdecl Sleep
diff --git a/src/platform/platform_log.c b/src/platform/platform_log.c
index d6ed1cde..fe349b70 100644
--- a/src/platform/platform_log.c
+++ b/src/platform/platform_log.c
@@ -32,7 +32,7 @@ void P_RegisterLogCallback(P_LogEventCallbackFunc *func, i32 level)
 {
     P_SharedLogState *ctx = &P_shared_log_state;
     if (!Atomic32Fetch(&ctx->initialized)) { return; }
-    P_Lock lock = P_LockE(&ctx->callbacks_mutex);
+    Lock lock = LockE(&ctx->callbacks_mutex);
     {
         LogEventCallback *callback = PushStruct(ctx->callbacks_arena, LogEventCallback);
         callback->func = func;
@@ -47,7 +47,7 @@ void P_RegisterLogCallback(P_LogEventCallbackFunc *func, i32 level)
         }
         ctx->last_callback = callback;
     }
-    P_Unlock(&lock);
+    Unlock(&lock);
 }
 
 ////////////////////////////////
@@ -149,12 +149,12 @@ void P_Log_(i32 level, String msg)
     }
 
 
-    u32 tid = P_GetThreadId();
+    u32 tid = ThreadId();
 
 
     //- Format message
     P_DateTime datetime = P_LocalTime();
-    i64 time_ns = P_TimeNs();
+    i64 time_ns = TimeNs();
     String shorthand = settings.shorthand;
 
 #if P_IncludeLogSourceLocation
@@ -217,7 +217,7 @@ void P_Log_(i32 level, String msg)
     event.line = line;
 #endif
     {
-        P_Lock lock = P_LockS(&ctx->callbacks_mutex);
+        Lock lock = LockS(&ctx->callbacks_mutex);
         for (LogEventCallback *callback = ctx->first_callback; callback; callback = callback->next)
         {
             if (level <= callback->level)
@@ -226,7 +226,7 @@ void P_Log_(i32 level, String msg)
                 callback->func(event);
             }
         }
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
 
     EndScratch(scratch);
diff --git a/src/platform/platform_log.h b/src/platform/platform_log.h
index 3200acf9..51fabc32 100644
--- a/src/platform/platform_log.h
+++ b/src/platform/platform_log.h
@@ -63,7 +63,7 @@ Struct(P_SharedLogState)
 {
     Atomic32 initialized;
 
-    P_Mutex callbacks_mutex;
+    Mutex callbacks_mutex;
     Arena *callbacks_arena;
     LogEventCallback *first_callback;
     LogEventCallback *last_callback;
diff --git a/src/platform/platform_snc.h b/src/platform/platform_snc.h
deleted file mode 100644
index 49bd1c3a..00000000
--- a/src/platform/platform_snc.h
+++ /dev/null
@@ -1,83 +0,0 @@
-////////////////////////////////
-//~ Mutex types
-
-#define P_DefaultMutexSpin 4000
-
-AlignedStruct(P_Mutex, 64)
-{
-    /* Bit 31   = Exclusive lock is held
-     * Bit 30   = Exclusive lock is pending
-     * Bit 0-30 = Shared locks count
-     */
-    Atomic32 v;
-
-#if RtcIsEnabled
-    Atomic32 exclusive_fiber_id;
-    u8 _pad[56];
-#else
-    u8 _pad[60];
-#endif
-};
-StaticAssert(sizeof(P_Mutex) == 64);  /* Padding validation */
-StaticAssert(alignof(P_Mutex) == 64);  /* Prevent false sharing */
-
-Struct(P_Lock)
-{
-    P_Mutex *mutex;
-    b32 exclusive;
-};
-
-////////////////////////////////
-//~ Condition variable types
-
-AlignedStruct(P_Cv, 64)
-{
-    Atomic64 wake_gen;
-    u8 _pad[56];
-};
-StaticAssert(sizeof(P_Cv) == 64);  /* Padding validation */
-StaticAssert(alignof(P_Cv) == 64);  /* Prevent false sharing */
-
-////////////////////////////////
-//~ Counter types
-
-AlignedStruct(Counter, 64)
-{
-    Atomic64 v;
-    u8 _pad[56];
-};
-StaticAssert(sizeof(Counter) == 64);  /* Padding validation */
-StaticAssert(alignof(Counter) == 64);  /* Prevent false sharing */
-
-////////////////////////////////
-//~ Mutex operations
-
-P_Lock P_LockSpinE(P_Mutex *m, i32 spin);
-P_Lock P_LockSpinS(P_Mutex *m, i32 spin);
-
-P_Lock P_LockE(P_Mutex *m);
-P_Lock P_LockS(P_Mutex *m);
-
-void P_Unlock(P_Lock *lock);
-
-//- Lock assertion
-#if RtcIsEnabled
-# define P_AssertLockedE(l, m) Assert((l)->mutex == (m) && (l)->exclusive == 1)
-# define P_AssertLockedES(l, m) Assert((l)->mutex == (m))
-#else
-# define P_AssertLockedE(l, m) LAX l
-# define P_AssertLockedES(l, m) LAX l
-#endif
-
-////////////////////////////////
-//~ Condition variable operations
-
-void P_WaitOnCv(P_Cv *cv, P_Lock *lock);
-void P_WaitOnCvTime(P_Cv *cv, P_Lock *l, i64 timeout_ns);
-void P_SignalCv(P_Cv *cv, i32 count);
-
-////////////////////////////////
-//~ Counter operations
-
-void P_CounterAdd(Counter *counter, i64 x);
-void P_WaitOnCounter(Counter *counter);
diff --git a/src/platform/platform_win32.c b/src/platform/platform_win32.c
index 75b17eec..f5625120 100644
--- a/src/platform/platform_win32.c
+++ b/src/platform/platform_win32.c
@@ -72,7 +72,7 @@ P_W32_Window *P_W32_AllocWindow(void)
     P_W32_SharedCtx *g = &P_W32_shared_ctx;
     P_W32_Window *window = 0;
     {
-        P_Lock lock = P_LockE(&g->windows_mutex);
+        Lock lock = LockE(&g->windows_mutex);
         if (g->first_free_window)
         {
             window = g->first_free_window;
@@ -82,7 +82,7 @@ P_W32_Window *P_W32_AllocWindow(void)
         {
             window = PushStructNoZero(g->windows_arena, P_W32_Window);
         }
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
     ZeroStruct(window);
 
@@ -93,9 +93,9 @@ P_W32_Window *P_W32_AllocWindow(void)
     /* NOTE: This thread must finish building for the window to actually be
      * created and receive a HWND, because on Windows a the event proc must run on
      * the same thread that created the window. */
-    P_CounterAdd(&window->ready_fence, 1);
-    window->window_thread = P_W32_AllocThread(&W32_WindowThreadEntryFunc, window, Lit("Window thread"), PROF_THREAD_GROUP_WINDOW);
-    P_WaitOnCounter(&window->ready_fence);
+    AddCounter(&window->ready_fence, 1);
+    window->window_thread = W32_AllocThread(&P_W32_WindowThreadEntryFunc, window, Lit("Window thread"), PROF_THREAD_GROUP_WINDOW);
+    WaitOnCounter(&window->ready_fence);
 
     return window;
 }
@@ -106,14 +106,14 @@ void P_W32_ReleaseWindow(P_W32_Window *window)
     Atomic32FetchSet(&window->shutdown, 1);
     P_W32_SharedCtx *g = &P_W32_shared_ctx;
     P_W32_WakeWindow(window);
-    P_W32_WaitReleaseThread(window->window_thread);
+    W32_WaitReleaseThread(window->window_thread);
 
-    P_Lock lock = P_LockE(&g->windows_mutex);
+    Lock lock = LockE(&g->windows_mutex);
     {
         window->next_free = g->first_free_window;
         g->first_free_window = window;
     }
-    P_Unlock(&lock);
+    Unlock(&lock);
 }
 
 HWND P_W32_InitWindow(P_W32_Window *window)
@@ -302,7 +302,7 @@ void P_W32_UpdateWindowFromSettings(P_W32_Window *window, P_WindowSettings *sett
 ////////////////////////////////
 //~ Win32 window thread
 
-W32_ThreadDef(W32_WindowThreadEntryFunc, arg)
+W32_ThreadDef(P_W32_WindowThreadEntryFunc, arg)
 {
     P_W32_Window *window = (P_W32_Window *)arg;
 
@@ -310,7 +310,7 @@ W32_ThreadDef(W32_WindowThreadEntryFunc, arg)
     window->hwnd = P_W32_InitWindow(window);
     P_W32_UpdateWindowFromSystem(window);
     BringWindowToTop(window->hwnd);
-    P_CounterAdd(&window->ready_fence, -1);
+    AddCounter(&window->ready_fence, -1);
 
     while (!Atomic32Fetch(&window->shutdown))
     {
@@ -335,11 +335,11 @@ W32_ThreadDef(W32_WindowThreadEntryFunc, arg)
 void P_W32_ProcessWindowEvent(P_W32_Window *window, P_WindowEvent event)
 {
     __prof;
-    P_Lock lock = P_LockE(&window->event_arena_swp_mutex);
+    Lock lock = LockE(&window->event_arena_swp_mutex);
     {
         *PushStruct(window->event_arenas[window->current_event_arena_index], P_WindowEvent) = event;
     }
-    P_Unlock(&lock);
+    Unlock(&lock);
 }
 
 void P_W32_WakeWindow(P_W32_Window *window)
@@ -785,15 +785,6 @@ P_DateTime P_LocalTime(void)
     return P_W32_DateTimeFromWin32SystemTime(lt);
 }
 
-i64 P_TimeNs(void)
-{
-    struct P_W32_SharedCtx *g = &P_W32_shared_ctx;
-    LARGE_INTEGER qpc;
-    QueryPerformanceCounter(&qpc);
-    i64 result = (qpc.QuadPart - g->timer_start_qpc) * g->ns_per_qpc;
-    return result;
-}
-
 ////////////////////////////////
 //~ @hookdef File system hooks
 
@@ -1166,7 +1157,7 @@ P_Watch *P_AllocWatch(String dir_path)
 
     P_W32_Watch *w32_watch = 0;
     {
-        P_Lock lock = P_LockE(&g->watches_mutex);
+        Lock lock = LockE(&g->watches_mutex);
         {
             if (g->watches_first_free)
             {
@@ -1178,7 +1169,7 @@ P_Watch *P_AllocWatch(String dir_path)
                 w32_watch = PushStructNoZero(g->watches_arena, P_W32_Watch);
             }
         }
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
     ZeroStruct(w32_watch);
 
@@ -1206,12 +1197,12 @@ void P_ReleaseWatch(P_Watch *dw)
     CloseHandle(w32_watch->dir_handle);
     CloseHandle(w32_watch->wake_handle);
 
-    P_Lock lock = P_LockE(&g->watches_mutex);
+    Lock lock = LockE(&g->watches_mutex);
     {
         w32_watch->next_free = g->watches_first_free;
         g->watches_first_free = w32_watch;
     }
-    P_Unlock(&lock);
+    Unlock(&lock);
 }
 
 P_WatchInfoList P_ReadWatchWait(Arena *arena, P_Watch *dw)
@@ -1369,10 +1360,10 @@ P_WindowEventArray P_PopWindowEvents(Arena *arena, P_Window *p_window)
     i32 event_arena_index = 0;
     {
         /* Swap event buffers */
-        P_Lock lock = P_LockE(&window->event_arena_swp_mutex);
+        Lock lock = LockE(&window->event_arena_swp_mutex);
         event_arena_index = window->current_event_arena_index;
         window->current_event_arena_index = 1 - window->current_event_arena_index;
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
     Arena *events_arena = window->event_arenas[event_arena_index];
     P_WindowEventArray events = ZI;
@@ -1388,11 +1379,11 @@ void P_UpdateWindowSettings(P_Window *p_window, P_WindowSettings *settings)
 {
     __prof;
     P_W32_Window *window = (P_W32_Window *)p_window;
-    P_Lock lock = P_LockE(&window->settings_mutex);
+    Lock lock = LockE(&window->settings_mutex);
     {
         P_W32_UpdateWindowFromSettings(window, settings);
     }
-    P_Unlock(&lock);
+    Unlock(&lock);
 }
 
 /* FIXME: Lock settings mutex for these functions */
@@ -1407,7 +1398,7 @@ void P_ShowWindow(P_Window *p_window)
 {
     P_W32_Window *window = (P_W32_Window *)p_window;
     HWND hwnd = window->hwnd;
-    P_Lock lock = P_LockE(&window->settings_mutex);
+    Lock lock = LockE(&window->settings_mutex);
     {
         i32 show_cmd = SW_NORMAL;
         P_WindowSettings *settings = &window->settings;
@@ -1423,7 +1414,7 @@ void P_ShowWindow(P_Window *p_window)
         ShowWindow(hwnd, show_cmd);
         BringWindowToTop(hwnd);
     }
-    P_Unlock(&lock);
+    Unlock(&lock);
 }
 
 void P_SetWindowCursorPos(P_Window *p_window, Vec2 pos)
@@ -1699,7 +1690,7 @@ P_Sock *P_AllocSock(u16 listen_port, u64 sndbuf_size, u64 rcvbuf_size)
     P_W32_SharedCtx *g = &P_W32_shared_ctx;
     P_W32_Sock *ws = 0;
     {
-        P_Lock lock = P_LockE(&g->socks_mutex);
+        Lock lock = LockE(&g->socks_mutex);
         if (g->first_free_sock)
         {
             ws = g->first_free_sock;
@@ -1709,7 +1700,7 @@ P_Sock *P_AllocSock(u16 listen_port, u64 sndbuf_size, u64 rcvbuf_size)
         {
             ws = PushStructNoZero(g->socks_arena, P_W32_Sock);
         }
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
     ZeroStruct(ws);
 
@@ -1734,12 +1725,12 @@ void P_ReleaseSock(P_Sock *sock)
     P_W32_SharedCtx *g = &P_W32_shared_ctx;
     P_W32_Sock *ws = (P_W32_Sock *)sock;
     closesocket(ws->sock);
-    P_Lock lock = P_LockE(&g->socks_mutex);
+    Lock lock = LockE(&g->socks_mutex);
     {
         ws->next_free = g->first_free_sock;
         g->first_free_sock = ws;
     }
-    P_Unlock(&lock);
+    Unlock(&lock);
 }
 
 P_SockReadResult P_ReadSock(Arena *arena, P_Sock *sock)
@@ -1885,22 +1876,6 @@ String P_GetClipboardText(Arena *arena)
     return result;
 }
 
-u32 P_GetLogicalProcessorCount(void)
-{
-    return GetActiveProcessorCount(ALL_PROCESSOR_GROUPS);
-}
-
-u32 P_GetThreadId(void)
-{
-    return GetCurrentThreadId();
-}
-
-i64 P_GetCurrentSchedulerPeriodNs(void)
-{
-    P_W32_SharedCtx *g = &P_W32_shared_ctx;
-    return Atomic64Fetch(&g->current_scheduler_cycle_period_ns.v);
-}
-
 ////////////////////////////////
 //~ @hookdef Sleep hooks
 
@@ -1908,19 +1883,19 @@ void P_SleepPrecise(i64 sleep_time_ns)
 {
     __prof;
 
-    i64 big_sleep = P_GetCurrentSchedulerPeriodNs();
+    i64 big_sleep = GetCurrentSchedulerPeriodNs();
     i64 tolerance = (f64)big_sleep * 0.5;
     //i64 tolerance = 1000000000;
 
-    i64 now_ns = P_TimeNs();
+    i64 now_ns = TimeNs();
     i64 target_ns = now_ns + sleep_time_ns;
 
     /* Sleep */
     while (now_ns < target_ns - big_sleep - tolerance)
     {
         __profn("Sleep part");
-        P_Wait(0, 0, 0, big_sleep);
-        now_ns = P_TimeNs();
+        FutexWait(0, 0, 0, big_sleep);
+        now_ns = TimeNs();
     }
 
     /* Spin */
@@ -1929,7 +1904,7 @@ void P_SleepPrecise(i64 sleep_time_ns)
         while (now_ns < target_ns)
         {
             _mm_pause();
-            now_ns = P_TimeNs();
+            now_ns = TimeNs();
         }
     }
 }
@@ -1938,7 +1913,7 @@ void P_SleepFrame(i64 last_frame_time_ns, i64 target_dt_ns)
 {
     if (last_frame_time_ns != 0 && target_dt_ns > 0)
     {
-        i64 now_ns = P_TimeNs();
+        i64 now_ns = TimeNs();
         i64 last_frame_dt_ns = now_ns - last_frame_time_ns;
         i64 sleep_time_ns = target_dt_ns - last_frame_dt_ns;
         if (sleep_time_ns > 0)
@@ -2073,7 +2048,7 @@ void P_W32_InitBtnTable(void)
     g->vk_btn_table[VK_OEM_1] = P_Btn_Semicolon;
 }
 
-JobDef(P_W32_AppStartupJob, UNUSED job)
+JobDef(P_W32_AppStartupJob, UNUSED sig, UNUSED id)
 {
     P_W32_SharedCtx *g = &P_W32_shared_ctx;
     TempArena scratch = BeginScratchNoConflict();
@@ -2085,7 +2060,7 @@ JobDef(P_W32_AppStartupJob, UNUSED job)
     EndScratch(scratch);
 }
 
-JobDef(P_W32_AppShutdownJob, _)
+JobDef(P_W32_AppShutdownJob, UNUSED sig, UNUSED id)
 {
     __prof;
     P_W32_SharedCtx *g = &P_W32_shared_ctx;
@@ -2188,18 +2163,6 @@ int CALLBACK wWinMain(_In_ HINSTANCE instance, _In_opt_ HINSTANCE prev_instance,
     g->exit_begin_event = CreateEventW(0, 1, 0, 0);
     g->exit_end_event = CreateEventW(0, 1, 0, 0);
 
-    /* Init timer */
-    {
-        LARGE_INTEGER qpf;
-        QueryPerformanceFrequency(&qpf);
-        g->ns_per_qpc = 1000000000 / qpf.QuadPart;
-    }
-    {
-        LARGE_INTEGER qpc;
-        QueryPerformanceCounter(&qpc);
-        g->timer_start_qpc = qpc.QuadPart;
-    }
-
     /* Convert main thread to fiber */
     W32_AllocFiber(0);
 
@@ -2253,9 +2216,6 @@ int CALLBACK wWinMain(_In_ HINSTANCE instance, _In_opt_ HINSTANCE prev_instance,
         LAX success;
     }
 
-    /* Init threads pool */
-    g->threads_arena = AllocArena(Gibi(64));
-
     /* Init watches pool */
     g->watches_arena = AllocArena(Gibi(64));
 
@@ -2271,7 +2231,7 @@ int CALLBACK wWinMain(_In_ HINSTANCE instance, _In_opt_ HINSTANCE prev_instance,
      /* Run app start job */
     if (!Atomic32Fetch(&g->panicking))
     {
-        P_Run(1, P_W32_AppStartupJob, 0, PoolKind_Floating, PriorityKind_High, 0);
+        RunJob(1, P_W32_AppStartupJob, 0, PoolKind_Floating, PriorityKind_High, 0);
     }
 
     /* Wait for startup end or panic */
@@ -2299,7 +2259,7 @@ int CALLBACK wWinMain(_In_ HINSTANCE instance, _In_opt_ HINSTANCE prev_instance,
     /* Run exit callbacks job */
     if (!Atomic32Fetch(&g->panicking))
     {
-        P_Run(1, P_W32_AppShutdownJob, 0, PoolKind_Floating, PriorityKind_High, 0);
+        RunJob(1, P_W32_AppShutdownJob, 0, PoolKind_Floating, PriorityKind_High, 0);
     }
 
     /* Wait for exit end or panic */
diff --git a/src/platform/platform_win32.h b/src/platform/platform_win32.h
index e734b57a..3068110b 100644
--- a/src/platform/platform_win32.h
+++ b/src/platform/platform_win32.h
@@ -39,7 +39,7 @@ Struct(P_W32_Window)
 
     u16 utf16_high_surrogate_last_input;
 
-    P_Mutex settings_mutex;
+    Mutex settings_mutex;
     P_WindowSettings settings;
 
     i32 monitor_width;
@@ -57,7 +57,7 @@ Struct(P_W32_Window)
     Atomic32 topmost_toggles;
     b32 is_topmost;
 
-    P_Mutex event_arena_swp_mutex;
+    Mutex event_arena_swp_mutex;
     i32 current_event_arena_index;
     Arena *event_arenas[2];
 
@@ -113,8 +113,6 @@ Struct(P_W32_Sock)
 Struct(P_W32_SharedCtx)
 {
     SYSTEM_INFO info;
-    i64 timer_start_qpc;
-    i64 ns_per_qpc;
     u32 main_thread_id;
     Atomic32 shutdown;
 
@@ -132,20 +130,20 @@ Struct(P_W32_SharedCtx)
     P_Btn vk_btn_table[256];
 
     //- Watches pool
-    P_Mutex watches_mutex;
+    Mutex watches_mutex;
     Arena *watches_arena;
     P_W32_Watch *watches_first_free;
 
     //- Windows pool
     WNDCLASSEXW window_class;
-    P_Mutex windows_mutex;
+    Mutex windows_mutex;
     Arena *windows_arena;
     P_W32_Window *first_free_window;
 
     //- Socket pool
     WSADATA wsa_data;
     Arena *socks_arena;
-    P_Mutex socks_mutex;
+    Mutex socks_mutex;
     P_W32_Sock *first_free_sock;
 
     //- Exit funcs
diff --git a/src/playback/playback_win32.c b/src/playback/playback_win32.c
index 5581be39..97cdfd3e 100644
--- a/src/playback/playback_win32.c
+++ b/src/playback/playback_win32.c
@@ -16,7 +16,7 @@ PB_StartupReceipt PB_Startup(MIX_StartupReceipt *mixer_sr)
     LAX mixer_sr;
     PB_WSP_InitializeWasapi();
     /* Start playback job */
-    P_Run(1, PB_WSP_PlaybackJob, 0, PoolKind_Audio, PriorityKind_High, &g->PB_WSP_PlaybackJob_counter);
+    RunJob(1, PB_WSP_PlaybackJob, 0, PoolKind_Audio, PriorityKind_High, &g->PB_WSP_PlaybackJob_counter);
     P_OnExit(&PB_WSP_Shutdown);
 
     return (PB_StartupReceipt) { 0 };
@@ -27,7 +27,7 @@ P_ExitFuncDef(PB_WSP_Shutdown)
     __prof;
     PB_WSP_SharedState *g = &PB_WSP_shared_state;
     Atomic32FetchSet(&g->shutdown, 1);
-    P_WaitOnCounter(&g->PB_WSP_PlaybackJob_counter);
+    WaitOnCounter(&g->PB_WSP_PlaybackJob_counter);
 }
 
 void PB_WSP_InitializeWasapi(void)
@@ -184,7 +184,7 @@ void PB_WSP_EndUpdate(PB_WSP_Buff *wspbuf, MIX_PcmF32 src)
 ////////////////////////////////
 //~ Playback job
 
-JobDef(PB_WSP_PlaybackJob, _)
+JobDef(PB_WSP_PlaybackJob, UNUSED sig, UNUSED id)
 {
     __prof;
     PB_WSP_SharedState *g = &PB_WSP_shared_state;
diff --git a/src/pp/pp_core.c b/src/pp/pp_core.c
index 1cb33ef3..b9eabc31 100644
--- a/src/pp/pp_core.c
+++ b/src/pp/pp_core.c
@@ -11,11 +11,9 @@ UserStartupReceipt StartupUser(S_StartupReceipt *sprite_sr,
                                String connect_address_str)
 {
     __prof;
-    LAX font_sr;
     LAX sprite_sr;
     LAX draw_sr;
     LAX asset_cache_sr;
-    LAX sound_sr;
     LAX mixer_sr;
     LAX sim_sr;
     SharedUserState *g = &shared_user_state;
@@ -23,7 +21,7 @@ UserStartupReceipt StartupUser(S_StartupReceipt *sprite_sr,
     SetGstat(GSTAT_DEBUG_STEPS, U64Max);
 
     g->arena = AllocArena(Gibi(64));
-    g->real_time_ns = P_TimeNs();
+    g->real_time_ns = TimeNs();
 
     /* TODO: Remove this */
     g->connect_address_str = PushString(g->arena, connect_address_str);
@@ -55,8 +53,8 @@ UserStartupReceipt StartupUser(S_StartupReceipt *sprite_sr,
     P_ShowWindow(g->window);
 
     /* Start jobs  */
-    P_Run(1, UpdateUserJob, 0, PoolKind_User, PriorityKind_High, &g->shutdown_job_counters);
-    P_Run(1, SimJob, 0, PoolKind_Sim, PriorityKind_High, &g->shutdown_job_counters);
+    RunJob(1, UpdateUserJob, 0, PoolKind_User, PriorityKind_High, &g->shutdown_job_counters);
+    RunJob(1, SimJob, 0, PoolKind_Sim, PriorityKind_High, &g->shutdown_job_counters);
     P_OnExit(&ShutdownUser);
 
     return (UserStartupReceipt) { 0 };
@@ -70,7 +68,7 @@ P_ExitFuncDef(ShutdownUser)
     __prof;
     SharedUserState *g = &shared_user_state;
     Atomic32FetchSet(&g->shutdown, 1);
-    P_WaitOnCounter(&g->shutdown_job_counters);
+    WaitOnCounter(&g->shutdown_job_counters);
     P_ReleaseWindow(g->window);
 }
 
@@ -234,7 +232,7 @@ P_LogEventCallbackFuncDef(ConsoleLogCallback, log)
 {
     __prof;
     SharedUserState *g = &shared_user_state;
-    P_Lock lock = P_LockE(&g->console_logs_mutex);
+    Lock lock = LockE(&g->console_logs_mutex);
     {
         ConsoleLog *clog = PushStruct(g->console_logs_arena, ConsoleLog);
         clog->level = log.level;
@@ -257,7 +255,7 @@ P_LogEventCallbackFuncDef(ConsoleLogCallback, log)
         }
         g->last_console_log = clog;
     }
-    P_Unlock(&lock);
+    Unlock(&lock);
 }
 
 //- Draw console
@@ -298,11 +296,11 @@ void DrawDebugConsole(i32 level, b32 minimized)
     }
     g->console_logs_height = 0;
 
-    i64 now_ns = P_TimeNs();
+    i64 now_ns = TimeNs();
     F_Font *font = F_LoadFontAsync(Lit("font/fixedsys.ttf"), 12.0f);
     if (font)
     {
-        P_Lock lock = P_LockE(&g->console_logs_mutex);
+        Lock lock = LockE(&g->console_logs_mutex);
         {
             for (ConsoleLog *log = g->last_console_log; log; log = log->prev)
             {
@@ -356,7 +354,7 @@ void DrawDebugConsole(i32 level, b32 minimized)
                 }
             }
         }
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
     if (bounds_top < F32Infinity && bounds_bottom > -F32Infinity)
     {
@@ -417,7 +415,7 @@ void UpdateUser(P_Window *window)
     TempArena scratch = BeginScratchNoConflict();
 
     //- Begin frame
-    g->real_dt_ns = P_TimeNs() - g->real_time_ns;
+    g->real_dt_ns = TimeNs() - g->real_time_ns;
     g->real_time_ns += g->real_dt_ns;
     g->screen_size = P_GetWindowSize(window);
     S_Scope *sprite_frame_scope = S_BeginScope();
@@ -426,7 +424,7 @@ void UpdateUser(P_Window *window)
 
     {
         __profn("Pull snapshot");
-        P_Lock lock = P_LockE(&g->local_to_user_client_mutex);
+        Lock lock = LockE(&g->local_to_user_client_mutex);
         u64 old_last_tick = g->user_unblended_client->last_tick;
         u64 last_tick = g->local_to_user_client->last_tick;
         if (last_tick > old_last_tick)
@@ -437,7 +435,7 @@ void UpdateUser(P_Window *window)
             g->average_local_to_user_snapshot_publish_dt_ns -= g->average_local_to_user_snapshot_publish_dt_ns / 50;
             g->average_local_to_user_snapshot_publish_dt_ns += g->local_to_user_client_publish_dt_ns / 50;
         }
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
 
     //- Create user world from blended snapshots
@@ -1832,7 +1830,7 @@ void UpdateUser(P_Window *window)
 
             /* Set user sim control */
             {
-                P_Lock lock = P_LockE(&g->user_sim_cmd_mutex);
+                Lock lock = LockE(&g->user_sim_cmd_mutex);
 
                 /* Reset flags */
                 if (g->user_sim_cmd_gen != g->last_user_sim_cmd_gen)
@@ -1845,7 +1843,7 @@ void UpdateUser(P_Window *window)
                 g->user_sim_cmd_control = control;
                 g->user_sim_cmd_control.flags |= old_flags;
                 g->user_hovered_ent = hovered_ent->id;
-                P_Unlock(&lock);
+                Unlock(&lock);
             }
         }
 
@@ -1869,7 +1867,7 @@ void UpdateUser(P_Window *window)
 
     {
         /* Update network usage stats */
-        i64 stat_now_ns = P_TimeNs();
+        i64 stat_now_ns = TimeNs();
         g->net_bytes_read.last_second_end = GetGstat(GSTAT_SOCK_BYTES_RECEIVED);
         g->net_bytes_sent.last_second_end = GetGstat(GSTAT_SOCK_BYTES_SENT);
         if (stat_now_ns - g->last_second_reset_ns > NsFromSeconds(1))
@@ -2066,10 +2064,10 @@ void UpdateUser(P_Window *window)
 
 //- User update job
 
-JobDef(UpdateUserJob, _)
+JobDef(UpdateUserJob, UNUSED sig, UNUSED id)
 {
     SharedUserState *g = &shared_user_state;
-    i64 time_ns = P_TimeNs();
+    i64 time_ns = TimeNs();
     while (!Atomic32Fetch(&g->shutdown))
     {
         P_Window *window = g->window;
@@ -2082,7 +2080,7 @@ JobDef(UpdateUserJob, _)
             {
                 __profn("Frame limiter wait");
                 P_SleepFrame(time_ns, 1000000000 / FPS_LIMIT);
-                time_ns = P_TimeNs();
+                time_ns = TimeNs();
             }
         }
         UpdateUser(window);
@@ -2109,7 +2107,7 @@ void GenerateuserInputCmds(Client *user_input_client, u64 tick)
         sim_ent_activate(control_cmd, user_input_ss->tick);
     }
     {
-        P_Lock lock = P_LockE(&g->user_sim_cmd_mutex);
+        Lock lock = LockE(&g->user_sim_cmd_mutex);
         /* Update control cmd */
         {
             control_cmd->cmd_control = g->user_sim_cmd_control;
@@ -2125,14 +2123,14 @@ void GenerateuserInputCmds(Client *user_input_client, u64 tick)
         }
 #endif
         ++g->user_sim_cmd_gen;
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
 }
 
 ////////////////////////////////
 //~ Sim update
 
-JobDef(SimJob, UNUSED job)
+JobDef(SimJob, UNUSED sig, UNUSED id)
 {
     SharedUserState *g = &shared_user_state;
 #if 0
@@ -2221,7 +2219,7 @@ JobDef(SimJob, UNUSED job)
         {
             __profn("Sim update");
 
-            real_dt_ns = P_TimeNs() - real_time_ns;
+            real_dt_ns = TimeNs() - real_time_ns;
             real_time_ns += real_dt_ns;
 
             N_EventList host_events = N_BeginUpdate(scratch.arena, host);
@@ -2798,9 +2796,9 @@ JobDef(SimJob, UNUSED job)
                 if (local_ss->valid)
                 {
                     /* TODO: Double buffer */
-                    P_Lock lock = P_LockE(&g->local_to_user_client_mutex);
+                    Lock lock = LockE(&g->local_to_user_client_mutex);
                     sim_snapshot_alloc(g->local_to_user_client, local_ss, local_ss->tick);
-                    i64 publish_ns = P_TimeNs();
+                    i64 publish_ns = TimeNs();
                     if (last_publish_to_user_ns == 0)
                     {
                         last_publish_to_user_ns = publish_ns - g->average_local_to_user_snapshot_publish_dt_ns;
@@ -2809,7 +2807,7 @@ JobDef(SimJob, UNUSED job)
                     g->local_to_user_client_publish_time_ns = publish_ns;
                     last_publish_to_user_ns = publish_ns;
                     sim_snapshot_release_ticks_in_range(g->local_to_user_client, 0, local_ss->tick - 1);
-                    P_Unlock(&lock);
+                    Unlock(&lock);
                 }
             }
 
diff --git a/src/pp/pp_core.h b/src/pp/pp_core.h
index 3619327a..6805d9aa 100644
--- a/src/pp/pp_core.h
+++ b/src/pp/pp_core.h
@@ -185,7 +185,7 @@ Struct(SharedUserState)
     b32 debug_draw;
 
     //- Debug console
-    P_Mutex console_logs_mutex;
+    Mutex console_logs_mutex;
     Arena *console_logs_arena;
     ConsoleLog *first_console_log;
     ConsoleLog *last_console_log;
@@ -194,11 +194,11 @@ Struct(SharedUserState)
     b32 debug_console;
 
     //- Window -> user
-    P_Mutex sys_window_events_mutex;
+    Mutex sys_window_events_mutex;
     Arena *sys_window_events_arena;
 
     //- User -> sim
-    P_Mutex user_sim_cmd_mutex;
+    Mutex user_sim_cmd_mutex;
     ControlData user_sim_cmd_control;
     EntityId user_hovered_ent;
     u64 last_user_sim_cmd_gen;
@@ -208,7 +208,7 @@ Struct(SharedUserState)
     Atomic32 user_paused_steps;
 
     //- Sim -> user
-    P_Mutex local_to_user_client_mutex;
+    Mutex local_to_user_client_mutex;
     ClientStore *local_to_user_client_store;
     Client *local_to_user_client;
     i64 local_to_user_client_publish_dt_ns;
diff --git a/src/pp/pp_sim.c b/src/pp/pp_sim.c
index 61f3b7e4..d8c77f5e 100644
--- a/src/pp/pp_sim.c
+++ b/src/pp/pp_sim.c
@@ -329,14 +329,14 @@ Snapshot *sim_snapshot_alloc(Client *client, Snapshot *src, u64 tick)
     if (ss->num_ents_reserved == 0) {
         /* Copying from nil snapshot, need to create blank & root entity */
 
-        /* PushStruct blank ent at index 0 (because index 0 is never valid anyway since it maps to sim_ent_nil()) */
+        /* Push blank ent at index 0 (because index 0 is never valid anyway since it maps to sim_ent_nil()) */
         {
             PushStruct(ss->ents_arena, Entity);
             ++ss->num_ents_allocated;
             ++ss->num_ents_reserved;
         }
 
-        /* PushStruct root ent with constant id */
+        /* Push root ent with constant id */
         {
             Entity *root = PushStructNoZero(ss->ents_arena, Entity);
             *root = *sim_ent_nil();
diff --git a/src/sound/sound_core.c b/src/sound/sound_core.c
index 1c0f9fa8..6f0559cd 100644
--- a/src/sound/sound_core.c
+++ b/src/sound/sound_core.c
@@ -1,17 +1,16 @@
 ////////////////////////////////
 //~ Load job
 
-JobDef(SND_LoadAssetJob, job)
+JobDef(SND_LoadJob, sig, UNUSED id)
 {
     __prof;
-    SND_LoadAssetJobSig *params = job.sig;
     TempArena scratch = BeginScratchNoConflict();
-    String path = STRING(params->path_len, (u8 *)params->path_cstr);
-    AC_Asset *asset = params->asset;
-    SND_SoundFlag flags = params->flags;
+    String path = sig->path;
+    AC_Asset *asset = sig->asset;
+    SND_SoundFlag flags = sig->flags;
 
     P_LogInfoF("Loading sound \"%F\"", FmtString(path));
-    i64 start_ns = P_TimeNs();
+    i64 start_ns = TimeNs();
 
     String error_msg = Lit("Unknown error");
 
@@ -61,7 +60,7 @@ JobDef(SND_LoadAssetJob, job)
         sound->samples = samples;
         CopyBytes(sound->samples, decoded.samples, decoded.samples_count * sizeof(*decoded.samples));
 
-        P_LogSuccessF("Loaded sound \"%F\" in %F seconds", FmtString(path), FmtFloat(SecondsFromNs(P_TimeNs() - start_ns)));
+        P_LogSuccessF("Loaded sound \"%F\" in %F seconds", FmtString(path), FmtFloat(SecondsFromNs(TimeNs() - start_ns)));
         AC_MarkReady(asset, sound);
     }
     else
@@ -101,22 +100,12 @@ AC_Asset *SND_LoadAsset(String path, SND_SoundFlag flags, b32 wait)
 
     if (is_first_touch)
     {
-        /* Assemble task params */
-        SND_LoadAssetJobSig *params = SND_AllocJobSig();
-        if (path.len > (sizeof(params->path_cstr) - 1))
-        {
-            P_Panic(StringFormat(scratch.arena,
-                                 Lit("Sound path \"%F\" too long!"),
-                                 FmtString(path)));
-        }
-        CstrBuffFromStringToBuff(StringFromArray(params->path_cstr), path);
-        params->path_len = path.len;
-        params->asset = asset;
-        params->flags = flags;
-
-        /* PushStruct task */
         AC_MarkLoading(asset);
-        P_Run(1, SND_LoadAssetJob, params, PoolKind_Background, PriorityKind_Low, &asset->counter);
+        SND_LoadJob_Desc *desc = PushJobDesc(SND_LoadJob, .pool = PoolKind_Background, .priority = PriorityKind_Low, .counter = &asset->counter);
+        desc->sig->path = PushString(desc->arena, path);
+        desc->sig->asset = asset;
+        desc->sig->flags = flags;
+        RunJobEx((GenericJobDesc *)desc);
         if (wait)
         {
             AC_WaitOnAssetReady(asset);
diff --git a/src/sound/sound_core.h b/src/sound/sound_core.h
index 31e6eb9e..f1f22fb9 100644
--- a/src/sound/sound_core.h
+++ b/src/sound/sound_core.h
@@ -19,7 +19,7 @@ Struct(SND_Sound)
 ////////////////////////////////
 //~ Sound load operations
 
-JobDecl(SND_LoadAssetJob, { SND_SoundFlag flags; AC_Asset *asset; String path; });
+JobDecl(SND_LoadJob, { SND_SoundFlag flags; AC_Asset *asset; String path; });
 AC_Asset *SND_LoadAsset(String path, SND_SoundFlag flags, b32 wait);
 SND_Sound *SND_LoadSoundAsync(String path, SND_SoundFlag flags);
 SND_Sound *SND_LoadSoundWait(String path, SND_SoundFlag flags);
diff --git a/src/sprite/sprite_core.c b/src/sprite/sprite_core.c
index 3c65487e..998351d1 100644
--- a/src/sprite/sprite_core.c
+++ b/src/sprite/sprite_core.c
@@ -40,11 +40,9 @@ S_StartupReceipt S_Startup(void)
     g->cache.arena = AllocArena(Gibi(64));
     g->cache.bins = PushStructs(g->cache.arena, S_CacheEntryBin, S_CacheBinsCount);
 
-    g->cmds_arena = AllocArena(Gibi(64));
-
     g->scopes_arena = AllocArena(Gibi(64));
 
-    P_Run(1, S_EvictorJob, 0, PoolKind_Background, PriorityKind_Low, &g->shutdown_counter);
+    RunJob(1, S_EvictorJob, 0, PoolKind_Background, PriorityKind_Low, &g->shutdown_counter);
 
     P_OnExit(&S_Shutdown);
 #if RESOURCE_RELOADING
@@ -63,13 +61,13 @@ P_ExitFuncDef(S_Shutdown)
     S_SharedState *g = &S_shared_state;
     /* Signal evictor shutdown */
     {
-        P_Lock lock = P_LockE(&g->evictor_scheduler_mutex);
+        Lock lock = LockE(&g->evictor_scheduler_mutex);
         g->evictor_scheduler_shutdown = 1;
-        P_SignalCv(&g->evictor_scheduler_shutdown_cv, I32Max);
-        P_Unlock(&lock);
+        SignalCv(&g->evictor_scheduler_shutdown_cv, I32Max);
+        Unlock(&lock);
     }
     /* Wait for evictor shutdown */
-    P_WaitOnCounter(&g->shutdown_counter);
+    WaitOnCounter(&g->shutdown_counter);
 }
 
 ////////////////////////////////
@@ -421,69 +419,35 @@ S_Sheet S_SheetFromAseResult(Arena *arena, ASE_DecodedSheet ase)
 ////////////////////////////////
 //~ Load job
 
-//- Job def
-JobDef(S_LoadSpriteJob, job)
+JobDef(S_LoadSpriteJob, sig, UNUSED id)
 {
     __prof;
-    S_SharedState *g = &S_shared_state;
-    S_Cmd *cmd = job.sig;
-    S_CacheEntryRef ref = cmd->ref;
-
+    S_CacheEntryRef ref = sig->ref;
+    S_Tag tag = sig->tag;
+    S_Scope *scope = sig->scope;
     switch (ref.e->kind)
     {
         case S_CacheEntryKind_Texture:
         {
-            S_LoadCacheEntryTexture(ref, cmd->tag);
+            S_LoadCacheEntryTexture(ref, tag);
         } break;
         case S_CacheEntryKind_Sheet:
         {
-            S_LoadCacheEntrySheet(ref, cmd->tag);
+            S_LoadCacheEntrySheet(ref, tag);
         } break;
         default: { P_Panic(Lit("Unknown sprite cache node kind")); } break;
     }
-
-    /* Free cmd */
-    P_Lock lock = P_LockE(&g->cmds_mutex);
-    {
-        S_EndScope(cmd->scope);
-        cmd->next_free = g->first_free_cmd;
-        g->first_free_cmd = cmd;
-    }
-    P_Unlock(&lock);
+    S_EndScope(scope);
 }
 
-//- Push
 void S_PushLoadJob(S_CacheEntryRef ref, S_Tag tag)
 {
-    S_SharedState *g = &S_shared_state;
-    S_Cmd *cmd = 0;
-    {
-        P_Lock lock = P_LockE(&g->cmds_mutex);
-        if (g->first_free_cmd)
-        {
-            cmd = g->first_free_cmd;
-            g->first_free_cmd = cmd->next_free;
-        }
-        else
-        {
-            cmd = PushStructNoZero(g->cmds_arena, S_Cmd);
-        }
-        P_Unlock(&lock);
-    }
-    ZeroStruct(cmd);
-
-    /* Initialize cmd */
-    cmd->scope = S_BeginScope();
-    cmd->ref = S_EnsureRefFromRef(cmd->scope, ref)->ref;
-    cmd->tag = tag;
-    {
-        u64 copy_len = MinU64(tag.path.len, countof(cmd->tag_path_buff));
-        cmd->tag.path.text = cmd->tag_path_buff;
-        CopyBytes(cmd->tag.path.text, tag.path.text, copy_len);
-    }
-
-    /* PushStruct work */
-    P_Run(1, S_LoadSpriteJob, cmd, PoolKind_Background, PriorityKind_Inherit, 0);
+    S_LoadSpriteJob_Desc *desc = PushJobDesc(S_LoadSpriteJob, .pool = PoolKind_Background, .priority = PriorityKind_Inherit);
+    desc->sig->scope = S_BeginScope();  /* Scope ended by job */
+    desc->sig->ref = S_EnsureRefFromRef(desc->sig->scope, ref)->ref;
+    desc->sig->tag = tag;
+    tag.path = PushString(desc->arena, tag.path);
+    RunJobEx((GenericJobDesc *)desc);
 }
 
 ////////////////////////////////
@@ -501,7 +465,7 @@ void S_LoadCacheEntryTexture(S_CacheEntryRef ref, S_Tag tag)
 
     P_LogInfoF("Loading sprite texture [%F] \"%F\"", FmtHex(e->hash.v), FmtString(path));
     b32 success = 0;
-    i64 start_ns = P_TimeNs();
+    i64 start_ns = TimeNs();
 
     Assert(StringEndsWith(path, Lit(".ase")));
     Assert(e->kind == S_CacheEntryKind_Texture);
@@ -549,7 +513,7 @@ void S_LoadCacheEntryTexture(S_CacheEntryRef ref, S_Tag tag)
         P_LogSuccessF("Loaded sprite texture [%F] \"%F\" in %F seconds (cache size: %F bytes).",
                       FmtHex(e->hash.v),
                       FmtString(path),
-                      FmtFloat(SecondsFromNs(P_TimeNs() - start_ns)),
+                      FmtFloat(SecondsFromNs(TimeNs() - start_ns)),
                       FmtUint(e->memory_usage));
     }
 
@@ -557,7 +521,7 @@ void S_LoadCacheEntryTexture(S_CacheEntryRef ref, S_Tag tag)
 
 #if RESOURCE_RELOADING
     S_CacheEntryBin *bin = &g->cache.bins[e->hash.v % S_CacheBinsCount];
-    P_Lock bin_lock = P_LockE(&bin->mutex);
+    Lock bin_lock = LockE(&bin->mutex);
     {
         for (S_CacheEntry *old_entry = bin->first; old_entry; old_entry = old_entry->next_in_bin)
         {
@@ -566,9 +530,9 @@ void S_LoadCacheEntryTexture(S_CacheEntryRef ref, S_Tag tag)
                 Atomic32FetchSet(&old_entry->out_of_date, 1);
             }
         }
-        e->load_time_ns = P_TimeNs();
+        e->load_time_ns = TimeNs();
     }
-    P_Unlock(&bin_lock);
+    Unlock(&bin_lock);
 #endif
 
     EndScratch(scratch);
@@ -586,7 +550,7 @@ void S_LoadCacheEntrySheet(S_CacheEntryRef ref, S_Tag tag)
 
     P_LogInfoF("Loading sprite sheet [%F] \"%F\"", FmtHex(e->hash.v), FmtString(path));
     b32 success = 0;
-    i64 start_ns = P_TimeNs();
+    i64 start_ns = TimeNs();
 
     Assert(e->kind == S_CacheEntryKind_Sheet);
 
@@ -632,7 +596,7 @@ void S_LoadCacheEntrySheet(S_CacheEntryRef ref, S_Tag tag)
         P_LogSuccessF("Loaded sprite sheet [%F] \"%F\" in %F seconds (cache size: %F bytes).",
                       FmtHex(e->hash.v),
                       FmtString(path),
-                      FmtFloat(SecondsFromNs(P_TimeNs() - start_ns)),
+                      FmtFloat(SecondsFromNs(TimeNs() - start_ns)),
                       FmtUint(e->memory_usage));
     }
 
@@ -640,7 +604,7 @@ void S_LoadCacheEntrySheet(S_CacheEntryRef ref, S_Tag tag)
 
 #if RESOURCE_RELOADING
     S_CacheEntryBin *bin = &g->cache.bins[e->hash.v % S_CacheBinsCount];
-    P_Lock bin_lock = P_LockE(&bin->mutex);
+    Lock bin_lock = LockE(&bin->mutex);
     {
         for (S_CacheEntry *old_entry = bin->first; old_entry; old_entry = old_entry->next_in_bin)
         {
@@ -649,9 +613,9 @@ void S_LoadCacheEntrySheet(S_CacheEntryRef ref, S_Tag tag)
                 Atomic32FetchSet(&old_entry->out_of_date, 1);
             }
         }
-        e->load_time_ns = P_TimeNs();
+        e->load_time_ns = TimeNs();
     }
-    P_Unlock(&bin_lock);
+    Unlock(&bin_lock);
 #endif
 
     EndScratch(scratch);
@@ -720,11 +684,11 @@ S_ScopeCacheEntryRef *S_EnsureRefUnsafely(S_Scope *scope, S_CacheEntry *e)
     return *slot;
 }
 
-S_ScopeCacheEntryRef *S_EnsureRefFromEntryLocked(S_Scope *scope, S_CacheEntry *e, P_Lock *bin_lock)
+S_ScopeCacheEntryRef *S_EnsureRefFromEntryLocked(S_Scope *scope, S_CacheEntry *e, Lock *bin_lock)
 {
     S_SharedState *g = &S_shared_state;
     /* Guaranteed safe if caller has lock on entry's bin, since entry may not have an existing reference and could otherwise be evicted while ensuring this reference */
-    P_AssertLockedES(bin_lock, &g->cache.bins[e->hash.v % S_CacheBinsCount].mutex);
+    AssertLockedES(bin_lock, &g->cache.bins[e->hash.v % S_CacheBinsCount].mutex);
     return S_EnsureRefUnsafely(scope, e);
 }
 
@@ -745,7 +709,7 @@ S_Scope *S_BeginScope(void)
     S_ScopeCacheEntryRef **bins = 0;
     S_ScopeCacheEntryRef *pool = 0;
     {
-        P_Lock lock = P_LockE(&g->scopes_mutex);
+        Lock lock = LockE(&g->scopes_mutex);
         {
             if (g->first_free_scope)
             {
@@ -761,7 +725,7 @@ S_Scope *S_BeginScope(void)
                 pool = PushStructsNoZero(g->scopes_arena, S_ScopeCacheEntryRef, S_MaxScopeReferences);
             }
         }
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
     ZeroStruct(result);
     ZeroBytes(bins, sizeof(*bins) * S_CacheBinsCount);
@@ -782,12 +746,12 @@ void S_EndScope(S_Scope *scope)
     }
 
     /* Release scope */
-    P_Lock lock = P_LockE(&g->scopes_mutex);
+    Lock lock = LockE(&g->scopes_mutex);
     {
         scope->next_free = g->first_free_scope;
         g->first_free_scope = scope;
     }
-    P_Unlock(&lock);
+    Unlock(&lock);
 }
 
 ////////////////////////////////
@@ -795,13 +759,13 @@ void S_EndScope(S_Scope *scope)
 
 //- Locked lookup
 
-S_ScopeCacheEntryRef *S_EntryFromHashLocked(S_Scope *scope, S_Hash hash, P_Lock *bin_lock)
+S_ScopeCacheEntryRef *S_EntryFromHashLocked(S_Scope *scope, S_Hash hash, Lock *bin_lock)
 {
     S_SharedState *g = &S_shared_state;
     S_ScopeCacheEntryRef *scope_ref = 0;
 
     S_CacheEntryBin *bin = &g->cache.bins[hash.v % S_CacheBinsCount];
-    P_AssertLockedES(bin_lock, &bin->mutex);  /* Lock required for iterating bin */
+    AssertLockedES(bin_lock, &bin->mutex);  /* Lock required for iterating bin */
 
 #if RESOURCE_RELOADING
     /* If resource reloading is enabled, then we want to find the
@@ -870,17 +834,17 @@ S_ScopeCacheEntryRef *S_EntryFromTag(S_Scope *scope, S_Tag tag, S_CacheEntryKind
         /* Search in cache */
         if (!force_new)
         {
-            P_Lock bin_lock = P_LockS(&bin->mutex);
+            Lock bin_lock = LockS(&bin->mutex);
             {
                 scope_ref = S_EntryFromHashLocked(scope, hash, &bin_lock);
             }
-            P_Unlock(&bin_lock);
+            Unlock(&bin_lock);
         }
 
         /* If not in cache, allocate new entry */
         if (!scope_ref)
         {
-            P_Lock bin_lock = P_LockE(&bin->mutex);
+            Lock bin_lock = LockE(&bin->mutex);
             {
                 /* Search cache one more time in case an entry was allocated between locks */
                 if (!force_new)
@@ -893,7 +857,7 @@ S_ScopeCacheEntryRef *S_EntryFromTag(S_Scope *scope, S_Tag tag, S_CacheEntryKind
                     /* Cache entry still absent, allocate new entry */
                     S_CacheEntry *entry = 0;
                     {
-                        P_Lock pool_lock = P_LockE(&g->cache.entry_pool_mutex);
+                        Lock pool_lock = LockE(&g->cache.entry_pool_mutex);
                         if (g->cache.entry_pool_first_free)
                         {
                             entry = g->cache.entry_pool_first_free;
@@ -903,7 +867,7 @@ S_ScopeCacheEntryRef *S_EntryFromTag(S_Scope *scope, S_Tag tag, S_CacheEntryKind
                         {
                             entry = PushStructNoZero(g->cache.arena, S_CacheEntry);
                         }
-                        P_Unlock(&pool_lock);
+                        Unlock(&pool_lock);
                     }
                     ZeroStruct(entry);
 
@@ -928,7 +892,7 @@ S_ScopeCacheEntryRef *S_EntryFromTag(S_Scope *scope, S_Tag tag, S_CacheEntryKind
                     scope_ref = S_EnsureRefFromEntryLocked(scope, entry, &bin_lock);
                 }
             }
-            P_Unlock(&bin_lock);
+            Unlock(&bin_lock);
         }
     }
 
@@ -1132,11 +1096,11 @@ void S_ReloadSpriteFromTag(S_Scope *scope, S_Tag tag, S_CacheEntryKind kind)
     S_Hash hash = S_CacheEntryFromTagHash(tag.hash, kind);
     S_CacheEntryBin *bin = &g->cache.bins[hash.v % S_CacheBinsCount];
     S_ScopeCacheEntryRef *existing_ref = 0;
-    P_Lock bin_lock = P_LockS(&bin->mutex);
+    Lock bin_lock = LockS(&bin->mutex);
     {
         existing_ref = S_EntryFromHashLocked(scope, hash, &bin_lock);
     }
-    P_Unlock(&bin_lock);
+    Unlock(&bin_lock);
 
     if (existing_ref)
     {
@@ -1189,7 +1153,7 @@ MergesortCompareFuncDef(S_EvictorSortCmp, arg_a, arg_b, _)
  *     - The cache is over its memory budget and the node's last reference is longer ago than the grace period
  *     - Resource reloading is enabled and the node is out of date due to a change to its original resource file
  */
-JobDef(S_EvictorJob, _)
+JobDef(S_EvictorJob, UNUSED sig, UNUSED job_id)
 {
     S_SharedState *g = &S_shared_state;
     b32 shutdown = 0;
@@ -1211,7 +1175,7 @@ JobDef(S_EvictorJob, _)
                     for (u64 i = 0; i < S_CacheBinsCount; ++i)
                     {
                         S_CacheEntryBin *bin = &g->cache.bins[i];
-                        P_Lock bin_lock = P_LockS(&bin->mutex);
+                        Lock bin_lock = LockS(&bin->mutex);
                         {
                             S_CacheEntry *n = bin->first;
                             while (n)
@@ -1244,7 +1208,7 @@ JobDef(S_EvictorJob, _)
                                 n = n->next_in_bin;
                             }
                         }
-                        P_Unlock(&bin_lock);
+                        Unlock(&bin_lock);
                     }
                 }
 
@@ -1269,7 +1233,7 @@ JobDef(S_EvictorJob, _)
                         S_CacheEntry *entry = en->cache_entry;
                         i32 last_ref_cycle = en->last_ref_cycle;
                         b32 cache_over_budget_target = Atomic64Fetch(&g->cache.memory_usage.v) > (i64)S_CacheMemoryBudgetTarget;
-                        P_Lock bin_lock = P_LockE(&bin->mutex);
+                        Lock bin_lock = LockE(&bin->mutex);
                         {
                             u64 refcount_uncast = Atomic64Fetch(&entry->refcount_struct.v);
                             S_Refcount refcount = *(S_Refcount *)&refcount_uncast;
@@ -1311,7 +1275,7 @@ JobDef(S_EvictorJob, _)
                                 stop_evicting = 1;
                             }
                         }
-                        P_Unlock(&bin_lock);
+                        Unlock(&bin_lock);
                     }
                 }
 
@@ -1334,14 +1298,14 @@ JobDef(S_EvictorJob, _)
                     /* Add evicted nodes to free list */
                     {
                         __profn("Evictor free list append");
-                        P_Lock pool_lock = P_LockE(&g->cache.entry_pool_mutex);
+                        Lock pool_lock = LockE(&g->cache.entry_pool_mutex);
                         for (S_EvictorNode *en = first_evicted; en; en = en->next_evicted)
                         {
                             S_CacheEntry *n = en->cache_entry;
                             n->next_free = g->cache.entry_pool_first_free;
                             g->cache.entry_pool_first_free = n;
                         }
-                        P_Unlock(&pool_lock);
+                        Unlock(&pool_lock);
                     }
                 }
             }
@@ -1351,15 +1315,15 @@ JobDef(S_EvictorJob, _)
 
         /* Evictor sleep */
         {
-            P_Lock lock = P_LockE(&g->evictor_scheduler_mutex);
+            Lock lock = LockE(&g->evictor_scheduler_mutex);
             {
                 if (!g->evictor_scheduler_shutdown)
                 {
-                    P_WaitOnCvTime(&g->evictor_scheduler_shutdown_cv, &lock, S_EvictorCycleIntervalNs);
+                    WaitOnCvTime(&g->evictor_scheduler_shutdown_cv, &lock, S_EvictorCycleIntervalNs);
                 }
                 shutdown = g->evictor_scheduler_shutdown;
             }
-            P_Unlock(&lock);
+            Unlock(&lock);
         }
     }
 }
diff --git a/src/sprite/sprite_core.h b/src/sprite/sprite_core.h
index 6803ce6e..2a9c4e52 100644
--- a/src/sprite/sprite_core.h
+++ b/src/sprite/sprite_core.h
@@ -150,7 +150,7 @@ Struct(S_CacheEntry)
 
 Struct(S_CacheEntryBin)
 {
-    P_Mutex mutex;
+    Mutex mutex;
     S_CacheEntry *first;
     S_CacheEntry *last;
 };
@@ -160,7 +160,7 @@ Struct(S_Cache)
     Atomic64Padded memory_usage;
     Arena *arena;
     S_CacheEntryBin *bins;
-    P_Mutex entry_pool_mutex;
+    Mutex entry_pool_mutex;
     S_CacheEntry *entry_pool_first_free;
 };
 
@@ -238,7 +238,7 @@ Struct(S_SharedState)
     S_Cache cache;
 
     /* Scopes */
-    P_Mutex scopes_mutex;
+    Mutex scopes_mutex;
     Arena *scopes_arena;
     S_Scope *first_free_scope;
 
@@ -246,8 +246,8 @@ Struct(S_SharedState)
     Atomic32Padded evictor_cycle;
     Counter shutdown_counter;
     b32 evictor_scheduler_shutdown;
-    P_Mutex evictor_scheduler_mutex;
-    P_Cv evictor_scheduler_shutdown_cv;
+    Mutex evictor_scheduler_mutex;
+    Cv evictor_scheduler_shutdown_cv;
 };
 
 extern S_SharedState S_shared_state;
@@ -298,7 +298,7 @@ void S_LoadCacheEntrySheet(S_CacheEntryRef ref, S_Tag tag);
 
 void S_AddRef(S_CacheEntry *e, i32 amount);
 S_ScopeCacheEntryRef *S_EnsureRefUnsafely(S_Scope *scope, S_CacheEntry *e);
-S_ScopeCacheEntryRef *S_EnsureRefFromEntryLocked(S_Scope *scope, S_CacheEntry *e, P_Lock *bin_lock);
+S_ScopeCacheEntryRef *S_EnsureRefFromEntryLocked(S_Scope *scope, S_CacheEntry *e, Lock *bin_lock);
 S_ScopeCacheEntryRef *S_EnsureRefFromRef(S_Scope *scope, S_CacheEntryRef ref);
 
 ////////////////////////////////
@@ -310,7 +310,7 @@ void S_EndScope(S_Scope *scope);
 ////////////////////////////////
 //~ Cache lookup operations
 
-S_ScopeCacheEntryRef *S_EntryFromHashLocked(S_Scope *scope, S_Hash hash, P_Lock *bin_lock);
+S_ScopeCacheEntryRef *S_EntryFromHashLocked(S_Scope *scope, S_Hash hash, Lock *bin_lock);
 S_ScopeCacheEntryRef *S_EntryFromTag(S_Scope *scope, S_Tag tag, S_CacheEntryKind kind, b32 force_new);
 void *S_DataFromTag(S_Scope *scope, S_Tag tag, S_CacheEntryKind kind, b32 await);
 
diff --git a/src/watch/watch_core.c b/src/watch/watch_core.c
index 2186ff9e..44236f29 100644
--- a/src/watch/watch_core.c
+++ b/src/watch/watch_core.c
@@ -10,8 +10,8 @@ void W_Startup(void)
 
     g->watch_events_arena = AllocArena(Gibi(64));
 
-    P_Run(1, W_MonitorJob, 0, PoolKind_Floating, PriorityKind_Low, &g->watch_jobs_counter);
-    P_Run(1, W_DispatcherJob, 0, PoolKind_Background, PriorityKind_Low, &g->watch_jobs_counter);
+    RunJob(1, W_MonitorJob, 0, PoolKind_Floating, PriorityKind_Low, &g->watch_jobs_counter);
+    RunJob(1, W_DispatcherJob, 0, PoolKind_Background, PriorityKind_Low, &g->watch_jobs_counter);
     P_OnExit(&W_Shutdown);
 }
 
@@ -21,12 +21,12 @@ P_ExitFuncDef(W_Shutdown)
     W_SharedState *g = &W_shared_state;
     Atomic32FetchSet(&g->W_Shutdown, 1);
     {
-        P_Lock lock = P_LockE(&g->watch_dispatcher_mutex);
-        P_SignalCv(&g->watch_dispatcher_cv, I32Max);
+        Lock lock = LockE(&g->watch_dispatcher_mutex);
+        SignalCv(&g->watch_dispatcher_cv, I32Max);
         P_WakeWatch(g->watch);
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
-    P_WaitOnCounter(&g->watch_jobs_counter);
+    WaitOnCounter(&g->watch_jobs_counter);
 }
 
 ////////////////////////////////
@@ -35,7 +35,7 @@ P_ExitFuncDef(W_Shutdown)
 void W_RegisterCallback(W_CallbackFunc *callback)
 {
     W_SharedState *g = &W_shared_state;
-    P_Lock lock = P_LockE(&g->watch_callbacks_mutex);
+    Lock lock = LockE(&g->watch_callbacks_mutex);
     {
         if (g->num_watch_callbacks < countof(g->watch_callbacks))
         {
@@ -46,15 +46,14 @@ void W_RegisterCallback(W_CallbackFunc *callback)
             P_Panic(Lit("Max resource watch callbacks reached"));
         }
     }
-    P_Unlock(&lock);
+    Unlock(&lock);
 }
 
-JobDef(W_RunCallbackJob, job)
+JobDef(W_RunCallbacksJob , sig, id)
 {
     __prof;
-    W_RunCallbackJobSig *sig = job.sig;
     String name = sig->name;
-    W_CallbackFunc *callback = sig->callbacks[job.id];
+    W_CallbackFunc *callback = sig->callbacks[id];
     callback(name);
 }
 
@@ -65,7 +64,7 @@ JobDef(W_RunCallbackJob, job)
  * & dispatching watch callbacks into two separate jobs so that we can delay
  * the dispatch, allowing for deduplication of file modification notifications. */
 
-JobDef(W_MonitorJob, _)
+JobDef(W_MonitorJob, UNUSED sig, UNUSED job_id)
 {
     TempArena scratch = BeginScratchNoConflict();
     W_SharedState *g = &W_shared_state;
@@ -81,7 +80,7 @@ JobDef(W_MonitorJob, _)
         P_WatchInfoList info_list = P_ReadWatchWait(temp.arena, g->watch);
         if (info_list.first && !Atomic32Fetch(&g->W_Shutdown))
         {
-            P_Lock lock = P_LockE(&g->watch_dispatcher_mutex);
+            Lock lock = LockE(&g->watch_dispatcher_mutex);
             {
                 for (P_WatchInfo *info = info_list.first; info; info = info->next)
                 {
@@ -111,8 +110,8 @@ JobDef(W_MonitorJob, _)
                     }
                 }
             }
-            P_SignalCv(&g->watch_dispatcher_cv, I32Max);
-            P_Unlock(&lock);
+            SignalCv(&g->watch_dispatcher_cv, I32Max);
+            Unlock(&lock);
         }
         EndTempArena(temp);
     }
@@ -123,7 +122,7 @@ JobDef(W_MonitorJob, _)
 ////////////////////////////////
 //~ Dispatcher job
 
-JobDef(W_DispatcherJob, _)
+JobDef(W_DispatcherJob, UNUSED sig, UNUSED job_id)
 {
     W_SharedState *g = &W_shared_state;
 
@@ -138,12 +137,12 @@ JobDef(W_DispatcherJob, _)
             /* Delay so that duplicate events pile up */
             {
                 __profn("Delay");
-                P_Wait(0, 0, 0, NsFromSeconds(W_DispatcherDelaySeconds));
+                FutexWait(0, 0, 0, NsFromSeconds(W_DispatcherDelaySeconds));
             }
 
             /* Pull watch events from queue */
             {
-                P_Lock lock = P_LockE(&g->watch_dispatcher_mutex);
+                Lock lock = LockE(&g->watch_dispatcher_mutex);
                 for (W_Event *src_event = g->first_watch_event; src_event; src_event = src_event->next)
                 {
                     W_Event *e = PushStruct(scratch.arena, W_Event);
@@ -161,13 +160,13 @@ JobDef(W_DispatcherJob, _)
                 g->first_watch_event = 0;
                 g->last_watch_event = 0;
                 ResetArena(g->watch_events_arena);
-                P_Unlock(&lock);
+                Unlock(&lock);
             }
 
             /* Build callbacks array */
             u64 num_callbacks = 0;
             W_CallbackFunc **callbacks = 0;
-            P_Lock callbacks_lock = P_LockS(&g->watch_callbacks_mutex);
+            Lock callbacks_lock = LockS(&g->watch_callbacks_mutex);
             {
                 num_callbacks = g->num_watch_callbacks;
                 callbacks = PushStructsNoZero(scratch.arena, W_CallbackFunc *, num_callbacks);
@@ -176,7 +175,7 @@ JobDef(W_DispatcherJob, _)
                     callbacks[i] = g->watch_callbacks[i];
                 }
             }
-            P_Unlock(&callbacks_lock);
+            Unlock(&callbacks_lock);
 
             /* Run callbacks */
             {
@@ -197,12 +196,9 @@ JobDef(W_DispatcherJob, _)
                     }
                     if (!skip)
                     {
-                        W_RunCallbackJobSig sig = ZI;
-                        sig.name = e->name;
-                        sig.callbacks = callbacks;
                         Counter counter = ZI;
-                        P_Run(num_callbacks, W_RunCallbackJob, &sig, PoolKind_Background, PriorityKind_Low, &counter);
-                        P_WaitOnCounter(&counter);
+                        RunJob(num_callbacks, W_RunCallbacksJob, PoolKind_Background, PriorityKind_Low, &counter, .name = e->name, .callbacks = callbacks);
+                        WaitOnCounter(&counter);
                     }
                 }
             }
@@ -211,15 +207,15 @@ JobDef(W_DispatcherJob, _)
         }
 
         /* Wait for event */
-        P_Lock lock = P_LockS(&g->watch_dispatcher_mutex);
+        Lock lock = LockS(&g->watch_dispatcher_mutex);
         {
             shutdown = Atomic32Fetch(&g->W_Shutdown);
             while (!shutdown && !g->first_watch_event)
             {
-                P_WaitOnCv(&g->watch_dispatcher_cv, &lock);
+                WaitOnCv(&g->watch_dispatcher_cv, &lock);
                 shutdown = Atomic32Fetch(&g->W_Shutdown);
             }
         }
-        P_Unlock(&lock);
+        Unlock(&lock);
     }
 }
diff --git a/src/watch/watch_core.h b/src/watch/watch_core.h
index 6aadb871..f3ec6cc3 100644
--- a/src/watch/watch_core.h
+++ b/src/watch/watch_core.h
@@ -4,12 +4,6 @@
 #define W_CallbackFuncDef(func_name, arg_name) void func_name(String arg_name)
 typedef W_CallbackFuncDef(W_CallbackFunc, name);
 
-Struct(W_RunCallbackJobSig)
-{
-    String name;
-    W_CallbackFunc **callbacks;
-};
-
 ////////////////////////////////
 //~ Event types
 
@@ -31,13 +25,13 @@ Struct(W_SharedState)
     Atomic32 W_Shutdown;
     Counter watch_jobs_counter;
 
-    P_Mutex watch_dispatcher_mutex;
+    Mutex watch_dispatcher_mutex;
     Arena *watch_events_arena;
     W_Event *first_watch_event;
     W_Event *last_watch_event;
-    P_Cv watch_dispatcher_cv;
+    Cv watch_dispatcher_cv;
 
-    P_Mutex watch_callbacks_mutex;
+    Mutex watch_callbacks_mutex;
     W_CallbackFunc *watch_callbacks[64];
     u64 num_watch_callbacks;
 };
@@ -58,7 +52,7 @@ void W_RegisterCallback(W_CallbackFunc *callback);
 ////////////////////////////////
 //~ Callback job
 
-JobDecl(W_RunCallbackJob, EmptySig);
+JobDecl(W_RunCallbacksJob, { String name; W_CallbackFunc **callbacks; });
 
 ////////////////////////////////
 //~ Long running jobs