From 8aa9f1402bed9140dba0997591009297cc2e2661 Mon Sep 17 00:00:00 2001
From: jacob <jacob@cagori.com>
Date: Wed, 28 May 2025 03:24:26 -0500
Subject: [PATCH] use float2x3 for xforms in shaders

---
 res/shaders/common.hlsl  | 5 -----
 res/shaders/grid.hlsl    | 8 +++-----
 res/shaders/texture.hlsl | 8 +++-----
 src/common.h             | 2 +-
 src/gpu_dx11.c           | 6 ++++--
 5 files changed, 11 insertions(+), 18 deletions(-)

diff --git a/res/shaders/common.hlsl b/res/shaders/common.hlsl
index 769bba61..71d77684 100644
--- a/res/shaders/common.hlsl
+++ b/res/shaders/common.hlsl
@@ -23,8 +23,3 @@ float4 linear_from_srgb32(uint srgb32)
     res.a = ((srgb32 >> 24) & 0xFF) / 255.0;
     return linear_from_srgb(res);
 }
-
-float2 xform_mul(struct xform xf, float2 v)
-{
-    return xf.bx * v.x + xf.by * v.y + xf.og;
-}
\ No newline at end of file
diff --git a/res/shaders/grid.hlsl b/res/shaders/grid.hlsl
index c01e7509..1c7b6e7a 100644
--- a/res/shaders/grid.hlsl
+++ b/res/shaders/grid.hlsl
@@ -1,7 +1,7 @@
 #include "shaders/common.hlsl"
 
 struct vs_instance {
-    struct xform xf;
+    float2x3 xf;
     float line_thickness;
     float line_spacing;
     float2 offset;
@@ -51,12 +51,10 @@ ps_input vs_main(uint instance_id : SV_InstanceID, uint vertex_id : SV_VertexID)
 {
     vs_instance instance = G_instance_buffer[G_instance_offset + instance_id];
     float2 vert = G_quad_verts[vertex_id];
-
-    float2 world_pos = xform_mul(instance.xf, vert);
-    float4 screen_pos = mul(G_projection, float4(world_pos, 0, 1));
+    float2 world_pos = mul(instance.xf, float3(vert, 1)).xy;
 
     ps_input output;
-    output.screen_pos = screen_pos;
+    output.screen_pos = mul(G_projection, float4(world_pos, 0, 1));
     output.line_thickness = instance.line_thickness;
     output.line_spacing = instance.line_spacing;
     output.offset = instance.offset;
diff --git a/res/shaders/texture.hlsl b/res/shaders/texture.hlsl
index 1a65e7a4..db910821 100644
--- a/res/shaders/texture.hlsl
+++ b/res/shaders/texture.hlsl
@@ -1,7 +1,7 @@
 #include "shaders/common.hlsl"
 
 struct vs_instance {
-    struct xform xf;
+    float2x3 xf;
     float2 uv0;
     float2 uv1;
     uint tint_srgb;
@@ -51,13 +51,11 @@ ps_input vs_main(uint instance_id : SV_InstanceID, uint vertex_id : SV_VertexID)
 {
     vs_instance instance = G_instance_buffer[G_instance_offset + instance_id];
     float2 vert = G_quad_verts[vertex_id];
+    float2 world_pos = mul(instance.xf, float3(vert, 1)).xy;
     float2 uv_factor  = G_uv_factors[vertex_id];
 
-    float2 world_pos = xform_mul(instance.xf, vert);
-    float4 screen_pos = mul(G_projection, float4(world_pos, 0, 1));
-
     ps_input output;
-    output.screen_pos = screen_pos;
+    output.screen_pos = mul(G_projection, float4(world_pos, 0, 1));
     output.uv = instance.uv0 + uv_factor * (instance.uv1 - instance.uv0);
     output.tint_lin = linear_from_srgb32(instance.tint_srgb);
 
diff --git a/src/common.h b/src/common.h
index 2e6cbcc2..682130e5 100644
--- a/src/common.h
+++ b/src/common.h
@@ -655,7 +655,7 @@ INLINE void __prof_dx11_zone_cleanup_func(TracyCD3D11ZoneCtx *ctx) { ___tracy_d3
 
 #define __profalloc(ptr, size)      TracyCAlloc((ptr), (size))
 #define __proffree(ptr)             TracyCFree((ptr))
-#define __profmsg(txt, len, col)    TracyCMessageC((txt), (len), (col));
+#define __profmsg(txt, len, col)    TracyCMessageC((txt), (len), BGR(col));
 #define __profframe(name)           TracyCFrameMarkNamed((name))
 #define __profthread(name)          TracyCSetThreadName((name))
 
diff --git a/src/gpu_dx11.c b/src/gpu_dx11.c
index 632c459b..9fb3fb27 100644
--- a/src/gpu_dx11.c
+++ b/src/gpu_dx11.c
@@ -731,7 +731,7 @@ INTERNAL struct string shader_alloc(struct arena *arena, struct dx11_shader *sha
 
     struct dx11_include_handler include_handler = dx11_include_handler_alloc(shader);
 
-    u32 flags = D3DCOMPILE_PACK_MATRIX_COLUMN_MAJOR;
+    u32 flags = 0;
 #if DX11_SHADER_DEBUG
     flags |= D3DCOMPILE_DEBUG | D3DCOMPILE_SKIP_OPTIMIZATION | D3DCOMPILE_ENABLE_STRICTNESS | D3DCOMPILE_WARNINGS_ARE_ERRORS;
 #else
@@ -952,7 +952,6 @@ INTERNAL void dx11_buffer_release(struct dx11_buffer *buffer)
 
 INTERNAL void *dx11_buffer_push(struct dx11_buffer *buffer, u64 size)
 {
-    __prof;
     void *data = arena_push_array_no_zero(&buffer->cpu_buffer_arena, u8, size);
     return data;
 }
@@ -1017,6 +1016,7 @@ INTERNAL void dx11_buffer_submit(struct dx11_buffer *buffer)
 
 struct gpu_cmd_store gpu_cmd_store_alloc(void)
 {
+    __prof;
     struct dx11_cmd_store *store = NULL;
     {
         struct arena cpu_cmds_arena = ZI;
@@ -1107,12 +1107,14 @@ struct gpu_cmd_store gpu_cmd_store_alloc(void)
 
 void gpu_cmd_store_release(struct gpu_cmd_store gpu_cmd_store)
 {
+    __prof;
     /* TODO */
     (UNUSED)gpu_cmd_store;
 }
 
 void gpu_push_cmd(struct gpu_cmd_store gpu_cmd_store, struct gpu_cmd_params params)
 {
+    __prof;
     struct dx11_cmd_store *store = (struct dx11_cmd_store *)gpu_cmd_store.handle;
 
     switch (params.kind) {