More control flow simplifications. (#2673)

* More control flow and Phi param simplifications.

* Fix.

* Fix gcc error.

* Fix.

* More IR cleanup.

* Fix bug in phi param dce + ifelse simplify.

* Propagate and DCE side-effect-free functions.

* Enhance CFG simplifcation to remove loops with no side effects.

* Fix.

* Fixes.

* Fix tests. Add [__AlwaysFoldIntoUseSite] for rayPayloadLocation.

* More cleanup.

* Fixes.

* Fix.

---------

Co-authored-by: Yong He <yhe@nvidia.com>

5 files changed, 47 insertions, 169 deletions

diff --git a/tests/pipeline/rasterization/fragment-shader-interlock.slang.glsl b/tests/pipeline/rasterization/fragment-shader-interlock.slang.glsl
index 84eba46f0..7f53576e9 100644
--- a/tests/pipeline/rasterization/fragment-shader-interlock.slang.glsl
+++ b/tests/pipeline/rasterization/fragment-shader-interlock.slang.glsl
@@ -1,10 +1,7 @@
-//TEST_IGNORE_FILE:
-
 #version 450
 #extension GL_ARB_fragment_shader_interlock : require
 layout(row_major) uniform;
 layout(row_major) buffer;
-
 layout(rgba32f)
 layout(binding = 0)
 uniform image2D entryPointParams_texture_0;
@@ -17,15 +14,12 @@ out vec4 _S2;
 
 void main()
 {
-    beginInvocationInterlockARB();
-
-    vec2 _S3 = _S1.xy;
-
-    vec4 _S4 = (imageLoad((entryPointParams_texture_0), ivec2((uvec2(_S3)))));
-    imageStore((entryPointParams_texture_0), ivec2((uvec2(_S3))), _S4 + _S1);
 
+    beginInvocationInterlockARB();
+    uvec2 _S3 = uvec2(_S1.xy);
+    vec4 _S4 = (imageLoad((entryPointParams_texture_0), ivec2((_S3))));
+    imageStore((entryPointParams_texture_0), ivec2((_S3)), _S4 + _S1);
     endInvocationInterlockARB();
-
     _S2 = _S4;
     return;
 }
diff --git a/tests/pipeline/rasterization/mesh/passing-outputs.slang.glsl b/tests/pipeline/rasterization/mesh/passing-outputs.slang.glsl
index 31c2f0db2..1102a838e 100644
--- a/tests/pipeline/rasterization/mesh/passing-outputs.slang.glsl
+++ b/tests/pipeline/rasterization/mesh/passing-outputs.slang.glsl
@@ -160,16 +160,10 @@ void main()
         d_0(gl_LocalInvocationIndex);
         e_0(gl_LocalInvocationIndex);
     }
-    else
-    {
-    }
     if(gl_LocalInvocationIndex < 1U)
     {
         gl_PrimitiveTriangleIndicesEXT[gl_LocalInvocationIndex] = uvec3(0U, 1U, 2U);
     }
-    else
-    {
-    }
     return;
 }
 
diff --git a/tests/pipeline/ray-tracing/acceleration-structure-in-compute.slang.glsl b/tests/pipeline/ray-tracing/acceleration-structure-in-compute.slang.glsl
index f95321039..83797d2d5 100644
--- a/tests/pipeline/ray-tracing/acceleration-structure-in-compute.slang.glsl
+++ b/tests/pipeline/ray-tracing/acceleration-structure-in-compute.slang.glsl
@@ -1,6 +1,7 @@
 #version 460
 #extension GL_EXT_ray_tracing : require
-
+layout(row_major) uniform;
+layout(row_major) buffer;
 int helper_0(accelerationStructureEXT a_0, int b_0)
 {
     return b_0;
diff --git a/tests/pipeline/ray-tracing/trace-ray-inline.slang.glsl b/tests/pipeline/ray-tracing/trace-ray-inline.slang.glsl
index 0364d2513..1c2bc8090 100644
--- a/tests/pipeline/ray-tracing/trace-ray-inline.slang.glsl
+++ b/tests/pipeline/ray-tracing/trace-ray-inline.slang.glsl
@@ -3,7 +3,6 @@
 #extension GL_EXT_ray_query : require
 layout(row_major) uniform;
 layout(row_major) buffer;
-
 struct SLANG_ParameterGroup_C_0
 {
     vec3 origin_0;
@@ -20,7 +19,6 @@ layout(std140) uniform _S1
 {
     SLANG_ParameterGroup_C_0 _data;
 } C_0;
-
 layout(binding = 0)
 uniform accelerationStructureEXT myAccelerationStructure_0;
 
@@ -70,40 +68,32 @@ void myMiss_0(inout MyRayPayload_0 payload_4)
 layout(local_size_x = 1, local_size_y = 1, local_size_z = 1) in;
 void main()
 {
+    MyProceduralHitAttrs_0 committedProceduralAttrs_0;
 
     rayQueryEXT query_0;
 
     MyRayPayload_0 payload_5;
     payload_5.value_1 = -1;
     rayQueryInitializeEXT((query_0), (myAccelerationStructure_0), (C_0._data.rayFlags_0 | 512), (C_0._data.instanceMask_0), (C_0._data.origin_0), (C_0._data.tMin_0), (C_0._data.direction_0), (C_0._data.tMax_0));
-
-    MyProceduralHitAttrs_0 committedProceduralAttrs_0;
-
     for(;;)
     {
-
         bool _S2 = rayQueryProceedEXT(query_0);
-
         if(!_S2)
         {
             break;
         }
         uint _S3 = (rayQueryGetIntersectionTypeEXT((query_0), false));
-
         switch(_S3)
         {
         case 1U:
             {
                 MyProceduralHitAttrs_0 candidateProceduralAttrs_0;
-
                 candidateProceduralAttrs_0.value_0 = 0;
                 float tHit_1 = 0.0;
                 bool _S4 = myProceduralIntersection_0(tHit_1, candidateProceduralAttrs_0);
-
                 if(_S4)
                 {
                     bool _S5 = myProceduralAnyHit_0(payload_5);
-
                     if(_S5)
                     {
                         rayQueryGenerateIntersectionEXT(query_0, tHit_1);
@@ -112,35 +102,22 @@ void main()
                         {
                             rayQueryTerminateEXT(query_0);
                         }
-                        else
-                        {
-                        }
-
                         committedProceduralAttrs_0 = _S6;
-
                     }
                     else
                     {
-
                         committedProceduralAttrs_0 = committedProceduralAttrs_0;
-
                     }
-
                 }
                 else
                 {
-
                     committedProceduralAttrs_0 = committedProceduralAttrs_0;
-
                 }
-
                 break;
             }
         case 0U:
             {
-
                 bool _S7 = myTriangleAnyHit_0(payload_5);
-
                 if(_S7)
                 {
                     rayQueryConfirmIntersectionEXT(query_0);
@@ -148,12 +125,6 @@ void main()
                     {
                         rayQueryTerminateEXT(query_0);
                     }
-                    else
-                    {
-                    }
-                }
-                else
-                {
                 }
                 break;
             }
@@ -162,13 +133,8 @@ void main()
                 break;
             }
         }
-
-        committedProceduralAttrs_0 = committedProceduralAttrs_0;
-
     }
-
     uint _S8 = (rayQueryGetIntersectionTypeEXT((query_0), true));
-
     switch(_S8)
     {
     case 1U:
diff --git a/tests/pipeline/ray-tracing/trace-ray-inline.slang.hlsl b/tests/pipeline/ray-tracing/trace-ray-inline.slang.hlsl
index 97d972328..b0c798b2d 100644
--- a/tests/pipeline/ray-tracing/trace-ray-inline.slang.hlsl
+++ b/tests/pipeline/ray-tracing/trace-ray-inline.slang.hlsl
@@ -1,11 +1,14 @@
-// trace-ray-inline.slang.hlsl
-//TEST_IGNORE_FILE:
+#pragma pack_matrix(column_major)
+#ifdef SLANG_HLSL_ENABLE_NVAPI
+#include "nvHLSLExtns.h"
+#endif
+#pragma warning(disable: 3557)
 
 struct SLANG_ParameterGroup_C_0
 {
-    vector<float,3> origin_0;
+    float3 origin_0;
     float tMin_0;
-    vector<float,3> direction_0;
+    float3 direction_0;
     float tMax_0;
     uint rayFlags_0;
     uint instanceMask_0;
@@ -16,7 +19,6 @@ cbuffer C_0 : register(b0)
 {
     SLANG_ParameterGroup_C_0 C_0;
 }
-
 RaytracingAccelerationStructure myAccelerationStructure_0 : register(t0);
 
 struct MyProceduralHitAttrs_0
@@ -62,170 +64,92 @@ void myMiss_0(inout MyRayPayload_0 payload_4)
     return;
 }
 
-
-[shader("compute")]
-[numthreads(1, 1, 1)]
-void main(vector<uint,3> tid_0 : SV_DISPATCHTHREADID)
+[shader("compute")][numthreads(1, 1, 1)]
+void main(uint3 tid_0 : SV_DISPATCHTHREADID)
 {
-    MyRayPayload_0 payload_5;
     MyProceduralHitAttrs_0 committedProceduralAttrs_0;
-    MyProceduralHitAttrs_0 committedProceduralAttrs_1;
-    MyRayPayload_0 payload_6;
-    MyProceduralHitAttrs_0 committedProceduralAttrs_2;
-    MyRayPayload_0 payload_7;
-    MyProceduralHitAttrs_0 committedProceduralAttrs_3;
 
     RayQuery<int(512) > query_0;
 
-    MyRayPayload_0 _S1 = { int(-1) };
+    MyRayPayload_0 payload_5;
+    payload_5.value_1 = int(-1);
     RayDesc ray_0 = { C_0.origin_0, C_0.tMin_0, C_0.direction_0, C_0.tMax_0 };
     query_0.TraceRayInline(myAccelerationStructure_0, C_0.rayFlags_0, C_0.instanceMask_0, ray_0);
-
-    MyProceduralHitAttrs_0 _S2;
-
-    payload_5 = _S1;
-    committedProceduralAttrs_0 = _S2;
     for(;;)
     {
-        bool _S3 = query_0.Proceed();
-
-        if(!_S3)
+        bool _S1 = query_0.Proceed();
+        if(!_S1)
         {
             break;
         }
-        uint _S4 = query_0.CandidateType();
-
-        switch(_S4)
+        uint _S2 = query_0.CandidateType();
+        switch(_S2)
         {
-        case (uint) int(1):
+        case 1U:
             {
-                MyProceduralHitAttrs_0 candidateProceduralAttrs_0 = { int(0) };
-
-                float _S5;
-
-                _S5 = 0.00000000000000000000;
-
-                MyProceduralHitAttrs_0 _S6;
-
-                _S6 = candidateProceduralAttrs_0;
-
-                bool _S7 = myProceduralIntersection_0(_S5, _S6);
-
-                float tHit_1 = _S5;
-
-                MyProceduralHitAttrs_0 candidateProceduralAttrs_1 = _S6;
-
-                if(_S7)
+                MyProceduralHitAttrs_0 candidateProceduralAttrs_0;
+                candidateProceduralAttrs_0.value_0 = int(0);
+                float tHit_1 = 0.0;
+                bool _S3 = myProceduralIntersection_0(tHit_1, candidateProceduralAttrs_0);
+                if(_S3)
                 {
-                    MyRayPayload_0 _S8;
-
-                    _S8 = payload_5;
-
-                    bool _S9 = myProceduralAnyHit_0(_S8);
-
-                    MyRayPayload_0 _S10 = _S8;
-
-                    if(_S9)
+                    bool _S4 = myProceduralAnyHit_0(payload_5);
+                    if(_S4)
                     {
                         query_0.CommitProceduralPrimitiveHit(tHit_1);
-
-                        if((bool) C_0.shouldStopAtFirstHit_0)
+                        MyProceduralHitAttrs_0 _S5 = candidateProceduralAttrs_0;
+                        if(C_0.shouldStopAtFirstHit_0 != 0U)
                         {
-
                             query_0.Abort();
                         }
-                        else
-                        {
-                        }
-
-                        committedProceduralAttrs_1 = candidateProceduralAttrs_1;
+                        committedProceduralAttrs_0 = _S5;
                     }
                     else
                     {
-                        committedProceduralAttrs_1 = committedProceduralAttrs_0;
+                        committedProceduralAttrs_0 = committedProceduralAttrs_0;
                     }
-
-                    payload_6 = _S10;
-                    committedProceduralAttrs_2 = committedProceduralAttrs_1;
                 }
                 else
                 {
-                    payload_6 = payload_5;
-                    committedProceduralAttrs_2 = committedProceduralAttrs_0;
+                    committedProceduralAttrs_0 = committedProceduralAttrs_0;
                 }
-
-                payload_7 = payload_6;
-                committedProceduralAttrs_3 = committedProceduralAttrs_2;
                 break;
             }
-        case (uint) int(0):
+        case 0U:
             {
-                MyRayPayload_0 _S11;
-                _S11 = payload_5;
-
-                bool _S12 = myTriangleAnyHit_0(_S11);
-                MyRayPayload_0 _S13 = _S11;
-
-                if(_S12)
+                bool _S6 = myTriangleAnyHit_0(payload_5);
+                if(_S6)
                 {
                     query_0.CommitNonOpaqueTriangleHit();
-                    if((bool) C_0.shouldStopAtFirstHit_0)
+                    if(C_0.shouldStopAtFirstHit_0 != 0U)
                     {
                         query_0.Abort();
                     }
-                    else
-                    {
-                    }
-                }
-                else
-                {
                 }
-
-                payload_7 = _S13;
-                committedProceduralAttrs_3 = committedProceduralAttrs_0;
                 break;
             }
         default:
             {
-                payload_7 = payload_5;
-                committedProceduralAttrs_3 = committedProceduralAttrs_0;
                 break;
             }
         }
-
-        payload_5 = payload_7;
-        committedProceduralAttrs_0 = committedProceduralAttrs_3;
     }
-
-    uint _S14 = query_0.CommittedStatus();
-
-    switch(_S14)
+    uint _S7 = query_0.CommittedStatus();
+    switch(_S7)
     {
-    case (uint) int(1):
+    case 1U:
         {
-            MyRayPayload_0 _S15;
-
-            _S15 = payload_5;
-
-            myTriangleClosestHit_0(_S15);
+            myTriangleClosestHit_0(payload_5);
             break;
         }
-    case (uint) int(2):
+    case 2U:
         {
-
-            MyRayPayload_0 _S16;
-            _S16 = payload_5;
-
-            myProceduralClosestHit_0(_S16, committedProceduralAttrs_0);
+            myProceduralClosestHit_0(payload_5, committedProceduralAttrs_0);
             break;
         }
-    case (uint) int(0):
+    case 0U:
         {
-            MyRayPayload_0 _S17;
-
-            _S17 = payload_5;
-
-            myMiss_0(_S17);
+            myMiss_0(payload_5);
             break;
         }
     default:
@@ -233,6 +157,5 @@ void main(vector<uint,3> tid_0 : SV_DISPATCHTHREADID)
             break;
         }
     }
-
     return;
 }