Implement for metal `SV_GroupIndex` (#4385)

* Implement for metal `SV_GroupIndex`

 1. If we don't have `sv_GroupThreadId` available we create one using `SV_GroupIndex`s location data.
 2. We emit code emulating `sv_GroupThreadId` from the same logic that CUDA/CPP uses.

* address most review comments

Addressed all but two: [1](https://github.com/shader-slang/slang/pull/4385#discussion_r1639058473) and [2](https://github.com/shader-slang/slang/pull/4385#issuecomment-2166934855)

I want to enable tests and be sure there is no bugs using CI before I redesign the code so I have a working fallback.

* address comment, enable tests

enable now functioning tests due to `SV_GroupIndex` working with metal

* syntax error with groupThreadID search

did `= param` instead of `= i.param`

* add `sv_groupid` for test + test fixes

* disable test that won't work regardless

11 files changed, 38 insertions, 10 deletions

diff --git a/tests/bugs/frexp-double.slang b/tests/bugs/frexp-double.slang
index 40623a17b..c2b4f21c7 100644
--- a/tests/bugs/frexp-double.slang
+++ b/tests/bugs/frexp-double.slang
@@ -4,7 +4,8 @@
 //TEST(smoke,compute):COMPARE_COMPUTE(filecheck-buffer=BUF):-vk -output-using-type  -render-feature double
 //TEST(smoke,compute):COMPARE_COMPUTE(filecheck-buffer=BUF):-vk -emit-spirv-directly -output-using-type  -render-feature double
 //TEST(smoke,compute):COMPARE_COMPUTE(filecheck-buffer=BUF):-cuda -output-using-type
-//DISABLE_TEST(compute):COMPARE_COMPUTE:-slang -shaderobj -mtl
+//metal currently does not support `double` 
+//DISABLE_TEST(smoke,compute):COMPARE_COMPUTE(filecheck-buffer=BUF):-slang -shaderobj -mtl -output-using-type
 
 // BUF:      type: int32_t
 // BUF-NEXT: 1
diff --git a/tests/bugs/frexp.slang b/tests/bugs/frexp.slang
index 01c345d1e..60912297c 100644
--- a/tests/bugs/frexp.slang
+++ b/tests/bugs/frexp.slang
@@ -4,7 +4,7 @@
 //TEST(smoke,compute):COMPARE_COMPUTE(filecheck-buffer=BUF):-vk -output-using-type
 //TEST(smoke,compute):COMPARE_COMPUTE(filecheck-buffer=BUF):-vk -emit-spirv-directly -output-using-type
 //TEST(smoke,compute):COMPARE_COMPUTE(filecheck-buffer=BUF):-cuda -output-using-type
-//DISABLE_TEST(compute):COMPARE_COMPUTE:-slang -shaderobj -mtl
+//TEST(smoke,compute):COMPARE_COMPUTE(filecheck-buffer=BUF):-slang -shaderobj -mtl -output-using-type
 
 // BUF:      type: int32_t
 // BUF-NEXT: 1
diff --git a/tests/bugs/gh-3980.slang b/tests/bugs/gh-3980.slang
index 509212ea9..57cd28dbe 100644
--- a/tests/bugs/gh-3980.slang
+++ b/tests/bugs/gh-3980.slang
@@ -4,7 +4,7 @@
 //TEST(compute):COMPARE_COMPUTE_EX(filecheck-buffer=BUF):-vk -compute -shaderobj -output-using-type -emit-spirv-directly
 //TEST(compute):COMPARE_COMPUTE_EX(filecheck-buffer=BUF):-cpu -compute -output-using-type
 //TEST(compute):COMPARE_COMPUTE_EX(filecheck-buffer=BUF):-cuda -compute -output-using-type
-//DISABLE_TEST(compute):COMPARE_COMPUTE:-slang -shaderobj -mtl
+//TEST(compute):COMPARE_COMPUTE(filecheck-buffer=BUF):-slang -shaderobj -mtl -output-using-type
 
 // Slang removes parentheses characters for the bitwise operators when they are not needed.
 // DXC prints warning messages even when the expression is correct.
diff --git a/tests/compute/texture-simpler.slang b/tests/compute/texture-simpler.slang
index 196e8e96c..0f9dfcbad 100644
--- a/tests/compute/texture-simpler.slang
+++ b/tests/compute/texture-simpler.slang
@@ -4,7 +4,7 @@
 //TEST(smoke,compute):COMPARE_COMPUTE_EX:-slang -compute -dx12 -profile cs_6_0 -use-dxil -shaderobj -output-using-type
 //TEST(smoke,compute, vulkan):COMPARE_COMPUTE_EX:-vk -compute -shaderobj -output-using-type -render-feature hardware-device
 //TEST(smoke,compute):COMPARE_COMPUTE_EX:-cuda -compute  -shaderobj -output-using-type
-//DISABLE_TEST(compute):COMPARE_COMPUTE:-slang -shaderobj -mtl
+//TEST(smoke,compute):COMPARE_COMPUTE:-slang -shaderobj -mtl -output-using-type
 
 //TEST_INPUT: Texture2D(size=4, content = one):name t2D
 Texture2D<float> t2D;
diff --git a/tests/language-feature/higher-order-functions/simple.slang b/tests/language-feature/higher-order-functions/simple.slang
index 8a3544b91..13fc16aa5 100644
--- a/tests/language-feature/higher-order-functions/simple.slang
+++ b/tests/language-feature/higher-order-functions/simple.slang
@@ -1,7 +1,7 @@
 //TEST(compute, vulkan):COMPARE_COMPUTE_EX():-vk -compute -shaderobj -output-using-type
 //TEST(compute):COMPARE_COMPUTE_EX():-slang -compute -shaderobj -output-using-type
 //TEST(compute):COMPARE_COMPUTE:-cpu -shaderobj -output-using-type
-//DISABLE_TEST(compute):COMPARE_COMPUTE:-slang -shaderobj -mtl
+//TEST(compute):COMPARE_COMPUTE:-slang -shaderobj -mtl -output-using-type
 
 //TEST_INPUT:ubuffer(data=[0 0 0 0], stride=4):out,name=outputBuffer
 RWStructuredBuffer<uint> outputBuffer;
diff --git a/tests/language-feature/swizzles/matrix-swizzle-write-array.slang b/tests/language-feature/swizzles/matrix-swizzle-write-array.slang
index e9510de06..616a19b19 100644
--- a/tests/language-feature/swizzles/matrix-swizzle-write-array.slang
+++ b/tests/language-feature/swizzles/matrix-swizzle-write-array.slang
@@ -1,6 +1,6 @@
 //TEST(compute):COMPARE_COMPUTE: -compute -shaderobj -output-using-type
 //TEST(compute, vulkan):COMPARE_COMPUTE: -vk -compute -shaderobj -output-using-type
-//DISABLE_TEST(compute):COMPARE_COMPUTE:-slang -shaderobj -mtl
+//TEST(compute):COMPARE_COMPUTE:-slang -shaderobj -mtl -output-using-type
 
 // Test that matrix swizzle writes work correctly
 // Matrix swizzles can either be one or zero indexed
diff --git a/tests/language-feature/swizzles/matrix-swizzle-write-single.slang b/tests/language-feature/swizzles/matrix-swizzle-write-single.slang
index d8bb11ea5..34d54ac55 100644
--- a/tests/language-feature/swizzles/matrix-swizzle-write-single.slang
+++ b/tests/language-feature/swizzles/matrix-swizzle-write-single.slang
@@ -1,6 +1,6 @@
 //TEST(compute):COMPARE_COMPUTE: -compute -shaderobj -output-using-type
 //TEST(compute, vulkan):COMPARE_COMPUTE: -vk -compute -shaderobj -output-using-type
-//DISABLE_TEST(compute):COMPARE_COMPUTE:-slang -shaderobj -mtl
+//TEST(compute):COMPARE_COMPUTE:-slang -shaderobj -mtl -output-using-type
 
 // Test that writes to single matrix elements with swizzles work
 
diff --git a/tests/language-feature/swizzles/matrix-swizzle-write-swizzle.slang b/tests/language-feature/swizzles/matrix-swizzle-write-swizzle.slang
index b73b14249..f1ad7bc6a 100644
--- a/tests/language-feature/swizzles/matrix-swizzle-write-swizzle.slang
+++ b/tests/language-feature/swizzles/matrix-swizzle-write-swizzle.slang
@@ -1,6 +1,6 @@
 //TEST(compute):COMPARE_COMPUTE: -compute -shaderobj -output-using-type
 //TEST(compute, vulkan):COMPARE_COMPUTE: -vk -compute -shaderobj -output-using-type
-//DISABLE_TEST(compute):COMPARE_COMPUTE:-slang -shaderobj -mtl
+//TEST(compute):COMPARE_COMPUTE:-slang -shaderobj -mtl -output-using-type
 
 // Test that writing to swizzles of matrix swizzles works correctly
 
diff --git a/tests/language-feature/swizzles/matrix-swizzle-write.slang b/tests/language-feature/swizzles/matrix-swizzle-write.slang
index d467d33db..6ccb4c29c 100644
--- a/tests/language-feature/swizzles/matrix-swizzle-write.slang
+++ b/tests/language-feature/swizzles/matrix-swizzle-write.slang
@@ -1,6 +1,6 @@
 //TEST(compute):COMPARE_COMPUTE: -compute -shaderobj -output-using-type
 //TEST(compute, vulkan):COMPARE_COMPUTE: -vk -compute -shaderobj -output-using-type
-//DISABLE_TEST(compute):COMPARE_COMPUTE:-slang -shaderobj -mtl
+//TEST(compute):COMPARE_COMPUTE:-slang -shaderobj -mtl -output-using-type
 
 // Test that matrix swizzle writes work correctly
 // Matrix swizzles can either be one or zero indexed
diff --git a/tests/metal/thread_position_in_threadgroup.slang b/tests/metal/thread_position_in_threadgroup.slang
new file mode 100644
index 000000000..a20fe2ec7
--- /dev/null
+++ b/tests/metal/thread_position_in_threadgroup.slang
@@ -0,0 +1,27 @@
+//TEST:SIMPLE(filecheck=CHECK): -entry computeMain -stage compute -target metal -D GROUPID
+//TEST:SIMPLE(filecheck=CHECK): -entry computeMain -stage compute -target metal
+
+//CHECK: computeMain
+
+// ensure we compute the SV_GroupIndex from SV_GroupThreadID and `numthreads`
+// CHECK: thread_position_in_threadgroup
+// CHECK-DAG: *{{.*}}2
+
+RWBuffer<uint> dst;
+
+void indirection(uint groupIndex)
+{
+    dst[groupIndex] = groupIndex;
+}
+
+#define THREAD_COUNT 2
+[numthreads(THREAD_COUNT, 1, 1)]
+#ifdef GROUPID
+void computeMain(uint GI : SV_GroupIndex, uint GTID : SV_GroupThreadID)
+#else
+void computeMain(uint GI : SV_GroupIndex)
+#endif
+{
+    dst[GI + THREAD_COUNT] = GI;
+    indirection(GI);
+} 
\ No newline at end of file
diff --git a/tests/pipeline/compute/compute-system-values.slang b/tests/pipeline/compute/compute-system-values.slang
index f7aef06ff..912827557 100644
--- a/tests/pipeline/compute/compute-system-values.slang
+++ b/tests/pipeline/compute/compute-system-values.slang
@@ -2,7 +2,7 @@
 
 //TEST(compute):COMPARE_COMPUTE: -shaderobj
 //TEST(compute):COMPARE_COMPUTE: -cpu -shaderobj
-//DISABLE_TEST(compute):COMPARE_COMPUTE:-slang -shaderobj -mtl
+//TEST(compute):COMPARE_COMPUTE:-slang -shaderobj -mtl
 
 //TEST_INPUT:ubuffer(data=[0 0 0 0 0 0 0 0], stride=4):out,name=outputBuffer
 RWStructuredBuffer<int> outputBuffer;