[clang] [AMDGPU] Add missing `__builtin_amdgcn_wavefrontsize` builtin (PR #80741)

2024-02-05 Thread Joseph Huber via cfe-commits 

https://github.com/jhuber6 closed 
https://github.com/llvm/llvm-project/pull/80741
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [AMDGPU] Add missing `__builtin_amdgcn_wavefrontsize` builtin (PR #80741)

2024-02-05 Thread Stanislav Mekhanoshin via cfe-commits 


@@ -832,6 +832,13 @@ void test_atomic_inc_dec(local uint *lptr, global uint 
*gptr, uint val) {
   res = __builtin_amdgcn_atomic_dec32((volatile global uint*)gptr, val, 
__ATOMIC_SEQ_CST, "");
 }
 
+// CHECK-LABEL test_wavefrontsize(
+unsigned test_wavefrontsize() {

rampitec wrote:

Ugh, it's inside the body. Unusual, but test above is he same.

https://github.com/llvm/llvm-project/pull/80741
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [AMDGPU] Add missing `__builtin_amdgcn_wavefrontsize` builtin (PR #80741)

2024-02-05 Thread Stanislav Mekhanoshin via cfe-commits 

https://github.com/rampitec approved this pull request.


https://github.com/llvm/llvm-project/pull/80741
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [AMDGPU] Add missing `__builtin_amdgcn_wavefrontsize` builtin (PR #80741)

2024-02-05 Thread Stanislav Mekhanoshin via cfe-commits 

https://github.com/rampitec edited 
https://github.com/llvm/llvm-project/pull/80741
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [AMDGPU] Add missing `__builtin_amdgcn_wavefrontsize` builtin (PR #80741)

2024-02-05 Thread Joseph Huber via cfe-commits 


@@ -832,6 +832,13 @@ void test_atomic_inc_dec(local uint *lptr, global uint 
*gptr, uint val) {
   res = __builtin_amdgcn_atomic_dec32((volatile global uint*)gptr, val, 
__ATOMIC_SEQ_CST, "");
 }
 
+// CHECK-LABEL test_wavefrontsize(
+unsigned test_wavefrontsize() {

jhuber6 wrote:

Which check? This just makes sure that we generate the appropriate LLVM 
intrinsic below.

https://github.com/llvm/llvm-project/pull/80741
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [AMDGPU] Add missing `__builtin_amdgcn_wavefrontsize` builtin (PR #80741)

2024-02-05 Thread Stanislav Mekhanoshin via cfe-commits 


@@ -832,6 +832,13 @@ void test_atomic_inc_dec(local uint *lptr, global uint 
*gptr, uint val) {
   res = __builtin_amdgcn_atomic_dec32((volatile global uint*)gptr, val, 
__ATOMIC_SEQ_CST, "");
 }
 
+// CHECK-LABEL test_wavefrontsize(
+unsigned test_wavefrontsize() {

rampitec wrote:

Missing check for the test.

https://github.com/llvm/llvm-project/pull/80741
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [AMDGPU] Add missing `__builtin_amdgcn_wavefrontsize` builtin (PR #80741)

2024-02-05 Thread via cfe-commits 

llvmbot wrote:



@llvm/pr-subscribers-clang

@llvm/pr-subscribers-backend-amdgpu

Author: Joseph Huber (jhuber6)


Changes

Summary:
The backend supports the wavefrontsize intrinsic, and suggests that it
is tied to a corresponding clang builtin, but it is not actually
present. This simply adds it in so it can be used from clang. This
attribute likely isn't the best to rely on, but for the `libc` use-case
we will need to detect a struct's differing size in a way that will
depend on the wavefront size.


---
Full diff: https://github.com/llvm/llvm-project/pull/80741.diff


2 Files Affected:

- (modified) clang/include/clang/Basic/BuiltinsAMDGPU.def (+1) 
- (modified) clang/test/CodeGenOpenCL/builtins-amdgcn.cl (+7) 


``diff
diff --git a/clang/include/clang/Basic/BuiltinsAMDGPU.def 
b/clang/include/clang/Basic/BuiltinsAMDGPU.def
index 5f8001e61a028..213311b96df74 100644
--- a/clang/include/clang/Basic/BuiltinsAMDGPU.def
+++ b/clang/include/clang/Basic/BuiltinsAMDGPU.def
@@ -69,6 +69,7 @@ BUILTIN(__builtin_amdgcn_s_dcache_inv, "v", "n")
 BUILTIN(__builtin_amdgcn_buffer_wbinvl1, "v", "n")
 BUILTIN(__builtin_amdgcn_fence, "vUicC*", "n")
 BUILTIN(__builtin_amdgcn_groupstaticsize, "Ui", "n")
+BUILTIN(__builtin_amdgcn_wavefrontsize, "Ui", "nc")
 
 BUILTIN(__builtin_amdgcn_atomic_inc32, "UZiUZiD*UZiUicC*", "n")
 BUILTIN(__builtin_amdgcn_atomic_inc64, "UWiUWiD*UWiUicC*", "n")
diff --git a/clang/test/CodeGenOpenCL/builtins-amdgcn.cl 
b/clang/test/CodeGenOpenCL/builtins-amdgcn.cl
index 8d9e4e018b12e..7d9010ee9067d 100644
--- a/clang/test/CodeGenOpenCL/builtins-amdgcn.cl
+++ b/clang/test/CodeGenOpenCL/builtins-amdgcn.cl
@@ -832,6 +832,13 @@ void test_atomic_inc_dec(local uint *lptr, global uint 
*gptr, uint val) {
   res = __builtin_amdgcn_atomic_dec32((volatile global uint*)gptr, val, 
__ATOMIC_SEQ_CST, "");
 }
 
+// CHECK-LABEL test_wavefrontsize(
+unsigned test_wavefrontsize() {
+
+  // CHECK: call i32 @llvm.amdgcn.wavefrontsize()
+  return __builtin_amdgcn_wavefrontsize();
+}
+
 // CHECK-DAG: [[$WI_RANGE]] = !{i32 0, i32 1024}
 // CHECK-DAG: [[$WS_RANGE]] = !{i16 1, i16 1025}
 // CHECK-DAG: attributes #[[$NOUNWIND_READONLY]] = { convergent mustprogress 
nocallback nofree nounwind willreturn memory(none) }

``




https://github.com/llvm/llvm-project/pull/80741
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [AMDGPU] Add missing `__builtin_amdgcn_wavefrontsize` builtin (PR #80741)

2024-02-05 Thread Joseph Huber via cfe-commits 

https://github.com/jhuber6 created 
https://github.com/llvm/llvm-project/pull/80741

Summary:
The backend supports the wavefrontsize intrinsic, and suggests that it
is tied to a corresponding clang builtin, but it is not actually
present. This simply adds it in so it can be used from clang. This
attribute likely isn't the best to rely on, but for the `libc` use-case
we will need to detect a struct's differing size in a way that will
depend on the wavefront size.


>From c747348cf0b6fd51ec67dbbea64e0bcbe4266faa Mon Sep 17 00:00:00 2001
From: Joseph Huber 
Date: Mon, 5 Feb 2024 14:59:36 -0600
Subject: [PATCH] [AMDGPU] Add missing `__builtin_amdgcn_wavefrontsize` builtin

Summary:
The backend supports the wavefrontsize intrinsic, and suggests that it
is tied to a corresponding clang builtin, but it is not actually
present. This simply adds it in so it can be used from clang. This
attribute likely isn't the best to rely on, but for the `libc` use-case
we will need to detect a struct's differing size in a way that will
depend on the wavefront size.
---
 clang/include/clang/Basic/BuiltinsAMDGPU.def | 1 +
 clang/test/CodeGenOpenCL/builtins-amdgcn.cl  | 7 +++
 2 files changed, 8 insertions(+)

diff --git a/clang/include/clang/Basic/BuiltinsAMDGPU.def 
b/clang/include/clang/Basic/BuiltinsAMDGPU.def
index 5f8001e61a0282..213311b96df74f 100644
--- a/clang/include/clang/Basic/BuiltinsAMDGPU.def
+++ b/clang/include/clang/Basic/BuiltinsAMDGPU.def
@@ -69,6 +69,7 @@ BUILTIN(__builtin_amdgcn_s_dcache_inv, "v", "n")
 BUILTIN(__builtin_amdgcn_buffer_wbinvl1, "v", "n")
 BUILTIN(__builtin_amdgcn_fence, "vUicC*", "n")
 BUILTIN(__builtin_amdgcn_groupstaticsize, "Ui", "n")
+BUILTIN(__builtin_amdgcn_wavefrontsize, "Ui", "nc")
 
 BUILTIN(__builtin_amdgcn_atomic_inc32, "UZiUZiD*UZiUicC*", "n")
 BUILTIN(__builtin_amdgcn_atomic_inc64, "UWiUWiD*UWiUicC*", "n")
diff --git a/clang/test/CodeGenOpenCL/builtins-amdgcn.cl 
b/clang/test/CodeGenOpenCL/builtins-amdgcn.cl
index 8d9e4e018b12e5..7d9010ee9067d6 100644
--- a/clang/test/CodeGenOpenCL/builtins-amdgcn.cl
+++ b/clang/test/CodeGenOpenCL/builtins-amdgcn.cl
@@ -832,6 +832,13 @@ void test_atomic_inc_dec(local uint *lptr, global uint 
*gptr, uint val) {
   res = __builtin_amdgcn_atomic_dec32((volatile global uint*)gptr, val, 
__ATOMIC_SEQ_CST, "");
 }
 
+// CHECK-LABEL test_wavefrontsize(
+unsigned test_wavefrontsize() {
+
+  // CHECK: call i32 @llvm.amdgcn.wavefrontsize()
+  return __builtin_amdgcn_wavefrontsize();
+}
+
 // CHECK-DAG: [[$WI_RANGE]] = !{i32 0, i32 1024}
 // CHECK-DAG: [[$WS_RANGE]] = !{i16 1, i16 1025}
 // CHECK-DAG: attributes #[[$NOUNWIND_READONLY]] = { convergent mustprogress 
nocallback nofree nounwind willreturn memory(none) }

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits