Skip to content

[AMDGPU] Add gfx12 run lines to fence MMRA tests #94333

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Merged
merged 1 commit into from
Jun 4, 2024
Merged

[AMDGPU] Add gfx12 run lines to fence MMRA tests #94333

merged 1 commit into from
Jun 4, 2024

Conversation

Pierre-vh
Copy link
Contributor

No description provided.

@Pierre-vh Pierre-vh requested review from jayfoad and arsenm June 4, 2024 10:38
@llvmbot
Copy link
Member

llvmbot commented Jun 4, 2024

@llvm/pr-subscribers-backend-amdgpu

Author: Pierre van Houtryve (Pierre-vh)

Changes

Patch is 34.69 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/94333.diff

2 Files Affected:

  • (modified) llvm/test/CodeGen/AMDGPU/memory-legalizer-fence-mmra-global.ll (+384)
  • (modified) llvm/test/CodeGen/AMDGPU/memory-legalizer-fence-mmra-local.ll (+218)
diff --git a/llvm/test/CodeGen/AMDGPU/memory-legalizer-fence-mmra-global.ll b/llvm/test/CodeGen/AMDGPU/memory-legalizer-fence-mmra-global.ll
index da9bc6b331134..7a68fb5ce508c 100644
--- a/llvm/test/CodeGen/AMDGPU/memory-legalizer-fence-mmra-global.ll
+++ b/llvm/test/CodeGen/AMDGPU/memory-legalizer-fence-mmra-global.ll
@@ -10,6 +10,8 @@
 ; RUN: llc -mtriple=amdgcn-amd-amdhsa -O0 -mcpu=gfx940 -mattr=+tgsplit < %s | FileCheck -check-prefixes=GFX940-TGSPLIT %s
 ; RUN: llc -mtriple=amdgcn-amd-amdhsa -O0 -mcpu=gfx1100 < %s | FileCheck --check-prefixes=GFX11-WGP %s
 ; RUN: llc -mtriple=amdgcn-amd-amdhsa -O0 -mcpu=gfx1100 -mattr=+cumode < %s | FileCheck --check-prefixes=GFX11-CU %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -O0 -mcpu=gfx1200 < %s | FileCheck --check-prefixes=GFX12-WGP %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -O0 -mcpu=gfx1200 -mattr=+cumode < %s | FileCheck --check-prefixes=GFX12-CU %s
 
 define amdgpu_kernel void @workgroup_acquire_fence() {
 ; GFX6-LABEL: workgroup_acquire_fence:
@@ -65,6 +67,19 @@ define amdgpu_kernel void @workgroup_acquire_fence() {
 ; GFX11-CU-LABEL: workgroup_acquire_fence:
 ; GFX11-CU:       ; %bb.0: ; %entry
 ; GFX11-CU-NEXT:    s_endpgm
+;
+; GFX12-WGP-LABEL: workgroup_acquire_fence:
+; GFX12-WGP:       ; %bb.0: ; %entry
+; GFX12-WGP-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_samplecnt 0x0
+; GFX12-WGP-NEXT:    s_wait_loadcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_storecnt 0x0
+; GFX12-WGP-NEXT:    global_inv scope:SCOPE_SE
+; GFX12-WGP-NEXT:    s_endpgm
+;
+; GFX12-CU-LABEL: workgroup_acquire_fence:
+; GFX12-CU:       ; %bb.0: ; %entry
+; GFX12-CU-NEXT:    s_endpgm
 entry:
   fence syncscope("workgroup") acquire, !mmra !{!"amdgpu-as", !"global"}
   ret void
@@ -120,6 +135,18 @@ define amdgpu_kernel void @workgroup_release_fence() {
 ; GFX11-CU-LABEL: workgroup_release_fence:
 ; GFX11-CU:       ; %bb.0: ; %entry
 ; GFX11-CU-NEXT:    s_endpgm
+;
+; GFX12-WGP-LABEL: workgroup_release_fence:
+; GFX12-WGP:       ; %bb.0: ; %entry
+; GFX12-WGP-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_samplecnt 0x0
+; GFX12-WGP-NEXT:    s_wait_loadcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_storecnt 0x0
+; GFX12-WGP-NEXT:    s_endpgm
+;
+; GFX12-CU-LABEL: workgroup_release_fence:
+; GFX12-CU:       ; %bb.0: ; %entry
+; GFX12-CU-NEXT:    s_endpgm
 entry:
   fence syncscope("workgroup") release, !mmra !{!"amdgpu-as", !"global"}
   ret void
@@ -179,6 +206,19 @@ define amdgpu_kernel void @workgroup_acq_rel_fence() {
 ; GFX11-CU-LABEL: workgroup_acq_rel_fence:
 ; GFX11-CU:       ; %bb.0: ; %entry
 ; GFX11-CU-NEXT:    s_endpgm
+;
+; GFX12-WGP-LABEL: workgroup_acq_rel_fence:
+; GFX12-WGP:       ; %bb.0: ; %entry
+; GFX12-WGP-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_samplecnt 0x0
+; GFX12-WGP-NEXT:    s_wait_loadcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_storecnt 0x0
+; GFX12-WGP-NEXT:    global_inv scope:SCOPE_SE
+; GFX12-WGP-NEXT:    s_endpgm
+;
+; GFX12-CU-LABEL: workgroup_acq_rel_fence:
+; GFX12-CU:       ; %bb.0: ; %entry
+; GFX12-CU-NEXT:    s_endpgm
 entry:
   fence syncscope("workgroup") acq_rel, !mmra !{!"amdgpu-as", !"global"}
   ret void
@@ -238,6 +278,19 @@ define amdgpu_kernel void @workgroup_seq_cst_fence() {
 ; GFX11-CU-LABEL: workgroup_seq_cst_fence:
 ; GFX11-CU:       ; %bb.0: ; %entry
 ; GFX11-CU-NEXT:    s_endpgm
+;
+; GFX12-WGP-LABEL: workgroup_seq_cst_fence:
+; GFX12-WGP:       ; %bb.0: ; %entry
+; GFX12-WGP-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_samplecnt 0x0
+; GFX12-WGP-NEXT:    s_wait_loadcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_storecnt 0x0
+; GFX12-WGP-NEXT:    global_inv scope:SCOPE_SE
+; GFX12-WGP-NEXT:    s_endpgm
+;
+; GFX12-CU-LABEL: workgroup_seq_cst_fence:
+; GFX12-CU:       ; %bb.0: ; %entry
+; GFX12-CU-NEXT:    s_endpgm
 entry:
   fence syncscope("workgroup") seq_cst, !mmra !{!"amdgpu-as", !"global"}
   ret void
@@ -297,6 +350,19 @@ define amdgpu_kernel void @workgroup_one_as_acquire_fence() {
 ; GFX11-CU-LABEL: workgroup_one_as_acquire_fence:
 ; GFX11-CU:       ; %bb.0: ; %entry
 ; GFX11-CU-NEXT:    s_endpgm
+;
+; GFX12-WGP-LABEL: workgroup_one_as_acquire_fence:
+; GFX12-WGP:       ; %bb.0: ; %entry
+; GFX12-WGP-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_samplecnt 0x0
+; GFX12-WGP-NEXT:    s_wait_loadcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_storecnt 0x0
+; GFX12-WGP-NEXT:    global_inv scope:SCOPE_SE
+; GFX12-WGP-NEXT:    s_endpgm
+;
+; GFX12-CU-LABEL: workgroup_one_as_acquire_fence:
+; GFX12-CU:       ; %bb.0: ; %entry
+; GFX12-CU-NEXT:    s_endpgm
 entry:
   fence syncscope("workgroup-one-as") acquire, !mmra !{!"amdgpu-as", !"global"}
   ret void
@@ -352,6 +418,18 @@ define amdgpu_kernel void @workgroup_one_as_release_fence() {
 ; GFX11-CU-LABEL: workgroup_one_as_release_fence:
 ; GFX11-CU:       ; %bb.0: ; %entry
 ; GFX11-CU-NEXT:    s_endpgm
+;
+; GFX12-WGP-LABEL: workgroup_one_as_release_fence:
+; GFX12-WGP:       ; %bb.0: ; %entry
+; GFX12-WGP-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_samplecnt 0x0
+; GFX12-WGP-NEXT:    s_wait_loadcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_storecnt 0x0
+; GFX12-WGP-NEXT:    s_endpgm
+;
+; GFX12-CU-LABEL: workgroup_one_as_release_fence:
+; GFX12-CU:       ; %bb.0: ; %entry
+; GFX12-CU-NEXT:    s_endpgm
 entry:
   fence syncscope("workgroup-one-as") release, !mmra !{!"amdgpu-as", !"global"}
   ret void
@@ -411,6 +489,19 @@ define amdgpu_kernel void @workgroup_one_as_acq_rel_fence() {
 ; GFX11-CU-LABEL: workgroup_one_as_acq_rel_fence:
 ; GFX11-CU:       ; %bb.0: ; %entry
 ; GFX11-CU-NEXT:    s_endpgm
+;
+; GFX12-WGP-LABEL: workgroup_one_as_acq_rel_fence:
+; GFX12-WGP:       ; %bb.0: ; %entry
+; GFX12-WGP-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_samplecnt 0x0
+; GFX12-WGP-NEXT:    s_wait_loadcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_storecnt 0x0
+; GFX12-WGP-NEXT:    global_inv scope:SCOPE_SE
+; GFX12-WGP-NEXT:    s_endpgm
+;
+; GFX12-CU-LABEL: workgroup_one_as_acq_rel_fence:
+; GFX12-CU:       ; %bb.0: ; %entry
+; GFX12-CU-NEXT:    s_endpgm
 entry:
   fence syncscope("workgroup-one-as") acq_rel, !mmra !{!"amdgpu-as", !"global"}
   ret void
@@ -470,6 +561,19 @@ define amdgpu_kernel void @workgroup_one_as_seq_cst_fence() {
 ; GFX11-CU-LABEL: workgroup_one_as_seq_cst_fence:
 ; GFX11-CU:       ; %bb.0: ; %entry
 ; GFX11-CU-NEXT:    s_endpgm
+;
+; GFX12-WGP-LABEL: workgroup_one_as_seq_cst_fence:
+; GFX12-WGP:       ; %bb.0: ; %entry
+; GFX12-WGP-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_samplecnt 0x0
+; GFX12-WGP-NEXT:    s_wait_loadcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_storecnt 0x0
+; GFX12-WGP-NEXT:    global_inv scope:SCOPE_SE
+; GFX12-WGP-NEXT:    s_endpgm
+;
+; GFX12-CU-LABEL: workgroup_one_as_seq_cst_fence:
+; GFX12-CU:       ; %bb.0: ; %entry
+; GFX12-CU-NEXT:    s_endpgm
 entry:
   fence syncscope("workgroup-one-as") seq_cst, !mmra !{!"amdgpu-as", !"global"}
   ret void
@@ -548,6 +652,24 @@ define amdgpu_kernel void @agent_acquire_fence() {
 ; GFX11-CU-NEXT:    buffer_gl1_inv
 ; GFX11-CU-NEXT:    buffer_gl0_inv
 ; GFX11-CU-NEXT:    s_endpgm
+;
+; GFX12-WGP-LABEL: agent_acquire_fence:
+; GFX12-WGP:       ; %bb.0: ; %entry
+; GFX12-WGP-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_samplecnt 0x0
+; GFX12-WGP-NEXT:    s_wait_loadcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_storecnt 0x0
+; GFX12-WGP-NEXT:    global_inv scope:SCOPE_DEV
+; GFX12-WGP-NEXT:    s_endpgm
+;
+; GFX12-CU-LABEL: agent_acquire_fence:
+; GFX12-CU:       ; %bb.0: ; %entry
+; GFX12-CU-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-CU-NEXT:    s_wait_samplecnt 0x0
+; GFX12-CU-NEXT:    s_wait_loadcnt 0x0
+; GFX12-CU-NEXT:    s_wait_storecnt 0x0
+; GFX12-CU-NEXT:    global_inv scope:SCOPE_DEV
+; GFX12-CU-NEXT:    s_endpgm
 entry:
   fence syncscope("agent") acquire, !mmra !{!"amdgpu-as", !"global"}
   ret void
@@ -614,6 +736,22 @@ define amdgpu_kernel void @agent_release_fence() {
 ; GFX11-CU-NEXT:    s_waitcnt vmcnt(0)
 ; GFX11-CU-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX11-CU-NEXT:    s_endpgm
+;
+; GFX12-WGP-LABEL: agent_release_fence:
+; GFX12-WGP:       ; %bb.0: ; %entry
+; GFX12-WGP-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_samplecnt 0x0
+; GFX12-WGP-NEXT:    s_wait_loadcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_storecnt 0x0
+; GFX12-WGP-NEXT:    s_endpgm
+;
+; GFX12-CU-LABEL: agent_release_fence:
+; GFX12-CU:       ; %bb.0: ; %entry
+; GFX12-CU-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-CU-NEXT:    s_wait_samplecnt 0x0
+; GFX12-CU-NEXT:    s_wait_loadcnt 0x0
+; GFX12-CU-NEXT:    s_wait_storecnt 0x0
+; GFX12-CU-NEXT:    s_endpgm
 entry:
   fence syncscope("agent") release, !mmra !{!"amdgpu-as", !"global"}
   ret void
@@ -694,6 +832,24 @@ define amdgpu_kernel void @agent_acq_rel_fence() {
 ; GFX11-CU-NEXT:    buffer_gl1_inv
 ; GFX11-CU-NEXT:    buffer_gl0_inv
 ; GFX11-CU-NEXT:    s_endpgm
+;
+; GFX12-WGP-LABEL: agent_acq_rel_fence:
+; GFX12-WGP:       ; %bb.0: ; %entry
+; GFX12-WGP-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_samplecnt 0x0
+; GFX12-WGP-NEXT:    s_wait_loadcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_storecnt 0x0
+; GFX12-WGP-NEXT:    global_inv scope:SCOPE_DEV
+; GFX12-WGP-NEXT:    s_endpgm
+;
+; GFX12-CU-LABEL: agent_acq_rel_fence:
+; GFX12-CU:       ; %bb.0: ; %entry
+; GFX12-CU-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-CU-NEXT:    s_wait_samplecnt 0x0
+; GFX12-CU-NEXT:    s_wait_loadcnt 0x0
+; GFX12-CU-NEXT:    s_wait_storecnt 0x0
+; GFX12-CU-NEXT:    global_inv scope:SCOPE_DEV
+; GFX12-CU-NEXT:    s_endpgm
 entry:
   fence syncscope("agent") acq_rel, !mmra !{!"amdgpu-as", !"global"}
   ret void
@@ -774,6 +930,24 @@ define amdgpu_kernel void @agent_seq_cst_fence() {
 ; GFX11-CU-NEXT:    buffer_gl1_inv
 ; GFX11-CU-NEXT:    buffer_gl0_inv
 ; GFX11-CU-NEXT:    s_endpgm
+;
+; GFX12-WGP-LABEL: agent_seq_cst_fence:
+; GFX12-WGP:       ; %bb.0: ; %entry
+; GFX12-WGP-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_samplecnt 0x0
+; GFX12-WGP-NEXT:    s_wait_loadcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_storecnt 0x0
+; GFX12-WGP-NEXT:    global_inv scope:SCOPE_DEV
+; GFX12-WGP-NEXT:    s_endpgm
+;
+; GFX12-CU-LABEL: agent_seq_cst_fence:
+; GFX12-CU:       ; %bb.0: ; %entry
+; GFX12-CU-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-CU-NEXT:    s_wait_samplecnt 0x0
+; GFX12-CU-NEXT:    s_wait_loadcnt 0x0
+; GFX12-CU-NEXT:    s_wait_storecnt 0x0
+; GFX12-CU-NEXT:    global_inv scope:SCOPE_DEV
+; GFX12-CU-NEXT:    s_endpgm
 entry:
   fence syncscope("agent") seq_cst, !mmra !{!"amdgpu-as", !"global"}
   ret void
@@ -852,6 +1026,24 @@ define amdgpu_kernel void @agent_one_as_acquire_fence() {
 ; GFX11-CU-NEXT:    buffer_gl1_inv
 ; GFX11-CU-NEXT:    buffer_gl0_inv
 ; GFX11-CU-NEXT:    s_endpgm
+;
+; GFX12-WGP-LABEL: agent_one_as_acquire_fence:
+; GFX12-WGP:       ; %bb.0: ; %entry
+; GFX12-WGP-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_samplecnt 0x0
+; GFX12-WGP-NEXT:    s_wait_loadcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_storecnt 0x0
+; GFX12-WGP-NEXT:    global_inv scope:SCOPE_DEV
+; GFX12-WGP-NEXT:    s_endpgm
+;
+; GFX12-CU-LABEL: agent_one_as_acquire_fence:
+; GFX12-CU:       ; %bb.0: ; %entry
+; GFX12-CU-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-CU-NEXT:    s_wait_samplecnt 0x0
+; GFX12-CU-NEXT:    s_wait_loadcnt 0x0
+; GFX12-CU-NEXT:    s_wait_storecnt 0x0
+; GFX12-CU-NEXT:    global_inv scope:SCOPE_DEV
+; GFX12-CU-NEXT:    s_endpgm
 entry:
   fence syncscope("agent-one-as") acquire, !mmra !{!"amdgpu-as", !"global"}
   ret void
@@ -918,6 +1110,22 @@ define amdgpu_kernel void @agent_one_as_release_fence() {
 ; GFX11-CU-NEXT:    s_waitcnt vmcnt(0)
 ; GFX11-CU-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX11-CU-NEXT:    s_endpgm
+;
+; GFX12-WGP-LABEL: agent_one_as_release_fence:
+; GFX12-WGP:       ; %bb.0: ; %entry
+; GFX12-WGP-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_samplecnt 0x0
+; GFX12-WGP-NEXT:    s_wait_loadcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_storecnt 0x0
+; GFX12-WGP-NEXT:    s_endpgm
+;
+; GFX12-CU-LABEL: agent_one_as_release_fence:
+; GFX12-CU:       ; %bb.0: ; %entry
+; GFX12-CU-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-CU-NEXT:    s_wait_samplecnt 0x0
+; GFX12-CU-NEXT:    s_wait_loadcnt 0x0
+; GFX12-CU-NEXT:    s_wait_storecnt 0x0
+; GFX12-CU-NEXT:    s_endpgm
 entry:
   fence syncscope("agent-one-as") release, !mmra !{!"amdgpu-as", !"global"}
   ret void
@@ -998,6 +1206,24 @@ define amdgpu_kernel void @agent_one_as_acq_rel_fence() {
 ; GFX11-CU-NEXT:    buffer_gl1_inv
 ; GFX11-CU-NEXT:    buffer_gl0_inv
 ; GFX11-CU-NEXT:    s_endpgm
+;
+; GFX12-WGP-LABEL: agent_one_as_acq_rel_fence:
+; GFX12-WGP:       ; %bb.0: ; %entry
+; GFX12-WGP-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_samplecnt 0x0
+; GFX12-WGP-NEXT:    s_wait_loadcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_storecnt 0x0
+; GFX12-WGP-NEXT:    global_inv scope:SCOPE_DEV
+; GFX12-WGP-NEXT:    s_endpgm
+;
+; GFX12-CU-LABEL: agent_one_as_acq_rel_fence:
+; GFX12-CU:       ; %bb.0: ; %entry
+; GFX12-CU-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-CU-NEXT:    s_wait_samplecnt 0x0
+; GFX12-CU-NEXT:    s_wait_loadcnt 0x0
+; GFX12-CU-NEXT:    s_wait_storecnt 0x0
+; GFX12-CU-NEXT:    global_inv scope:SCOPE_DEV
+; GFX12-CU-NEXT:    s_endpgm
 entry:
   fence syncscope("agent-one-as") acq_rel, !mmra !{!"amdgpu-as", !"global"}
   ret void
@@ -1078,6 +1304,24 @@ define amdgpu_kernel void @agent_one_as_seq_cst_fence() {
 ; GFX11-CU-NEXT:    buffer_gl1_inv
 ; GFX11-CU-NEXT:    buffer_gl0_inv
 ; GFX11-CU-NEXT:    s_endpgm
+;
+; GFX12-WGP-LABEL: agent_one_as_seq_cst_fence:
+; GFX12-WGP:       ; %bb.0: ; %entry
+; GFX12-WGP-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_samplecnt 0x0
+; GFX12-WGP-NEXT:    s_wait_loadcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_storecnt 0x0
+; GFX12-WGP-NEXT:    global_inv scope:SCOPE_DEV
+; GFX12-WGP-NEXT:    s_endpgm
+;
+; GFX12-CU-LABEL: agent_one_as_seq_cst_fence:
+; GFX12-CU:       ; %bb.0: ; %entry
+; GFX12-CU-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-CU-NEXT:    s_wait_samplecnt 0x0
+; GFX12-CU-NEXT:    s_wait_loadcnt 0x0
+; GFX12-CU-NEXT:    s_wait_storecnt 0x0
+; GFX12-CU-NEXT:    global_inv scope:SCOPE_DEV
+; GFX12-CU-NEXT:    s_endpgm
 entry:
   fence syncscope("agent-one-as") seq_cst, !mmra !{!"amdgpu-as", !"global"}
   ret void
@@ -1158,6 +1402,24 @@ define amdgpu_kernel void @system_acquire_fence() {
 ; GFX11-CU-NEXT:    buffer_gl1_inv
 ; GFX11-CU-NEXT:    buffer_gl0_inv
 ; GFX11-CU-NEXT:    s_endpgm
+;
+; GFX12-WGP-LABEL: system_acquire_fence:
+; GFX12-WGP:       ; %bb.0: ; %entry
+; GFX12-WGP-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_samplecnt 0x0
+; GFX12-WGP-NEXT:    s_wait_loadcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_storecnt 0x0
+; GFX12-WGP-NEXT:    global_inv scope:SCOPE_SYS
+; GFX12-WGP-NEXT:    s_endpgm
+;
+; GFX12-CU-LABEL: system_acquire_fence:
+; GFX12-CU:       ; %bb.0: ; %entry
+; GFX12-CU-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-CU-NEXT:    s_wait_samplecnt 0x0
+; GFX12-CU-NEXT:    s_wait_loadcnt 0x0
+; GFX12-CU-NEXT:    s_wait_storecnt 0x0
+; GFX12-CU-NEXT:    global_inv scope:SCOPE_SYS
+; GFX12-CU-NEXT:    s_endpgm
 entry:
   fence acquire, !mmra !{!"amdgpu-as", !"global"}
   ret void
@@ -1226,6 +1488,22 @@ define amdgpu_kernel void @system_release_fence() {
 ; GFX11-CU-NEXT:    s_waitcnt vmcnt(0)
 ; GFX11-CU-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX11-CU-NEXT:    s_endpgm
+;
+; GFX12-WGP-LABEL: system_release_fence:
+; GFX12-WGP:       ; %bb.0: ; %entry
+; GFX12-WGP-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_samplecnt 0x0
+; GFX12-WGP-NEXT:    s_wait_loadcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_storecnt 0x0
+; GFX12-WGP-NEXT:    s_endpgm
+;
+; GFX12-CU-LABEL: system_release_fence:
+; GFX12-CU:       ; %bb.0: ; %entry
+; GFX12-CU-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-CU-NEXT:    s_wait_samplecnt 0x0
+; GFX12-CU-NEXT:    s_wait_loadcnt 0x0
+; GFX12-CU-NEXT:    s_wait_storecnt 0x0
+; GFX12-CU-NEXT:    s_endpgm
 entry:
   fence release, !mmra !{!"amdgpu-as", !"global"}
   ret void
@@ -1310,6 +1588,24 @@ define amdgpu_kernel void @system_acq_rel_fence() {
 ; GFX11-CU-NEXT:    buffer_gl1_inv
 ; GFX11-CU-NEXT:    buffer_gl0_inv
 ; GFX11-CU-NEXT:    s_endpgm
+;
+; GFX12-WGP-LABEL: system_acq_rel_fence:
+; GFX12-WGP:       ; %bb.0: ; %entry
+; GFX12-WGP-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_samplecnt 0x0
+; GFX12-WGP-NEXT:    s_wait_loadcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_storecnt 0x0
+; GFX12-WGP-NEXT:    global_inv scope:SCOPE_SYS
+; GFX12-WGP-NEXT:    s_endpgm
+;
+; GFX12-CU-LABEL: system_acq_rel_fence:
+; GFX12-CU:       ; %bb.0: ; %entry
+; GFX12-CU-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-CU-NEXT:    s_wait_samplecnt 0x0
+; GFX12-CU-NEXT:    s_wait_loadcnt 0x0
+; GFX12-CU-NEXT:    s_wait_storecnt 0x0
+; GFX12-CU-NEXT:    global_inv scope:SCOPE_SYS
+; GFX12-CU-NEXT:    s_endpgm
 entry:
   fence acq_rel, !mmra !{!"amdgpu-as", !"global"}
   ret void
@@ -1394,6 +1690,24 @@ define amdgpu_kernel void @system_seq_cst_fence() {
 ; GFX11-CU-NEXT:    buffer_gl1_inv
 ; GFX11-CU-NEXT:    buffer_gl0_inv
 ; GFX11-CU-NEXT:    s_endpgm
+;
+; GFX12-WGP-LABEL: system_seq_cst_fence:
+; GFX12-WGP:       ; %bb.0: ; %entry
+; GFX12-WGP-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_samplecnt 0x0
+; GFX12-WGP-NEXT:    s_wait_loadcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_storecnt 0x0
+; GFX12-WGP-NEXT:    global_inv scope:SCOPE_SYS
+; GFX12-WGP-NEXT:    s_endpgm
+;
+; GFX12-CU-LABEL: system_seq_cst_fence:
+; GFX12-CU:       ; %bb.0: ; %entry
+; GFX12-CU-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-CU-NEXT:    s_wait_samplecnt 0x0
+; GFX12-CU-NEXT:    s_wait_loadcnt 0x0
+; GFX12-CU-NEXT:    s_wait_storecnt 0x0
+; GFX12-CU-NEXT:    global_inv scope:SCOPE_SYS
+; GFX12-CU-NEXT:    s_endpgm
 entry:
   fence seq_cst, !mmra !{!"amdgpu-as", !"global"}
   ret void
@@ -1474,6 +1788,24 @@ define amdgpu_kernel void @system_one_as_acquire_fence() {
 ; GFX11-CU-NEXT:    buffer_gl1_inv
 ; GFX11-CU-NEXT:    buffer_gl0_inv
 ; GFX11-CU-NEXT:    s_endpgm
+;
+; GFX12-WGP-LABEL: system_one_as_acquire_fence:
+; GFX12-WGP:       ; %bb.0: ; %entry
+; GFX12-WGP-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_samplecnt 0x0
+; GFX12-WGP-NEXT:    s_wait_loadcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_storecnt 0x0
+; GFX12-WGP-NEXT:    global_inv scope:SCOPE_SYS
+; GFX12-WGP-NEXT:    s_endpgm
+;
+; GFX12-CU-LABEL: system_one_as_acquire_fence:
+; GFX12-CU:       ; %bb.0: ; %entry
+; GFX12-CU-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-CU-NEXT:    s_wait_samplecnt 0x0
+; GFX12-CU-NEXT:    s_wait_loadcnt 0x0
+; GFX12-CU-NEXT:    s_wait_storecnt 0x0
+; GFX12-CU-NEXT:    global_inv scope:SCOPE_SYS
+; GFX12-CU-NEXT:    s_endpgm
 entry:
   fence syncscope("one-as") acquire, !mmra !{!"amdgpu-as", !"global"}
   ret void
@@ -1542,6 +1874,22 @@ define amdgpu_kernel void @system_one_as_release_fence() {
 ; GFX11-CU-NEXT:    s_waitcnt vmcnt(0)
 ; GFX11-CU-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GFX11-CU-NEXT:    s_endpgm
+;
+; GFX12-WGP-LABEL: system_one_as_release_fence:
+; GFX12-WGP:       ; %bb.0: ; %entry
+; GFX12-WGP-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_samplecnt 0x0
+; GFX12-WGP-NEXT:    s_wait_loadcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_storecnt 0x0
+; GFX12-WGP-NEXT:    s_endpgm
+;
+; GFX12-CU-LABEL: system_one_as_release_fence:
+; GFX12-CU:       ; %bb.0: ; %entry
+; GFX12-CU-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-CU-NEXT:    s_wait_samplecnt 0x0
+; GFX12-CU-NEXT:    s_wait_loadcnt 0x0
+; GFX12-CU-NEXT:    s_wait_storecnt 0x0
+; GFX12-CU-NEXT:    s_endpgm
 entry:
   fence syncscope("one-as") release, !mmra !{!"amdgpu-as", !"global"}
   ret void
@@ -1626,6 +1974,24 @@ define amdgpu_kernel void @system_one_as_acq_rel_fence() {
 ; GFX11-CU-NEXT:    buffer_gl1_inv
 ; GFX11-CU-NEXT:    buffer_gl0_inv
 ; GFX11-CU-NEXT:    s_endpgm
+;
+; GFX12-WGP-LABEL: system_one_as_acq_rel_fence:
+; GFX12-WGP:       ; %bb.0: ; %entry
+; GFX12-WGP-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_samplecnt 0x0
+; GFX12-WGP-NEXT:    s_wait_loadcnt 0x0
+; GFX12-WGP-NEXT:    s_wait_storecnt 0x0
+; GFX12-WGP-NEXT:    global_inv scope:SCOPE_SYS
+; GFX12-WGP-NEXT:    s_endpgm
+;
+; GFX12-CU-LABEL: system_one_as_acq_rel_fence:
+; GFX12-CU:       ; %bb.0: ; %entry
+; GFX12-CU-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-CU-NEXT:    s_wait_samplecnt 0x0
+; GFX12-CU-NEXT:    s_wait_loadcnt 0x0
+; GFX12-CU-NEXT:    s_wait_storecnt 0x0
+; GFX12-CU-NEXT:    global_inv scope:SCOPE_SYS
+; GFX12-CU-NEXT:    s_endpgm
 entry:
   fence syncscope("one-as") acq_rel, !mmra !{!"amdgpu-as", !"global"}
   ret void
@@ -1710,6 +2076,24 @@ define amdgpu_kernel void @system_one_as_seq_cst_fence() {
 ; GFX11-CU-NEXT:    buffer_gl1_inv
 ; GFX11-CU-NE...
[truncated]

jayfoad
jayfoad approved these changes Jun 4, 2024
View reviewed changes
Copy link
Contributor

@jayfoad jayfoad left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Seems obviously fine.

@Pierre-vh Pierre-vh merged commit 9372e1a into llvm:main Jun 4, 2024
6 of 8 checks passed
@Pierre-vh Pierre-vh deleted the amdgpu-as-gfx12-test branch June 4, 2024 10:49
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@jayfoad jayfoad jayfoad approved these changes

@arsenm arsenm Awaiting requested review from arsenm

Assignees
No one assigned
Labels
backend:AMDGPU
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
3 participants
@Pierre-vh @llvmbot @jayfoad