AMDGPU: Remove floating point dot4 and dot2 from gfx1250 support #145751
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
We have FeatureDot9Insts for dot2 and FeatureDot11Insts for dot4. And gfx1250 does not support these two features. So everthing was in place except that we need to copy SubtargetPredicate from pseudo to real for VOP3_DPP8_t16.
|
@llvm/pr-subscribers-mc Author: Changpeng Fang (changpeng) ChangesWe have FeatureDot9Insts for dot2 and FeatureDot11Insts for dot4. And gfx1250 does not support these two features. So everything was in place except that we need to copy SubtargetPredicate from pseudo to real for VOP3_DPP8_t16. Full diff: https://github.com/llvm/llvm-project/pull/145751.diff 2 Files Affected:
diff --git a/llvm/lib/Target/AMDGPU/VOPInstructions.td b/llvm/lib/Target/AMDGPU/VOPInstructions.td
index 0b64b504466c8..1e47acb5fde4f 100644
--- a/llvm/lib/Target/AMDGPU/VOPInstructions.td
+++ b/llvm/lib/Target/AMDGPU/VOPInstructions.td
@@ -1679,6 +1679,7 @@ class Base_VOP3_DPP8_t16<bits<10> op, VOP_Pseudo ps, string opName = ps.OpName>
let SchedRW = ps.SchedRW;
let Uses = ps.Uses;
+ let SubtargetPredicate = ps.SubtargetPredicate;
let OtherPredicates = ps.OtherPredicates;
let True16Predicate = ps.True16Predicate;
}
diff --git a/llvm/test/MC/AMDGPU/gfx1250_asm_unsupported.s b/llvm/test/MC/AMDGPU/gfx1250_asm_unsupported.s
index 55cb9dab3c048..2111e91cd5ef2 100644
--- a/llvm/test/MC/AMDGPU/gfx1250_asm_unsupported.s
+++ b/llvm/test/MC/AMDGPU/gfx1250_asm_unsupported.s
@@ -1,5 +1,67 @@
; RUN: not llvm-mc -triple=amdgcn -mcpu=gfx1250 -show-encoding %s 2>&1 | FileCheck --check-prefix=GFX1250-ERR --implicit-check-not=error: --strict-whitespace %s
+;; DOT4_F32_*, DOT2_F32_*, DOT2_F16 and DOT2_BF16
+
+v_dot4_f32_fp8_fp8 v0, v1, v2, v3
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot4_f32_fp8_fp8 v0, v1, v2, v3 row_mirror
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot4_f32_fp8_fp8 v0, v1, v2, v3 dpp8:[0,1,2,3,4,5,6,7]
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot4_f32_fp8_bf8 v0, v1, v2, v3
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot4_f32_fp8_bf8 v0, v1, v2, v3 quad_perm:[3,2,1,0]
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot4_f32_fp8_bf8 v0, v1, v2, v3 dpp8:[0,1,2,3,4,5,6,7]
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot4_f32_bf8_fp8 v0, v1, v2, v3
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot4_f32_bf8_fp8 v0, v1, v2, v3 row_shl:15
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot4_f32_bf8_fp8 v0, v1, v2, v3 dpp8:[0,1,2,3,4,5,6,7]
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot4_f32_bf8_bf8 v0, v1, v2, v3
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot4_f32_bf8_bf8 v0, v1, v2, v3 row_share:15
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot4_f32_bf8_bf8 v0, v1, v2, v3 dpp8:[0,1,2,3,4,5,6,7]
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot2_f16_f16 v5, v1, v2, s3
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot2_f16_f16_e64_dpp v0, v1, v2, v3 quad_perm:[0,1,2,3] row_mask:0x0 bank_mask:0x0 fi:1
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot2_f16_f16_e64_dpp v0, v1, v2, v3 dpp8:[0,1,2,3,4,4,4,4]
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot2_bf16_bf16 v5, v1, v2, s3
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot2_bf16_bf16_e64_dpp v0, v1, v2, v3 quad_perm:[0,1,2,3] row_mask:0x0 bank_mask:0x0 fi:1
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot2_bf16_bf16_e64_dpp v0, v1, v2, v3 dpp8:[0,1,2,3,4,4,4,4]
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot2_f32_bf16 v5, v1, v2, v3
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot2_f32_f16 v5, v1, v2, s3
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
;; LDS-direct and parameter-load, VINTERP
ds_direct_load v1 wait_va_vdst:15
|
|
@llvm/pr-subscribers-backend-amdgpu Author: Changpeng Fang (changpeng) ChangesWe have FeatureDot9Insts for dot2 and FeatureDot11Insts for dot4. And gfx1250 does not support these two features. So everything was in place except that we need to copy SubtargetPredicate from pseudo to real for VOP3_DPP8_t16. Full diff: https://github.com/llvm/llvm-project/pull/145751.diff 2 Files Affected:
diff --git a/llvm/lib/Target/AMDGPU/VOPInstructions.td b/llvm/lib/Target/AMDGPU/VOPInstructions.td
index 0b64b504466c8..1e47acb5fde4f 100644
--- a/llvm/lib/Target/AMDGPU/VOPInstructions.td
+++ b/llvm/lib/Target/AMDGPU/VOPInstructions.td
@@ -1679,6 +1679,7 @@ class Base_VOP3_DPP8_t16<bits<10> op, VOP_Pseudo ps, string opName = ps.OpName>
let SchedRW = ps.SchedRW;
let Uses = ps.Uses;
+ let SubtargetPredicate = ps.SubtargetPredicate;
let OtherPredicates = ps.OtherPredicates;
let True16Predicate = ps.True16Predicate;
}
diff --git a/llvm/test/MC/AMDGPU/gfx1250_asm_unsupported.s b/llvm/test/MC/AMDGPU/gfx1250_asm_unsupported.s
index 55cb9dab3c048..2111e91cd5ef2 100644
--- a/llvm/test/MC/AMDGPU/gfx1250_asm_unsupported.s
+++ b/llvm/test/MC/AMDGPU/gfx1250_asm_unsupported.s
@@ -1,5 +1,67 @@
; RUN: not llvm-mc -triple=amdgcn -mcpu=gfx1250 -show-encoding %s 2>&1 | FileCheck --check-prefix=GFX1250-ERR --implicit-check-not=error: --strict-whitespace %s
+;; DOT4_F32_*, DOT2_F32_*, DOT2_F16 and DOT2_BF16
+
+v_dot4_f32_fp8_fp8 v0, v1, v2, v3
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot4_f32_fp8_fp8 v0, v1, v2, v3 row_mirror
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot4_f32_fp8_fp8 v0, v1, v2, v3 dpp8:[0,1,2,3,4,5,6,7]
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot4_f32_fp8_bf8 v0, v1, v2, v3
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot4_f32_fp8_bf8 v0, v1, v2, v3 quad_perm:[3,2,1,0]
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot4_f32_fp8_bf8 v0, v1, v2, v3 dpp8:[0,1,2,3,4,5,6,7]
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot4_f32_bf8_fp8 v0, v1, v2, v3
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot4_f32_bf8_fp8 v0, v1, v2, v3 row_shl:15
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot4_f32_bf8_fp8 v0, v1, v2, v3 dpp8:[0,1,2,3,4,5,6,7]
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot4_f32_bf8_bf8 v0, v1, v2, v3
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot4_f32_bf8_bf8 v0, v1, v2, v3 row_share:15
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot4_f32_bf8_bf8 v0, v1, v2, v3 dpp8:[0,1,2,3,4,5,6,7]
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot2_f16_f16 v5, v1, v2, s3
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot2_f16_f16_e64_dpp v0, v1, v2, v3 quad_perm:[0,1,2,3] row_mask:0x0 bank_mask:0x0 fi:1
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot2_f16_f16_e64_dpp v0, v1, v2, v3 dpp8:[0,1,2,3,4,4,4,4]
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot2_bf16_bf16 v5, v1, v2, s3
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot2_bf16_bf16_e64_dpp v0, v1, v2, v3 quad_perm:[0,1,2,3] row_mask:0x0 bank_mask:0x0 fi:1
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot2_bf16_bf16_e64_dpp v0, v1, v2, v3 dpp8:[0,1,2,3,4,4,4,4]
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot2_f32_bf16 v5, v1, v2, v3
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
+v_dot2_f32_f16 v5, v1, v2, s3
+// GFX1250-ERR: :[[@LINE-1]]:{{[0-9]+}}: error: instruction not supported on this GPU
+
;; LDS-direct and parameter-load, VINTERP
ds_direct_load v1 wait_va_vdst:15
|
rampitec
approved these changes
Jun 25, 2025
anthonyhatran
pushed a commit
to anthonyhatran/llvm-project
that referenced
this pull request
Jun 26, 2025
…m#145751) We have FeatureDot9Insts for dot2 and FeatureDot11Insts for dot4. And gfx1250 does not support these two features. So everything was in place except that we need to copy SubtargetPredicate from pseudo to real for VOP3_DPP8_t16.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
We have FeatureDot9Insts for dot2 and FeatureDot11Insts for dot4. And gfx1250 does not support these two features. So everything was in place except that we need to copy SubtargetPredicate from pseudo to real for VOP3_DPP8_t16.