Skip to content

[AMDGPU] Update strict floating point tests to be more comprehensive #169578

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

Merged
merged 3 commits into from
Nov 26, 2025
Merged

[AMDGPU] Update strict floating point tests to be more comprehensive #169578

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
ae8adc5
[AMDGPU] Update strict floating point tests to be more comprehensive
chinmaydd Nov 25, 2025
0fe914e
[AMDGPU] Update fsub.f32 test
chinmaydd Nov 25, 2025
bb312b4
[AMDGPU] Update RUN lines to use common prefixes
chinmaydd Nov 26, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
677 changes: 533 additions & 144 deletions llvm/test/CodeGen/AMDGPU/strict_fadd.f16.ll
View file
Open in desktop

Large diffs are not rendered by default.

516 changes: 444 additions & 72 deletions llvm/test/CodeGen/AMDGPU/strict_fadd.f32.ll
View file
Open in desktop

Large diffs are not rendered by default.

88 changes: 28 additions & 60 deletions llvm/test/CodeGen/AMDGPU/strict_fadd.f64.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,20 +1,19 @@
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=amdgcn-mesa-mesa3d -mcpu=gfx900 < %s | FileCheck -check-prefix=GCN %s
; RUN: llc -mtriple=amdgcn-mesa-mesa3d -mcpu=gfx1010 < %s | FileCheck -check-prefix=GFX10 %s
; RUN: llc -mtriple=amdgcn-mesa-mesa3d -mcpu=gfx1100 -amdgpu-enable-delay-alu=0 < %s | FileCheck -check-prefix=GFX10 %s
; RUN: llc -global-isel=0 -mtriple=amdgcn-mesa-mesa3d -mcpu=gfx900 < %s | FileCheck -check-prefixes=GCN,GCN-SDAG %s
; RUN: llc -global-isel=1 -mtriple=amdgcn-mesa-mesa3d -mcpu=gfx900 < %s | FileCheck -check-prefixes=GCN,GCN-GISEL %s

; RUN: llc -global-isel=0 -mtriple=amdgcn-mesa-mesa3d -mcpu=gfx1010 < %s | FileCheck -check-prefixes=GCN,GFX10PLUS,GFX10 %s
; RUN: llc -global-isel=1 -mtriple=amdgcn-mesa-mesa3d -mcpu=gfx1010 < %s | FileCheck -check-prefixes=GCN,GFX10PLUS,GFX10 %s

; RUN: llc -global-isel=0 -mtriple=amdgcn-mesa-mesa3d -mcpu=gfx1100 -amdgpu-enable-delay-alu=0 < %s | FileCheck -check-prefixes=GCN,GFX10PLUS,GFX11 %s
; RUN: llc -global-isel=1 -mtriple=amdgcn-mesa-mesa3d -mcpu=gfx1100 -amdgpu-enable-delay-alu=0 < %s | FileCheck -check-prefixes=GCN,GFX10PLUS,GFX11 %s

define double @v_constained_fadd_f64_fpexcept_strict(double %x, double %y) #0 {
; GCN-LABEL: v_constained_fadd_f64_fpexcept_strict:
; GCN: ; %bb.0:
; GCN-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GCN-NEXT: v_add_f64 v[0:1], v[0:1], v[2:3]
; GCN-NEXT: s_setpc_b64 s[30:31]
;
; GFX10-LABEL: v_constained_fadd_f64_fpexcept_strict:
; GFX10: ; %bb.0:
; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX10-NEXT: v_add_f64 v[0:1], v[0:1], v[2:3]
; GFX10-NEXT: s_setpc_b64 s[30:31]
%val = call double @llvm.experimental.constrained.fadd.f64(double %x, double %y, metadata !"round.tonearest", metadata !"fpexcept.strict")
ret double %val
}
Expand All @@ -25,12 +24,6 @@ define double @v_constained_fadd_f64_fpexcept_ignore(double %x, double %y) #0 {
; GCN-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GCN-NEXT: v_add_f64 v[0:1], v[0:1], v[2:3]
; GCN-NEXT: s_setpc_b64 s[30:31]
;
; GFX10-LABEL: v_constained_fadd_f64_fpexcept_ignore:
; GFX10: ; %bb.0:
; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX10-NEXT: v_add_f64 v[0:1], v[0:1], v[2:3]
; GFX10-NEXT: s_setpc_b64 s[30:31]
%val = call double @llvm.experimental.constrained.fadd.f64(double %x, double %y, metadata !"round.tonearest", metadata !"fpexcept.ignore")
ret double %val
}
Expand All @@ -41,12 +34,6 @@ define double @v_constained_fadd_f64_fpexcept_maytrap(double %x, double %y) #0 {
; GCN-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GCN-NEXT: v_add_f64 v[0:1], v[0:1], v[2:3]
; GCN-NEXT: s_setpc_b64 s[30:31]
;
; GFX10-LABEL: v_constained_fadd_f64_fpexcept_maytrap:
; GFX10: ; %bb.0:
; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX10-NEXT: v_add_f64 v[0:1], v[0:1], v[2:3]
; GFX10-NEXT: s_setpc_b64 s[30:31]
%val = call double @llvm.experimental.constrained.fadd.f64(double %x, double %y, metadata !"round.tonearest", metadata !"fpexcept.maytrap")
ret double %val
}
Expand All @@ -58,13 +45,6 @@ define <2 x double> @v_constained_fadd_v2f64_fpexcept_strict(<2 x double> %x, <2
; GCN-NEXT: v_add_f64 v[0:1], v[0:1], v[4:5]
; GCN-NEXT: v_add_f64 v[2:3], v[2:3], v[6:7]
; GCN-NEXT: s_setpc_b64 s[30:31]
;
; GFX10-LABEL: v_constained_fadd_v2f64_fpexcept_strict:
; GFX10: ; %bb.0:
; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX10-NEXT: v_add_f64 v[0:1], v[0:1], v[4:5]
; GFX10-NEXT: v_add_f64 v[2:3], v[2:3], v[6:7]
; GFX10-NEXT: s_setpc_b64 s[30:31]
%val = call <2 x double> @llvm.experimental.constrained.fadd.v2f64(<2 x double> %x, <2 x double> %y, metadata !"round.tonearest", metadata !"fpexcept.strict")
ret <2 x double> %val
}
Expand All @@ -76,13 +56,6 @@ define <2 x double> @v_constained_fadd_v2f64_fpexcept_ignore(<2 x double> %x, <2
; GCN-NEXT: v_add_f64 v[0:1], v[0:1], v[4:5]
; GCN-NEXT: v_add_f64 v[2:3], v[2:3], v[6:7]
; GCN-NEXT: s_setpc_b64 s[30:31]
;
; GFX10-LABEL: v_constained_fadd_v2f64_fpexcept_ignore:
; GFX10: ; %bb.0:
; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX10-NEXT: v_add_f64 v[0:1], v[0:1], v[4:5]
; GFX10-NEXT: v_add_f64 v[2:3], v[2:3], v[6:7]
; GFX10-NEXT: s_setpc_b64 s[30:31]
%val = call <2 x double> @llvm.experimental.constrained.fadd.v2f64(<2 x double> %x, <2 x double> %y, metadata !"round.tonearest", metadata !"fpexcept.ignore")
ret <2 x double> %val
}
Expand All @@ -94,13 +67,6 @@ define <2 x double> @v_constained_fadd_v2f64_fpexcept_maytrap(<2 x double> %x, <
; GCN-NEXT: v_add_f64 v[0:1], v[0:1], v[4:5]
; GCN-NEXT: v_add_f64 v[2:3], v[2:3], v[6:7]
; GCN-NEXT: s_setpc_b64 s[30:31]
;
; GFX10-LABEL: v_constained_fadd_v2f64_fpexcept_maytrap:
; GFX10: ; %bb.0:
; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX10-NEXT: v_add_f64 v[0:1], v[0:1], v[4:5]
; GFX10-NEXT: v_add_f64 v[2:3], v[2:3], v[6:7]
; GFX10-NEXT: s_setpc_b64 s[30:31]
%val = call <2 x double> @llvm.experimental.constrained.fadd.v2f64(<2 x double> %x, <2 x double> %y, metadata !"round.tonearest", metadata !"fpexcept.maytrap")
ret <2 x double> %val
}
Expand All @@ -113,30 +79,29 @@ define <3 x double> @v_constained_fadd_v3f64_fpexcept_strict(<3 x double> %x, <3
; GCN-NEXT: v_add_f64 v[2:3], v[2:3], v[8:9]
; GCN-NEXT: v_add_f64 v[4:5], v[4:5], v[10:11]
; GCN-NEXT: s_setpc_b64 s[30:31]
;
; GFX10-LABEL: v_constained_fadd_v3f64_fpexcept_strict:
; GFX10: ; %bb.0:
; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX10-NEXT: v_add_f64 v[0:1], v[0:1], v[6:7]
; GFX10-NEXT: v_add_f64 v[2:3], v[2:3], v[8:9]
; GFX10-NEXT: v_add_f64 v[4:5], v[4:5], v[10:11]
; GFX10-NEXT: s_setpc_b64 s[30:31]
%val = call <3 x double> @llvm.experimental.constrained.fadd.v3f64(<3 x double> %x, <3 x double> %y, metadata !"round.tonearest", metadata !"fpexcept.strict")
ret <3 x double> %val
}

define amdgpu_ps <2 x float> @s_constained_fadd_f64_fpexcept_strict(double inreg %x, double inreg %y) #0 {
; GCN-LABEL: s_constained_fadd_f64_fpexcept_strict:
; GCN: ; %bb.0:
; GCN-NEXT: v_mov_b32_e32 v0, s4
; GCN-NEXT: v_mov_b32_e32 v1, s5
; GCN-NEXT: v_add_f64 v[0:1], s[2:3], v[0:1]
; GCN-NEXT: ; return to shader part epilog
; GCN-SDAG-LABEL: s_constained_fadd_f64_fpexcept_strict:
; GCN-SDAG: ; %bb.0:
; GCN-SDAG-NEXT: v_mov_b32_e32 v0, s4
; GCN-SDAG-NEXT: v_mov_b32_e32 v1, s5
; GCN-SDAG-NEXT: v_add_f64 v[0:1], s[2:3], v[0:1]
; GCN-SDAG-NEXT: ; return to shader part epilog
;
; GCN-GISEL-LABEL: s_constained_fadd_f64_fpexcept_strict:
; GCN-GISEL: ; %bb.0:
; GCN-GISEL-NEXT: v_mov_b32_e32 v0, s4
; GCN-GISEL-NEXT: v_mov_b32_e32 v1, s5
; GCN-GISEL-NEXT: v_add_f64 v[0:1], s[2:3], v[0:1]
; GCN-GISEL-NEXT: ; return to shader part epilog
;
; GFX10-LABEL: s_constained_fadd_f64_fpexcept_strict:
; GFX10: ; %bb.0:
; GFX10-NEXT: v_add_f64 v[0:1], s[2:3], s[4:5]
; GFX10-NEXT: ; return to shader part epilog
; GFX10PLUS-LABEL: s_constained_fadd_f64_fpexcept_strict:
; GFX10PLUS: ; %bb.0:
; GFX10PLUS-NEXT: v_add_f64 v[0:1], s[2:3], s[4:5]
; GFX10PLUS-NEXT: ; return to shader part epilog
%val = call double @llvm.experimental.constrained.fadd.f64(double %x, double %y, metadata !"round.tonearest", metadata !"fpexcept.strict")
%cast = bitcast double %val to <2 x float>
ret <2 x float> %cast
Expand All @@ -148,3 +113,6 @@ declare <3 x double> @llvm.experimental.constrained.fadd.v3f64(<3 x double>, <3

attributes #0 = { strictfp }
attributes #1 = { inaccessiblememonly nounwind willreturn }
;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:
; GFX10: {{.*}}
; GFX11: {{.*}}
Loading