derek.holloway
7502bf64a4
CodeQL / Analyze (actions) (push) Canceled after 0s
CodeQL / Analyze (python) (push) Canceled after 0s
pre-commit / pre-commit (push) Canceled after 0s
Multi-Arch CI / setup (push) Canceled after 0s
Publish build_manylinux_x86_64 images / publish_build_manylinux_x86_64 (push) Canceled after 0s
Setup (Multi-Arch) / setup (push) Canceled after 0s
Multi-Arch CI / Linux::${{ fromJSON(needs.setup.outputs.linux_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubi10 image / publish_no_rocm_image_ubi10 (push) Canceled after 0s
Multi-Arch CI / Windows::${{ fromJSON(needs.setup.outputs.windows_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Multi-Arch CI / Manifest Diff (push) Canceled after 0s
TheRock Manifest Diff Report / Generate Manifest Diff Report (push) Canceled after 0s
Multi-Arch CI / CI Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_ocl_rt images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_rocgdb images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Unit Tests / Unit Tests :: ubuntu-24.04 (push) Canceled after 0s
Unit Tests / Unit Tests :: windows-2022 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_media (push) Canceled after 0s
Unit Tests / Unit Tests Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_openmpi (push) Canceled after 0s
117 lines
5.1 KiB
LLVM
117 lines
5.1 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: llc -global-isel -new-reg-bank-select -mtriple=amdgcn -mcpu=gfx900 < %s | FileCheck -check-prefixes=GFX9-32BANK %s
|
|
; RUN: llc -global-isel -new-reg-bank-select -mtriple=amdgcn -mcpu=fiji < %s | FileCheck -check-prefixes=GFX8-32BANK %s
|
|
; RUN: llc -global-isel -new-reg-bank-select -mtriple=amdgcn -mcpu=gfx810 < %s | FileCheck -check-prefixes=GFX8-16BANK %s
|
|
|
|
define amdgpu_ps float @interp_f16(float %i, i32 inreg %m0) #0 {
|
|
; GFX9-32BANK-LABEL: interp_f16:
|
|
; GFX9-32BANK: ; %bb.0:
|
|
; GFX9-32BANK-NEXT: s_mov_b32 m0, s0
|
|
; GFX9-32BANK-NEXT: s_setreg_imm32_b32 hwreg(HW_REG_MODE, 2, 2), 3
|
|
; GFX9-32BANK-NEXT: v_interp_p1ll_f16 v0, v0, attr2.y
|
|
; GFX9-32BANK-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX8-32BANK-LABEL: interp_f16:
|
|
; GFX8-32BANK: ; %bb.0:
|
|
; GFX8-32BANK-NEXT: s_mov_b32 m0, s0
|
|
; GFX8-32BANK-NEXT: s_setreg_imm32_b32 hwreg(HW_REG_MODE, 2, 2), 3
|
|
; GFX8-32BANK-NEXT: v_interp_p1ll_f16 v0, v0, attr2.y
|
|
; GFX8-32BANK-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX8-16BANK-LABEL: interp_f16:
|
|
; GFX8-16BANK: ; %bb.0:
|
|
; GFX8-16BANK-NEXT: s_mov_b32 m0, s0
|
|
; GFX8-16BANK-NEXT: v_interp_mov_f32_e32 v1, p0, attr2.y
|
|
; GFX8-16BANK-NEXT: s_setreg_imm32_b32 hwreg(HW_REG_MODE, 2, 2), 3
|
|
; GFX8-16BANK-NEXT: v_interp_p1lv_f16 v0, v0, attr2.y, v1
|
|
; GFX8-16BANK-NEXT: ; return to shader part epilog
|
|
%res = call float @llvm.amdgcn.interp.p1.f16(float %i, i32 1, i32 2, i1 false, i32 %m0)
|
|
ret float %res
|
|
}
|
|
|
|
define amdgpu_ps float @interp_f16_high(float %i, i32 inreg %m0) #0 {
|
|
; GFX9-32BANK-LABEL: interp_f16_high:
|
|
; GFX9-32BANK: ; %bb.0:
|
|
; GFX9-32BANK-NEXT: s_mov_b32 m0, s0
|
|
; GFX9-32BANK-NEXT: s_setreg_imm32_b32 hwreg(HW_REG_MODE, 2, 2), 3
|
|
; GFX9-32BANK-NEXT: v_interp_p1ll_f16 v0, v0, attr2.y high
|
|
; GFX9-32BANK-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX8-32BANK-LABEL: interp_f16_high:
|
|
; GFX8-32BANK: ; %bb.0:
|
|
; GFX8-32BANK-NEXT: s_mov_b32 m0, s0
|
|
; GFX8-32BANK-NEXT: s_setreg_imm32_b32 hwreg(HW_REG_MODE, 2, 2), 3
|
|
; GFX8-32BANK-NEXT: v_interp_p1ll_f16 v0, v0, attr2.y high
|
|
; GFX8-32BANK-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX8-16BANK-LABEL: interp_f16_high:
|
|
; GFX8-16BANK: ; %bb.0:
|
|
; GFX8-16BANK-NEXT: s_mov_b32 m0, s0
|
|
; GFX8-16BANK-NEXT: v_interp_mov_f32_e32 v1, p0, attr2.y
|
|
; GFX8-16BANK-NEXT: s_setreg_imm32_b32 hwreg(HW_REG_MODE, 2, 2), 3
|
|
; GFX8-16BANK-NEXT: v_interp_p1lv_f16 v0, v0, attr2.y, v1 high
|
|
; GFX8-16BANK-NEXT: ; return to shader part epilog
|
|
%res = call float @llvm.amdgcn.interp.p1.f16(float %i, i32 1, i32 2, i1 true, i32 %m0)
|
|
ret float %res
|
|
}
|
|
|
|
define amdgpu_ps float @interp_f16_0_0(float %i, i32 inreg %m0) #0 {
|
|
; GFX9-32BANK-LABEL: interp_f16_0_0:
|
|
; GFX9-32BANK: ; %bb.0:
|
|
; GFX9-32BANK-NEXT: s_mov_b32 m0, s0
|
|
; GFX9-32BANK-NEXT: s_setreg_imm32_b32 hwreg(HW_REG_MODE, 2, 2), 3
|
|
; GFX9-32BANK-NEXT: v_interp_p1ll_f16 v0, v0, attr0.x
|
|
; GFX9-32BANK-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX8-32BANK-LABEL: interp_f16_0_0:
|
|
; GFX8-32BANK: ; %bb.0:
|
|
; GFX8-32BANK-NEXT: s_mov_b32 m0, s0
|
|
; GFX8-32BANK-NEXT: s_setreg_imm32_b32 hwreg(HW_REG_MODE, 2, 2), 3
|
|
; GFX8-32BANK-NEXT: v_interp_p1ll_f16 v0, v0, attr0.x
|
|
; GFX8-32BANK-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX8-16BANK-LABEL: interp_f16_0_0:
|
|
; GFX8-16BANK: ; %bb.0:
|
|
; GFX8-16BANK-NEXT: s_mov_b32 m0, s0
|
|
; GFX8-16BANK-NEXT: v_interp_mov_f32_e32 v1, p0, attr0.x
|
|
; GFX8-16BANK-NEXT: s_setreg_imm32_b32 hwreg(HW_REG_MODE, 2, 2), 3
|
|
; GFX8-16BANK-NEXT: v_interp_p1lv_f16 v0, v0, attr0.x, v1
|
|
; GFX8-16BANK-NEXT: ; return to shader part epilog
|
|
%res = call float @llvm.amdgcn.interp.p1.f16(float %i, i32 0, i32 0, i1 false, i32 %m0)
|
|
ret float %res
|
|
}
|
|
|
|
; Copy needed to legalize %i
|
|
define amdgpu_ps float @interp_f16_sgpr_i(float inreg %i,i32 inreg %m0) #0 {
|
|
; GFX9-32BANK-LABEL: interp_f16_sgpr_i:
|
|
; GFX9-32BANK: ; %bb.0:
|
|
; GFX9-32BANK-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX9-32BANK-NEXT: s_mov_b32 m0, s1
|
|
; GFX9-32BANK-NEXT: s_setreg_imm32_b32 hwreg(HW_REG_MODE, 2, 2), 3
|
|
; GFX9-32BANK-NEXT: v_interp_p1ll_f16 v0, v0, attr2.y
|
|
; GFX9-32BANK-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX8-32BANK-LABEL: interp_f16_sgpr_i:
|
|
; GFX8-32BANK: ; %bb.0:
|
|
; GFX8-32BANK-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX8-32BANK-NEXT: s_mov_b32 m0, s1
|
|
; GFX8-32BANK-NEXT: s_setreg_imm32_b32 hwreg(HW_REG_MODE, 2, 2), 3
|
|
; GFX8-32BANK-NEXT: v_interp_p1ll_f16 v0, v0, attr2.y
|
|
; GFX8-32BANK-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX8-16BANK-LABEL: interp_f16_sgpr_i:
|
|
; GFX8-16BANK: ; %bb.0:
|
|
; GFX8-16BANK-NEXT: s_mov_b32 m0, s1
|
|
; GFX8-16BANK-NEXT: v_mov_b32_e32 v0, s0
|
|
; GFX8-16BANK-NEXT: v_interp_mov_f32_e32 v1, p0, attr2.y
|
|
; GFX8-16BANK-NEXT: s_setreg_imm32_b32 hwreg(HW_REG_MODE, 2, 2), 3
|
|
; GFX8-16BANK-NEXT: v_interp_p1lv_f16 v0, v0, attr2.y, v1
|
|
; GFX8-16BANK-NEXT: ; return to shader part epilog
|
|
%res = call float @llvm.amdgcn.interp.p1.f16(float %i, i32 1, i32 2, i1 false, i32 %m0)
|
|
ret float %res
|
|
}
|
|
|
|
declare float @llvm.amdgcn.interp.p1.f16(float, i32 immarg, i32 immarg, i1 immarg, i32) #0
|
|
|
|
attributes #0 = { nounwind readnone speculatable }
|