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
107 lines
3.3 KiB
LLVM
107 lines
3.3 KiB
LLVM
; RUN: llc < %s -mtriple=s390x-linux-gnu -mcpu=z10 \
|
|
; RUN: | FileCheck -check-prefix=CHECK -check-prefix=CHECK-SCALAR %s
|
|
; RUN: llc < %s -mtriple=s390x-linux-gnu -mcpu=z13 \
|
|
; RUN: | FileCheck -check-prefix=CHECK -check-prefix=CHECK-VECTOR %s
|
|
|
|
declare double @llvm.fma.f64(double %f1, double %f2, double %f3)
|
|
|
|
define double @f1(double %f1, double %f2, double %acc) {
|
|
; CHECK-LABEL: f1:
|
|
; CHECK-SCALAR: madbr %f4, %f0, %f2
|
|
; CHECK-SCALAR: ldr %f0, %f4
|
|
; CHECK-VECTOR: wfmadb %f0, %f0, %f2, %f4
|
|
; CHECK: br %r14
|
|
%res = call double @llvm.fma.f64 (double %f1, double %f2, double %acc)
|
|
ret double %res
|
|
}
|
|
|
|
define double @f2(double %f1, ptr %ptr, double %acc) {
|
|
; CHECK-LABEL: f2:
|
|
; CHECK: madb %f2, %f0, 0(%r2)
|
|
; CHECK: ldr %f0, %f2
|
|
; CHECK: br %r14
|
|
%f2 = load double, ptr %ptr
|
|
%res = call double @llvm.fma.f64 (double %f1, double %f2, double %acc)
|
|
ret double %res
|
|
}
|
|
|
|
define double @f3(double %f1, ptr %base, double %acc) {
|
|
; CHECK-LABEL: f3:
|
|
; CHECK: madb %f2, %f0, 4088(%r2)
|
|
; CHECK: ldr %f0, %f2
|
|
; CHECK: br %r14
|
|
%ptr = getelementptr double, ptr %base, i64 511
|
|
%f2 = load double, ptr %ptr
|
|
%res = call double @llvm.fma.f64 (double %f1, double %f2, double %acc)
|
|
ret double %res
|
|
}
|
|
|
|
define double @f4(double %f1, ptr %base, double %acc) {
|
|
; The important thing here is that we don't generate an out-of-range
|
|
; displacement. Other sequences besides this one would be OK.
|
|
;
|
|
; CHECK-LABEL: f4:
|
|
; CHECK: aghi %r2, 4096
|
|
; CHECK: madb %f2, %f0, 0(%r2)
|
|
; CHECK: ldr %f0, %f2
|
|
; CHECK: br %r14
|
|
%ptr = getelementptr double, ptr %base, i64 512
|
|
%f2 = load double, ptr %ptr
|
|
%res = call double @llvm.fma.f64 (double %f1, double %f2, double %acc)
|
|
ret double %res
|
|
}
|
|
|
|
define double @f5(double %f1, ptr %base, double %acc) {
|
|
; Here too the important thing is that we don't generate an out-of-range
|
|
; displacement. Other sequences besides this one would be OK.
|
|
;
|
|
; CHECK-LABEL: f5:
|
|
; CHECK: aghi %r2, -8
|
|
; CHECK: madb %f2, %f0, 0(%r2)
|
|
; CHECK: ldr %f0, %f2
|
|
; CHECK: br %r14
|
|
%ptr = getelementptr double, ptr %base, i64 -1
|
|
%f2 = load double, ptr %ptr
|
|
%res = call double @llvm.fma.f64 (double %f1, double %f2, double %acc)
|
|
ret double %res
|
|
}
|
|
|
|
define double @f6(double %f1, ptr %base, i64 %index, double %acc) {
|
|
; CHECK-LABEL: f6:
|
|
; CHECK: sllg %r1, %r3, 3
|
|
; CHECK: madb %f2, %f0, 0(%r1,%r2)
|
|
; CHECK: ldr %f0, %f2
|
|
; CHECK: br %r14
|
|
%ptr = getelementptr double, ptr %base, i64 %index
|
|
%f2 = load double, ptr %ptr
|
|
%res = call double @llvm.fma.f64 (double %f1, double %f2, double %acc)
|
|
ret double %res
|
|
}
|
|
|
|
define double @f7(double %f1, ptr %base, i64 %index, double %acc) {
|
|
; CHECK-LABEL: f7:
|
|
; CHECK: sllg %r1, %r3, 3
|
|
; CHECK: madb %f2, %f0, 4088({{%r1,%r2|%r2,%r1}})
|
|
; CHECK: ldr %f0, %f2
|
|
; CHECK: br %r14
|
|
%index2 = add i64 %index, 511
|
|
%ptr = getelementptr double, ptr %base, i64 %index2
|
|
%f2 = load double, ptr %ptr
|
|
%res = call double @llvm.fma.f64 (double %f1, double %f2, double %acc)
|
|
ret double %res
|
|
}
|
|
|
|
define double @f8(double %f1, ptr %base, i64 %index, double %acc) {
|
|
; CHECK-LABEL: f8:
|
|
; CHECK: sllg %r1, %r3, 3
|
|
; CHECK: lay %r1, 4096({{%r1,%r2|%r2,%r1}})
|
|
; CHECK: madb %f2, %f0, 0(%r1)
|
|
; CHECK: ldr %f0, %f2
|
|
; CHECK: br %r14
|
|
%index2 = add i64 %index, 512
|
|
%ptr = getelementptr double, ptr %base, i64 %index2
|
|
%f2 = load double, ptr %ptr
|
|
%res = call double @llvm.fma.f64 (double %f1, double %f2, double %acc)
|
|
ret double %res
|
|
}
|