Files
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
Flatten subprojects into a single base monorepo
2026-09-02 16:34:47 -07:00

132 lines
3.4 KiB
LLVM

; Test 32-bit multiplication in which the second operand is a sign-extended
; i16 memory value.
;
; RUN: llc < %s -mtriple=s390x-linux-gnu | FileCheck %s
; Check the low end of the MH range.
define i32 @f1(i32 %lhs, ptr %src) {
; CHECK-LABEL: f1:
; CHECK: mh %r2, 0(%r3)
; CHECK: br %r14
%half = load i16, ptr %src
%rhs = sext i16 %half to i32
%res = mul i32 %lhs, %rhs
ret i32 %res
}
; Check the high end of the aligned MH range.
define i32 @f2(i32 %lhs, ptr %src) {
; CHECK-LABEL: f2:
; CHECK: mh %r2, 4094(%r3)
; CHECK: br %r14
%ptr = getelementptr i16, ptr %src, i64 2047
%half = load i16, ptr %ptr
%rhs = sext i16 %half to i32
%res = mul i32 %lhs, %rhs
ret i32 %res
}
; Check the next halfword up, which should use MHY instead of MH.
define i32 @f3(i32 %lhs, ptr %src) {
; CHECK-LABEL: f3:
; CHECK: mhy %r2, 4096(%r3)
; CHECK: br %r14
%ptr = getelementptr i16, ptr %src, i64 2048
%half = load i16, ptr %ptr
%rhs = sext i16 %half to i32
%res = mul i32 %lhs, %rhs
ret i32 %res
}
; Check the high end of the aligned MHY range.
define i32 @f4(i32 %lhs, ptr %src) {
; CHECK-LABEL: f4:
; CHECK: mhy %r2, 524286(%r3)
; CHECK: br %r14
%ptr = getelementptr i16, ptr %src, i64 262143
%half = load i16, ptr %ptr
%rhs = sext i16 %half to i32
%res = mul i32 %lhs, %rhs
ret i32 %res
}
; Check the next halfword up, which needs separate address logic.
; Other sequences besides this one would be OK.
define i32 @f5(i32 %lhs, ptr %src) {
; CHECK-LABEL: f5:
; CHECK: agfi %r3, 524288
; CHECK: mh %r2, 0(%r3)
; CHECK: br %r14
%ptr = getelementptr i16, ptr %src, i64 262144
%half = load i16, ptr %ptr
%rhs = sext i16 %half to i32
%res = mul i32 %lhs, %rhs
ret i32 %res
}
; Check the high end of the negative aligned MHY range.
define i32 @f6(i32 %lhs, ptr %src) {
; CHECK-LABEL: f6:
; CHECK: mhy %r2, -2(%r3)
; CHECK: br %r14
%ptr = getelementptr i16, ptr %src, i64 -1
%half = load i16, ptr %ptr
%rhs = sext i16 %half to i32
%res = mul i32 %lhs, %rhs
ret i32 %res
}
; Check the low end of the MHY range.
define i32 @f7(i32 %lhs, ptr %src) {
; CHECK-LABEL: f7:
; CHECK: mhy %r2, -524288(%r3)
; CHECK: br %r14
%ptr = getelementptr i16, ptr %src, i64 -262144
%half = load i16, ptr %ptr
%rhs = sext i16 %half to i32
%res = mul i32 %lhs, %rhs
ret i32 %res
}
; Check the next halfword down, which needs separate address logic.
; Other sequences besides this one would be OK.
define i32 @f8(i32 %lhs, ptr %src) {
; CHECK-LABEL: f8:
; CHECK: agfi %r3, -524290
; CHECK: mh %r2, 0(%r3)
; CHECK: br %r14
%ptr = getelementptr i16, ptr %src, i64 -262145
%half = load i16, ptr %ptr
%rhs = sext i16 %half to i32
%res = mul i32 %lhs, %rhs
ret i32 %res
}
; Check that MH allows an index.
define i32 @f9(i32 %lhs, i64 %src, i64 %index) {
; CHECK-LABEL: f9:
; CHECK: mh %r2, 4094({{%r4,%r3|%r3,%r4}})
; CHECK: br %r14
%add1 = add i64 %src, %index
%add2 = add i64 %add1, 4094
%ptr = inttoptr i64 %add2 to ptr
%half = load i16, ptr %ptr
%rhs = sext i16 %half to i32
%res = mul i32 %lhs, %rhs
ret i32 %res
}
; Check that MHY allows an index.
define i32 @f10(i32 %lhs, i64 %src, i64 %index) {
; CHECK-LABEL: f10:
; CHECK: mhy %r2, 4096({{%r4,%r3|%r3,%r4}})
; CHECK: br %r14
%add1 = add i64 %src, %index
%add2 = add i64 %add1, 4096
%ptr = inttoptr i64 %add2 to ptr
%half = load i16, ptr %ptr
%rhs = sext i16 %half to i32
%res = mul i32 %lhs, %rhs
ret i32 %res
}