This patch adds support for the next-generation arch15 CPU architecture to the SystemZ backend. This includes: - Basic support for the new processor and its features. - Detection of arch15 as host processor. - Assembler/disassembler support for new instructions. - Exploitation of new instructions for code generation. - New vector (signed|unsigned|bool) __int128 data types. - New LLVM intrinsics for certain new instructions. - Support for low-level builtins mapped to new LLVM intrinsics. - New high-level intrinsics in vecintrin.h. - Indicate support by defining __VEC__ == 10305. Note: No currently available Z system supports the arch15 architecture. Once new systems become available, the official system name will be added as supported -march name.
59 lines
1.8 KiB
LLVM
59 lines
1.8 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 2
|
|
; Test high-part i128->i256 multiplications on arch15.
|
|
;
|
|
; RUN: llc < %s -mtriple=s390x-linux-gnu -mcpu=arch15 | FileCheck %s
|
|
|
|
; Multiply high signed.
|
|
define i128 @f1(i128 %a, i128 %b) {
|
|
; CHECK-LABEL: f1:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vl %v0, 0(%r4), 3
|
|
; CHECK-NEXT: vl %v1, 0(%r3), 3
|
|
; CHECK-NEXT: vmhq %v0, %v1, %v0
|
|
; CHECK-NEXT: vst %v0, 0(%r2), 3
|
|
; CHECK-NEXT: br %r14
|
|
%exta = sext i128 %a to i256
|
|
%extb = sext i128 %b to i256
|
|
%extres = mul i256 %exta, %extb
|
|
%shiftres = lshr i256 %extres, 128
|
|
%res = trunc i256 %shiftres to i128
|
|
ret i128 %res
|
|
}
|
|
|
|
; Multiply high unsigned.
|
|
define i128 @f2(i128 %a, i128 %b) {
|
|
; CHECK-LABEL: f2:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vl %v0, 0(%r4), 3
|
|
; CHECK-NEXT: vl %v1, 0(%r3), 3
|
|
; CHECK-NEXT: vmlhq %v0, %v1, %v0
|
|
; CHECK-NEXT: vst %v0, 0(%r2), 3
|
|
; CHECK-NEXT: br %r14
|
|
%exta = zext i128 %a to i256
|
|
%extb = zext i128 %b to i256
|
|
%extres = mul i256 %exta, %extb
|
|
%shiftres = lshr i256 %extres, 128
|
|
%res = trunc i256 %shiftres to i128
|
|
ret i128 %res
|
|
}
|
|
|
|
;; ; Multiply-and-add high signed.
|
|
;; define i128 @f3(i128 %a, i128 %b, i128 %add) {
|
|
;; ; CHECX-LABEL: f3:
|
|
;; ; CHECX: # %bb.0:
|
|
;; ; CHECX-NEXT: vl %v0, 0(%r3), 3
|
|
;; ; CHECX-NEXT: vl %v1, 0(%r4), 3
|
|
;; ; CHECX-NEXT: vl %v2, 0(%r5), 3
|
|
;; ; CHECX-NEXT: vmhq %v0, %v0, %v1, %v2
|
|
;; ; CHECX-NEXT: vst %v0, 0(%r2), 3
|
|
;; ; CHECX-NEXT: br %r14
|
|
;; %exta = sext i128 %a to i256
|
|
;; %extb = sext i128 %b to i256
|
|
;; %extadd = sext i128 %add to i256
|
|
;; %extmul = mul i256 %exta, %extb
|
|
;; %extres = add i256 %extmul, %extadd
|
|
;; %shiftres = lshr i256 %extres, 128
|
|
;; %res = trunc i256 %shiftres to i128
|
|
;; ret i128 %res
|
|
;; }
|