7fff1fbd3c
This hashing scheme has been useful out of tree, and I want to start experimenting with it. Specifically I want to experiment on the MIRVRegNamer, MIRCanononicalizer, and eventually the MachineOutliner. This diff is a first step, that optionally brings stable hashing to the MIRVRegNamer (and as a result, the MIRCanonicalizer). We've tested this hashing scheme on a lot of MachineOperand types that llvm::hash_value can not handle in a stable manner. This stable hashing was also the basis for "Global Machine Outliner for ThinLTO" in EuroLLVM 2020 http://llvm.org/devmtg/2020-04/talks.html#TechTalk_58 Credits: Kyungwoo Lee, Nikolai Tillmann Differential Revision: https://reviews.llvm.org/D86952
34 lines
1.4 KiB
YAML
34 lines
1.4 KiB
YAML
# RUN: llc -mtriple=aarch64-none-linux-gnu -run-pass none -verify-machineinstrs -o - %s | FileCheck %s
|
|
# RUN: llc -mtriple=aarch64-none-linux-gnu -run-pass mir-canonicalizer -verify-machineinstrs -o - %s
|
|
# RUN: llc -mtriple=aarch64-none-linux-gnu -run-pass mir-canonicalizer -mir-vreg-namer-use-stable-hash -verify-machineinstrs -o - %s
|
|
|
|
--- |
|
|
|
|
define void @addrspace_memoperands() {
|
|
ret void
|
|
}
|
|
|
|
...
|
|
---
|
|
name: addrspace_memoperands
|
|
body: |
|
|
bb.0:
|
|
|
|
; CHECK-LABEL: name: addrspace_memoperands
|
|
; CHECK: [[COPY:%[0-9]+]]:_(p0) = COPY $x0
|
|
; CHECK: [[LOAD:%[0-9]+]]:_(s64) = G_LOAD [[COPY]](p0) :: (load 8, addrspace 1)
|
|
; CHECK: [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD [[COPY]](p0) :: (load 4, align 2, addrspace 3)
|
|
; CHECK: G_STORE [[LOAD]](s64), [[COPY]](p0) :: (store 8, addrspace 1)
|
|
; CHECK: G_STORE [[LOAD1]](s32), [[COPY]](p0) :: (store 4, align 2, addrspace 3)
|
|
; CHECK: G_STORE [[LOAD1]](s32), [[COPY]](p0) :: (store 4)
|
|
; CHECK: RET_ReallyLR
|
|
%0:_(p0) = COPY $x0
|
|
%1:_(s64) = G_LOAD %0(p0) :: (load 8, addrspace 1)
|
|
%2:_(s32) = G_LOAD %0(p0) :: (load 4, align 2, addrspace 3)
|
|
G_STORE %1(s64), %0(p0) :: (store 8, addrspace 1)
|
|
G_STORE %2(s32), %0(p0) :: (store 4, align 2, addrspace 3)
|
|
; addrspace 0 is accepted by the parser but not printed
|
|
G_STORE %2(s32), %0(p0) :: (store 4, addrspace 0)
|
|
RET_ReallyLR
|
|
...
|