RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-25 20:23:11 +01:00
Code Issues Projects Releases Wiki Activity
da14179563
llvm-mirror/test/CodeGen/X86/rdseed.ll

49 lines
1.4 KiB
LLVM
Raw Normal View History

llvm/test/CodeGen/X86: Add explicit -mtriple=x86_64-unknown-unknown. llvm-svn: 184731
2013-06-24 15:19:59 +02:00
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=core-avx-i -mattr=+rdseed | FileCheck %s
Add support of RDSEED defined in AVX2 extension llvm-svn: 178314
2013-03-29 00:41:26 +01:00
declare {i16, i32} @llvm.x86.rdseed.16()
declare {i32, i32} @llvm.x86.rdseed.32()
declare {i64, i32} @llvm.x86.rdseed.64()
define i32 @_rdseed16_step(i16* %random_val) {
%call = call {i16, i32} @llvm.x86.rdseed.16()
%randval = extractvalue {i16, i32} %call, 0
store i16 %randval, i16* %random_val
%isvalid = extractvalue {i16, i32} %call, 1
ret i32 %isvalid
Mass update to CodeGen tests to use CHECK-LABEL for labels corresponding to function definitions for more informative error messages. No functionality change and all updated tests passed locally. This update was done with the following bash script: find test/CodeGen -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc.*debug" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_-]*\):\( *\)$FUNC: *\$/;\1\2-LABEL:\3$FUNC:/g" $TEMP done sed -i '' "s/;\(.*\)-LABEL-LABEL:/;\1-LABEL:/" $TEMP sed -i '' "s/;\(.*\)-NEXT-LABEL:/;\1-NEXT:/" $TEMP sed -i '' "s/;\(.*\)-NOT-LABEL:/;\1-NOT:/" $TEMP sed -i '' "s/;\(.*\)-DAG-LABEL:/;\1-DAG:/" $TEMP mv $TEMP $NAME fi done llvm-svn: 186280
2013-07-14 08:24:09 +02:00
; CHECK-LABEL: _rdseed16_step:
Add support of RDSEED defined in AVX2 extension llvm-svn: 178314
2013-03-29 00:41:26 +01:00
; CHECK: rdseedw %ax
; CHECK: movzwl %ax, %ecx
; CHECK: movl $1, %eax
; CHECK: cmovael %ecx, %eax
Enable MI Sched for x86. This changes the SelectionDAG scheduling preference to source order. Soon, the SelectionDAG scheduler can be bypassed saving a nice chunk of compile time. Performance differences that result from this change are often a consequence of register coalescing. The register coalescer is far from perfect. Bugs can be filed for deficiencies. On x86 SandyBridge/Haswell, the source order schedule is often preserved, particularly for small blocks. Register pressure is generally improved over the SD scheduler's ILP mode. However, we are still able to handle large blocks that require latency hiding, unlike the SD scheduler's BURR mode. MI scheduler also attempts to discover the critical path in single-block loops and adjust heuristics accordingly. The MI scheduler relies on the new machine model. This is currently unimplemented for AVX, so we may not be generating the best code yet. Unit tests are updated so they don't depend on SD scheduling heuristics. llvm-svn: 192750
2013-10-16 01:33:07 +02:00
; CHECK: movw %cx, (%r[[A0:di|cx]])
Add support of RDSEED defined in AVX2 extension llvm-svn: 178314
2013-03-29 00:41:26 +01:00
; CHECK: ret
}
define i32 @_rdseed32_step(i32* %random_val) {
%call = call {i32, i32} @llvm.x86.rdseed.32()
%randval = extractvalue {i32, i32} %call, 0
store i32 %randval, i32* %random_val
%isvalid = extractvalue {i32, i32} %call, 1
ret i32 %isvalid
Mass update to CodeGen tests to use CHECK-LABEL for labels corresponding to function definitions for more informative error messages. No functionality change and all updated tests passed locally. This update was done with the following bash script: find test/CodeGen -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc.*debug" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_-]*\):\( *\)$FUNC: *\$/;\1\2-LABEL:\3$FUNC:/g" $TEMP done sed -i '' "s/;\(.*\)-LABEL-LABEL:/;\1-LABEL:/" $TEMP sed -i '' "s/;\(.*\)-NEXT-LABEL:/;\1-NEXT:/" $TEMP sed -i '' "s/;\(.*\)-NOT-LABEL:/;\1-NOT:/" $TEMP sed -i '' "s/;\(.*\)-DAG-LABEL:/;\1-DAG:/" $TEMP mv $TEMP $NAME fi done llvm-svn: 186280
2013-07-14 08:24:09 +02:00
; CHECK-LABEL: _rdseed32_step:
Add support of RDSEED defined in AVX2 extension llvm-svn: 178314
2013-03-29 00:41:26 +01:00
; CHECK: rdseedl %e[[T0:[a-z]+]]
; CHECK: movl $1, %eax
; CHECK: cmovael %e[[T0]], %eax
Enable MI Sched for x86. This changes the SelectionDAG scheduling preference to source order. Soon, the SelectionDAG scheduler can be bypassed saving a nice chunk of compile time. Performance differences that result from this change are often a consequence of register coalescing. The register coalescer is far from perfect. Bugs can be filed for deficiencies. On x86 SandyBridge/Haswell, the source order schedule is often preserved, particularly for small blocks. Register pressure is generally improved over the SD scheduler's ILP mode. However, we are still able to handle large blocks that require latency hiding, unlike the SD scheduler's BURR mode. MI scheduler also attempts to discover the critical path in single-block loops and adjust heuristics accordingly. The MI scheduler relies on the new machine model. This is currently unimplemented for AVX, so we may not be generating the best code yet. Unit tests are updated so they don't depend on SD scheduling heuristics. llvm-svn: 192750
2013-10-16 01:33:07 +02:00
; CHECK: movl %e[[T0]], (%r[[A0]])
Add support of RDSEED defined in AVX2 extension llvm-svn: 178314
2013-03-29 00:41:26 +01:00
; CHECK: ret
}
define i32 @_rdseed64_step(i64* %random_val) {
%call = call {i64, i32} @llvm.x86.rdseed.64()
%randval = extractvalue {i64, i32} %call, 0
store i64 %randval, i64* %random_val
%isvalid = extractvalue {i64, i32} %call, 1
ret i32 %isvalid
Mass update to CodeGen tests to use CHECK-LABEL for labels corresponding to function definitions for more informative error messages. No functionality change and all updated tests passed locally. This update was done with the following bash script: find test/CodeGen -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc.*debug" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_-]*\):\( *\)$FUNC: *\$/;\1\2-LABEL:\3$FUNC:/g" $TEMP done sed -i '' "s/;\(.*\)-LABEL-LABEL:/;\1-LABEL:/" $TEMP sed -i '' "s/;\(.*\)-NEXT-LABEL:/;\1-NEXT:/" $TEMP sed -i '' "s/;\(.*\)-NOT-LABEL:/;\1-NOT:/" $TEMP sed -i '' "s/;\(.*\)-DAG-LABEL:/;\1-DAG:/" $TEMP mv $TEMP $NAME fi done llvm-svn: 186280
2013-07-14 08:24:09 +02:00
; CHECK-LABEL: _rdseed64_step:
Add support of RDSEED defined in AVX2 extension llvm-svn: 178314
2013-03-29 00:41:26 +01:00
; CHECK: rdseedq %r[[T1:[a-z]+]]
; CHECK: movl $1, %eax
; CHECK: cmovael %e[[T1]], %eax
Enable MI Sched for x86. This changes the SelectionDAG scheduling preference to source order. Soon, the SelectionDAG scheduler can be bypassed saving a nice chunk of compile time. Performance differences that result from this change are often a consequence of register coalescing. The register coalescer is far from perfect. Bugs can be filed for deficiencies. On x86 SandyBridge/Haswell, the source order schedule is often preserved, particularly for small blocks. Register pressure is generally improved over the SD scheduler's ILP mode. However, we are still able to handle large blocks that require latency hiding, unlike the SD scheduler's BURR mode. MI scheduler also attempts to discover the critical path in single-block loops and adjust heuristics accordingly. The MI scheduler relies on the new machine model. This is currently unimplemented for AVX, so we may not be generating the best code yet. Unit tests are updated so they don't depend on SD scheduling heuristics. llvm-svn: 192750
2013-10-16 01:33:07 +02:00
; CHECK: movq %r[[T1]], (%r[[A0]])
Add support of RDSEED defined in AVX2 extension llvm-svn: 178314
2013-03-29 00:41:26 +01:00
; CHECK: ret
}
Reference in New Issue Copy Permalink