Summary:
For flat loop, even if it is hot, it is not a good idea to unroll in runtime, thus we set a lower partial unroll threshold.
For hot loop, we set a higher unroll threshold and allows expensive tripcount computation to allow more aggressive unrolling.
Reviewers: davidxl, mzolotukhin
Subscribers: sanjoy, mehdi_amini, llvm-commits
Differential Revision: https://reviews.llvm.org/D26527
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@287186 91177308-0d34-0410-b5e6-96231b3b80d8
This is a remnant of an on-chip unit testing tool that has since been
moved out-of-tree.
It was accidentally committed in r287162.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@287180 91177308-0d34-0410-b5e6-96231b3b80d8
The '-fpermissive' compiler flag complains if the template
specializations used in the class are used in a different namespace.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@287176 91177308-0d34-0410-b5e6-96231b3b80d8
We save an inter-register file move this way. If there's any CPU where
the FP logic is slower, we could transform this back to int-logic in
MachineCombiner.
This helps, but doesn't solve, PR6137:
https://llvm.org/bugs/show_bug.cgi?id=6137
The 'andn' test shows that we're missing a pattern match to
recognize the xor with -1 constant as a 'not' op.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@287171 91177308-0d34-0410-b5e6-96231b3b80d8
This should prevent stack overflows in non-optimized builds on
.ll files with lots of consecutive commented-out lines.
Instead of recursing into LexToken(), continue into a 'while (true)'.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@287170 91177308-0d34-0410-b5e6-96231b3b80d8
This reverts commit r286962.
We want to avoid depending on SelectionDAG, and AddLandingPadInfo
lives in CodeGen now.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@287168 91177308-0d34-0410-b5e6-96231b3b80d8
They're not SelectionDAG- or FunctionLoweringInfo-specific. They
are, however, specific to building MMI from IR.
We could make them members, but it's nice having MMI be a "simple" data
structure and this logic kept separate.
This also lets us reuse them from GlobalISel.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@287167 91177308-0d34-0410-b5e6-96231b3b80d8
No real functional change with this commit.
The problem with report_fatal_error() is it does not include the tool name
and the file name the for which the error message was generated.
Uses of report_fatal_error() were change to report_error() or error()
to get a better error and to make the code smaller and cleaner.
Also changed things like error(errorToErrorCode(SOrErr.takeError())) to
use report_error() with a file name and the llvm::Error (as well as the
ArchitectureName if available) so the error message is printed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@287163 91177308-0d34-0410-b5e6-96231b3b80d8
Summary:
A lot of the pseudo instructions are required because LLVM assumes that
all integers of the same size as the pointer size are legal. This means
that it will not currently expand 16-bit instructions to their 8-bit
variants because it thinks 16-bit types are legal for the operations.
This also adds all of the CodeGen tests that required the pass to run.
Reviewers: arsenm, kparzysz
Subscribers: wdng, mgorny, modocache, llvm-commits
Differential Revision: https://reviews.llvm.org/D26577
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@287162 91177308-0d34-0410-b5e6-96231b3b80d8
We only ever create TargetConstantPool, TargetJumpTable, TargetExternalSymbol,
TargetGlobalAddress, TargetGlobalTLSAddress, MCSymbol and TargetBlockAddress
nodes as operands of X86ISD::Wrapper nodes, so we can remove one check and
invert the other.
Also update the documentation comment for X86ISD::Wrapper.
Differential Revision: https://reviews.llvm.org/D26731
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@287160 91177308-0d34-0410-b5e6-96231b3b80d8
We don't track callee clobbered registers correctly, so avoid hoisting
across calls.
Note: for this bug to trigger we need a `readonly` call target, since we
already have logic to not hoist across potentially storing instructions
either.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@287159 91177308-0d34-0410-b5e6-96231b3b80d8
One half of the shifts obviously needed conditional selection based on whether
the shift amount is more than 32-bits, but leaving the other half as the
natural shift isn't acceptable either: it's undefined behaviour to shift a
32-bit value by more than 31.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@287149 91177308-0d34-0410-b5e6-96231b3b80d8
We fail to produce bit-to-bit matching stage2 and stage3 compiler in PGO
bootstrap build. The reason is because LoopBlockSet is of SmallPtrSet type
whose iterating order depends on the pointer value.
This patch fixes this issue by changing to use SmallSetVector.
Differential Revision: http://reviews.llvm.org/D26634
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@287148 91177308-0d34-0410-b5e6-96231b3b80d8
Summary:
Extend replaceZeroVectorStore to handle more vector type stores,
floating point zero vectors and set alignment more accurately on split
stores.
This is a follow-up change to r286875.
This change fixes PR31038.
Reviewers: MatzeB
Subscribers: mcrosier, aemerson, llvm-commits, rengolin
Differential Revision: https://reviews.llvm.org/D26682
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@287142 91177308-0d34-0410-b5e6-96231b3b80d8
Summary:
1. Don't try to copy values to and from the same register class.
2. Replace copies with of registers with immediate values with v_mov/s_mov
instructions.
The main purpose of this change is to make MachineSink do a better job of
determining when it is beneficial to split a critical edge, since the pass
assumes that copies will become move instructions.
This prevents a regression in uniform-cfg.ll if we enable critical edge
splitting for AMDGPU.
Reviewers: arsenm
Subscribers: arsenm, kzhuravl, llvm-commits
Differential Revision: https://reviews.llvm.org/D23408
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@287131 91177308-0d34-0410-b5e6-96231b3b80d8
We can replace "scalar" FP-bitwise-logic with other forms of bitwise-logic instructions.
Scalar SSE/AVX FP-logic instructions only exist in your imagination and/or the bowels of
compilers, but logically equivalent int, float, and double variants of bitwise-logic
instructions are reality in x86, and the float variant may be a shorter instruction
depending on which flavor (SSE or AVX) of vector ISA you have...so just prefer float all
the time.
This is a preliminary step towards solving PR6137:
https://llvm.org/bugs/show_bug.cgi?id=6137
Differential Revision:
https://reviews.llvm.org/D26712
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@287122 91177308-0d34-0410-b5e6-96231b3b80d8
This unit test infinite-looped on s390x due to a thread_yield being optimized
out. I've updated the QueueChannel class (where thread_yield was called) to use
a condition variable instead. This should cause the unit test to behave
correctly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@287121 91177308-0d34-0410-b5e6-96231b3b80d8
Both the (V)CVTDQ2PD (i32 to f64) and (V)CVTUDQ2PD (u32 to f64) conversion instructions are lossless and can be safely represented as generic SINT_TO_FP/UINT_TO_FP calls instead of x86 intrinsics without affecting final codegen.
LLVM counterpart to D26686
Differential Revision: https://reviews.llvm.org/D26736
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@287108 91177308-0d34-0410-b5e6-96231b3b80d8
MipsFastISel uses a a class to represent addresses with a signed member
to represent the offset. MipsFastISel::emitStore, emitLoad and computeAddress
all treated the offset as being positive. In cases where the offset was
actually negative and a frame pointer was used, this would cause the constant
synthesis routine to crash as it would generate an unexpected instruction
sequence when frame indexes are replaced.
Reviewers: vkalintiris
Differential Revision: https://reviews.llvm.org/D26192
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@287099 91177308-0d34-0410-b5e6-96231b3b80d8
This patch adds the single operand form of the not alias to microMIPS and
MIPS along with additional tests.
This partially resolves PR/30381.
Thanks to Sean Bruno for reporting the issue!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@287097 91177308-0d34-0410-b5e6-96231b3b80d8
