♟ bkramer

[MC] Use a byte_swap in emitIntValue instead of doing it in a loop. NFCI.

[RuntimeDyld] Replace hand-rolled endian conversion with support::endian. NFCI.

[LTO] Replace hand-rolled endian conversion with support::endian. NFCI.

[llvm-bcanalyzer] Simplify code. NFCI.

[X86] Roll some loops. NFCI.

[LoopOps] Return ArrayRefs from accessors instead of iterator_ranges

Make helpers static. NFC.

Drop unused diagnostic. NFC.

[LoopDataPrefetch] Remove unused include that's a layering violation

Revert "[SimplifyLibCalls] Erase replaced instructions"

[ARM] Silence warning in Release builds

[GVNSink] Simplify code. NFC.

Replace std::string::find == 0 with StringRef::startswith

[LLDB] Initialize temporary token

[ARMMVE] Create fewer temporary SmallVectors

[MC] Move deprecation infos from MCTargetDesc to MCInstrInfo

Put back initializers that were dropped in…

[COFF] Stabilize sort

[MDBuilder] Don't use stable sort for sorting integers.

[ADT] Automatically forward llvm::sort to array_pod_sort if safe

[AMDGPU] Stabilize sort order

Upgrade some instances of std::sort to llvm::sort. NFC.

Remove constexpr that MSVC doesn't like

Const-initialize ParsedAttrInfos

In D76007#1944582, @lewis-revill wrote:

I'm lacking an understanding of how concurrency is expected to work within LLVM, but since there is only one RegisterBankInfo constructed per subtarget I don't see how there is a problem with updating the global RegBanks array? Likewise the HwMode is constant for the subtarget, so the write will be of the same value.

[SelectionDAG] Don't crash when freezing illegal float types

Make helpers static. NFC.

[Attributor] Fold single-use variable into assert

[Matrix] Fold single-use variable into assert

Purge unused diagnostics. NFC.

Right, we don't have names for values in release builds.

[AArch64] Remove unused variable

[AVR] Make helper functions static. NFC.

[tblgen] Remove unused private field. NFC.

Give helpers internal linkage. NFC.

Fix pessimizing move. NFC.

Revert "Fix regression in bdad0a1: force rebuilding of StmtExpr nodes in"…

[x86][SLH] Rm liveness check from data invariance check

[llvm-ar] Remove unused variable. NFC.

ArrayRef'ize restoreCalleeSavedRegisters. NFCI.

Add a pass that specializes parallel loops for easier unrolling and…

Renamed the thing ParallelLoopSpecialization
Match the canonicalized AffineMap form, not the one produced by ParallelLoopTiling

Make test not write to the source directory

[ORC] Remove spammy debug print

Add a basic tiling pass for parallel loops

Silence compiler warnings

Address moar comments

Make tileSizes a list
Fix min again

One round of addressed review comments

use pass option instead of global option
fix the recursive walker to walk recursively

Drop a constexpr in favor of const, MSVC complains.

Move DFA tables into the read-only data segmant.

Hide implementation details. NFC>

[wasm] Unbreak after 5fc5c7db38672c8962879b6fdce68393181c5e08. NFCI.

Strength reduce vectors into arrays. NFCI.

[ADT] Use inherited ctors to forward to base. NFCI.

In the long term I'd prefer to not initialize backends inside of pass at all, but for the short term this is better than pulling in all backends.

<courage wolf>Push and watch the bots</courage wolf>

Put back makeArrayRef to make GCC 5 happy

Drop some uses of StringLiteral in favor of StringRef

Use heterogenous lookup for std;:map<std::string with a StringRef. NFCI.

ArrayRef'ize spillCalleeSavedRegisters. NFCI.

Was this ever committed? Loop flattening looks like something that's useful to have.

[mlir-tblgen] Stop leaking PredNodes

Rewrite test not to rely on StrEq with StringRef

[clang-tidy] Initialize token before handing it to the lexer

Fix an implicit conversion in clang-tidy. GCC 5 complains about it.

[ASTMatchers] StringRef'ify hasName

Fix clang unnittest build with GCC 5

One more bugpoitn fix for GCC5

Try harder to fix bugpoint with GCC5

Make bugpoint work with gcc5 again.

Fix more implicit conversions. Getting closer to having clang working with gcc…

Fix conversions in clang and examples

GCC5 buildbot made it to clang. Fix implicit conversions it found.

Another stab at making the gold plugin compile again

Another round of GCC5 fixes.

Fix one round of implicit conversions found by g++5.

Address implicit conversions detected by g++ 5 only.

One more batch of things found by g++ 6

bkramer (Benjamin Kramer)
Senior Procrastinator

Projects

User Details

Recent Activity
View All

Today

Yesterday

Fri, Apr 3

Thu, Apr 2

Wed, Apr 1

Tue, Mar 31

Mon, Mar 30

Sun, Mar 29

Sat, Mar 28

Thu, Mar 26

Tue, Mar 24

Mon, Mar 23

Thu, Mar 19

Tue, Mar 17

Mon, Mar 16

Sun, Mar 15

Tue, Mar 10

Mar 5 2020

Mar 4 2020

Mar 3 2020

Feb 29 2020

Feb 28 2020

Feb 27 2020

Feb 25 2020

Feb 24 2020

Feb 21 2020

Feb 18 2020

Feb 17 2020

Feb 14 2020

Feb 12 2020

Feb 11 2020

Feb 8 2020

Feb 7 2020

Feb 6 2020

Jan 29 2020

Jan 28 2020

bkramer (Benjamin Kramer)Senior Procrastinator