Ecosyste.ms: OpenCollective
An open API service for software projects hosted on Open Collective.
github.com/swiftwasm/swift-llvm
This repository is no longer in use, please refer to the LLVM monorepo https://github.com/swiftwasm/llvm-project
https://github.com/swiftwasm/swift-llvm
This patch really contains two pieces:
Teach SCEV how to fold a phi in the header of a loop ...
different range, in preparation for use in Clang.
git-svn-id: https://llvm.org/svn/llvm-project/...
b31d1763674f4464952feafc04b22f81dd372fa6 authored over 5 years ago by Richard Smith <[email protected]>
Summary:
Change the way we deal with iterator invalidation in the extload combines as it
was sti...
AMDGPUPropagateAttributes will not work on function bitcatsts,
so move AMDGPUFixFunctionBitcasts...
Recommit r363289 with a bug fix for crash identified in pr42279. Issue was that a loop exit tes...
d773cb7e1cd21435f0b7d380b0d0d57f59bf1f02 authored over 5 years ago by Philip Reames <[email protected]>git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@363610 91177308-0d34-0410-b5e6-96231b3b...
e6bda13f369a91a0e785f4231c3a2b03ab9762bc authored over 5 years ago by Peter Collingbourne <[email protected]>git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@363609 91177308-0d34-0410-b5e6-96231b3b...
d07c0627e3cecca93db3e36260365f87cda1a7fc authored over 5 years ago by Peter Collingbourne <[email protected]>
Summary:
The purpose of the padding is to guard against stale code being
fetched into the instru...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@363601 91177308-0d34-0410-b5e6-96231b...
185d55ed9f67126c37290f32bd94c03df69f0e05 authored over 5 years ago by Philip Reames <[email protected]>
Summary:
This is related to the changes to the groupstaticsize intrinsic in
D61494 which would o...
Summary:
Update compare normalization in SimpleValue hashing to break ties (when
the same value ...
Summary:
LoopRotate doesn't create a faithful clone of an instruction, it may
simplify it before...
Basically porting over the behaviour in AArch64ISelLowering to GISel. See
emitComparison for ref...
We don't know if its safe to unfold if we're in 32-bit mode.
This is simlar to what was done to...
6750ca7ad7952dfce15082339c98d7ec8df0f85c authored over 5 years ago by Craig Topper <[email protected]>Differential revision: https://reviews.llvm.org/D62411
git-svn-id: https://llvm.org/svn/llvm-pr...
97e04d1cf3037b97457707f05203c27d71fcb93c authored over 5 years ago by Valery Pykhtin <[email protected]>If a XMM non-temporal store has less than natural alignment, scalarize the vector - with SSE4A w...
22af01a41402861dd1eaf9a2210a247b7c5a7f65 authored over 5 years ago by Simon Pilgrim <[email protected]>git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@363591 91177308-0d34-0410-b5e6-96231b3b...
fcf015b682f0e6dcf5e2f3c213b0f5d3ddae2856 authored over 5 years ago by Matt Arsenault <[email protected]>
Summary:
Add all MemoryPhis in IDF before filling in their incomign values.
Otherwise, a new Phi...
This behavior was added in r130928 for both FastISel and SD, and then
disabled in r131156 for Fa...
Differential Revision: https://reviews.llvm.org/D63206
git-svn-id: https://llvm.org/svn/llvm-pr...
95dd0d84d21ea397636287892bf0df588711566a authored over 5 years ago by Stanislav Mekhanoshin <[email protected]>git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@363587 91177308-0d34-0410-b5e6-96231b3b...
55ce7b4ea98ce61e1d4371bf0e42d3ae3890c967 authored over 5 years ago by Matt Arsenault <[email protected]>The pass works in two modes:
Mode 1: Just set attributes starting from kernels. This can work a...
f1a5ef5a39d61698f7da9c4f592d5ea1803624a2 authored over 5 years ago by Stanislav Mekhanoshin <[email protected]>git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@363583 91177308-0d34-0410-b5e6-96231b3b...
c0de78b22080e541a867162203de479367f54219 authored over 5 years ago by Nico Weber <[email protected]>If a YMM/ZMM non-temporal store has less than natural alignment, split the vector - either they ...
1d6c2fbdd1f8dfe6534b81df84755ee580bb50f7 authored over 5 years ago by Simon Pilgrim <[email protected]>
When considering a loop containing nontemporal stores or loads for
vectorization, suppress the v...
A target intrinsic may be defined as possibly reading memory, but the
call site may have additio...
I keep using the wrong instruction when manually writing tests. This
really needs to check the n...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@363578 91177308-0d34-0410-b5e6-96231b3b...
33482eea4c6ddeaacbbd04b67159af6a1966321e authored over 5 years ago by Matt Arsenault <[email protected]>Differential Revision: https://reviews.llvm.org/D63207
git-svn-id: https://llvm.org/svn/llvm-pr...
c33d0e66505b286414f9dfaf3348388fdaf6747c authored over 5 years ago by Stanislav Mekhanoshin <[email protected]>Reviewers: arsenm
Subscribers: kzhuravl, jvesely, wdng, nhaehnle, yaxunl, rovka, kristof.beyls,...
2d928b87dee907eb069a68e54105a1688d0733c4 authored over 5 years ago by Tom Stellard <[email protected]>
Use -fsave-optimization-record=<format> to specify a different format
than the default, which is...
This is currently only used for ymm->xmm splitting but we shouldn't hardcode the offsets/alignme...
3ea8c2ac6a9772267d3ad442756cffca66c95f83 authored over 5 years ago by Simon Pilgrim <[email protected]>
Summary:
There is PHINode::getBasicBlockIndex() and PHINode::setIncomingValue()
but no function ...
Test both 'unaligned' (which we should just use regular unaligned loads) and 'subvector aligned'...
dbb1cec2e70078e863cd9c210da60dd22227ec53 authored over 5 years ago by Simon Pilgrim <[email protected]>For loads, pre-SSE41 we can't perform NT loads at all, and after that we can only perform vector...
bfe7eb96c9736b21e7790d493e39b19c514ed478 authored over 5 years ago by Simon Pilgrim <[email protected]>
If an addrspacecast needed to be inserted again, this was creating a
clone of the original cast ...
Even if the target doesn't have flat instructions, addrspace(0) is
still flat. It just happens t...
Should also be marked writeonly, but I think that would require
splitting the version with done ...
Tests will be in future commits when new intrinsics are handled here.
git-svn-id: https://llvm....
baa4b8244b7eb151d21b41d4f3c5ba78a0b17027 authored over 5 years ago by Matt Arsenault <[email protected]>
Use separate enums for each kind, avoid repeating overloads, and add
missing classof implementat...
The HardwareLoops pass finds exit blocks with a scevable exit count.
If the target specifies to ...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@363552 91177308-0d34-0410-b5e6-96231b3b...
208eae0feb1d4b20b10f8226f885e603f5e23040 authored over 5 years ago by Simon Pilgrim <[email protected]>Change-Id: I376fcbd58f84a2aac6aaf744bc1665c92d312b25
git-svn-id: https://llvm.org/svn/llvm-proj...
ce390642d524b587b0aefb6ffd0237d2e9bbbe5e authored over 5 years ago by Nicolai Haehnle <[email protected]>git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@363549 91177308-0d34-0410-b5e6-96231b3b...
f95bc3347d67efe3229ea677ee1f98994bf4adc0 authored over 5 years ago by Nico Weber <[email protected]>
Summary:
Avoid that loop vectorizer creates loads/stores of vectors
with "irregular" types when ...
This patch tests the forward-referencing added in D62995 by changing
some existing intrinsics to...
Some GEPs were not being split, presumably because that split would just be
undone by the DAGCo...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@363543 91177308-0d34-0410-b5e6-96231b3b...
ba4b2d50e144b6cfbed01ccc25629c34b228db87 authored over 5 years ago by Fangrui Song <[email protected]>This is already done in DAGCombiner::visitINSERT_SUBVECTOR, but this helps a number of shuffles ...
ec69a27d0a634a618f01cf1d3edc509b73dc483b authored over 5 years ago by Simon Pilgrim <[email protected]>
Second functional change following on from rL362687. Pass the
NoWrapFlags from the MulExpr to In...
Reviewed By: grimar
Differential Revision: https://reviews.llvm.org/D63398
git-svn-id: https:/...
400438c3c0fe0d233819a493ccc52d30bffd99cc authored over 5 years ago by Fangrui Song <[email protected]>9290bd1e343c5eb57bbf1c5b0420a83c401976f5 authored over 5 years ago by Automerger <Automerger@Swift>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@363538 91177308-0d34-0410-b5e6-96231b3b...
1002960b9df562de440b30d307ac1c50602e5a1b authored over 5 years ago by Fangrui Song <[email protected]>git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@363537 91177308-0d34-0410-b5e6-96231b3b...
e3afbf5368f03be68c43ae7f05b4362f1b3da372 authored over 5 years ago by Roman Lebedev <[email protected]>git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@363535 91177308-0d34-0410-b5e6-96231b3b...
992d474561f1299312510aa3db0ec1fb85516c78 authored over 5 years ago by Fangrui Song <[email protected]>git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@363534 91177308-0d34-0410-b5e6-96231b3b...
679d03e384e821fc52c5b0a13c94dd6e91bdc21f authored over 5 years ago by Fangrui Song <[email protected]>
This patch changes MIR stack-id from an integer to an enum,
and adds printing/parsing support fo...
Forgot to remove file!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@363532 9117730...
82ad3d12cc52409acae3930934ba1e8034c32cae authored over 5 years ago by Sam Parker <[email protected]>Forgot to add file!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@363531 91177308-0...
659a32042076a406b2957e95cc55f8740a189656 authored over 5 years ago by Sam Parker <[email protected]>
Create the ARMBasicBlockUtils class for tracking and querying basic
blocks sizes so we can use t...
Third time's the charm.
This was reverted in r363220 due to being suspected of an internal benc...
8daad51954f1be30f1c8b83618a6485d6f8f57af authored over 5 years ago by Hans Wennborg <[email protected]>
SimplifyCFG has a bug that results in inconsistent prof branch_weights metadata
if unreachable s...
Summary:
SPE passes doubles the same as soft-float, in register pairs as i32
types. This is all...
Summary:
Currently, MachOWriter::writeSectionData writes dummy data (0xdeadbeef) to fill section...
Summary:
The "sparc"/"sparcel" architectures appears in ArchMap (used by -B option) but not in O...
It would not be safe to unfold the memory form the register form
without checking that we are co...
Fix folds of addo and subo with an undef operand to be:
`@llvm.{u,s}{add,sub}.with.overflow` al...
d09f140cd1ebdfcabc333c45fe7e7fe59752bdf9 authored over 5 years ago by Roman Lebedev <[email protected]>
Summary:
This allows target to implement custom emit of global variables if
required. See subseq...
Summary:
We will use absolute relocations for LDS symbols.
Change-Id: I9a32795ed0ea835e433a7871...
6435d005d68411f87ef42bbda5c4d48f6edfa54b authored over 5 years ago by Nicolai Haehnle <[email protected]>
Summary:
Instead of encoding a high-word of 0 using a fake TargetGlobalAddress,
just use a liter...
Summary: Change-Id: Ie4c971462a7749740938c687144e77441dac2539
Reviewers: rampitec, arsenm
Subs...
d116e8400b55d43623fd8350a69f7a1e42c7b8f6 authored over 5 years ago by Nicolai Haehnle <[email protected]>
This is cpp source part of wave32 support, excluding overriden
getRegClass().
Differential Revi...
c6fce1250e5007a497fe3c0e53dbaa00d820e029 authored over 5 years ago by Stanislav Mekhanoshin <[email protected]>
This is based on the example/discussion in PR37428:
https://bugs.llvm.org/show_bug.cgi?id=37428
...
This is similar logic/motivation to the select splitting in D62969.
In D63233, the pattern chan...
78f1b314644d2968590e41c71d3eec2b853642f8 authored over 5 years ago by Sanjay Patel <[email protected]>Insert the shorter vector source into an undef vector of the longer vector source's type.
git-s...
06e81472730158e50cbc9f660d5db486f8a57f61 authored over 5 years ago by Simon Pilgrim <[email protected]>a40651b56ee4c9dc56b2e8e09539efd889107fea authored over 5 years ago by Automerger <Automerger@Swift>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@363505 91177308-0d34-0410-b5e6-96231b3b...
29fd1c2a34d2b420d22edfb2aaa48eb35e73dd03 authored over 5 years ago by Nico Weber <[email protected]>git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@363501 91177308-0d34-0410-b5e6-96231b3b...
0f6694464b8f52cf92b3d1edf953dd757ffffecc authored over 5 years ago by Simon Pilgrim <[email protected]>Pull out the existing (non)lane-crossing fold into a helper lambda and use for lane-crossing una...
170bbaa5a54e2da4c02c47104921bad5ee64a4d1 authored over 5 years ago by Simon Pilgrim <[email protected]>This mostly happens due to SimplifyDemandedVectorElts reducing a vector to insert_subvector(unde...
2720bd6686713de365b97d8c19776552c50979f9 authored over 5 years ago by Simon Pilgrim <[email protected]>git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@363498 91177308-0d34-0410-b5e6-96231b3b...
dc0aa24fcdfe2123df70ecf03635351b7ee75d4e authored over 5 years ago by Roman Lebedev <[email protected]>Looking into sched model for that CPU ...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/tr...
fb0a468b49cb537a507553e3e4221078b2351557 authored over 5 years ago by Roman Lebedev <[email protected]>
Summary:
With Split DWARF the resulting object file (then called skeleton CU)
contains the file ...
Summary:
If the nested loop is an innermost loop, prefer to a 32-byte alignment, so that
we can ...
Summary:
When using ConstantExpr we often need the result of the expression to be kept in the AS...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@363492 91177308-0d34-0410-b5e6-96231b3b...
dc69eb3bf3d11f065a9f4963f87a1120231c6533 authored over 5 years ago by Fangrui Song <[email protected]>
Based on D59959, this switches SCEV to use unsigned/signed range
intersection based on the sign ...
1717645743ef3a90c8899cbba19cf36638596beb authored over 5 years ago by Automerger <Automerger@Swift>
If we can detect that saturating math that depends on an IV cannot
overflow, replace it with sim...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@363487 91177308-0d34-0410-b5e6-96231b3b...
5eaae6c4c2b4364378dc8a2be3ca8993514f4edf authored over 5 years ago by Fangrui Song <[email protected]>git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@363486 91177308-0d34-0410-b5e6-96231b3b...
c5ac9abee742f9d8cda89294645726ae7c04fdc2 authored over 5 years ago by Fangrui Song <[email protected]>Summary: llvm-objcopy/strip now error when -p is specified when reading from stdin or writing to...
1f1f241e1685dddad2db4749a915f07cb1eaca26 authored over 5 years ago by Alex Brachet <[email protected]>git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@363484 91177308-0d34-0410-b5e6-96231b3b...
d71bd8db7b8c3f5e256f131cc83285ec1f0677ab authored over 5 years ago by Michael Berg <[email protected]>
This reverts rL363474. -debug-only=isel was added to some tests that
don't specify `REQUIRES: as...
Summary:
For icmp pred (and (sh X, Y), C), 0
When C is signbit, expect to fold (X << Y) & sig...
34a27682ed02e6273e91b49bacf6bfc75e8da157 authored over 5 years ago by Huihui Zhang <[email protected]>
This reapplies r363410, avoiding null dereference if there is no
AltRegBank.
git-svn-id: https:...
daecbabf55cf11efe6edcd40df96ae7e9359ecf5 authored over 5 years ago by Matt Arsenault <[email protected]>In preparation for use in Clang.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36347...
4000c17fc9b553dbe113251bb0ba2a68f6dc8ad7 authored over 5 years ago by Richard Smith <[email protected]>
This patch breaks UBSan build bots. See
https://github.com/google/sanitizers/wiki/SanitizerBotRe...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@363474 91177308-0d34-0410-b5e6-96231b3b...
aad9d5e5dd9bccf64421d544db81a8b55108667f authored over 5 years ago by Michael Berg <[email protected]>Current findBestLoopTop can find and move one kind of block to top, a latch block has one succes...
7eae8125c63bb72c2103f17ba01d6eb1e7ca5052 authored over 5 years ago by Guozhi Wei <[email protected]>