WMMA = "Warp Level Matrix Multiply-Accumulate".
These are the new instructions introduced in PTX6.0 and available
on sm_70 GPUs.
Differential Revision: https://reviews.llvm.org/D38645
llvm-svn: 315601
Building with BUILD_SHARED_LIBS makes it tricky to copy around
executables at will, since they won't be able to find the LLVM
libraries any more. This makes testing a feature that's based on the
executable name problematic, so we'll just disable these two tests in
that configuration.
We could potentially fix this by symlinking the lib directory into the
test directory, but that wouldn't work on windows, and losing testing
on windows would be far worse than losing testing on a configuration
that's barely even supported.
llvm-svn: 315599
Add profitability checks for modifying counted loops to use the mtctr instruction.
The latency of mtctr is only justified if there are more than 4 comparisons that
will be removed as a result. Usually counted loops are formed relatively early
and before unrolling, so most low trip count loops often don't survive. However
we want to ensure that if they do, we do not mistakenly update them to mtctr loops.
Use CodeMetrics to ensure we are only doing this for small loops with small trip counts.
Differential Revision: https://reviews.llvm.org/D38212
llvm-svn: 315592
Summary:
The interpolation mode workaround ensures that at least one
interpolation mode is enabled in PSInputAddr. It does not also check
PSInputEna on the basis that the user might enable bits in that
depending on run-time state.
However, for amdpal os type, the user does not enable some bits after
compilation based on run-time states; the register values being
generated here are the final ones set in the hardware. Therefore, apply
the workaround to PSInputAddr and PSInputEnable together. (The case
where a bit is set in PSInputAddr but not in PSInputEnable is where the
frontend set up an input arg for a particular interpolation mode, but
nothing uses that input arg. Really we should have an earlier pass that
removes such an arg.)
Reviewers: arsenm, nhaehnle, dstuttard
Subscribers: kzhuravl, wdng, yaxunl, t-tye, llvm-commits
Differential Revision: https://reviews.llvm.org/D37758
llvm-svn: 315591
Summary:
The comments in the code said
// Remove <def,read-undef> flags. This def is now a partial redef.
but the code didn't just remove read-undef, it could introduce new ones which
could cause errors.
E.g. if we have something like
%vreg1<def> = IMPLICIT_DEF
%vreg2:subreg1<def, read-undef> = op %vreg3, %vreg4
%vreg2:subreg2<def> = op %vreg6, %vreg7
and we merge %vreg1 and %vreg2 then we should not set undef on the second subreg
def, which the old code did.
Now we solve this by actually do what the code comment says. We remove
read-undef flags rather than remove or introduce them.
Reviewers: qcolombet, MatzeB
Reviewed By: MatzeB
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D38616
llvm-svn: 315564
Here we add a secondary option parser to llvm-isel-fuzzer (and provide
it for use with other fuzzers). With this, you can copy the fuzzer to
a name like llvm-isel-fuzzer=aarch64-gisel for a fuzzer that fuzzer
AArch64 with GlobalISel enabled, or fuzzer=x86_64 to fuzz x86, with no
flags required. This should be useful for running these in OSS-Fuzz.
Note that this handrolls a subset of cl::opts to recognize, rather
than embedding a complete command parser for argv[0]. If we find we
really need the flexibility of handling arbitrary options at some
point we can rethink this.
This re-applies 315545 using "=" instead of ":" as a separator for
arguments.
llvm-svn: 315557
It broke some tests on Windows:
Failing Tests (4):
LLVM :: tools/llvm-isel-fuzzer/execname-options.ll
LLVM :: tools/llvm-isel-fuzzer/missing-triple.ll
LLVM :: tools/llvm-isel-fuzzer/x86-empty-bc.ll
LLVM :: tools/llvm-isel-fuzzer/x86-empty.ll
> llvm-isel-fuzzer: Handle a subset of backend flags in the executable name
>
> Here we add a secondary option parser to llvm-isel-fuzzer (and provide
> it for use with other fuzzers). With this, you can copy the fuzzer to
> a name like llvm-isel-fuzzer:aarch64-gisel for a fuzzer that fuzzer
> AArch64 with GlobalISel enabled, or fuzzer:x86_64 to fuzz x86, with no
> flags required. This should be useful for running these in OSS-Fuzz.
>
> Note that this handrolls a subset of cl::opts to recognize, rather
> than embedding a complete command parser for argv[0]. If we find we
> really need the flexibility of handling arbitrary options at some
> point we can rethink this.
llvm-svn: 315554
Here we add a secondary option parser to llvm-isel-fuzzer (and provide
it for use with other fuzzers). With this, you can copy the fuzzer to
a name like llvm-isel-fuzzer:aarch64-gisel for a fuzzer that fuzzer
AArch64 with GlobalISel enabled, or fuzzer:x86_64 to fuzz x86, with no
flags required. This should be useful for running these in OSS-Fuzz.
Note that this handrolls a subset of cl::opts to recognize, rather
than embedding a complete command parser for argv[0]. If we find we
really need the flexibility of handling arbitrary options at some
point we can rethink this.
llvm-svn: 315545
- Move PAL metadata definitions to AMDGPUMetadata
- Make naming consistent with HSA metadata
Differential Revision: https://reviews.llvm.org/D38745
llvm-svn: 315523
- Rename AMDGPUCodeObjectMetadata to AMDGPUMetadata (PAL metadata will be included in this file in the follow up change)
- Rename AMDGPUCodeObjectMetadataStreamer to AMDGPUHSAMetadataStreamer
- Introduce HSAMD namespace
- Other minor name changes in function and test names
llvm-svn: 315522
Summary:
This adds a set of new directives that describe 32-bit x86 prologues.
The directives are limited and do not expose the full complexity of
codeview FPO data. They are merely a convenience for the compiler to
generate more readable assembly so we don't need to generate tons of
labels in CodeGen. If our prologue emission changes in the future, we
can change the set of available directives to suit our needs. These are
modelled after the .seh_ directives, which use a different format that
interacts with exception handling.
The directives are:
.cv_fpo_proc _foo
.cv_fpo_pushreg ebp/ebx/etc
.cv_fpo_setframe ebp/esi/etc
.cv_fpo_stackalloc 200
.cv_fpo_endprologue
.cv_fpo_endproc
.cv_fpo_data _foo
I tried to follow the implementation of ARM EHABI CFI directives by
sinking most directives out of MCStreamer and into X86TargetStreamer.
This helps avoid polluting non-X86 code with WinCOFF specific logic.
I used cdb to confirm that this can show locals in parent CSRs in a few
cases, most importantly the one where we use ESI as a frame pointer,
i.e. the one in http://crbug.com/756153#c28
Once we have cdb integration in debuginfo-tests, we can add integration
tests there.
Reviewers: majnemer, hans
Subscribers: aemerson, mgorny, kristof.beyls, llvm-commits, hiraditya
Differential Revision: https://reviews.llvm.org/D38776
llvm-svn: 315513
Summary:
Fixes a bogus iterator resulting from the removal of a block's first instruction at the point that incremental update is enabled.
Patch by Paul Walker.
Reviewers: fhahn, Gerolf, efriedma, MatzeB
Reviewed By: fhahn
Subscribers: aemerson, javed.absar, llvm-commits
Differential Revision: https://reviews.llvm.org/D38734
llvm-svn: 315502
Currently we produce a bunch of unnecessary code when emitting the
prologue/epilogue for spills/restores. Namely, if the load from stack
slot/store to stack slot instruction is an X-Form instruction, we will
always produce an LIS/ORI sequence for the stack offset.
Furthermore, we have not exploited the P9 vector D-Form loads/stores for this
purpose.
This patch address both issues.
Specifying the D-Form load as the instruction to use for stack spills/reloads
should be safe because:
1. The stack should be aligned according to the ABI
2. If the stack isn't aligned, PPCRegisterInfo::eliminateFrameIndex() will
check for the offset being a multiple of 16 and will convert it to an
X-Form instruction if it isn't.
Differential Revision : https://reviews.llvm.org/D38758
llvm-svn: 315500
Previously we would only look in the current directory for a
resource, which might not be the same as the directory of the
rc file. Furthermore, MSVC rc supports a /I option, and can
also look in the system environment. This patch adds support
for this search algorithm.
Differential Revision: https://reviews.llvm.org/D38740
llvm-svn: 315499
Legalization of fp128 assumes things that we should have asserts for,
so that's another potential improvement.
Differential Revision: https://reviews.llvm.org/D38771
llvm-svn: 315485
ubsan caught an issue I made where I was converting a null pointer to a
reference.
elf utils implements a particularly extreme form of stripping that I'd
like to support. eu-strip has an option called "strip-sections" that
removes all section headers and leaves only program headers and the
segment data. I have implemented this option partly as a test but mainly
because in Fuchsia we would like to use this option to minimize the size
of our executables. The other strip options that are on my list include
--strip-all and --strip-debug. This is a preliminary implementation that
I'd like to start using in Fuchsia builds if possible. This change
implements such a stripping option for llvm-objcopy
Differential Revision: https://reviews.llvm.org/D38335
llvm-svn: 315484
The pipeliner is generating a serial sequence that causes poor
register allocation when a post-increment instruction appears
prior to the use of the post-increment register. This occurs when
there is a circular set of dependences involved with a sequence
of instructions in the same cycle. In this case, there is no
serialization of the parallel semantics that will not cause an
additional register to be allocated.
This patch fixes the problem by changing the instructions so that
the post-increment instruction is used by the subsequent
instruction, which enables the register allocator to make a
better decision and not require another register.
Patch by Brendon Cahoon.
llvm-svn: 315466
Eg:
insert v4i32 V, (v2i16 X), 2 --> shuffle v8i16 V', X', {0,1,2,3,8,9,6,7}
This is a generalization of the IR fold in D38316 to handle insertion into a non-undef vector.
We may want to abandon that one if we can't find value in squashing the more specific pattern sooner.
We're using the existing legal shuffle target hook to avoid AVX512 horror with vXi1 shuffles.
There may be room for improvement in the shuffle lowering here, but that would be follow-up work.
Differential Revision: https://reviews.llvm.org/D38388
llvm-svn: 315460
This patch adds timestamp verification for swiftmodule files.
- A new flag is provided to allows us to continue testing of the code
for embedding the__swift_ast. (git doesn't maintain timestamps)
- Adds a new test for fat (arm) binaries.
Differential revision: https://reviews.llvm.org/D38686
llvm-svn: 315456
Adding this test files now so after another commit that will add a new pattern for
TESTM and TESTNM instructions will show the improvemnts that have been done.
Change-Id: If3908b7f91897d764053312365a2bc1de78b291d
llvm-svn: 315443
This patch fixes the miscompile that happens when PRE hoists loads across guards and
other instructions that don't always pass control flow to their successors. PRE is now prohibited
to hoist across such instructions because there is no guarantee that the load standing after such
instruction is still valid before such instruction. For example, a load from under a guard may be
invalid before the guard in the following case:
int array[LEN];
...
guard(0 <= index && index < LEN);
use(array[index]);
Differential Revision: https://reviews.llvm.org/D37460
llvm-svn: 315440
Sinking of unordered atomic load into loop must be disallowed because it turns
a single load into multiple loads. The relevant section of the documentation
is: http://llvm.org/docs/Atomics.html#unordered, specifically the Notes for
Optimizers section. Here is the full text of this section:
> Notes for optimizers
> In terms of the optimizer, this **prohibits any transformation that
> transforms a single load into multiple loads**, transforms a store into
> multiple stores, narrows a store, or stores a value which would not be
> stored otherwise. Some examples of unsafe optimizations are narrowing
> an assignment into a bitfield, rematerializing a load, and turning loads
> and stores into a memcpy call. Reordering unordered operations is safe,
> though, and optimizers should take advantage of that because unordered
> operations are common in languages that need them.
Patch by Daniil Suchkov!
Reviewed By: reames
Differential Revision: https://reviews.llvm.org/D38392
llvm-svn: 315438
IRCE should not apply when the safe iteration range is proved to be empty.
In this case we do unneeded job creating pre/post loops and then never
go to the main loop.
This patch makes IRCE not apply to empty safe ranges, adds test for this
situation and also modifies one of existing tests where it used to happen
slightly.
Reviewed By: anna
Differential Revision: https://reviews.llvm.org/D38577
llvm-svn: 315437
elf utils implements a particularly extreme form of stripping that I'd
like to support. eu-strip has an option called "strip-sections" that
removes all section headers and leaves only program headers and the
segment data. I have implemented this option partly as a test but mainly
because in Fuchsia we would like to use this option to minimize the size
of our executables. The other strip options that are on my list include
--strip-all and --strip-debug. This is a preliminary implementation that
I'd like to start using in Fuchsia builds if possible. This change
implements such a stripping option for llvm-objcopy
Differential Revision: https://reviews.llvm.org/D38335
llvm-svn: 315412
This change adds the ability to use the "-R"/"-remove-section" option
multiple times.
Differential Revision: https://reviews.llvm.org/D38332
llvm-svn: 315385
Summary: In the current implementation, we only have accurate profile count for standalone symbols. For inlined functions, we do not have entry count data because it's not available in LBR. In this patch, we use the first instruction's frequency to estimiate the function's entry count, especially for inlined functions. This may be inaccurate due to debug info in optimized code. However, this is a better estimate than the static 80/20 estimation we have in the current implementation.
Reviewers: tejohnson, davidxl
Reviewed By: tejohnson
Subscribers: sanjoy, llvm-commits, aprantl
Differential Revision: https://reviews.llvm.org/D38478
llvm-svn: 315369
Rather than using the AdditionalPredicates mechanism to guard
the microMIPS instructions, use the existing predicates to properly
guard those instructions.
This also resolves a case where an instruction pattern was incorrectly
available for microMIPS32R6, which caused a register allocation failure
as the registers specified in the pattern were not available.
Reviewers: nitesh.jain, atanasyan
Differential Revision: https://reviews.llvm.org/D38451
llvm-svn: 315362
Craig Topper