Each function summary has an attached list of type identifier GUIDs. The
idea is that during the regular LTO phase we would match these GUIDs to type
identifiers defined by the regular LTO module and store the resolutions in
a top-level "type identifier summary" (which will be implemented separately).
Differential Revision: https://reviews.llvm.org/D27967
llvm-svn: 290280
argument even if the expression is value-dependent (we need to suppress the
final portion of the narrowing check, but the rest of the checking can still be
done eagerly).
This affects template template argument validity and partial ordering under
p0522r0.
llvm-svn: 290276
In order for the llvm DWARF parser to be used in LLDB we will need to be able to get the parent of a DIE. This patch adds that functionality by changing the DWARFDebugInfoEntry class to store a depth field instead of a sibling index. Using a depth field allows us to easily calculate the sibling and the parent without increasing the size of DWARFDebugInfoEntry.
I tested llvm-dsymutil on a debug version of clang where this fully parses DWARF in over 1200 .o files to verify there was no serious regression in performance.
Added a full suite of unit tests to test this functionality.
Differential Revision: https://reviews.llvm.org/D27995
llvm-svn: 290274
This patch adds the last bit of support to get LLVM_DISTRIBUTION_COMPONENTS working with libLLDB when built as a framework.
This patch adds dummy install targets for binaries built into the framework's Resources directory, and makes the framework's install target depend on all the binaries that get installed with the framework.
llvm-svn: 290273
This sets USES_TERMINAL for the native llvm-config build, so that it
doesn't run at the same time as builds of other native tools (namely,
tablegen). Without this, if you're very unlucky with the timing it's
possible to be relinking libSupport as one of the tools is linking,
causing a spurious failure.
The tablegen build adopted USES_TERMINAL for this same reason in
r280748.
llvm-svn: 290271
RTDyldMemoryManager.cpp describes the differing __register_frame
API between libunwind and libgcc, with a mailing list posting URL.
The original link was 404; replace it with what I believe is the
intended post, as well as a reference to the "OS X" implementation in
libunwind.
Differential Revision: https://reviews.llvm.org/D27965
llvm-svn: 290269
The constantexpr parsing was too constrained and rejected legal vector GEPs.
This relaxes it to be similar to the ones for instruction parsing.
This fixes PR30816.
Differential Revision: https://reviews.llvm.org/D28013
llvm-svn: 290261
For vector GEPs, CastGEPIndices can end up in an infinite recursion, because
we compare the vector type to the scalar pointer type, find them different,
and then try to cast a type to itself.
Differential Revision: https://reviews.llvm.org/D28009
llvm-svn: 290260
Members that are themselves wrapped in fake parentheses would lead to
AvoidBinPacking be set on the wrong ParenState.
After:
vector<int> aaaa = {
aaaaaa.aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa,
aaaaaa.aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa,
aaaaaa.aaaaaaa,
aaaaaa.aaaaaaa,
aaaaaa.aaaaaaa,
aaaaaa.aaaaaaa,
};
Before we were falling back to bin-packing these.
llvm-svn: 290259
Typically processor architectures do not include an L3 cache, which means that
Nc, the parameter of the micro-kernel, is, for all practical purposes,
redundant ([1]). However, its small values can cause the redundant packing of
the same elements of the matrix A, the first operand of the matrix
multiplication. At the same time, big values of the parameter Nc can cause
segmentation faults in case the available stack is exceeded.
This patch adds an option to specify the parameter Nc as a multiple of
the parameter of the micro-kernel Nr.
In case of Intel Core i7-3820 SandyBridge and the following options,
clang -O3 gemm.c -I utilities/ utilities/polybench.c -DPOLYBENCH_TIME
-march=native -mllvm -polly -mllvm -polly-pattern-matching-based-opts=true
-DPOLYBENCH_USE_SCALAR_LB -mllvm -polly-target-cache-level-associativity=8,8
-mllvm -polly-target-cache-level-sizes=32768,262144 -mllvm
-polly-target-latency-vector-fma=8
it helps to improve the performance from 11.303 GFlops/sec (39,247% of
theoretical peak) to 17.896 GFlops/sec (62,14% of theoretical peak).
Refs.:
[1] - http://www.cs.utexas.edu/users/flame/pubs/TOMS-BLIS-Analytical.pdf
Reviewed-by: Tobias Grosser <tobias@grosser.es>
Differential Revision: https://reviews.llvm.org/D28019
llvm-svn: 290256
Aligning data to cache lines boundaries helps to avoid overheads related to
an access to it ([1]). This patch aligns newly created arrays and adds an
option to specify the first level cache line size. By default we use 64 bytes,
which is a typical cache-line size ([2]).
In case of Intel Core i7-3820 SandyBridge and the following options,
clang -O3 gemm.c -I utilities/ utilities/polybench.c -DPOLYBENCH_TIME
-march=native -mllvm -polly -mllvm -polly-pattern-matching-based-opts=true
-DPOLYBENCH_USE_SCALAR_LB -mllvm -polly-target-cache-level-associativity=8,8
-mllvm -polly-target-cache-level-sizes=32768,262144 -mllvm
-polly-target-latency-vector-fma=8
it helps to improve the performance from 11.303 GFlops/sec (39,247% of
theoretical peak) to 12.63 GFlops/sec (43,8542% of theoretical peak).
Refs.:
[1] - http://www.alexonlinux.com/aligned-vs-unaligned-memory-access
[2] - http://igoro.com/archive/gallery-of-processor-cache-effects/
Differential Revision: https://reviews.llvm.org/D28020
Reviewed-by: Tobias Grosser <tobias@grosser.es>
llvm-svn: 290253
OpenBSD's cpio does not accept the -t option without -i.
Apparently some systems implement cpio -t as a shortcut
for cpio -it, the latter is the only thing that's documented.
This change avoids test failures on OpenBSD.
Patch by Mark Kettenis!
Differential Revision: https://reviews.llvm.org/D28002
llvm-svn: 290252
multiplication
Previously we had two-dimensional accesses to store packed operands of
the matrix multiplication for the sake of simplicity of the packed arrays.
However, addition of the third dimension helps to simplify the corresponding
memory access, reduce the execution time of isl operations applied to it, and
consequently reduce the compile-time of Polly. For example, in case of
Intel Core i7-3820 SandyBridge and the following options,
clang -O3 gemm.c -I utilities/ utilities/polybench.c -DPOLYBENCH_TIME
-march=native -mllvm -polly -mllvm -polly-pattern-matching-based-opts=true
-DPOLYBENCH_USE_SCALAR_LB -mllvm -polly-target-cache-level-associativity=8,8
-mllvm -polly-target-cache-level-sizes=32768,262144 -mllvm
-polly-target-latency-vector-fma=7
it helps to reduce the compile-time from about 361.456 seconds to about 0.816
seconds.
Reviewed-by: Michael Kruse <llvm@meinersbur.de>,
Tobias Grosser <tobias@grosser.es>
Differential Revision: https://reviews.llvm.org/D27878
llvm-svn: 290251
I added API for creation a target specific memory node in DAG. Today, all memory nodes are common for all targets and their constructors are located in SelectionDAG.cpp.
There are some cases in X86 where we need to create a special node - truncation-with-saturation store, float-to-half-store.
In the current patch I added truncation-with-saturation nodes and I'm using them for intrinsics. In the future I plan to implement DAG lowering for truncation-with-saturation pattern.
Differential Revision: https://reviews.llvm.org/D27899
llvm-svn: 290250
DefinedSynthetic is not created for a real ELF object, so it doesn't
have to be a template function. It has a virtual st_value, which is
either 32 bit or 64 bit, but we can simply use 64 bit.
llvm-svn: 290241
The vectorcall calling convention specifies that arguments to functions are to be passed in registers, when possible.
vectorcall uses more registers for arguments than fastcall or the default x64 calling convention use.
The vectorcall calling convention is only supported in native code on x86 and x64 processors that include Streaming SIMD Extensions 2 (SSE2) and above.
The current implementation does not handle Homogeneous Vector Aggregates (HVAs) correctly and this review attempts to fix it.
This aubmit also includes additional lit tests to cover better HVAs corner cases.
Differential Revision: https://reviews.llvm.org/D27392
llvm-svn: 290240
We probably would want to avoid fatal() if we can in context of librarification,
but for me reason of that patch is to help D27900 go.
D27900 changes errors reporting to something like
error: text1
note: text2
note: text3
where hint used to provide additional information about location. In that case
I can't just call fatal() because user will not see notes after that what adds additional complication to handle.
So It is good to switch fatal() to error() where it is possible.
Also it adds testcase with broken relocation number.
Previously we did not have any, It checks that error() instead of fatal() works fine.
Differential revision: https://reviews.llvm.org/D27973
llvm-svn: 290239
It was revealed by D27831.
If we have linkerscript that includes another one that sets OUTPUT for example:
RUN: echo "INCLUDE \"foo.script\"" > %t.script
RUN: echo "OUTPUT(\"%t.out\")" > %T/foo.script
then we do:
void ScriptParser::readInclude() {
...
std::unique_ptr<MemoryBuffer> &MB = *MBOrErr;
tokenize(MB->getMemBufferRef());
OwningMBs.push_back(std::move(MB));
}
void ScriptParser::readOutput() {
...
Config->OutputFile = unquote(Tok);
...
}
Problem is that OwningMBs are destroyed after script parser do its job.
So all Toks are dead and Config->OutputFile points to destroyed data.
Patch suggests to save all included scripts into using string Saver.
Differential revision: https://reviews.llvm.org/D27987
llvm-svn: 290238
Older versions of BFD generate libraries with .MIPS.abiflags that only
concatenate the individual .MIPS.abiflags sections instead of merging.
Patch by Alexander Richardson.
Differential revision: https://reviews.llvm.org/D27770
llvm-svn: 290237
Summary:
This follows up to r290217, and makes functions on ASTRecordReader consistent
and valid style.
Reviewers: rsmith
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D28008
llvm-svn: 290236
In r267672, where the loop distribution pragma was introduced, I tried
it hard to keep the old behavior for opt: when opt is invoked
with -loop-distribute, it should distribute the loop (it's off by
default when ran via the optimization pipeline).
As MichaelZ has discovered this has the unintended consequence of
breaking a very common developer work-flow to reproduce compilations
using opt: First you print the pass pipeline of clang
with -debug-pass=Arguments and then invoking opt with the returned
arguments.
clang -debug-pass will include -loop-distribute but the pass is invoked
with default=off so nothing happens unless the loop carries the pragma.
While through opt (default=on) we will try to distribute all loops.
This changes opt's default to off as well to match clang. The tests are
modified to explicitly enable the transformation.
llvm-svn: 290235
Ahmed Bougacha