After loop interchange, the (old) outer loop header should not jump to
`LoopExit`. Note that the old outer loop becomes the new inner loop
after interchange. If we branched to `LoopExit` then after interchange
we would jump directly from the (new) inner loop header to `LoopExit`
without executing the rest of (new) outer loop.
This patch modifies adjustLoopBranches() such that the old outer
loop header (which becomes the new inner loop header) jumps to the
old inner loop latch which becomes the new outer loop latch after
interchange.
Reviewed By: bmahjour
Differential Revision: https://reviews.llvm.org/D98475
We don't need to instantiate single multiclasses inside of
other multiclasses. We can use inheritance and save writing 'defm ""'.
Reviewed By: khchen
Differential Revision: https://reviews.llvm.org/D100074
Benchmarking chromium_framework on a 3.2 GHz 16-Core Intel Xeon W Mac Pro:
N Min Max Median Avg Stddev
x 20 4.33 4.42 4.37 4.37 0.021026299
+ 20 4.12 4.23 4.18 4.175 0.035318103
Difference at 95.0% confidence
-0.195 +/- 0.0186025
-4.46224% +/- 0.425686%
(Student's t, pooled s = 0.0290644)
Reviewed By: #lld-macho, gkm
Differential Revision: https://reviews.llvm.org/D99998
Allow pass to work separately with SGPR, VGPR registers or both.
This is NFC now but will be needed to split RA for separate
SGPR and VGPR passes.
Differential Revision: https://reviews.llvm.org/D100063
When looking up user specified formatters qualifiers are removed from types before matching,
I have added a clarifying example to the document and added an example to a relevant test to demonstrate this behavior.
Differential Revision: https://reviews.llvm.org/D99827
If the constants have a difference of 1 we can convert one to
the other by adding or subtracting the condition.
We have a DAG combine for this, but it only runs before type
legalization. If the select is introduced later during type
legalization or op legalization we will miss it.
We don't need a specific condition, but some conditions are
harder to materialize than others on RISCV. I know that SETLT
will be a single instruction and it is what is used by the
motivating pattern from signed saturating add/sub.
Differential Revision: https://reviews.llvm.org/D99021
6fe7de90b9 changed GNU toolchain,
and added new RUN line to test expected behavior.
The change is for GNU toolchain only, so this will fail other toolchain,
eg: AIX.
Update the test with `-target` to test GNU tool chain only.
Reviewed By: MaskRay
Differential Revision: https://reviews.llvm.org/D99901
If either `time-trace-granularity` or `time-trace-file` is specified, then don't make users specify `-time-trace`.
It seems silly that I have to type all three options, eg, `-time-trace -time-trace-file=- -time-trace-granularity=...`.
Differential Revision: https://reviews.llvm.org/D100011
When allocLikeOp is updated in alloc constant folding,
alighnment attribute was ignored. This patch fixes it.
Signed-off-by: Haruki Imai <imaihal@jp.ibm.com>
Reviewed By: mehdi_amini
Differential Revision: https://reviews.llvm.org/D99882
When using the large code model with FastISel (for example via
clang -O0 which adds the optnone attribute), FP constants could
still be materialized using adrp + ldr. Unconditionally enable
the existing path for MachO to materialize the constant in code.
For testing, restore literal_pools_float.ll to exercise the constant
pool and add two optnone-functions that return a float and a double,
respectively. Consolidate fpimm.ll and add a new fast-isel-fpimm.ll
to check the code paths taken with FastISel.
Differential Revision: https://reviews.llvm.org/D99607
Since we have created a new OF_TextWithCRLF flag, we no longer need to worry about OF_Text flag turning on CRLF translation. I can remove this workaround I added to globally open all ToolOutputFiles as binary on Windows.
Reviewed By: rnk
Differential Revision: https://reviews.llvm.org/D100034
Just because an attribute is a statement attribute doesn't mean it's
not also a declaration attribute. In Clang, there are not currently any
DeclOrStmtAttr attributes that require mutual exclusion checking, but
downstream clients discovered this issue.
- use "empty()" instead of "size()"
- refactor the re-export code so it doesn't create a new vector every time.
Differential Revision: https://reviews.llvm.org/D100019
An empty history entry can happen by entering the expression evaluator an immediately hitting enter:
```
$ lldb
(lldb) e
Enter expressions, then terminate with an empty line to evaluate:
1: <hit enter>
```
The next time the user enters the expression evaluator, if they hit the up arrow to load the previous expression, lldb crashes. This patch treats empty history sessions as a single expression of zero length, instead of an empty list of expressions.
Fixes http://llvm.org/PR49845.
Differential Revision: https://reviews.llvm.org/D100048
This can't use our normal strategy of splatting the scalar and using
a .vv operation instead of .vx.
Instead this patch bitcasts the vector to the equivalent SEW=32
vector and inserts the scalar parts using two vslide1up/down. We
do that unmasked and apply the mask separately at the end with
a vmerge.
For vslide1up there maybe some other options here like getting
i64 into element 0 and using vslideup.vi with this vector as
vd and the original source as vs1. Masking would still need to
be done afterwards.
That idea doesn't work for vslide1down. We need to slidedown and
then insert a single scalar at vl-1 which we could do with a
vslideup, but that assumes vl > 0 which I don't think we can assume.
The i32 double slide1down implemented here is the best I could come
up with and I just made vslide1up consistent.
Reviewed By: frasercrmck
Differential Revision: https://reviews.llvm.org/D99910
The ROCm installation directory may be another
directory, llvm/ inside the build directory.
Reviewed By: yaxunl
Differential Revision: https://reviews.llvm.org/D100045
This allows FoldConstantArithmetic to handle SPLAT_VECTOR in
addition to BUILD_VECTOR. This allows it to support scalable
vectors. I'm also allowing fixed length SPLAT_VECTOR which is
used by some targets, but I'm not familiar enough to write tests
for those targets.
I had to block this function from running on CONCAT_VECTORS to
avoid calling getNode for a CONCAT_VECTORS of 2 scalars.
This can happen because the 2 operand getNode calls this
function for any opcode. Previously we were protected because
CONCAT_VECTORs of BUILD_VECTOR is folded to a larger BUILD_VECTOR
before that call. But it's not always possible to fold a CONCAT_VECTORS
of SPLAT_VECTORs, and we don't even try.
This fixes PR49781 where DAG combine thought constant folding
should be possible, but FoldConstantArithmetic couldn't do it.
Reviewed By: david-arm
Differential Revision: https://reviews.llvm.org/D99682
-Make sure of the CreateShl/LShr/AShr methods that take a uint64_t
instead of creating a ConstantInt for 1 ourselves.
-Use Builder.getInt1 or ConstantInt::getBool instead of a conditional.
-Pull out repeated calls to getType.
Some sparse matrices operate on integral values (in contrast with the common
f32 and f64 values). This CL expands the compiler and runtime support to deal
with several common type combinations.
Reviewed By: bixia
Differential Revision: https://reviews.llvm.org/D99999
As D99834 was meant specifically for FreeBSD, which still uses the older
non-trivial std::pair copy constructors, test for `__FreeBSD__` instead
of relying on a macro which is an internal detail of libc++.
Noted by Louis Dionne.
All of the code that handles general constant here (other than the more
restrictive APInt-dealing code) expects that it is an immediate,
because otherwise we won't actually fold the constants, and increase
instruction count. And it isn't obvious why we'd be okay with
increasing the number of constant expressions,
those still will have to be run..
But after 2829094a8e
this could also cause endless combine loops.
So actually properly restrict this code to immediates.
The document has the following updates:
- Rename 'feature test' to 'feature-test', the latter is the spelling
used in the Standard.
- Add information how an ABI list can be downloaded from Buildkite.
Differential Revision: https://reviews.llvm.org/D99290
This fixes the examples from
D99674 and
https://llvm.org/PR49878
The matchers succeed on partial undef/poison vector constants,
but the transform creates a full 'not' (-1) constant, so it
would undo a demanded vector elements change triggered by the
extractelement.
Differential Revision: https://reviews.llvm.org/D100044
* `std::predicate`
* `std::relation`
* `std::equivalence_relation`
* `std::strict_weak_order`
Implements parts of:
- P0898R3 Standard Library Concepts
- P1754 Rename concepts to standard_case for C++20, while we still can
Differential Revision: https://reviews.llvm.org/D96477
We had been giving them a string index of zero, which actually corresponds to a
string with a single space due to {D89639}.
This was far from obvious in the old test because llvm-nm doesn't quote the
symbol names, making the empty string look identical to a string of a single
space. `dsymutil -s` quotes its strings, so I've changed the test accordingly.
Fixes llvm.org/PR48714. Thanks @clayborg for the tips!
Reviewed By: clayborg
Differential Revision: https://reviews.llvm.org/D100003
We see a regression related to low probe factor(0.01) which prevents some callsites being promoted in ICPPass and later cause the missing inline in CGSCC inliner. The root cause is due to redundant(the second) multiplication of the probe factor and this change try to fix it.
`Sum` does multiply a factor right after findCallSamples but later when using as the parameter in setProbeDistributionFactor, it multiplies one again.
This change could get ~2% perf back on mcf benchmark. In mcf, previously the corresponding factor is 1 and it's the recent feature introducing the <1 factor then trigger this bug.
Reviewed By: hoy, wenlei
Differential Revision: https://reviews.llvm.org/D99787
This used to work before rG77d625f8d8aa, but we now merge the shuffles across the fadd resulting in a hadd that requires a lane crossing post shuffle, which we don't permit on AVX1 targets
After D98856 these tests will by default break (fatal_error) if any of
the wrong interfaces are used, so there's no longer a need to have a
RUN line that checks for a warning message emitted by the compiler.
Use report_fatal_error here since this is an internal error, and not
something the user can/should be trying to fix.
Also distinguish between the symbol being missing and the symbol having
the wrong type.
We have a failure internally where the symbol is missing. Currently
trying to reduce the test case to something we can attach to an llvm
bug.
Differential Revision: https://reviews.llvm.org/D99960
The struct is used for both, callee and caller-save registers now.
The frame index is not set for entrypoints, as we do not need to save
the registers then.
Update the struct name to reflect that.
Differential Revision: https://reviews.llvm.org/D99722
Extend D94856 to handle 'and', 'or' and 'xor' instructions as well
We still fail on many i8/i16 cases as the test and the logic-op are performed on different widths
CongzheUalberta