| Commit message (Collapse) | Author | Age | Files | Lines |
| ... | |
| |
|
|
|
|
| |
optimizing moves to use 2 byte VEX prefix.
llvm-svn: 270394
|
| |
|
|
|
|
|
| |
This reverts commit r270106. It results in certain function types
omitted in the output.
llvm-svn: 270389
|
| |
|
|
|
|
|
|
| |
subvectors using XMM or YMM stores instead of the vector extract instructions.
Similar is already done for AVX and we had lost it going to AVX512VL.
llvm-svn: 270383
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
(PR27823)
This isn't the complete fix, but it handles the trivial examples of duplicate vzero* ops in PR27823:
https://llvm.org/bugs/show_bug.cgi?id=27823
...and amusingly, the bogus cases already exist as regression tests, so let's take this baby step.
We'll need to do more in the general case where there's legitimate AVX usage in the function + there's
already a vzero in the code.
Differential Revision: http://reviews.llvm.org/D20477
llvm-svn: 270378
|
| |
|
|
| |
llvm-svn: 270372
|
| |
|
|
| |
llvm-svn: 270370
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
vector splat constants
We could try harder to handle non-splat vector constants too,
but that seems much rarer to me.
Note that the div test isn't resolved because there's a check
for isIntegerTy() guarding that transform.
Differential Revision: http://reviews.llvm.org/D20497
llvm-svn: 270369
|
| |
|
|
|
|
| |
Differential Revision: http://reviews.llvm.org/D20513
llvm-svn: 270357
|
| |
|
|
|
|
| |
the source is 512-bits. The 256-bit source patterns were redundant with AVX.
llvm-svn: 270356
|
| |
|
|
|
|
| |
index 0 patterns. This gives them higher priority than the memory patterns. This matches AVX1/2.
llvm-svn: 270355
|
| |
|
|
|
|
| |
equivalents. This helps group them close together in the isel tables and enable table compression.
llvm-svn: 270354
|
| |
|
|
|
|
| |
inversions could appear in a row.
llvm-svn: 270344
|
| |
|
|
| |
llvm-svn: 270343
|
| |
|
|
|
|
| |
Differential Revision: http://reviews.llvm.org/D20459
llvm-svn: 270336
|
| |
|
|
|
|
| |
for integer types when only AVX1 is supported.
llvm-svn: 270335
|
| |
|
|
| |
llvm-svn: 270334
|
| |
|
|
|
|
| |
used to indicating the zero masking behavior which is not the case here. NFC
llvm-svn: 270333
|
| |
|
|
|
|
| |
reflect the fact that memory is the destination.
llvm-svn: 270332
|
| |
|
|
| |
llvm-svn: 270331
|
| |
|
|
| |
llvm-svn: 270329
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
repairIntervalsInRange
This fixes a bug introduced in:
r262115 - CodeGen: Take MachineInstr& in SlotIndexes and LiveIntervals, NFC
The iterator End here might == MBB->end(), and so we can't unconditionally
dereference it. This often goes unnoticed (I don't have a test case that always
crashes, and ASAN does not catch it either) because the function call arguments are
turned right back into iterators. MachineInstrBundleIterator's constructor,
however, does have an assert which might randomly fire.
llvm-svn: 270323
|
| |
|
|
|
|
| |
Differential Revision: http://reviews.llvm.org/D20438
llvm-svn: 270322
|
| |
|
|
|
|
| |
Differential Revision: http://reviews.llvm.org/D20324
llvm-svn: 270321
|
| |
|
|
|
|
| |
AVX2 versions of vector extract when AVX512VL is enabled.
llvm-svn: 270318
|
| |
|
|
|
|
| |
AVX512VL is enabled. Also add shuffle comment printing for AVX512VL VPERMPD/VPERMQ to keep some tests that now use these instructions instead of the AVX2 ones.
llvm-svn: 270317
|
| |
|
|
|
|
| |
is enabled.
llvm-svn: 270316
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
lifetime.end
A cleanuppad is not cheap, they turn into many instructions and result
in additional spills and fills. It is not worth keeping a cleanuppad
around if all it does is hold a lifetime.end instruction.
N.B. We first try to merge the cleanuppad with another cleanuppad to
avoid dropping the lifetime and debug info markers.
llvm-svn: 270314
|
| |
|
|
|
|
|
|
|
|
| |
Allocating larger register classes first should give better allocation
results (and more importantly for myself, make the lit tests more stable
with respect to scheduler changes).
Patch by Matthias Braun
llvm-svn: 270312
|
| |
|
|
|
|
| |
AVX512VL/AVX512BWI equivalents are available.
llvm-svn: 270311
|
| |
|
|
|
|
| |
instead of pattern matching the intrinsics. This unifies handling with AVX512 and allows these intrinsics to select EVEX encoded instructions to increase available registers.
llvm-svn: 270310
|
| |
|
|
|
|
|
|
| |
The InductiveRangeCheck struct is only five words long; so passing these
around value is fine. The allocator makes the code look more complex
than it is.
llvm-svn: 270309
|
| |
|
|
| |
llvm-svn: 270308
|
| |
|
|
|
|
|
|
| |
These are kind of a mess and hard to follow, particularly
for loads and stores. Fix various redundant, unnecessary
and dead settings.
llvm-svn: 270307
|
| |
|
|
|
|
|
|
|
|
|
| |
I had used `std::remove_if` under the assumption that it moves the
predicate matching elements to the end, but actaully the elements
remaining towards the end (after the iterator returned by
`std::remove_if`) are indeterminate. Fix the bug (and make the code
more straightforward) by using a temporary SmallVector, and add a test
case demonstrating the issue.
llvm-svn: 270306
|
| |
|
|
|
|
|
| |
This is essentially doing a 24-bit signed division with FP.
We need to truncate to the N bit result.
llvm-svn: 270305
|
| |
|
|
|
|
|
| |
Prior to this patch, we were using 1 for all the repairing costs.
Now, we use the information from the target to get this information.
llvm-svn: 270304
|
| |
|
|
| |
llvm-svn: 270302
|
| |
|
|
|
|
|
|
|
|
|
| |
The current SGPR spilling test does not stress this
because it is using s_buffer_load instructions to
increase SGPR pressure and spill, but their output
operands have the same SReg_32_XM0 constraint. This fixes
an error when the SReg_32 output from most instructions
is spilled.
llvm-svn: 270301
|
| |
|
|
| |
llvm-svn: 270300
|
| |
|
|
| |
llvm-svn: 270299
|
| |
|
|
|
|
|
| |
Everything now compiles successfully, but there are still undefined
references.
llvm-svn: 270298
|
| |
|
|
| |
llvm-svn: 270297
|
| |
|
|
| |
llvm-svn: 270296
|
| |
|
|
|
|
| |
Original patch by Tom Stellard
llvm-svn: 270295
|
| |
|
|
|
|
|
|
|
| |
This saves a small amount of code size, and is a first small step toward
passing values on the stack across block boundaries.
Differential Review: http://reviews.llvm.org/D20450
llvm-svn: 270294
|
| |
|
|
|
|
|
| |
This should not be making assumptions on the value of
the casted pointer.
llvm-svn: 270293
|
| |
|
|
| |
llvm-svn: 270292
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We now use LiveRangeCalc::extendToUses() instead of a specially designed
algorithm in constructMainRangeFromSubranges():
- The original motivation for constructMainRangeFromSubranges() were
differences between the main liverange and subranges because of hidden
dead definitions. This case however cannot happen anymore with the
DetectDeadLaneMasks pass in place.
- It simplifies the code.
- This fixes a longstanding bug where we did not properly create new SSA
values on merging control flow (the MachineVerifier missed most of
these cases).
- Move constructMainRangeFromSubranges() to LiveIntervalAnalysis and
LiveRangeCalc to better match the implementation/available helper
functions.
This re-applies r269016. The fixes from r270290 and r270259 should avoid
the machine verifier problems this time.
llvm-svn: 270291
|
| |
|
|
|
|
|
|
|
|
|
| |
It is fine for subregister ranges to be undefined on some CFG paths as
we may have a "vregX:other_subreg<read-undef> =" def on that path. We
do not (and should not) have live segments for the subregister ranges.
The MachineVerifier should not complain about this.
This is a slight variant of http://llvm.org/PR27705
llvm-svn: 270290
|
| |
|
|
| |
llvm-svn: 270289
|