Summary:
This patch adds the following
1. Adds a skeleton scheduler model for AMD Znver1.
2. Introduces the znver1 execution units and pipes.
3. Caters the instructions based on the generic scheduler classes.
4. Further additions to the scheduler model with instruction itineraries will be carried out incrementally based on
a. Instructions types
b. Registers used
5. Since itineraries are not added based on instructions, throughput information are bound to change when incremental changes are added.
6. Scheduler testcases are modified accordingly to suit the new model.
Patch by Ganesh Gopalasubramanian. With minor formatting tweaks from me.
Reviewers: craig.topper, RKSimon
Subscribers: javed.absar, shivaram, ddibyend, vprasad
Differential Revision: https://reviews.llvm.org/D35293
llvm-svn: 308411
It should be a win to avoid going out to the system lib for all small memcmp() calls using scalar ops. For x86 32-bit, this means most everything up to 16 bytes. For 64-bit, that doubles because we can do 8-byte loads.
Notes:
Reduced from 4 to 2 loads for -Os behavior, which might not be optimal in all cases. It's effectively a question of how much do we trust the system implementation. Linux and macOS (and Windows I assume, but did not test) have optimized memcmp() code for x86, so it's probably not bad either way? PPC is using 8/4 for defaults on these. We do not expand at all for -Oz.
There are still potential improvements to make for the CGP expansion IR and/or lowering such as avoiding select-of-constants (D34904) and not doing zexts to the max load type before doing a compare.
We have special-case SSE/AVX codegen for (memcmp(x, y, 16/32) == 0) that will no longer be produced after this patch. I've shown the experimental justification for that change in PR33329:
https://bugs.llvm.org/show_bug.cgi?id=33329#c12
TLDR: While the vector code is a likely winner, we can't guarantee that it's a winner in all cases on all CPUs, so I'm willing to sacrifice it for the greater good of expanding all small memcmp(). If we want to resurrect that codegen, it can be done by adjusting the CGP params or poking a hole to let those fall-through the CGP expansion.
Committed on behalf of Sanjay Patel
Differential Revision: https://reviews.llvm.org/D35067
llvm-svn: 308322
This isn't legal code, but we shouldn't crash on it. Now we just don't convert the gather intrinsic if the scale isn't constant and let it go through to isel where we'll report an isel failure.
Fixes PR33772.
llvm-svn: 308267
Rename the enum value from X86_64_Win64 to plain Win64.
The symbol exposed in the textual IR is changed from 'x86_64_win64cc'
to 'win64cc', but the numeric value is kept, keeping support for
old bitcode.
Differential Revision: https://reviews.llvm.org/D34474
llvm-svn: 308208
LLVM compiler recognizes opportunities to transform a branch into IR select instruction(s) - later it will be lowered into X86::CMOV instruction, assuming no other optimization eliminated the SelectInst.
However, it is not always profitable to emit X86::CMOV instruction. For example, branch is preferable over an X86::CMOV instruction when:
1. Branch is well predicted
2. Condition operand is expensive, compared to True-value and the False-value operands
In CodeGenPrepare pass there is a shallow optimization that tries to convert SelectInst into branch, but it is not enough.
This commit, implements machine optimization pass that converts X86::CMOV instruction(s) into branch, based on a conservative heuristic.
Differential Revision: https://reviews.llvm.org/D34769
llvm-svn: 308142
FastIsel can't handle them, so we would end up crashing during
register class selection.
Fixes PR26522.
Differential Revision: https://reviews.llvm.org/D35272
llvm-svn: 307797
The issue is not if the value is pcrel. It is whether we have a
relocation or not.
If we have a relocation, the static linker will select the upper
bits. If we don't have a relocation, we have to do it.
llvm-svn: 307730
OpenCL 2.0 introduces the notion of memory scopes in atomic operations to
global and local memory. These scopes restrict how synchronization is
achieved, which can result in improved performance.
This change extends existing notion of synchronization scopes in LLVM to
support arbitrary scopes expressed as target-specific strings, in addition to
the already defined scopes (single thread, system).
The LLVM IR and MIR syntax for expressing synchronization scopes has changed
to use *syncscope("<scope>")*, where <scope> can be "singlethread" (this
replaces *singlethread* keyword), or a target-specific name. As before, if
the scope is not specified, it defaults to CrossThread/System scope.
Implementation details:
- Mapping from synchronization scope name/string to synchronization scope id
is stored in LLVM context;
- CrossThread/System and SingleThread scopes are pre-defined to efficiently
check for known scopes without comparing strings;
- Synchronization scope names are stored in SYNC_SCOPE_NAMES_BLOCK in
the bitcode.
Differential Revision: https://reviews.llvm.org/D21723
llvm-svn: 307722
The SandyBridge architects have provided us with a more accurate information about each instruction latency, number of uOPs and used ports and I used it to replace the existing estimated SNB instructions scheduling and to add missing scheduling information.
Please note that the patch extensively affects the X86 MC instr scheduling for SNB.
Also note that this patch will be followed by additional patches for the remaining target architectures HSW, IVB, BDW, SKL and SKX.
The updated and extended information about each instruction includes the following details:
•static latency of the instruction
•number of uOps from which the instruction consists of
•all ports used by the instruction's' uOPs
For example, the following code dictates that instructions, ADC64mr, ADC8mr, SBB64mr, SBB8mr have a static latency of 9 cycles. Each of these instructions is decoded into 6 micro operations which use ports 4, ports 2 or 3 and port 0 and ports 0 or 1 or 5:
def SBWriteResGroup94 : SchedWriteRes<[SBPort4,SBPort23,SBPort0,SBPort015]> {
let Latency = 9;
let NumMicroOps = 6;
let ResourceCycles = [1,2,2,1];
}
def: InstRW<[SBWriteResGroup94], (instregex "ADC64mr")>;
def: InstRW<[SBWriteResGroup94], (instregex "ADC8mr")>;
def: InstRW<[SBWriteResGroup94], (instregex "SBB64mr")>;
def: InstRW<[SBWriteResGroup94], (instregex "SBB8mr")>;
Note that apart for the header, most of the X86SchedSandyBridge.td file was generated by a script.
Reviewers: zvi, chandlerc, RKSimon, m_zuckerman, craig.topper, igorb
Differential Revision: https://reviews.llvm.org/D35019#inline-304691
llvm-svn: 307529
Summary:
Mark G_ZEXT/G_SEXT i1 to i8/i16, i8 to i16 as legal.
Support G_ZEXT i1 to i8/i16 instruction selection ( C++ code).
This patch requred to support G_LOAD/G_STORE i1.
Reviewers: zvi, guyblank
Reviewed By: guyblank
Subscribers: rovka, llvm-commits, kristof.beyls
Differential Revision: https://reviews.llvm.org/D35177
llvm-svn: 307526
GHC 8.4 will know how to use YMM and ZMM registers for calls.
Submitted on behalf of @bgamari (Ben Gamari)
Differential Revision: https://reviews.llvm.org/D34854
llvm-svn: 307504
x86 scalar select-of-constants (Cond ? C1 : C2) combining/lowering is a mess
with missing optimizations. We handle some patterns, but miss logical variants.
To clean that up, we should convert all select-of-constants to logic/math and
enhance the combining for the expected patterns from that. Selecting 0 or -1
needs extra attention to produce the optimal code as shown here.
Attempt to verify that all of these IR forms are logically equivalent:
http://rise4fun.com/Alive/plxs
Earlier steps in this series:
rL306040
rL306072
rL307404 (D34652)
As acknowledged in the earlier review, there's a possibility that some Intel
uarch would prefer to produce an xor to clear the fake register operand with
sbb %eax, %eax. This will likely need to be addressed in a separate pass.
llvm-svn: 307471
x86 scalar select-of-constants (Cond ? C1 : C2) combining/lowering is a mess
with missing optimizations. We handle some patterns, but miss logical variants.
To clean that up, we should convert all select-of-constants to logic/math and
enhance the combining for the expected patterns from that. DAGCombiner already
has the foundation to allow the transforms, so we just need to fill in the holes
for x86 math op lowering. Selecting 0 or -1 needs extra attention to produce the
optimal code as shown here.
Attempt to verify that all of these IR forms are logically equivalent:
http://rise4fun.com/Alive/plxs
Earlier steps in this series:
rL306040
rL306072
Differential Revision: https://reviews.llvm.org/D34652
llvm-svn: 307404
Summary:
Replace the matcher if-statements for each rule with a state-machine. This
significantly reduces compile time, memory allocations, and cumulative memory
allocation when compiling AArch64InstructionSelector.cpp.o after r303259 is
recommitted.
The following patches will expand on this further to fully fix the regressions.
Reviewers: rovka, ab, t.p.northover, qcolombet, aditya_nandakumar
Reviewed By: ab
Subscribers: vitalybuka, aemerson, javed.absar, igorb, llvm-commits, kristof.beyls
Differential Revision: https://reviews.llvm.org/D33758
llvm-svn: 307079
Summary:
When broadcasting from the constant pool its useful to print out the final vector similar to what we do for normal moves from the constant pool.
I changed only a couple tests that were broadcast focused. One of them had been previously hand tweaked after running the script so that it could check the constant pool declaration. But I think this patch makes that unnecessary now since we can check the comment instead.
Reviewers: spatel, RKSimon, zvi
Reviewed By: spatel
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D34923
llvm-svn: 307062
Summary: I believe this should be supported on GLM since RDSEED is.
Reviewers: m_zuckerman, zvi, RKSimon
Reviewed By: RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D34828
llvm-svn: 307060
Summary:
Add a combine for creating a truncate to replace a build_vector composed of extracts with
indices that form a stride-2^N series.
Example:
v8i32 V = ...
v4i32 build_vector((extract_elt V, 0), (extract_elt V, 2), (extract_elt V, 4), (extract_elt V, 6))
-->
v4i32 truncate (bitcast V to v4i64)
Related discussion in llvm-dev about canonicalizing shuffles to
truncates in LLVM IR:
http://lists.llvm.org/pipermail/llvm-dev/2017-January/108936.html.
Reviewers: spatel, RKSimon, efriedma, igorb, craig.topper, wolfgangp, delena
Reviewed By: delena
Subscribers: guyblank, delena, javed.absar, llvm-commits
Differential Revision: https://reviews.llvm.org/D34077
llvm-svn: 307036
We are combining shuffles to bit shifts before unary permutes, which means we can't fold loads plus the destination register is destructive
llvm-svn: 306978
We are combining shuffles to bit shifts before unary permutes, which means we can't fold loads plus the destination register is destructive
The 32-bit shuffles are a bit tricky and will be dealt with in a later patch
llvm-svn: 306977
this patch updates the cost of addq\subq (add\subtract of vectors of 64bits)
based on the performance numbers of SLM arch.
Differential Revision: https://reviews.llvm.org/D33983
llvm-svn: 306974
Summary: Support G_GLOBAL_VALUE operation. For now most of the PIC configurations not implemented yet.
Reviewers: zvi, guyblank
Reviewed By: guyblank
Subscribers: rovka, kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D34738
Conflicts:
test/CodeGen/X86/GlobalISel/regbankselect-X86_64.mir
llvm-svn: 306972
Summary:
Support vector type G_UNMERGE_VALUES selection.
For now G_UNMERGE_VALUES marked as legal for any type, so nothing to do in legalizer.
Reviewers: t.p.northover, qcolombet, zvi, guyblank
Reviewed By: guyblank
Subscribers: rovka, kristof.beyls, guyblank, llvm-commits
Differential Revision: https://reviews.llvm.org/D33665
llvm-svn: 306971
Summary:
Support vector type G_MERGE_VALUES selection. For now G_MERGE_VALUES marked as legal for any type, so nothing to do in legalizer.
Split from https://reviews.llvm.org/D33665
Reviewers: qcolombet, t.p.northover, zvi, guyblank
Reviewed By: guyblank
Subscribers: rovka, kristof.beyls, guyblank, llvm-commits
Differential Revision: https://reviews.llvm.org/D33958
llvm-svn: 306665
CFI instructions that set appropriate cfa offset and cfa register are now
inserted in emitEpilogue() in X86FrameLowering.
Majority of the changes in this patch:
1. Ensure that CFI instructions do not affect code generation.
2. Enable maintaining correct information about cfa offset and cfa register
in a function when basic blocks are reordered, merged, split, duplicated.
These changes are target independent and described below.
Changed CFI instructions so that they:
1. are duplicable
2. are not counted as instructions when tail duplicating or tail merging
3. can be compared as equal
Add information to each MachineBasicBlock about cfa offset and cfa register
that are valid at its entry and exit (incoming and outgoing CFI info). Add
support for updating this information when basic blocks are merged, split,
duplicated, created. Add a verification pass (CFIInfoVerifier) that checks
that outgoing cfa offset and register of predecessor blocks match incoming
values of their successors.
Incoming and outgoing CFI information is used by a late pass
(CFIInstrInserter) that corrects CFA calculation rule for a basic block if
needed. That means that additional CFI instructions get inserted at basic
block beginning to correct the rule for calculating CFA. Having CFI
instructions in function epilogue can cause incorrect CFA calculation rule
for some basic blocks. This can happen if, due to basic block reordering,
or the existence of multiple epilogue blocks, some of the blocks have wrong
cfa offset and register values set by the epilogue block above them.
Patch by Violeta Vukobrat.
Differential Revision: https://reviews.llvm.org/D18046
llvm-svn: 306529
•static latency
•number of uOps from which the instructions consists
•all ports used by the instruction
Reviewers:
RKSimon
zvi
aymanmus
m_zuckerman
Differential Revision: https://reviews.llvm.org/D33897
llvm-svn: 306414
[X86][AVX512] Improve lowering of AVX512 compare intrinsics (remove redundant shift left+right instructions).
AVX512 compare instructions return v*i1 types.
In cases where the number of elements in the returned value are less than 8, clang adds zeroes to get a mask of v8i1 type.
Later on it's replaced with CONCAT_VECTORS, which then is lowered to many DAG nodes including insert/extract element and shift right/left nodes.
The fact that AVX512 compare instructions put the result in a k register and zeroes all its upper bits allows us to remove the extra nodes simply by copying the result to the required register class.
When lowering, identify these cases and transform them into an INSERT_SUBVECTOR node (marked legal), then catch this pattern in instructions selection phase and transform it into one avx512 cmp instruction.
Differential Revision: https://reviews.llvm.org/D33188
llvm-svn: 306402
The non-AVX-512 behavior was changed in r248266 to match N1778
(C bindings for IEEE-754 (2008)), which defined the four functions
to not raise the inexact exception ("rint" is still defined as raising
it).
Update the AVX-512 lowering of these functions to match that: it should
not be different.
llvm-svn: 306299
Convert vector increment or decrement to sub/add with an all-ones constant:
add X, <1, 1...> --> sub X, <-1, -1...>
sub X, <1, 1...> --> add X, <-1, -1...>
The all-ones vector constant can be materialized using a pcmpeq instruction that is
commonly recognized as an idiom (has no register dependency), so that's better than
loading a splat 1 constant.
AVX512 uses 'vpternlogd' for 512-bit vectors because there is apparently no better
way to produce 512 one-bits.
The general advantages of this lowering are:
1. pcmpeq has lower latency than a memop on every uarch I looked at in Agner's tables,
so in theory, this could be better for perf, but...
2. That seems unlikely to affect any OOO implementation, and I can't measure any real
perf difference from this transform on Haswell or Jaguar, but...
3. It doesn't look like it from the diffs, but this is an overall size win because we
eliminate 16 - 64 constant bytes in the case of a vector load. If we're broadcasting
a scalar load (which might itself be a bug), then we're replacing a scalar constant
load + broadcast with a single cheap op, so that should always be smaller/better too.
4. This makes the DAG/isel output more consistent - we use pcmpeq already for padd x, -1
and psub x, -1, so we should use that form for +1 too because we can. If there's some
reason to favor a constant load on some CPU, let's make the reverse transform for all
of these cases (either here in the DAG or in a later machine pass).
This should fix:
https://bugs.llvm.org/show_bug.cgi?id=33483
Differential Revision: https://reviews.llvm.org/D34336
llvm-svn: 306289
Summary:
Support vector type G_EXTRACT selection. For now G_EXTRACT marked as legal for any type, so nothing to do in legalizer.
Split from https://reviews.llvm.org/D33665
Reviewers: qcolombet, t.p.northover, zvi, guyblank
Reviewed By: guyblank
Subscribers: guyblank, rovka, llvm-commits, kristof.beyls
Differential Revision: https://reviews.llvm.org/D33957
llvm-svn: 306240
The cost of an interleaved access was only implemented for AVX512. For other
X86 targets an overly conservative Base cost was returned, resulting in
avoiding vectorization where it is actually profitable to vectorize.
This patch starts to add costs for AVX2 for most prominent cases of
interleaved accesses (stride 3,4 chars, for now).
Note1: Improvements of up to ~4x were observed in some of EEMBC's rgb
workloads; There is also a known issue of 15-30% degradations on some of these
workloads, associated with an interleaved access followed by type
promotion/widening; the resulting shuffle sequence is currently inefficient and
will be improved by a series of patches that extend the X86InterleavedAccess pass
(such as D34601 and more to follow).
Note 2: The costs in this patch do not reflect port pressure penalties which can
be very dominant in the case of interleaved accesses since most of the shuffle
operations are restricted to a single port. Further tuning, that may incorporate
these considerations, will be done on top of the upcoming improved shuffle
sequences (that is, along with the abovementioned work to extend
X86InterleavedAccess pass).
Differential Revision: https://reviews.llvm.org/D34023
llvm-svn: 306238
processFixupValue is called on every relaxation iteration. applyFixup
is only called once at the very end. applyFixup is then the correct
place to do last minute changes and value checks.
While here, do proper range checks again for fixup_arm_thumb_bl. We
used to do it, but dropped because of thumb2. We now do it again, but
use the thumb2 range.
llvm-svn: 306177
Commit r306010 adjusted the condition as follows:
- if (Is64Bit) {
+ if (!STI.isTargetWin32()) {
The intent was to preserve the behavior on all Windows platforms
but extend the behavior on 64-bit Windows platforms to every
other one. (Before r306010, emitStackProbeCall only ever executed
when emitting code for Windows triples.)
Unfortunately,
if (Is64Bit && STI.isOSWindows())
is not the same as
if (!STI.isTargetWin32())
because of the way isTargetWin32() is defined:
bool isTargetWin32() const {
return !In64BitMode && (isTargetCygMing() ||
isTargetKnownWindowsMSVC());
}
In practice this broke the JIT tests on 32-bit Windows, which did not
satisfy the new condition:
LLVM :: ExecutionEngine/MCJIT/2003-01-15-AlignmentTest.ll
LLVM :: ExecutionEngine/MCJIT/2003-08-15-AllocaAssertion.ll
LLVM :: ExecutionEngine/MCJIT/2003-08-23-RegisterAllocatePhysReg.ll
LLVM :: ExecutionEngine/MCJIT/test-loadstore.ll
LLVM :: ExecutionEngine/OrcMCJIT/2003-01-15-AlignmentTest.ll
LLVM :: ExecutionEngine/OrcMCJIT/2003-08-15-AllocaAssertion.ll
LLVM :: ExecutionEngine/OrcMCJIT/2003-08-23-RegisterAllocatePhysReg.ll
LLVM :: ExecutionEngine/OrcMCJIT/test-loadstore.ll
because %esp was not updated correctly. The failures are only visible
on a MSVC 2017 Debug build, for which we do not have bots.
llvm-svn: 306142
X86_64 COFF only has support for 32 bit pcrel relocations. Produce an
error on all others.
Note that gnu as has extended the relocation values to support
this. It is not clear if we should support the gnu extension.
llvm-svn: 306082
Details: There was a use but it was in the assert which was not
exercised during product build.
Reviewers: Andrew Kaylor
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D32658
llvm-svn: 306073
This is very similar to the transform in:
https://reviews.llvm.org/rL306040
...but in this case, we use cmp X, 1 to set the carry bit as needed.
Again, we can show that all of these are logically equivalent (although
InstCombine currently canonicalizes to a form not seen here), and if
we believe IACA, then this is the smallest/fastest code. Eg, with SNB:
| Num Of | Ports pressure in cycles | |
| Uops | 0 - DV | 1 | 2 - D | 3 - D | 4 | 5 | |
---------------------------------------------------------------------
| 1 | 1.0 | | | | | | | cmp edi, 0x1
| 2 | | 1.0 | | | | 1.0 | CP | sbb eax, eax
The larger motivation is to clean up all select-of-constants combining/lowering
because we're missing some common cases.
llvm-svn: 306072
Summary:
These intrinsics aren't used by clang and haven't been for a while.
There's some really terrible codegen in the 32-bit target for avx512bw due to i64 not being legal. But as I said these intrinsics aren't used by clang even before this patch so this codegen reflects our clang behavior today.
Reviewers: spatel, RKSimon, zvi, igorb
Reviewed By: RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D34389
llvm-svn: 306047
Our handling of select-of-constants is lumpy in IR (https://reviews.llvm.org/D24480),
lumpy in DAGCombiner, and lumpy in X86ISelLowering. That's why we only had the 'sbb'
codegen in 1 out of the 4 tests. This is a step towards smoothing that out.
First, show that all of these IR forms are equivalent:
http://rise4fun.com/Alive/mx
Second, show that the 'sbb' version is faster/smaller. IACA output for SandyBridge
(later Intel and AMD chips are similar based on Agner's tables):
This is the "obvious" x86 codegen (what gcc appears to produce currently):
| Num Of | Ports pressure in cycles | |
| Uops | 0 - DV | 1 | 2 - D | 3 - D | 4 | 5 | |
---------------------------------------------------------------------
| 1* | | | | | | | | xor eax, eax
| 1 | 1.0 | | | | | | CP | test edi, edi
| 1 | | | | | | 1.0 | CP | setnz al
| 1 | | 1.0 | | | | | CP | neg eax
This is the adc version:
| 1* | | | | | | | | xor eax, eax
| 1 | 1.0 | | | | | | CP | cmp edi, 0x1
| 2 | | 1.0 | | | | 1.0 | CP | adc eax, 0xffffffff
And this is sbb:
| 1 | 1.0 | | | | | | | neg edi
| 2 | | 1.0 | | | | 1.0 | CP | sbb eax, eax
If IACA is trustworthy, then sbb became a single uop in Broadwell, so this will be
clearly better than the alternatives going forward.
llvm-svn: 306040
Masked gather for vector length 2 is lowered incorrectly for element type i32.
The type <2 x i32> was automatically extended to <2 x i64> and we generated VPGATHERQQ instead of VPGATHERQD.
The type <2 x float> is extended to <4 x float>, so there is no bug for this type, but the sequence may be more optimal.
In this patch I'm fixing <2 x i32>bug and optimizing <2 x float> sequence for GATHERs only. The same fix should be done for Scatters as well.
Differential revision: https://reviews.llvm.org/D34343
llvm-svn: 305987
There are a couple of potential improvements as seen in the IR and asm:
1. We're unnecessarily extending to a larger type to compare values.
2. The codegen for (select cond, 1, -1) could avoid a cmov.
(or we could change the order of the compares, so we have a select with 0 operand)
llvm-svn: 305802
Target shuffle combining now supports the matching of INSERT_VECTOR_ELT/PINSRW/PINSRB for merging multiple insertions into shuffles/bitmasks.
llvm-svn: 305788
This seems to be interacting badly with ASan somehow, causing false reports of
heap-buffer overflows: PR33514.
> Summary:
> The patch makes instruction count the highest priority for
> LSR solution for X86 (previously registers had highest priority).
>
> Reviewers: qcolombet
>
> Differential Revision: http://reviews.llvm.org/D30562
>
> From: Evgeny Stupachenko <evstupac@gmail.com>
llvm-svn: 305720
Summary: Implement some of the simplest addressing modes.It should help to test ABI.
Reviewers: zvi, guyblank
Reviewed By: guyblank
Subscribers: rovka, llvm-commits, kristof.beyls
Differential Revision: https://reviews.llvm.org/D33888
llvm-svn: 305691
Use llvm::make_unique to avoid ambiguity with MSVC.
This patch adds a generic MacroFusion pass, that is used on X86 and
AArch64, which both define target-specific shouldScheduleAdjacent
functions. This generic pass should make it easier for other targets to
implement macro fusion and I intend to add macro fusion for ARM shortly.
Differential Revision: https://reviews.llvm.org/D34144
llvm-svn: 305690
Summary:
This patch adds a generic MacroFusion pass, that is used on X86 and
AArch64, which both define target-specific shouldScheduleAdjacent
functions. This generic pass should make it easier for other targets to
implement macro fusion and I intend to add macro fusion for ARM shortly.
Reviewers: craig.topper, evandro, t.p.northover, atrick, MatzeB
Reviewed By: MatzeB
Subscribers: atrick, aemerson, mgorny, javed.absar, kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D34144
llvm-svn: 305677
AVX512 compare instructions return v*i1 types.
In cases where the number of elements in the returned value are less than 8, clang adds zeroes to get a mask of v8i1 type.
Later on it's replaced with CONCAT_VECTORS, which then is lowered to many DAG nodes including insert/extract element and shift right/left nodes.
The fact that AVX512 compare instructions put the result in a k register and zeroes all its upper bits allows us to remove the extra nodes simply by copying the result to the required register class.
When lowering, identify these cases and transform them into an INSERT_SUBVECTOR node (marked legal), then catch this pattern in instructions selection phase and transform it into one avx512 cmp instruction.
Differential Revision: https://reviews.llvm.org/D33188
llvm-svn: 305465
We know that shuffle masks are power-of-2 sizes, but there's no way (?) for LLVM to know that,
so hack combineX86ShufflesRecursively() to be much faster by replacing div/rem with shift/mask.
This makes the motivating compile-time bug in PR32037 ( https://bugs.llvm.org/show_bug.cgi?id=32037 )
about 9% faster overall.
Differential Revision: https://reviews.llvm.org/D34174
llvm-svn: 305398
Much of PR32037's compile time regression is due to getTargetConstantBitsFromNode always creating large (>64bit) APInts during the bitcasting from the source data to the destination bitwidth.
This commit avoids this bitcast stage if the data is already the correct bitwidth.
llvm-svn: 305284
This step is just intended to reduce code duplication rather than change any functionality.
A follow-up would be to replace PPCTargetLowering::spliceIntoChain() usage with this new helper.
Differential Revision: https://reviews.llvm.org/D33649
llvm-svn: 305192
First possible step towards merging SSE/AVX memory folding pattern fragments.
Also allows us to remove the duplicate non-temporal load logic.
Differential Revision: https://reviews.llvm.org/D33902
llvm-svn: 305184
I was looking closer at the x86 test diffs in D33866, and the first change seems like it
shouldn't happen in the first place. So this patch will resolve that.
Using Agner's tables and AMD docs, vperm2f128 and vinsertf128 have identical timing for
any given CPU model, so we should be able to interchange those without affecting perf.
But as we can see in some of the diffs here, using vperm2f128 allows load folding, so
we should take that opportunity to reduce code size and register pressure.
A secondary advantage is making AVX1 and AVX2 codegen more similar. Given that vperm2f128
was introduced with AVX1, we should be selecting it in all of the same situations that we
would with AVX2. If there's some reason that an AVX1 CPU would not want to use this
instruction, that should be fixed up in a later pass.
Differential Revision: https://reviews.llvm.org/D33938
llvm-svn: 305171
If we know that both operands of an unsigned integer vector comparison are non-negative,
then it's safe to directly use a signed-compare-greater-than instruction (the only non-equality
integer vector compare predicate provided by SSE/AVX).
We're intentionally not changing the condition code to signed in order to preserve the
existing transforms that use min/max/psubus below here.
This should solve PR33276:
https://bugs.llvm.org/show_bug.cgi?id=33276
Differential Revision: https://reviews.llvm.org/D33862
llvm-svn: 304909
This creates a new library called BinaryFormat that has all of
the headers from llvm/Support containing structure and layout
definitions for various types of binary formats like dwarf, coff,
elf, etc as well as the code for identifying a file from its
magic.
Differential Revision: https://reviews.llvm.org/D33843
llvm-svn: 304864
Summary:
The patch makes instruction count the highest priority for
LSR solution for X86 (previously registers had highest priority).
Reviewers: qcolombet
Differential Revision: http://reviews.llvm.org/D30562
From: Evgeny Stupachenko <evstupac@gmail.com>
llvm-svn: 304824
Summary:
Expanding the loop idiom test for memcpy to also recognize
unordered atomic memcpy. The only difference for recognizing
an unordered atomic memcpy and instead of a normal memcpy is
that the loads and/or stores involved are unordered atomic operations.
Background: http://lists.llvm.org/pipermail/llvm-dev/2017-May/112779.html
Patch by Daniel Neilson!
Reviewers: reames, anna, skatkov
Reviewed By: reames, anna
Subscribers: llvm-commits, mzolotukhin
Differential Revision: https://reviews.llvm.org/D33243
llvm-svn: 304806
I did this a long time ago with a janky python script, but now
clang-format has built-in support for this. I fed clang-format every
line with a #include and let it re-sort things according to the precise
LLVM rules for include ordering baked into clang-format these days.
I've reverted a number of files where the results of sorting includes
isn't healthy. Either places where we have legacy code relying on
particular include ordering (where possible, I'll fix these separately)
or where we have particular formatting around #include lines that
I didn't want to disturb in this patch.
This patch is *entirely* mechanical. If you get merge conflicts or
anything, just ignore the changes in this patch and run clang-format
over your #include lines in the files.
Sorry for any noise here, but it is important to keep these things
stable. I was seeing an increasing number of patches with irrelevant
re-ordering of #include lines because clang-format was used. This patch
at least isolates that churn, makes it easy to skip when resolving
conflicts, and gets us to a clean baseline (again).
llvm-svn: 304787
In testing, we've found yet another miscompile caused by the new tables.
And this one is even less clear how to fix (we could teach it to fold
a 16-bit load instead of the 32-bit load it wants, or block folding
entirely).
Also, the approach to excluding instructions seems increasingly to not
scale well.
I have left a more detailed analysis on the review log for the original
patch (https://reviews.llvm.org/D32684) along with suggested path
forward. I will land an additional test case that I wrote which covers
the code that was miscompiling (folding into the output of `pextrw`) in
a subsequent commit to keep this a pure revert.
For each commit reverted here, I've restricted the revert to the
non-test code touching the x86 fold table emission until the last commit
where I did revert the test updates. This means the *new* test cases
added for `insertps` and `xchg` remain untouched (and continue to pass).
Reverted commits:
r304540: [X86] Don't fold into memory operands into insertps in the ...
r304347: [TableGen] Adapt more places to getValueAsString now ...
r304163: [X86] Don't fold away the memory operand of an xchg.
r304123: Don't capture a temporary std::string in a StringRef.
r304122: Resubmit "[X86] Adding new LLVM TableGen backend that ..."
Original commit was in r304088, and after a string of fixes was reverted
previously in r304121 to fix build bots, and then re-landed in r304122.
llvm-svn: 304762
We currently generate BUILD_VECTOR as a tree of UNPCKL shuffles of the same type:
e.g. for v4f32:
Step 1: unpcklps 0, 2 ==> X: <?, ?, 2, 0>
: unpcklps 1, 3 ==> Y: <?, ?, 3, 1>
Step 2: unpcklps X, Y ==> <3, 2, 1, 0>
The issue is because we are not placing sequential vector elements together early enough, we fail to recognise many combinable patterns - consecutive scalar loads, extractions etc.
Instead, this patch unpacks progressively larger sequential vector elements together:
e.g. for v4f32:
Step 1: unpcklps 0, 2 ==> X: <?, ?, 1, 0>
: unpcklps 1, 3 ==> Y: <?, ?, 3, 2>
Step 2: unpcklpd X, Y ==> <3, 2, 1, 0>
This does mean that we are creating UNPCKL shuffle of different value types, but the relevant combines that benefit from this are quite capable of handling the additional BITCASTs that are now included in the shuffle tree.
Differential Revision: https://reviews.llvm.org/D33864
llvm-svn: 304688
Since r288804, we try to lower build_vectors on AVX using broadcasts of
float/double. However, when we broadcast integer values that happen to
have a NaN float bitpattern, we lose the NaN payload, thereby changing
the integer value being broadcast.
This is caused by ConstantFP::get, to which we pass the splat i32 as
a float (by bitcasting it using bitsToFloat). ConstantFP::get takes
a double parameter, so we end up lossily converting a single-precision
NaN to double-precision.
Instead, avoid any kinds of conversions by directly building an APFloat
from the splatted APInt.
Note that this also fixes another piece of code (broadcast of
subvectors), that currently isn't susceptible to the same problem.
Also note that we could really just use APInt and ConstantInt
throughout: the constant pool type doesn't matter much. Still, for
consistency, use the appropriate type.
llvm-svn: 304590
This might give a few better opportunities to optimize these to memcpy
rather than loops - also a few minor cleanups (StringRef-izing,
templating (to avoid std::function indirection), etc).
The SmallVector::assign(iter, iter) could be improved with the use of
SFINAE, but the (iter, iter) ctor and append(iter, iter) need it to and
don't have it - so, workaround it for now rather than bothering with the
added complexity.
(also, as noted in the added FIXME, these assign ops could potentially
be optimized better at least for non-trivially-copyable types)
llvm-svn: 304566
Summary:
Add an early combine to match patterns such as:
(i16 bitcast (v16i1 x))
->
(i16 movmsk (v16i8 sext (v16i1 x)))
This combine needs to happen early enough before
type-legalization scalarizes the result of the setcc.
Reviewers: igorb, craig.topper, RKSimon
Subscribers: delena, llvm-commits
Differential Revision: https://reviews.llvm.org/D33311
llvm-svn: 304406
Summary:
This is a continuation of the work started in D29872 . Passing the carry down as a value rather than as a glue allows for further optimizations. Introducing setcccarry makes the use of addc/subc unecessary and we can start the removal process.
This patch only introduce the optimization strictly required to get the same level of optimization as was available before nothing more.
Reviewers: jyknight, nemanjai, mkuper, spatel, RKSimon, zvi, bkramer
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D33374
llvm-svn: 304404
Summary: LiveRangeShrink pass moves instruction right after the definition with the same BB if the instruction and its operands all have more than one use. This pass is inexpensive and guarantees optimal live-range within BB.
Reviewers: davidxl, wmi, hfinkel, MatzeB, andreadb
Reviewed By: MatzeB, andreadb
Subscribers: hiraditya, jyknight, sanjoy, skatkov, gberry, jholewinski, qcolombet, javed.absar, krytarowski, atrick, spatel, RKSimon, andreadb, MatzeB, mehdi_amini, mgorny, efriedma, davide, dberlin, llvm-commits
Differential Revision: https://reviews.llvm.org/D32563
llvm-svn: 304371
After transforming FP to ST registers:
- Do not add the ST register to the livein lists, they are reserved so
we do not need to track their liveness.
- Remove the FP registers from the livein lists, they don't have defs or
uses anymore and so are not live.
- (The setKillFlags() call is moved to an earlier place as it relies on
the FP registers still being present in the livein list.)
llvm-svn: 304342
This adds a callback to the LLVMTargetMachine that lets target indicate
that they do not pass the machine verifier checks in all cases yet.
This is intended to be a temporary measure while the targets are fixed
allowing us to enable the machine verifier by default with
EXPENSIVE_CHECKS enabled!
Differential Revision: https://reviews.llvm.org/D33696
llvm-svn: 304320
Craig Topper