Kernel arguments likely read by all workitems and should not bypass
cache. Fixes performance hit in sub-dword argument loads.
Differential Revision: https://reviews.llvm.org/D43249
llvm-svn: 325146
When creating high MachineMemOperand for MSTORE/MLOAD we supply
it with the original PointerInfo, while the pointer itself had been incremented.
The patch adds the proper offset to the PointerInfo.
llvm-svn: 325135
If a load follows a store and reloads data that the store has written to memory, Intel microarchitectures can in many cases forward the data directly from the store to the load, This "store forwarding" saves cycles by enabling the load to directly obtain the data instead of accessing the data from cache or memory.
A "store forward block" occurs in cases that a store cannot be forwarded to the load. The most typical case of store forward block on Intel Core microarchiticutre that a small store cannot be forwarded to a large load.
The estimated penalty for a store forward block is ~13 cycles.
This pass tries to recognize and handle cases where "store forward block" is created by the compiler when lowering memcpy calls to a sequence
of a load and a store.
The pass currently only handles cases where memcpy is lowered to XMM/YMM registers, it tries to break the memcpy into smaller copies.
breaking the memcpy should be possible since there is no atomicity guarantee for loads and stores to XMM/YMM.
Change-Id: Ic41aa9ade6512e0478db66e07e2fde41b4fb35f9
llvm-svn: 325128
While the AVX512 VTRUNCS/VTRUNCUS instructions require legal types, truncateVectorWithPACK handles cases with multiples of legal types through splitting/concatenation. So we just need to ensure that the src/dst scalar types are correct and leave truncateVectorWithPACK to handle the rest of it.
llvm-svn: 325127
Summary:
Instead of solving the hard problem of how to pass the callee to the indirect
jump thunk without a register, just use a CSR. At a call boundary, there's
nothing stopping us from using a CSR to hold the callee as long as we save and
restore it in the prologue.
Also, add tests for this mregparm=3 case. I wrote execution tests for
__llvm_retpoline_push, but they never got committed as lit tests, either
because I never rewrote them or because they got lost in merge conflicts.
Reviewers: chandlerc, dwmw2
Subscribers: javed.absar, kristof.beyls, hiraditya, llvm-commits
Differential Revision: https://reviews.llvm.org/D43214
llvm-svn: 325049
Summary:
If the and has an additional use we shouldn't invert it. That creates an additional instruction.
While there add a one use check to the transform above that looked similar.
Reviewers: spatel, RKSimon
Reviewed By: RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D43225
llvm-svn: 325019
Change ARMConstantIslandPass to:
- accept f16 literals as litpool entries,
- if the litpool needs to be inserted in the middle of a big block, then we
need to 4-byte align the next instruction in ARM mode.
Differential Revision: https://reviews.llvm.org/D42784
llvm-svn: 325012
The bug has been lying dormant, but apparently was never exposed, until
after rL324941 because we didn't return the correct result
for shifts with undef operands.
llvm-svn: 325010
This started by noticing that scalar and vector types were producing different results with div ops in PR36305:
https://bugs.llvm.org/show_bug.cgi?id=36305
...but the problem is bigger. I couldn't keep it straight without a table, so I'm attaching that as a PDF to
the review. The x86 tests in undef-ops.ll correspond to that table.
Green means that instsimplify and the DAG agree on the result for all types.
Red means the DAG was returning undef when IR was not.
Yellow means the DAG was returning a non-undef result when IR returned undef.
This patch assumes that we're currently doing the right thing in IR.
Note: I couldn't find any problems with lowering vector constants as the code comments were warning,
but those comments were written long ago in rL36413 .
Differential Revision: https://reviews.llvm.org/D43141
llvm-svn: 324941
If merging them, the dllexport attribute needs to be brought along
to the new GlobalAlias.
Differential Revision: https://reviews.llvm.org/D43192
llvm-svn: 324937
Armv8.1-A added an atomic load-clear instruction (which performs bitwise
and with the complement of it's operand), but not a load-and
instruction. Our current code-generation for atomic load-and always
inserts an MVN instruction to invert its argument, even if it could be
folded into a constant or another instruction.
This adds lowering early in selection DAG to convert a load-and
operation into an xor with -1 and a load-clear, allowing the normal DAG
optimisations to work on it.
To do this, I've had to add a new ISD opcode, ATOMIC_LOAD_CLR. I don't
see any easy way to do this with an AArch64-specific ISD node, because
the code-generation for atomic operations assumes the SDNodes are of
type AtomicSDNode.
I've left the old tablegen patterns in because they are still needed for
global isel.
Differential revision: https://reviews.llvm.org/D42478
llvm-svn: 324908
We only tagged it with the itinerary class, so completeness checks were erroneously passed (PR35639).
AMD targets can perform these a lot quicker than WriteMicrocoded so will need an override in the models.
llvm-svn: 324897
Armv8.1-A added an atomic load-add instruction, but not a load-subtract
instruction. Our current code-generation for atomic load-subtract always
inserts a NEG instruction to negate it's argument, even if it could be
folded into a constant or another instruction.
This adds lowering early in selection DAG to convert a load-subtract
operation into a subtract and a load-add, allowing the normal DAG
optimisations to work on it.
I've left the old tablegen patterns in because they are still needed for
global isel.
Some of the tests in this patch are copied from D35375 by Chad Rosier (which
was abandoned).
Differential revision: https://reviews.llvm.org/D42477
llvm-svn: 324892
It asserts building Chromium; see PR36346.
(This also reverts the follow-up r324836.)
> If a load follows a store and reloads data that the store has written to memory, Intel microarchitectures can in many cases forward the data directly from the store to the load, This "store forwarding" saves cycles by enabling the load to directly obtain the data instead of accessing the data from cache or memory.
> A "store forward block" occurs in cases that a store cannot be forwarded to the load. The most typical case of store forward block on Intel Core microarchiticutre that a small store cannot be forwarded to a large load.
> The estimated penalty for a store forward block is ~13 cycles.
>
> This pass tries to recognize and handle cases where "store forward block" is created by the compiler when lowering memcpy calls to a sequence
> of a load and a store.
>
> The pass currently only handles cases where memcpy is lowered to XMM/YMM registers, it tries to break the memcpy into smaller copies.
> breaking the memcpy should be possible since there is no atomicity guarantee for loads and stores to XMM/YMM.
llvm-svn: 324887
In case of correct using of the 'l' constraint llvm now generates valid
code; otherwise it shows an error message. Initially these triggers an
assertion.
This commit is the same as r324869 with fixed the test's file name.
llvm-svn: 324885
Add a common -trap-unreachable option, similar to the target
specific hexagon equivalent, which has been replaced. This
turns unreachable instructions into traps, which is useful for
debugging.
Differential Revision: https://reviews.llvm.org/D42965
llvm-svn: 324880
In case of correct using of the 'l' constraint llvm now generates valid
code; otherwise it shows an error message. Initially these triggers an
assertion.
llvm-svn: 324869
Previously we just emitted this as a MOV8rm which would likely get folded during the peephole pass anyway. This just makes it explicit earlier.
The gpr-to-mask.ll test changed because the kaddb instruction has no memory form.
llvm-svn: 324860
Summary:
Currently we only use min/max to help with ule/uge compares because it removes an invert of the result that would otherwise be needed. But we can also use it for ult/ugt compares if it will prevent the need for a sign bit flip needed to use pcmpgt at the cost of requiring an invert after the compare.
I also refactored the code so that the max/min code is self contained and does its own return instead of setting up a flag to manipulate the rest of the function's behavior.
Most of the test cases look ok with this. I did notice that we added instructions when one of the operands being sign flipped is a constant vector that we were able to constant fold the flip into.
I also noticed that sometimes the SSE min/max clobbers a register that is needed after the compare. This resulted in an extra move being inserted before the min/max to preserve the register. We could try to detect this and switch from min to max and change the compare operands to use the operand that gets reused in the compare.
Reviewers: spatel, RKSimon
Reviewed By: RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D42935
llvm-svn: 324842
This reverses instcombine's demanded bits' transform which always tries to clear bits in constants.
As noted in PR35792 and shown in the test diffs:
https://bugs.llvm.org/show_bug.cgi?id=35792
...we can do better in codegen by trying to form -1. The x86 sub test shows a missed opportunity.
I did investigate changing instcombine's behavior, but it would be more work to change
canonicalization in IR. Clearing bits / shrinking constants can allow killing instructions,
so we'd have to figure out how to not regress those cases.
Differential Revision: https://reviews.llvm.org/D42986
llvm-svn: 324839
This allows us to recognise more saturation patterns and also simplify some MINMAX codegen that was failing to combine CMPGE comparisons to a legal CMPGT.
Differential Revision: https://reviews.llvm.org/D43014
llvm-svn: 324837
If a load follows a store and reloads data that the store has written to memory, Intel microarchitectures can in many cases forward the data directly from the store to the load, This "store forwarding" saves cycles by enabling the load to directly obtain the data instead of accessing the data from cache or memory.
A "store forward block" occurs in cases that a store cannot be forwarded to the load. The most typical case of store forward block on Intel Core microarchiticutre that a small store cannot be forwarded to a large load.
The estimated penalty for a store forward block is ~13 cycles.
This pass tries to recognize and handle cases where "store forward block" is created by the compiler when lowering memcpy calls to a sequence
of a load and a store.
The pass currently only handles cases where memcpy is lowered to XMM/YMM registers, it tries to break the memcpy into smaller copies.
breaking the memcpy should be possible since there is no atomicity guarantee for loads and stores to XMM/YMM.
Change-Id: I620b6dc91583ad9a1444591e3ddc00dd25d81748
llvm-svn: 324835
This patch adds a new function attribute "required-vector-width" that can be set by the frontend to indicate the maximum vector width present in the original source code. The idea is that this would be set based on ABI requirements, intrinsics or explicit vector types being used, maybe simd pragmas, etc. The backend will then use this information to determine if its save to make 512-bit vectors illegal when the preference is for 256-bit vectors.
For code that has no vectors in it originally and only get vectors through the loop and slp vectorizers this allows us to generate code largely similar to our AVX2 only output while still enabling AVX512 features like mask registers and gather/scatter. The loop vectorizer doesn't always obey TTI and will create oversized vectors with the expectation the backend will legalize it. In order to avoid changing the vectorizer and potentially harm our AVX2 codegen this patch tries to make the legalizer behavior similar.
This is restricted to CPUs that support AVX512F and AVX512VL so that we have good fallback options to use 128 and 256-bit vectors and still get masking.
I've qualified every place I could find in X86ISelLowering.cpp and added tests cases for many of them with 2 different values for the attribute to see the codegen differences.
We still need to do frontend work for the attribute and teach the inliner how to merge it, etc. But this gets the codegen layer ready for it.
Differential Revision: https://reviews.llvm.org/D42724
llvm-svn: 324834
