llc would crash for (store (fptosi-f128-i32)) when -mcpu=pwr8, we should
not generate FP_TO_(S|U)INT_IN_VSR for f128 types at this time. This
patch fixes it.
Reviewed By: steven.zhang
Differential Revision: https://reviews.llvm.org/D86686
Writing the .note.gnu.property manually is error prone and hard to
maintain in the assembly files.
The -mmark-bti-property is for the assembler to emit the section with the
GNU_PROPERTY_AARCH64_FEATURE_1_BTI. To be used when C/C++ is compiled
with -mbranch-protection=bti.
This patch refactors the .note.gnu.property handling.
Reviewed By: chill, nickdesaulniers
Differential Revision: https://reviews.llvm.org/D81930
Reland with test dependency on aarch64 target.
Writing the .note.gnu.property manually is error prone and hard to
maintain in the assembly files.
The -mmark-bti-property is for the assembler to emit the section with the
GNU_PROPERTY_AARCH64_FEATURE_1_BTI. To be used when C/C++ is compiled
with -mbranch-protection=bti.
This patch refactors the .note.gnu.property handling.
Reviewed By: chill, nickdesaulniers
Differential Revision: https://reviews.llvm.org/D81930
eliminateFrameIndex won't fix up the offset register when the direct
frame index reference is moved to a separate move instruction. Switch
the offset to a base 0 (which it probably should be to begin with).
Additional sanity checks were added to get.active.lane.mask's second argument,
the loop tripcount/elementcount, in rG635b87511ec3. Like the other (overflow)
checks, skip this if tail-predication is forced.
Differential Revision: https://reviews.llvm.org/D87769
WeakRefDirective should specify a directive to declare "a global as being a weak undefined symbol".
The directive used by AMDGPU was incorrect - ".weakref" was intended for other purposes.
The correct directive is ".weak" and it is already defined as default for ELF.
So the redefinition was removed.
Reviewers: arsenm, rampitec
Differential Revision: https://reviews.llvm.org/D87762
Also renamed the fields to follow style guidelines.
Accessors help with readability - weight mutation, in particular,
is easier to follow this way.
Differential Revision: https://reviews.llvm.org/D87725
Fix lowering and instruction selection for v3x16 types
and enable InstCombine to emit them.
This patch only implements it for the selection dag.
GlobalISel tests in GlobalISel/llvm.amdgcn.image.load.1d.d16.ll and
GlobalISel/llvm.amdgcn.image.store.2d.d16.ll still don't work.
Differential Revision: https://reviews.llvm.org/D84420
Pre-gfx10 all MODE-setting instructions were S_SETREG_B32 which is
marked as having unmodeled side effects, which makes the machine
scheduler treat it as a barrier. Now that we have proper implicit $mode
operands we can use a no-side-effects S_SETREG_B32_mode pseudo instead
for setregs that only touch the FP MODE bits, to give the scheduler more
freedom.
Differential Revision: https://reviews.llvm.org/D87446
Now that we're getting better at combining shuffles of different vector widths, this can now be performed as part of the standard target shuffle combines and isn't required for cleanup.
Exposed a minor issue in combineX86ShufflesRecursively where we failed to check if a shuffle's src ops were simple types.
Clear the CurrentPredicate when we find an instruction which would
completely overwrite the VPR. This fix essentially means we're back
to not really being able to handle VPT instructions when tail
predicating.
Differential Revision: https://reviews.llvm.org/D87610
Modify the unit test to inspect all MVE instructions and mark the
load/store/move of vpr/p0 as valid, as well as the remaining scalar
shifts.
Differential Revision: https://reviews.llvm.org/D87753
These functions were extremely similar:
- `emitADD`
- `emitADDS`
- `emitCMN`
Refactor them a little, introducing a more generic `emitInstr` function to
do most of the work.
Also add support for the immediate + shifted register addressing modes in each
of them.
Update select-uaddo.mir to show that selecing ADDS now supports folding
immediates + shifts. (I don't think this can impact CMN, because the CMN checks
require a G_SUB with a non-constant on the RHS.)
This is around a 0.02% code size improvement on CTMark at -O3.
Differential Revision: https://reviews.llvm.org/D87529
We have a single noret intrinsic an a lot of special handling
around it. Declare it just as any other but do not define rtn
instructions itself instead.
Differential Revision: https://reviews.llvm.org/D87719
PR47534 exposes a case where calling lowerShuffleWithSHUFPS directly from a derived repeated mask (found by is128BitLaneRepeatedShuffleMask) results in us using an non-canonicalized mask.
The missed canonicalization in this case is trivial - just commute the mask so we have more (swapped) LHS than RHS references so lowerShuffleWithSHUFPS can handle it.
GetElementPtrInst::Create returns a GetElementPtrInst* so we don't need to cast. Similarly IntegerType inherits from the Type base class.
Also, I've used auto* in a few places to cleanup the code.
Helps fix some clang-tidy warnings which saw the dyn_casts and warned that these can return null.
Drop the pow2 vector limitation for AVG generation by padding the vector to the next pow2, creating the PAVG nodes and then extracting the final subvector.
Fixes some poor codegen that has been annoying me for years.....
The versions that take 'unsigned' will be removed in the future.
I tried to use getOriginalAlign instead of getAlign in some
places. getAlign factors in the minimum alignment implied by
the offset in the pointer info. Since we're also passing the
pointer info we can use the original alignment.
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D87592
This patch is the initial support for the Local Exec Thread Local
Storage model to produce code sequence and relocations correct
to the ABI for the model when using PC relative memory operations.
Patch by: Kamau Bridgeman
Differential Revision: https://reviews.llvm.org/D83404
This adds SoftenFloatRes, PromoteFloatRes and SoftPromoteHalfRes
legalizations for VECREDUCE, to fill the remaining hole in the SDAG
legalization. These legalizations simply expand the reduction and
let it be recursively legalized. For the PromoteFloatRes case at
least it is possible to do better than that, but it's pretty tricky
(because we need to consider the interaction of three different
vector legalizations and the type promotion) and probably not
really worthwhile.
I haven't added ExpandFloatRes support, as I am not familiar with
ppc_fp128.
Differential Revision: https://reviews.llvm.org/D87569
Summary:
In small code model, AIX assembler could not deal with labels that
could not be reached within the [-0x8000, 0x8000) range from TOC base.
So when generating the assembly, we would need to help the assembler
by subtracting an offset from the label to keep the actual value
within [-0x8000, 0x8000).
Reviewed By: hubert.reinterpretcast, Xiangling_L
Differential Revision: https://reviews.llvm.org/D86879
Clustering loads has caching benefits, but as far as I know there is no
advantage to clustering stores on any AMDGPU subtargets.
The disadvantage is that it tends to increase register pressure and
restricts scheduling freedom.
Differential Revision: https://reviews.llvm.org/D85530
LLVM will canonicalize conditional selectors to a different pattern than the old code that was used.
This is updating the function to match the new expected patterns and select SSAT or USAT when successful.
Tests have also been updated to use the new patterns.
Differential Review: https://reviews.llvm.org/D87379
This adds additional checks for the original scalar loop tripcount value, i.e.
get.active.lane.mask second argument, and perform several sanity checks to see
if it is of the form that we expect similarly like we already do for the IV
which is the first argument of get.active.lane.
Differential Revision: https://reviews.llvm.org/D86074
Check for NoNaNsFPMath function attribute in isKnownNeverSNaN.
Function attributes are in held in 'TargetMachine.Options'.
Among other things, this allows selection of some patterns imported
in D87351 since G_FCANONICALIZE is not generated when isKnownNeverSNaN
returns true in lowerFMinNumMaxNum.
However we notice some incorrect results since function attributes are
not correctly written in TargetMachine.Options when next function is
processed. Take a look at @v_test_no_global_nnans_med3_f32_pat0_srcmod0,
it has "no-nans-fp-math"="false" but TargetMachine.Options still has it
set to true since first function in test file had this attribute set to
true. This will be fixed in D87511.
Differential Revision: https://reviews.llvm.org/D87456
Predicates with 'let PredicateCodeUsesOperands = 1' want to examine
matched operands. When we encounter predicate code that uses operands,
analyze its named operand arguments and create a map between argument
index and name. Later, when leaf node with name is encountered, emit
GIM_RecordNamedOperand that will store that operand at its argument
index in operand list. This operand list will be an argument to c++
code of the predicate.
Differential Revision: https://reviews.llvm.org/D87285
Treating an SoImm offset as a multiple of 4 between -1020 and 1020
mis-handles the second of a pair of 16-bit constants where the offset is a multiple of 2 but not a multiple of 4,
leading to an LLVM ERROR: out of range pc-relative fixup value
For 32-bit and larger (64-bit) constants, continue to treat an SoImm offset as a multiple of 4 between -1020 and 1020.
For smaller (16-bit) constants, treat an SoImm offset as a multiple of 1 between -255 and 255.
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D86949
Clang emits (and (ctpop X), 1) for __builtin_parity. If ctpop
isn't natively supported by the target, this leads to poor codegen
due to the expansion of ctpop being more complex than what is needed
for parity.
This adds a DAG combine to convert the pattern to ISD::PARITY
before operation legalization. Type legalization is updated
to handled Expanding and Promoting this operation. If after type
legalization, CTPOP is supported for this type, LegalizeDAG will
turn it back into CTPOP+AND. Otherwise LegalizeDAG will emit a
series of shifts and xors followed by an AND with 1.
I've avoided vectors in this patch to avoid more legalization
complexity for this patch.
X86 previously had a custom DAG combiner for this. This is now
moved to Custom lowering for the new opcode. There is a minor
regression in vector-reduce-xor-bool.ll, but a follow up patch
can easily fix that.
Fixes PR47433
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D87209
This allows the backend to tell the vectorizer to produce inloop
reductions through a TTI hook.
For the moment on ARM under MVE this means allowing integer add
reductions of the correct size. In the future this can include integer
min/max too, under -Os.
Differential Revision: https://reviews.llvm.org/D75512
As detailed on PR11210, if the mask is known to come from a (sign extended) bool vector (e.g. comparisons) then we can represent with a generic masked load/store without losing anything.
We already do something similar for BLENDV -> SELECT conversion.
This fixes a complication on top of D87276. If we are sign extending
around a mul with the two operands that are the same, instcombine will
helpfully convert one of the sext to a zext. Reverse that so that we
again generate a reduction.
Differnetial Revision: https://reviews.llvm.org/D87287
As discussed on llvm-dev:
http://lists.llvm.org/pipermail/llvm-dev/2020-April/140729.html
This is hopefully the final remaining showstopper before we can remove
the 'experimental' from the reduction intrinsics.
No behavior was specified for the FP min/max reductions, so we have a
mess of different interpretations.
There are a few potential options for the semantics of these max/min ops.
I think this is the simplest based on current behavior/implementation:
make the reductions inherit from the existing llvm.maxnum/minnum intrinsics.
These correspond to libm fmax/fmin, and those are similar to the (now
deprecated?) IEEE-754 maxNum/minNum functions (NaNs are treated as missing
data). So the default expansion creates calls to libm functions.
Another option would be to inherit from llvm.maximum/minimum (NaNs propagate),
but most targets just crash in codegen when given those nodes because no
default expansion was ever implemented AFAICT.
We could also just assume 'nnan' semantics by default (we are already
assuming 'nsz' semantics in the maxnum/minnum intrinsics), but some targets
(AArch64, PowerPC) support the more defined behavior, so it doesn't make much
sense to not allow a tighter spec. Fast-math-flags (nnan) can be used to
loosen the semantics.
(Note that D67507 was proposed to update the LangRef to acknowledge the more
recent IEEE-754 2019 standard, but that patch seems to have stalled. If we do
update based on the new standard, the reduction instructions can seamlessly
inherit from whatever updates are made to the max/min intrinsics.)
x86 sees a regression here on 'nnan' tests because we have underlying,
longstanding bugs in FMF creation/propagation. Those need to be fixed apart
from this change (for example: https://llvm.org/PR35538). The expansion
sequence before this patch may not have been correct.
Differential Revision: https://reviews.llvm.org/D87391
We can sometimes get code that does:
xe = zext i16 x to i32
ye = zext i16 y to i32
m = mul i32 xe, ye
me = zext i32 m to i64
r = vecreduce.add(me)
This "double extend" can trip up the reduction identification, but
should give identical results.
This extends the pattern matching to handle them.
Differential Revision: https://reviews.llvm.org/D87276
Follow up to D86429 to handle the remaining regressions.
This patch generalizes lowerShuffleAsDecomposedShuffleBlend to lowerShuffleAsDecomposedShuffleMerge, and attempts to use an UNPCKL shuffle mask instead of a blend for the cases where the inputs are coming from alternating vXi8/vXi16 sources. Technically they don't have to be alternating (just as long as they can fit into a lower lane half for the unpack) but I didn't find as many general cases and it needed a lot more of the function to be altered.
For vXi32/vXi64 cases this could still be beneficial but in most cases the existing permute+blend approach was better.
Differential Revision: https://reviews.llvm.org/D87405
with P9 Model
Enable the pre-ra and post-ra scheduler strategy for Power10 as we want
to customize the heuristic later. And switch the scheduler model with P9
model before P10 Model is available. The NoSchedModel is modelled as
in-order cpu and the pre-ra scheduler is not bi-directional which will
have big impact on the scheduler.
Reviewed By: jji
Differential Revision: https://reviews.llvm.org/D86865
From ISA, fcmpu will raise the Floating-Point Invalid Operation
Exception (SNaN) if either of the operands is a Signaling NaN by setting
the bit VXSNAN. But the instruction description didn't set the
mayRaiseFPException which might have impact on the scheduling or some
backend optimization.
Reviewed By: qiucf
Differential Revision: https://reviews.llvm.org/D83937
This adds and optional ", immutable" to the end of a `.globaltype`
declaration. I would have prefered to match the `.wat` syntax
where immutable is the default and `mut` is the signifier for
mutable globals. Sadly changing the default would break backwards
compat with existing assembly in the wild so I think its best
to stick with this approach.
Differential Revision: https://reviews.llvm.org/D87515
Convert 2-byte opcodes to equivalent 1-byte ones.
Adjust the existing exhaustive testcase to avoid being altered by
the simplification rules (to keep that test exercising all individual
opcodes).
Fix the assembler parser limits for register pairs; for .seh_save_regp
and .seh_save_regp_x, we can allow up to x29, for a x29+x30 pair
(which gets remapped to the UOP_SaveFPLR(X) opcodes), for .seh_save_fregp
and .seh_save_fregpx, allow up to d14+d15.
Not creating .seh_save_next for float register pairs, as the
actual unwinder implementation in current versions of Windows is buggy
for that case.
This gives a minimal but measurable size reduction. (For a 6.5 MB
DLL with 300 KB .xdata, the .xdata shrinks by 48 bytes. The opcode
sequences are padded to a 4 byte boundary, so very small improvements
might not end up mattering directly.)
Differential Revision: https://reviews.llvm.org/D87367
With optimizations we leave the decision to eliminate fallthrough branches to
bock placement, but at -O0 we should do it in the selector to save code size.
This regressed -O0 with a recent change to a combiner.
Currently, using llvm-objdump to disassemble a function containing
unreachable will trigger an assertion while decoding the opcode, since both
unreachable and debug_unreachable have the same encoding. To avoid this, set
unreachable as the canonical decoding.
Differential Revision: https://reviews.llvm.org/D87431
This is the first in a series of patches to make implicit null checks
more general. This patch identifies instructions that preserves zero
value of a register and considers that as a valid instruction to hoist
along with the faulting load. See added testcases.
Reviewed-By: reames, dantrushin
Differential Revision: https://reviews.llvm.org/D87108
This adds the initial GlobalISel skeleton for PowerPC. It can only run
ir-translator and legalizer for `ret void`.
This is largely based on the initial GlobalISel patch for RISCV
(https://reviews.llvm.org/D65219).
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D83100
This prevents us from doing things like LICM'ing it out of a loop,
which is usually a net loss because we end up having to spill a
callee-saved FPR to accomodate it.
This does perturb instruction scheduling around this instruction,
so a number of tests had to be updated to account for it.
Reviewed By: t.p.northover
Differential Revision: https://reviews.llvm.org/D87316
Other types can be handled in future patches but their uniform / non-uniform costs are more similar and don't appear to cause many vectorization issues.
lowerShuffleAsSplitOrBlend always returns a target shuffle result (and is the default operation for lowering some shuffle types), so we don't need to check for null.
Truncating from an illegal SVE type to a legal type, e.g.
`trunc <vscale x 4 x i64> %in to <vscale x 4 x i32>`
fails after PromoteIntOp_CONCAT_VECTORS attempts to
create a BUILD_VECTOR.
This patch changes the promote function to create a sequence of
INSERT_SUBVECTORs if the return type is scalable, and replaces
these with UNPK+UZP1 for AArch64.
Reviewed By: paulwalker-arm
Differential Revision: https://reviews.llvm.org/D86548
values
The effects of unpredicated vector instruction with unknown
lanes cannot be predicted and therefore cannot be tail predicated. This
does not apply to predicated vector instructions and so this patch
allows tail predication on them.
Differential Revision: https://reviews.llvm.org/D87376
22a0edd0 introduced a config IsStrictFPEnabled, which controls the
strict floating point mutation (transforming some strict-fp operations
into non-strict in ISel). This patch disables the mutation by default
since we've finished PowerPC strict-fp enablement in backend.
Reviewed By: uweigand
Differential Revision: https://reviews.llvm.org/D87222
It was found some packed immediate operands (e.g. `<half 1.0, half 2.0>`) are
incorrectly processed so one of two packed values were lost.
Introduced new function to check immediate 32-bit operand can be folded.
Converted condition about current op_sel flags value to fall-through.
Fixes: SWDEV-247595
Reviewed By: rampitec
Differential Revision: https://reviews.llvm.org/D87158
We were missing support for the G_ADD_LOW + ADRP folding optimization in the
manual selection code for G_LOAD, G_STORE, and G_ZEXTLOAD.
As a result, we were missing cases like this:
```
@foo = external hidden global i32*
define void @baz(i32* %0) {
store i32* %0, i32** @foo
ret void
}
```
https://godbolt.org/z/16r7ad
This functionality already existed in the addressing mode functions for the
importer. So, this patch makes the manual selection code use
`selectAddrModeIndexed` rather than duplicating work.
This is a 0.2% geomean code size improvement for CTMark at -O3.
There is one code size increase (0.1% on lencod) which is likely because
`selectAddrModeIndexed` doesn't look through constants.
Differential Revision: https://reviews.llvm.org/D87397
We weren't using this before, so none of the MachineFunction CFG edges had the
branch probability information added. As a result, block placement later in the
pipeline was flying blind.
This is enabled only with optimizations enabled like SelectionDAG.
Differential Revision: https://reviews.llvm.org/D86824
This combine previously tried to take sequences like:
%cond = G_ICMP pred, a, b
G_BRCOND %cond, %truebb
G_BR %falsebb
%truebb:
...
%falsebb:
...
and by inverting the compare predicate and swapping branch targets, delete the
G_BR and instead have a single conditional branch to the falsebb. Since in an
earlier patch we have a combine to fold not(icmp) into just an inverted icmp,
we don't need this combine to do as much. This patch instead generalizes the
combine by just looking for:
G_BRCOND %cond, %truebb
G_BR %falsebb
%truebb:
...
%falsebb:
...
and then inverting the condition using a not (xor). The xor can be folded away
in a separate combine. This change also lets us avoid some optimization code
in the IRTranslator.
I also think that deleting G_BRs in the combiner is unnecessary. That's
something that targets can decide to do at selection time and could simplify
generic code in future.
Differential Revision: https://reviews.llvm.org/D86664
In standard C library, both rint and nearbyint returns rounding result
in current rounding mode. But nearbyint never raises inexact exception.
On PowerPC, x(v|s)r(d|s)pic may modify FPSCR XX, raising inexact
exception. So we can't select constrained fnearbyint into xvrdpic.
One exception here is xsrqpi, which will not raise inexact exception, so
fnearbyint f128 is okay here.
Reviewed By: uweigand
Differential Revision: https://reviews.llvm.org/D87220
Add subtarget feature check to avoid using ds_read/write_b96/128 with too
low alignment if a bug is present on that specific hardware.
Add this "feature" to GFX 10.1.1 as it is also affected.
Add global-isel test.
We really want to try and avoid spilling P0, which can be difficult
since there's only one register, so try to rematerialize any VCTP
instructions.
Differential Revision: https://reviews.llvm.org/D87280
This removes the after the fact FMF handling from D46854 in favor of passing fast math flags to getNode. This should be a superset of D87130.
This required adding a SDNodeFlags to SelectionDAG::getSetCC.
Now we manage to contant fold some stuff undefs during the
initial getNode that we don't do in later DAG combines.
Differential Revision: https://reviews.llvm.org/D87200
Failing example: v8i8 = truncate v8i32. v8i8 is legal, but v8i32 was
widened to HVX. Make sure that v8i8 does not get altered (even if it's
changed to another legal type).
When the function return type is non-void and `end` instructions are at
the very end of a function, CFGStackify's `fixEndsAtEndOfFunction`
function fixes the corresponding block/loop/try's type to match the
function's return type. This is applied to consecutive `end` markers at
the end of a function. For example, when the function return type is
`i32`,
```
block i32 ;; return type is fixed to i32
...
loop i32 ;; return type is fixed to i32
...
end_loop
end_block
end_function
```
But try-catch is a little different, because it consists of two parts:
a try part and a catch part, and both parts' return type should satisfy
the function's return type. Which means,
```
try i32 ;; return type is fixed to i32
...
block i32 ;; this should be changed i32 too!
...
end_block
catch
...
end_try
end_function
```
As you can see in this example, it is not sufficient to only `end`
instructions at the end of a function; in case of `try`, we should
check instructions before `catch`es, in case their corresponding `try`'s
type has been fixed.
This changes `fixEndsAtEndOfFunction`'s algorithm to use a worklist
that contains a reverse iterator, each of which is a starting point for
a new backward `end` instruction search.
Fixes https://bugs.llvm.org/show_bug.cgi?id=47413.
Reviewed By: dschuff, tlively
Differential Revision: https://reviews.llvm.org/D87207
count register
After my patch at D86087, code that now uses the mov operand rather than
the vctp operand will no longer remove modifications to the vctp operand
as they should. This patch fixes that by explicitly removing
modifications to the vctp operand rather than the register used as the
element count.
Commit 3c0b3250 introduced memory cluster under pwr10 target, but a
check for operands was unexpectedly removed. This adds it back to avoid
regression.
Implement AArch64 variant of shouldCoalesce() to detect a known failing case
and prevent the coalescing of a 32-bit copy into a 64-bit sign-extending load.
Do not coalesce in the following case:
COPY where source is bottom 32 bits of a 64-register,
and destination is a 32-bit subregister of a 64-bit register,
ie it causes the rest of the register to be implicitly set to zero.
A mir test has been added.
In the test case, the 32-bit copy implements a 32 to 64 bit zero extension
and relies on the upper 32 bits being zeroed.
Coalescing to the result of the 64-bit load meant overwriting
the upper 32 bits incorrectly when the loaded byte was negative.
Reviewed By: john.brawn
Differential Revision: https://reviews.llvm.org/D85956
Without gcc 7.4 warns with
../lib/Target/PowerPC/PPCInstrInfo.cpp:2284:25: warning: suggest parentheses around '&&' within '||' [-Wparentheses]
BaseOp1.isFI() &&
~~~~~~~~~~~~~~~^~
"Only base registers and frame indices are supported.");
~
On Power10, it's profitable to schedule some stores with adjacent target
address together. This patch implements this feature.
Reviewed By: steven.zhang
Differential Revision: https://reviews.llvm.org/D86754
This was reverted in 503deec218
because it caused gigantic increase (3x) in branch mispredictions
in certain benchmarks on certain CPU's,
see https://reviews.llvm.org/D84108#2227365.
It has since been investigated and here are the results:
https://lists.llvm.org/pipermail/llvm-commits/Week-of-Mon-20200907/827578.html
> It's an amazingly severe regression, but it's also all due to branch
> mispredicts (about 3x without this). The code layout looks ok so there's
> probably something else to deal with. I'm not sure there's anything we can
> reasonably do so we'll just have to take the hit for now and wait for
> another code reorganization to make the branch predictor a bit more happy :)
>
> Thanks for giving us some time to investigate and feel free to recommit
> whenever you'd like.
>
> -eric
So let's just reland this.
Original commit message:
I've been looking at missed vectorizations in one codebase.
One particular thing that stands out is that some of the loops
reach vectorizer in a rather mangled form, with weird PHI's,
and some of the loops aren't even in a rotated form.
After taking a more detailed look, that happened because
the loop's headers were too big by then. It is evident that
SimplifyCFG's common code hoisting transform is at fault there,
because the pattern it handles is precisely the unrotated
loop basic block structure.
Surprizingly, `SimplifyCFGOpt::HoistThenElseCodeToIf()` is enabled
by default, and is always run, unlike it's friend, common code sinking
transform, `SinkCommonCodeFromPredecessors()`, which is not enabled
by default and is only run once very late in the pipeline.
I'm proposing to harmonize this, and disable common code hoisting
until //late// in pipeline. Definition of //late// may vary,
here currently i've picked the same one as for code sinking,
but i suppose we could enable it as soon as right after
loop rotation happens.
Experimentation shows that this does indeed unsurprizingly help,
more loops got rotated, although other issues remain elsewhere.
Now, this undoubtedly seriously shakes phase ordering.
This will undoubtedly be a mixed bag in terms of both compile- and
run- time performance, codesize. Since we no longer aggressively
hoist+deduplicate common code, we don't pay the price of said hoisting
(which wasn't big). That may allow more loops to be rotated,
so we pay that price. That, in turn, that may enable all the transforms
that require canonical (rotated) loop form, including but not limited to
vectorization, so we pay that too. And in general, no deduplication means
more [duplicate] instructions going through the optimizations. But there's still
late hoisting, some of them will be caught late.
As per benchmarks i've run {F12360204}, this is mostly within the noise,
there are some small improvements, some small regressions.
One big regression i saw i fixed in rG8d487668d09fb0e4e54f36207f07c1480ffabbfd, but i'm sure
this will expose many more pre-existing missed optimizations, as usual :S
llvm-compile-time-tracker.com thoughts on this:
http://llvm-compile-time-tracker.com/compare.php?from=e40315d2b4ed1e38962a8f33ff151693ed4ada63&to=c8289c0ecbf235da9fb0e3bc052e3c0d6bff5cf9&stat=instructions
* this does regress compile-time by +0.5% geomean (unsurprizingly)
* size impact varies; for ThinLTO it's actually an improvement
The largest fallout appears to be in GVN's load partial redundancy
elimination, it spends *much* more time in
`MemoryDependenceResults::getNonLocalPointerDependency()`.
Non-local `MemoryDependenceResults` is widely-known to be, uh, costly.
There does not appear to be a proper solution to this issue,
other than silencing the compile-time performance regression
by tuning cut-off thresholds in `MemoryDependenceResults`,
at the cost of potentially regressing run-time performance.
D84609 attempts to move in that direction, but the path is unclear
and is going to take some time.
If we look at stats before/after diffs, some excerpts:
* RawSpeed (the target) {F12360200}
* -14 (-73.68%) loops not rotated due to the header size (yay)
* -272 (-0.67%) `"Number of live out of a loop variables"` - good for vectorizer
* -3937 (-64.19%) common instructions hoisted
* +561 (+0.06%) x86 asm instructions
* -2 basic blocks
* +2418 (+0.11%) IR instructions
* vanilla test-suite + RawSpeed + darktable {F12360201}
* -36396 (-65.29%) common instructions hoisted
* +1676 (+0.02%) x86 asm instructions
* +662 (+0.06%) basic blocks
* +4395 (+0.04%) IR instructions
It is likely to be sub-optimal for when optimizing for code size,
so one might want to change tune pipeline by enabling sinking/hoisting
when optimizing for size.
Reviewed By: mkazantsev
Differential Revision: https://reviews.llvm.org/D84108
This reverts commit 503deec218.
Rather than using SELECT instructions, use SRA, UADDO/ADDCARRY and
XORs to expand ABS. This is the multi-part version of the sequence
we use in LegalizeDAG.
It's also the same as the Custom sequence uses for i64 on 32-bit
and i128 on 64-bit. So we can remove the X86 customization.
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D87215
optimizeEndCF removes EXEC restoring instruction case this instruction is the only one except the branch to the single successor and that successor contains EXEC mask restoring instruction that was lowered from END_CF belonging to IF_ELSE.
As a result of such optimization we get the basic block with the only one instruction that is a branch to the single successor.
In case the control flow can reach such an empty block from S_CBRANCH_EXEZ/EXECNZ it might happen that spill/reload instructions that were inserted later by register allocator are placed under exec == 0 condition and never execute.
Removing empty block solves the problem.
This change require further work to re-implement LIS updates. Recently, LIS is always nullptr in this pass. To enable it we need another patch to fix many places across the codegen.
Reviewed By: rampitec
Differential Revision: https://reviews.llvm.org/D86634
We already simplify the unsigned comparisons if we've found the operands are non-negative, but we were still calling LowerVSETCCWithSUBUS which resulted in the PR47448 regressions.
lowerShuffleWithPERMV allows us to use the ZMM variants for 128/256-bit variable shuffles on non-VLX AVX512 targets.
This is another step towards shuffle combining through between vector widths - we still end up with an annoying regression (combine_vpermilvar_vperm2f128_zero_8f32) but we're going in the right direction....
When optimising for size, make the cost of i1 logical operations
relatively expensive so that optimisations don't try to combine
predicates.
Differential Revision: https://reviews.llvm.org/D86525
Fixes PR47375, in which an assertion was triggering because
WebAssemblyTargetLowering::isVectorLoadExtDesirable was improperly
assuming the use of simple value types.
Differential Revision: https://reviews.llvm.org/D87110
This patch implements the vec_expandm function prototypes in altivec.h in order
to utilize the vector expand with mask instructions introduced in Power10.
Differential Revision: https://reviews.llvm.org/D82727
rGabd33bf5eff2 enabled us to pad 128/256-bit shuffles to 512-bit on non-VLX targets, but wasn't updating binary shuffles to account for the new vector width.
During the PEI pass, the dead TargetStackID::SGPRSpill spill slots
are not being removed while spilling the FP/BP to memory.
Fixes: SWDEV-250393
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D87032
Qiu Chaofan