This reverts commit bb8850d34d601d4edd75fd30c07821c05a726c42.
It broke 3 check-llvm-transforms-loopfusion tests in an ASAN build.
LoopFuse.cpp `for (BasicBlock *Pred : predecessors(BB)) {` may operate on a deleted BB.
Summary:
This patch adds the ability to peel off iterations of the first loop in loop
fusion. This can allow for both loops to have the same trip count, making it
legal for them to be fused together.
Here is a simple scenario peeling can be used in loop fusion:
for (i = 0; i < 10; ++i)
a[i] = a[i] + 3;
for (j = 1; j < 10; ++j)
b[j] = b[j] + 5;
Here is we can make use of peeling, and then fuse the two loops together. We can
peel off the 0th iteration of the loop i, and then combine loop i and j for
i = 1 to 10.
a[0] = a[0] +3;
for (i = 1; i < 10; ++i) {
a[i] = a[i] + 3;
b[i] = b[i] + 5;
}
Currently peeling with loop fusion is only supported for loops with constant
trip counts and a single exit point. Both unguarded and guarded loops are
supported.
Author: sidbav (Sidharth Baveja)
Reviewers: kbarton, Meinersbur, bkramer, Whitney, skatkov, ashlykov, fhahn, bmahjour
Reviewed By: bmahjour
Subscribers: bmahjour, mgorny, hiraditya, zzheng
Tags: LLVM
Differential Revision: https://reviews.llvm.org/D82927
If we inferred a range for the function return value, we can add !range
at all call-sites of the function, if the range does not include undef.
Reviewers: efriedma, davide, nikic
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D83952
This patch uses the TileInfo introduced in D77550 to generate a loop
nest for tiled matrix multiplication, instead of generating the
unrolled code for the whole multiplication. This makes code-generation
more scalable for larger matrixes.
Initially loops are only used if both the number of rows and columns are
divisible by the tile size. Other cases will be added as follow-up.
Reviewers: anemet, Gerolf, hfinkel, andrew.w.kaylor, LuoYuanke, nicolasvasilache
Reviewed By: anemet
Differential Revision: https://reviews.llvm.org/D81308
Remove the temporary flag PGSOIRPassOrTestOnly and the guard code which was used
for the staged rollout. This is a cleanup (NFC) as it's now false by default.
Differential Revision: https://reviews.llvm.org/D84057
This patch adds a TileInfo abstraction and utilities to
create a 3-level loop nest for tiling.
Reviewers: anemet
Reviewed By: anemet
Differential Revision: https://reviews.llvm.org/D77550
This allows tracking the in-memory type of a pointer argument to a
function for ABI purposes. This is essentially a stripped down version
of byval to remove some of the stack-copy implications in its
definition.
This includes the base IR changes, and some tests for places where it
should be treated similarly to byval. Codegen support will be in a
future patch.
My original attempt at solving some of these problems was to repurpose
byval with a different address space from the stack. However, it is
technically permitted for the callee to introduce a write to the
argument, although nothing does this in reality. There is also talk of
removing and replacing the byval attribute, so a new attribute would
need to take its place anyway.
This is intended avoid some optimization issues with the current
handling of aggregate arguments, as well as fixes inflexibilty in how
frontends can specify the kernel ABI. The most honest representation
of the amdgpu_kernel convention is to expose all kernel arguments as
loads from constant memory. Today, these are raw, SSA Argument values
and codegen is responsible for turning these into loads.
Background:
There currently isn't a satisfactory way to represent how arguments
for the amdgpu_kernel calling convention are passed. In reality,
arguments are passed in a single, flat, constant memory buffer
implicitly passed to the function. It is also illegal to call this
function in the IR, and this is only ever invoked by a driver of some
kind.
It does not make sense to have a stack passed parameter in this
context as is implied by byval. It is never valid to write to the
kernel arguments, as this would corrupt the inputs seen by other
dispatches of the kernel. These argumets are also not in the same
address space as the stack, so a copy is needed to an alloca. From a
source C-like language, the kernel parameters are invisible.
Semantically, a copy is always required from the constant argument
memory to a mutable variable.
The current clang calling convention lowering emits raw values,
including aggregates into the function argument list, since using
byval would not make sense. This has some unfortunate consequences for
the optimizer. In the aggregate case, we end up with an aggregate
store to alloca, which both SROA and instcombine turn into a store of
each aggregate field. The optimizer never pieces this back together to
see that this is really just a copy from constant memory, so we end up
stuck with expensive stack usage.
This also means the backend dictates the alignment of arguments, and
arbitrarily picks the LLVM IR ABI type alignment. By allowing an
explicit alignment, frontends can make better decisions. For example,
there's real no advantage to an aligment higher than 4, so a frontend
could choose to compact the argument layout. Similarly, there is a
high penalty to using an alignment lower than 4, so a frontend could
opt into more padding for small arguments.
Another design consideration is when it is appropriate to expose the
fact that these arguments are all really passed in adjacent
memory. Currently we have a late IR optimization pass in codegen to
rewrite the kernel argument values into explicit loads to enable
vectorization. In most programs, unrelated argument loads can be
merged together. However, exposing this property directly from the
frontend has some disadvantages. We still need a way to track the
original argument sizes and alignments to report to the driver. I find
using some side-channel, metadata mechanism to track this
unappealing. If the kernel arguments were exposed as a single buffer
to begin with, alias analysis would be unaware that the padding bits
betewen arguments are meaningless. Another family of problems is there
are still some gaps in replacing all of the available parameter
attributes with metadata equivalents once lowered to loads.
The immediate plan is to start using this new attribute to handle all
aggregate argumets for kernels. Long term, it makes sense to migrate
all kernel arguments, including scalars, to be passed indirectly in
the same manner.
Additional context is in D79744.
This patch adds a new variant of the matrix lowering pass that only does
a minimal lowering and only depends on TTI. The main purpose of this pass
is to have a pass with minimal dependencies to run as part of the backend
pipeline.
At the moment, the only difference to the regular lowering pass is that it
does not support remarks. But in subsequent patches add support for tiling
to the lowering pass which will require more analysis, which we do not want
to run in the backend, as the lowering should happen in the middle-end in
practice and running it in the backend is mostly for convenience when
running llc.
Reviewers: anemet, Gerolf, efriedma, hfinkel
Reviewed By: anemet
Differential Revision: https://reviews.llvm.org/D76867
Common code sinking is already guarded with a (with default-off!) flag,
so add a flag for hoisting, too.
D84108 will hopefully make hoisting off-by-default too.
This patch
- adds `canCreateUndefOrPoison`
- refactors `canCreatePoison` so it can deal with constantexprs
`canCreateUndefOrPoison` will be used at D83926.
Reviewed By: nikic, jdoerfert
Differential Revision: https://reviews.llvm.org/D84007
Summary:
This change added a new inline advisor that takes optimization remarks from previous inlining as input, and provides the decision as advice so current inlining can replay inline decisions of a different compilation. Dwarf inline stack with line and discriminator is used as anchor for call sites including call context. The change can be useful for Inliner tuning as it provides a channel to allow external input for tweaking inline decisions. Existing alternatives like alwaysinline attribute is per-function, not per-callsite. Per-callsite inline intrinsic can be another solution (not yet existing), but it's intrusive to implement and also does not differentiate call context.
A switch -sample-profile-inline-replay=<inline_remarks_file> is added to hook up the new inline advisor with SampleProfileLoader's inline decision for replay. Since SampleProfileLoader does top-down inlining, inline decision can be specialized for each call context, hence we should be able to replay inlining accurately. However with a bottom-up inliner like CGSCC inlining, the replay can be limited due to lack of specialization for different call context. Apart from that limitation, the new inline advisor can still be used by regular CGSCC inliner later if needed for tuning purpose.
Subscribers: mgorny, aprantl, hiraditya, llvm-commits
Tags: #llvm
Resubmit for https://reviews.llvm.org/D84086
This reverts commit 4500db8c59621a31c622862a2946457fdee481ce,
which was reverted because lower thresholds exposed a new issue (PR46680).
Now that it was resolved by d12ec0f752e7f2c7f7252539da2d124264ec33f7,
we can reinstate lower limits and wait for a new bugreport before
reverting this again...
Both users of predicteinfo (NewGVN and SCCP) are interested in
getting a cmp constraint on the predicated value. They currently
implement separate logic for this. This patch adds a common method
for this in PredicateBase.
This enables a missing bit of PredicateInfo handling in SCCP: Now
the predicate on the condition itself is also used. For switches
it means we know that the switched-on value is the same as the case
value. For assumes/branches we know that the condition is true or
false.
Differential Revision: https://reviews.llvm.org/D83640
Fixes https://bugs.llvm.org/show_bug.cgi?id=46680.
Just like insertions through IRBuilder, InsertNewInstBefore()
should be using the deferred worklist mechanism, so that processing
of newly added instructions is prioritized.
There's one side-effect of the worklist order change which could be
classified as a regression. An add op gets pushed through a select
that at the time is not a umax. We could add a reverse transform
that tries to push adds in the reverse direction to restore a min/max,
but that seems like a sure way of getting infinite loops... Seems
like something that should best wait on min/max intrinsics.
Differential Revision: https://reviews.llvm.org/D84109
GCC r187297 (2012-05) introduced `__gcov_dump` and `__gcov_reset`.
`__gcov_flush = __gcov_dump + __gcov_reset`
The resolution to https://gcc.gnu.org/PR93623 ("No need to dump gcdas when forking" target GCC 11.0) removed the unuseful and undocumented __gcov_flush.
Close PR38064.
Reviewed By: calixte, serge-sans-paille
Differential Revision: https://reviews.llvm.org/D83149
Each concrete instance of a predicate has a condition (also noted in the
original PredicateBase comment) and to me it seems like there is no
clear benefit of having both PredicateBase and PredicateWithCondition
and they can be folded together.
Reviewers: nikic, efriedma
Reviewed By: nikic
Differential Revision: https://reviews.llvm.org/D84089
Yes, if operands are non-positive this comes at the extra cost
of two extra negations. But a. division is already just
ridiculously costly, two more subtractions can't hurt much :)
and b. we have better/more analyzes/folds for an unsigned division,
we could end up narrowing it's bitwidth, converting it to lshr, etc.
This is essentially a take two on 0fdcca07ad2c0bdc2cdd40ba638109926f4f513b,
which didn't fix the potential regression i was seeing,
because ValueTracking's computeKnownBits() doesn't make use
of dominating conditions in it's analysis.
While i could teach it that, this seems like the more general fix.
This big hammer actually does catch said potential regression.
Over vanilla test-suite + RawSpeed + darktable
(10M IR instrs, 1M IR BB, 1M X86 ASM instrs), this fires/converts 5 more
(+2%) SDiv's, the total instruction count at the end of middle-end pipeline
is only +6, so out of +10 extra negations, ~half are folded away,
and asm instr count is only +1, so practically speaking all extra
negations are folded away and are therefore free.
Sadly, all these new UDiv's remained, none folded away.
But there are two less basic blocks.
https://rise4fun.com/Alive/VS6
Name: v0
Pre: C0 >= 0 && C1 >= 0
%r = sdiv i8 C0, C1
=>
%r = udiv i8 C0, C1
Name: v1
Pre: C0 <= 0 && C1 >= 0
%r = sdiv i8 C0, C1
=>
%t0 = udiv i8 -C0, C1
%r = sub i8 0, %t0
Name: v2
Pre: C0 >= 0 && C1 <= 0
%r = sdiv i8 C0, C1
=>
%t0 = udiv i8 C0, -C1
%r = sub i8 0, %t0
Name: v3
Pre: C0 <= 0 && C1 <= 0
%r = sdiv i8 C0, C1
=>
%r = udiv i8 -C0, -C1
This is needed because macOS on Apple Silicon has some reserved pages inside the "regular" shadow memory location, and mapping over that location fails.
Differential Revision: https://reviews.llvm.org/D82912
This reverts commit d76e62fdb7a93d9a33f642b6b528f2562cc3c3f4.
Reverting since this can lead to linker errors:
```
ld.lld: error: undefined hidden symbol: __start_asan_globals
```
when using --gc-sections. The linker can discard __start_asan_globals
once there are no more `asan_globals` sections left, which can lead to
this error if we have external linkages to them.
as it is failing the inline-replay.ll test as well as sanitizers/Werror
from returning a stack local variable.
This reverts commit 029946b112684c27b27f7c2d7554f22b33ae1e0b.
Summary:
This change added a new inline advisor that takes optimization remarks for previous inlining as input, and provide the decision as advice so current inlining can replay inline decision of a different compilation. Dwarf inline stack with line and discriminator is used as anchor for call sites. The change can be useful for Inliner tuning.
A switch -sample-profile-inline-replay=<inline_remarks_file> is added to hook up the new inliner advisor with SampleProfileLoader's inline decision for replay. The new inline advisor can also be used by regular CGSCC inliner later if needed.
Reviewers: davidxl, mtrofin, wmi, hoy
Subscribers: aprantl, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D83743
This is the one i'm seeing as missed optimization,
although there are likely other possibilities, as usual.
There are 4 variants of a general sdiv->udiv fold:
https://rise4fun.com/Alive/VS6
Name: v0
Pre: C0 >= 0 && C1 >= 0
%r = sdiv i8 C0, C1
=>
%r = udiv i8 C0, C1
Name: v1
Pre: C0 <= 0 && C1 >= 0
%r = sdiv i8 C0, C1
=>
%t0 = udiv i8 -C0, C1
%r = sub i8 0, %t0
Name: v2
Pre: C0 >= 0 && C1 <= 0
%r = sdiv i8 C0, C1
=>
%t0 = udiv i8 C0, -C1
%r = sub i8 0, %t0
Name: v3
Pre: C0 <= 0 && C1 <= 0
%r = sdiv i8 C0, C1
=>
%r = udiv i8 -C0, -C1
If we really don't like sdiv (more than udiv that is),
and are okay with increasing instruction count (2 new negations),
and we ensure that we don't undo the fold,
then we could just implement these..
I've got the report clang11 issues signed/unsigned mismatch
warning here. For some reason only clang11 seems to issue
this warning.
Differential Revision: https://reviews.llvm.org/D83916
Summary:
This patch resolves an issue where the metadata of a loop is not added to the
new loop latch, and not removed from the old loop latch. This issue occurs in
the SplitBlockPredecessors function, which adds a new block in a loop, and
in the case that the block passed into this function is the header of the loop,
the loop can be modified such that the latch of the loop is replaced.
This patch applies to the Loop Simplify pass since it ensures that each loop
has exit blocks which only have predecessors that are inside of the loop. In
the case that this is not true, the pass will create a new exit block for the
loop. This guarantees that the loop preheader/header will dominate the exit blocks.
Author: sidbav (Sidharth Baveja)
Reviewers: asbirlea (Alina Sbirlea), chandlerc (Chandler Carruth), Whitney (Whitney Tsang), bmahjour (Bardia Mahjour)
Reviewed By: asbirlea (Alina Sbirlea)
Subscribers: hiraditya (Aditya Kumar), llvm-commits
Tag: LLVM
Differential Revision: https://reviews.llvm.org/D83869
This patch enables the LoopVectorizer to build a phi of pointer
type and provide the vector loads and stores with vector type
getelementptrs built from the pointer induction variable, which
produces much less instructions than the previous approach of
creating scalar getelementpointers and glue them together to a
vector.
Differential Revision: https://reviews.llvm.org/D81267
This adds option -tsan-compound-read-before-write to emit different
instrumentation for the write if the read before that write is omitted
from instrumentation. The default TSan runtime currently does not
support the different instrumentation, and the option is disabled by
default.
Alternative runtimes, such as the Kernel Concurrency Sanitizer (KCSAN)
can make use of the feature. Indeed, the initial motivation is for use
in KCSAN as it was determined that due to the Linux kernel having a
large number of unaddressed data races, it makes sense to improve
performance and reporting by distinguishing compounded operations. E.g.
the compounded instrumentation is typically emitted for compound
operations such as ++, +=, |=, etc. By emitting different reports, such
data races can easily be noticed, and also automatically bucketed
differently by CI systems.
Reviewed By: dvyukov, glider
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D83867
```
define i32 @test(i1 %cond) {
entry:
br i1 %cond, label %exit, label %exit
exit:
%result = select i1 %cond, i32 123, i32 456
ret i32 %result
}
```
In this test, after applying transformation of replacing select with Phis,
the result will be:
```
define i32 @test(i1 %cond) {
entry:
br i1 %cond, label %exit, label %exit
exit:
%result = i32 phi [123, %exit], [123, %exit]
ret i32 %result
}
```
That is, select is transformed into an invalid Phi, which will then be
reduced to 123 and the second value will be lost. But it is worth
noting that this problem will arise only if select is in the InstCombine
worklist will be before the branch. Otherwise, InstCombine will replace
the branch condition with false and transformation will not be applied.
The fix is to check the target labels in the branch condition for equality.
Patch By: Kirill Polushin
Differential Revision: https://reviews.llvm.org/D84003
Reviewed By: mkazantsev
SimplifyCFG was incorrectly reporting to the pass manager that it had not made
changes after folding away a PHI. This is detected in the EXPENSIVE_CHECKS
build when the function's hash changes.
Differential Revision: https://reviews.llvm.org/D83985
due to the performance bugs filed in https://bugs.llvm.org/show_bug.cgi?id=46753.
An SROA change soon may obviate some of these problems.
This reverts commit 8d09f20798ac180b1749276bff364682ce0196ab.
When the byref attribute is added, there will need to be two similar
functions for the existing cases which have an associate value copy,
and byref which does not. Most, but not all of the existing uses will
use the existing version.
The associated size function added by D82679 also needs to
contextually differ, and will help eliminate a few places still
relying on pointee element types.
