[Unroll/UnrollAndJam/Vectorizer/Distribute] Add followup loop attributes.

When multiple loop transformation are defined in a loop's metadata, their order of execution is defined by the order of thei

[Unroll/UnrollAndJam/Vectorizer/Distribute] Add followup loop attributes.

When multiple loop transformation are defined in a loop's metadata, their order of execution is defined by the order of their respective passes in the pass pipeline. For instance, e.g.

#pragma clang loop unroll_and_jam(enable)
#pragma clang loop distribute(enable)

is the same as

#pragma clang loop distribute(enable)
#pragma clang loop unroll_and_jam(enable)

and will try to loop-distribute before Unroll-And-Jam because the LoopDistribute pass is scheduled after UnrollAndJam pass. UnrollAndJamPass only supports one inner loop, i.e. it will necessarily fail after loop distribution. It is not possible to specify another execution order. Also,t the order of passes in the pipeline is subject to change between versions of LLVM, optimization options and which pass manager is used.

This patch adds 'followup' attributes to various loop transformation passes. These attributes define which attributes the resulting loop of a transformation should have. For instance,

!0 = !{!0, !1, !2}
!1 = !{!"llvm.loop.unroll_and_jam.enable"}
!2 = !{!"llvm.loop.unroll_and_jam.followup_inner", !3}
!3 = !{!"llvm.loop.distribute.enable"}

defines a loop ID (!0) to be unrolled-and-jammed (!1) and then the attribute !3 to be added to the jammed inner loop, which contains the instruction to distribute the inner loop.

Currently, in both pass managers, pass execution is in a fixed order and UnrollAndJamPass will not execute again after LoopDistribute. We hope to fix this in the future by allowing pass managers to run passes until a fixpoint is reached, use Polly to perform these transformations, or add a loop transformation pass which takes the order issue into account.

For mandatory/forced transformations (e.g. by having been declared by #pragma omp simd), the user must be notified when a transformation could not be performed. It is not possible that the responsible pass emits such a warning because the transformation might be 'hidden' in a followup attribute when it is executed, or it is not present in the pipeline at all. For this reason, this patche introduces a WarnMissedTransformations pass, to warn about orphaned transformations.

Since this changes the user-visible diagnostic message when a transformation is applied, two test cases in the clang repository need to be updated.

To ensure that no other transformation is executed before the intended one, the attribute `llvm.loop.disable_nonforced` can be added which should disable transformation heuristics before the intended transformation is applied. E.g. it would be surprising if a loop is distributed before a #pragma unroll_and_jam is applied.

With more supported code transformations (loop fusion, interchange, stripmining, offloading, etc.), transformations can be used as building blocks for more complex transformations (e.g. stripmining+stripmining+interchange -> tiling).

Reviewed By: hfinkel, dmgreen

Differential Revision: https://reviews.llvm.org/D49281
Differential Revision: https://reviews.llvm.org/D55288

llvm-svn: 348944

show more ...

Attempt to pacify buildbots after r280217

These clang tests check diagnostics from the backend by giving it an unvectorizable loop. This loop is now vectorized :/

Make it really unvectorizable by m

Attempt to pacify buildbots after r280217

These clang tests check diagnostics from the backend by giving it an unvectorizable loop. This loop is now vectorized :/

Make it really unvectorizable by making it unprofitable to ifconvert.

llvm-svn: 280220

show more ...

Add a loop's debug location to its llvm.loop metadata

Getting accurate locations for loops is important, because those locations are
used by the frontend to generate optimization remarks. Currently,

Add a loop's debug location to its llvm.loop metadata

Getting accurate locations for loops is important, because those locations are
used by the frontend to generate optimization remarks. Currently, optimization
remarks for loops often appear on the wrong line, often the first line of the
loop body instead of the loop itself. This is confusing because that line might
itself be another loop, or might be somewhere else completely if the body was
an inlined function call. This happens because of the way we find the loop's
starting location. First, we look for a preheader, and if we find one, and its
terminator has a debug location, then we use that. Otherwise, we look for a
location on an instruction in the loop header.

The fallback heuristic is not bad, but will almost always find the beginning of
the body, and not the loop statement itself. The preheader location search
often fails because there's often not a preheader, and even when there is a
preheader, depending on how it was formed, it sometimes carries the location of
some preceeding code.

I don't see any good theoretical way to fix this problem. On the other hand,
this seems like a straightforward solution: Put the debug location in the
loop's llvm.loop metadata. When emitting debug information, this commit causes
us to add the debug location as an operand to each loop's llvm.loop metadata.
Thus, we now generate this metadata for all loops (not just loops with
optimization hints) when we're otherwise generating debug information.

The remark test case changes depend on the companion LLVM commit r270771.

llvm-svn: 270772

show more ...

Stop messing with the 'g' group of options in CompilerInvocation.

With this change, most 'g' options are rejected by CompilerInvocation.
They remain only as Driver options. The new way to request de

Stop messing with the 'g' group of options in CompilerInvocation.

With this change, most 'g' options are rejected by CompilerInvocation.
They remain only as Driver options. The new way to request debug info
from cc1 is with "-debug-info-kind={line-tables-only|limited|standalone}"
and "-dwarf-version={2|3|4}". In the absence of a command-line option
to specify Dwarf version, the Toolchain decides it, rather than placing
Toolchain-specific logic in CompilerInvocation.

Also fix a bug in the Windows compatibility argument parsing
in which the "rightmost argument wins" principle failed.

Differential Revision: http://reviews.llvm.org/D13221

llvm-svn: 249655

show more ...

Fix test case.

r230921 broke backend-optimization-failure.cpp: after
r230921, LLVM no longer emits an expression to compute 'Length - 1'
and this perturbs LoopSimplify enough to emit the warning on

Fix test case.

r230921 broke backend-optimization-failure.cpp: after
r230921, LLVM no longer emits an expression to compute 'Length - 1'
and this perturbs LoopSimplify enough to emit the warning on line 10
instead of line 9. This is a review request to fix the test case once
I re-land r230921.

llvm-svn: 231020

show more ...

Adjust test to handle fallout from r217102.

llvm-svn: 217104

clang/test/Misc/backend-optimization-failure.cpp: Appease to add -triple=x86_64.

FIXME: Could this be made generic?
llvm-svn: 213441

Recommit: Handle diagnostic warnings in Frontend diagnostic handler.

Clang uses a diagnostic handler to grab diagnostic messages so it can print them
with the line of source code they refer to. This

Recommit: Handle diagnostic warnings in Frontend diagnostic handler.

Clang uses a diagnostic handler to grab diagnostic messages so it can print them
with the line of source code they refer to. This patch extends this to handle
optimization failures that were added to llvm to produce a warning when
loop vectorization is explicitly specified (using a pragma clang loop directive)
but fails.

Update renames warning flag name to avoid indicating the flag's severity and
adds a test.

Reviewed by Alp Toker

llvm-svn: 213400

show more ...