3bb969f3eb
Inlining `hlfir.matmul` as `hlfir.eval_in_mem` does not allow
to get rid of a temporary array in many cases, but it may still be
much better allowing to:
* Get rid of any overhead related to calling runtime MATMUL
(such as descriptors creation).
* Use CPU-specific vectorization cost model for matmul loops,
which Fortran runtime cannot currently do.
* Optimize matmul of known-size arrays by complete unrolling.
One of the drawbacks of `hlfir.eval_in_mem` inlining is that
the ops inside it with store memory effects block the current
MLIR CSE, so I decided to run this inlining late in the pipeline.
There is a source commen explaining the CSE issue in more detail.
Straightforward inlining of `hlfir.matmul` as an `hlfir.elemental`
is not good for performance, and I got performance regressions
with it comparing to Fortran runtime implementation. I put it
under an enigneering option for experiments.
At the same time, inlining `hlfir.matmul_transpose` as `hlfir.elemental`
seems to be a good approach, e.g. it allows getting rid of a temporay
array in cases like: `A(:)=B(:)+MATMUL(TRANSPOSE(C(:,:)),D(:))`.
This patch improves performance of galgel and tonto a little bit.
151 lines
5.6 KiB
Plaintext
151 lines
5.6 KiB
Plaintext
// RUN: tco %s | FileCheck %s
|
|
// RUN: tco %s --mlir-pass-statistics --mlir-pass-statistics-display=pipeline 2>&1 | FileCheck %s --check-prefix=PASSES
|
|
|
|
// REQUIRES: asserts
|
|
|
|
// Check that tco is working with a basic test.
|
|
// Also check the passes in the default pipeline.
|
|
|
|
func.func @_QQmain() {
|
|
return
|
|
}
|
|
|
|
// CHECK: ; ModuleID = 'FIRModule'
|
|
// CHECK-LABEL: define void @_QQmain()
|
|
// CHECK: ret void
|
|
|
|
// PASSES: Pass statistics report
|
|
|
|
// PASSES: Canonicalizer
|
|
// PASSES-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'omp.declare_reduction', 'omp.private']
|
|
// PASSES-NEXT: 'fir.global' Pipeline
|
|
// PASSES-NEXT: SimplifyHLFIRIntrinsics
|
|
// PASSES-NEXT: InlineElementals
|
|
// PASSES-NEXT: 'func.func' Pipeline
|
|
// PASSES-NEXT: SimplifyHLFIRIntrinsics
|
|
// PASSES-NEXT: InlineElementals
|
|
// PASSES-NEXT: 'omp.declare_reduction' Pipeline
|
|
// PASSES-NEXT: SimplifyHLFIRIntrinsics
|
|
// PASSES-NEXT: InlineElementals
|
|
// PASSES-NEXT: 'omp.private' Pipeline
|
|
// PASSES-NEXT: SimplifyHLFIRIntrinsics
|
|
// PASSES-NEXT: InlineElementals
|
|
// PASSES-NEXT: Canonicalizer
|
|
// PASSES-NEXT: CSE
|
|
// PASSES-NEXT: (S) 0 num-cse'd - Number of operations CSE'd
|
|
// PASSES-NEXT: (S) 0 num-dce'd - Number of operations DCE'd
|
|
// PASSES-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'omp.declare_reduction', 'omp.private']
|
|
// PASSES-NEXT: 'fir.global' Pipeline
|
|
// PASSES-NEXT: SimplifyHLFIRIntrinsics
|
|
// PASSES-NEXT: OptimizedBufferization
|
|
// PASSES-NEXT: InlineHLFIRAssign
|
|
// PASSES-NEXT: 'func.func' Pipeline
|
|
// PASSES-NEXT: SimplifyHLFIRIntrinsics
|
|
// PASSES-NEXT: OptimizedBufferization
|
|
// PASSES-NEXT: InlineHLFIRAssign
|
|
// PASSES-NEXT: 'omp.declare_reduction' Pipeline
|
|
// PASSES-NEXT: SimplifyHLFIRIntrinsics
|
|
// PASSES-NEXT: OptimizedBufferization
|
|
// PASSES-NEXT: InlineHLFIRAssign
|
|
// PASSES-NEXT: 'omp.private' Pipeline
|
|
// PASSES-NEXT: SimplifyHLFIRIntrinsics
|
|
// PASSES-NEXT: OptimizedBufferization
|
|
// PASSES-NEXT: InlineHLFIRAssign
|
|
// PASSES-NEXT: LowerHLFIROrderedAssignments
|
|
// PASSES-NEXT: LowerHLFIRIntrinsics
|
|
// PASSES-NEXT: BufferizeHLFIR
|
|
// PASSES-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'omp.declare_reduction', 'omp.private']
|
|
// PASSES-NEXT: 'fir.global' Pipeline
|
|
// PASSES-NEXT: InlineHLFIRAssign
|
|
// PASSES-NEXT: 'func.func' Pipeline
|
|
// PASSES-NEXT: InlineHLFIRAssign
|
|
// PASSES-NEXT: 'omp.declare_reduction' Pipeline
|
|
// PASSES-NEXT: InlineHLFIRAssign
|
|
// PASSES-NEXT: 'omp.private' Pipeline
|
|
// PASSES-NEXT: InlineHLFIRAssign
|
|
// PASSES-NEXT: ConvertHLFIRtoFIR
|
|
// PASSES-NEXT: LowerWorkshare
|
|
// PASSES-NEXT: CSE
|
|
// PASSES-NEXT: (S) 0 num-cse'd - Number of operations CSE'd
|
|
// PASSES-NEXT: (S) 0 num-dce'd - Number of operations DCE'd
|
|
|
|
// PASSES-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'omp.declare_reduction', 'omp.private']
|
|
// PASSES-NEXT: 'fir.global' Pipeline
|
|
// PASSES-NEXT: CharacterConversion
|
|
// PASSES-NEXT: 'func.func' Pipeline
|
|
// PASSES-NEXT: ArrayValueCopy
|
|
// PASSES-NEXT: CharacterConversion
|
|
// PASSES-NEXT: 'omp.declare_reduction' Pipeline
|
|
// PASSES-NEXT: CharacterConversion
|
|
// PASSES-NEXT: 'omp.private' Pipeline
|
|
// PASSES-NEXT: CharacterConversion
|
|
|
|
// PASSES-NEXT: Canonicalizer
|
|
// PASSES-NEXT: SimplifyRegionLite
|
|
// PASSES-NEXT: SimplifyIntrinsics
|
|
// PASSES-NEXT: AlgebraicSimplification
|
|
// PASSES-NEXT: CSE
|
|
// PASSES-NEXT: (S) 0 num-cse'd - Number of operations CSE'd
|
|
// PASSES-NEXT: (S) 0 num-dce'd - Number of operations DCE'd
|
|
|
|
// PASSES-NEXT: 'func.func' Pipeline
|
|
// PASSES-NEXT: MemoryAllocationOpt
|
|
|
|
// PASSES-NEXT: Inliner
|
|
// PASSES-NEXT: SimplifyRegionLite
|
|
// PASSES-NEXT: CSE
|
|
// PASSES-NEXT: (S) 0 num-cse'd - Number of operations CSE'd
|
|
// PASSES-NEXT: (S) 0 num-dce'd - Number of operations DCE'd
|
|
|
|
// PASSES-NEXT: PolymorphicOpConversion
|
|
// PASSES-NEXT: AssumedRankOpConversion
|
|
// PASSES-NEXT: AddAliasTags
|
|
|
|
// PASSES-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'omp.declare_reduction', 'omp.private']
|
|
// PASSES-NEXT: 'fir.global' Pipeline
|
|
// PASSES-NEXT: StackReclaim
|
|
// PASSES-NEXT: CFGConversion
|
|
// PASSES-NEXT: 'func.func' Pipeline
|
|
// PASSES-NEXT: StackReclaim
|
|
// PASSES-NEXT: CFGConversion
|
|
// PASSES-NEXT: 'omp.declare_reduction' Pipeline
|
|
// PASSES-NEXT: StackReclaim
|
|
// PASSES-NEXT: CFGConversion
|
|
// PASSES-NEXT: 'omp.private' Pipeline
|
|
// PASSES-NEXT: StackReclaim
|
|
// PASSES-NEXT: CFGConversion
|
|
|
|
// PASSES-NEXT: SCFToControlFlow
|
|
// PASSES-NEXT: Canonicalizer
|
|
// PASSES-NEXT: SimplifyRegionLite
|
|
// PASSES-NEXT: CSE
|
|
// PASSES-NEXT: (S) 0 num-cse'd - Number of operations CSE'd
|
|
// PASSES-NEXT: (S) 0 num-dce'd - Number of operations DCE'd
|
|
// PASSES-NEXT: BoxedProcedurePass
|
|
|
|
// PASSES-NEXT: Pipeline Collection : ['fir.global', 'func.func', 'gpu.module', 'omp.declare_reduction', 'omp.private']
|
|
// PASSES-NEXT: 'fir.global' Pipeline
|
|
// PASSES-NEXT: AbstractResultOpt
|
|
// PASSES-NEXT: 'func.func' Pipeline
|
|
// PASSES-NEXT: AbstractResultOpt
|
|
// PASSES-NEXT: 'gpu.module' Pipeline
|
|
// PASSES-NEXT: Pipeline Collection : ['func.func', 'gpu.func']
|
|
// PASSES-NEXT: 'func.func' Pipeline
|
|
// PASSES-NEXT: AbstractResultOpt
|
|
// PASSES-NEXT: 'gpu.func' Pipeline
|
|
// PASSES-NEXT: AbstractResultOpt
|
|
// PASSES-NEXT: 'omp.declare_reduction' Pipeline
|
|
// PASSES-NEXT: AbstractResultOpt
|
|
// PASSES-NEXT: 'omp.private' Pipeline
|
|
// PASSES-NEXT: AbstractResultOpt
|
|
|
|
// PASSES-NEXT: CodeGenRewrite
|
|
// PASSES-NEXT: (S) 0 num-dce'd - Number of operations eliminated
|
|
// PASSES-NEXT: TargetRewrite
|
|
// PASSES-NEXT: CompilerGeneratedNamesConversion
|
|
// PASSES-NEXT: 'func.func' Pipeline
|
|
// PASSES-NEXT: FunctionAttr
|
|
// PASSES-NEXT: FIRToLLVMLowering
|
|
// PASSES-NEXT: ReconcileUnrealizedCasts
|
|
// PASSES-NEXT: LLVMIRLoweringPass
|