aboutsummaryrefslogtreecommitdiff
path: root/polly
AgeCommit message (Collapse)AuthorFilesLines
2022-09-14[CMake] Avoid `LLVM_BINARY_DIR` when other more specific variable are ↵John Ericson2-2/+2
better-suited, part 2 A simple sed doing these substitutions: - `${LLVM_BINARY_DIR}/lib${LLVM_LIBDIR_SUFFIX}\>` -> `${LLVM_LIBRARY_DIR}` - `${LLVM_BINARY_DIR}/bin\>` -> `${LLVM_TOOLS_BINARY_DIR}` where `\>` means "word boundary". The only manual modifications were reverting changes in - `runtimes/CMakeLists.txt` because these were "entry points" where we wanted to tread carefully not not introduce a "loop" which would end with an undefined variable being expanded to nothing. There are some `${LLVM_BINARY_DIR}/lib` without the `${LLVM_LIBDIR_SUFFIX}`, but these refer to the lib subdirectory of the source (`llvm/lib`). That `lib` is automatically appended to make the local `CMAKE_CURRENT_BINARY_DIR` value by `add_subdirectory`; since the directory name in the source tree is fixed without any suffix, the corresponding `CMAKE_CURRENT_BINARY_DIR` will also be. We therefore do not replace it but leave it as-is. This picks up where D133828 left off, getting the occurrences with*out* `CMAKE_CFG_INTDIR`. But this is difficult to do correctly and so not done in the (retroactively) previous diff. This hopefully increases readability overall, and also decreases the usages of `LLVM_LIBDIR_SUFFIX`, preparing us for D130586. Reviewed By: sebastian-ne Differential Revision: https://reviews.llvm.org/D132316
2022-09-14[AA] Tracking per-location ModRef info in FunctionModRefBehavior (NFCI)Nikita Popov2-47/+25
Currently, FunctionModRefBehavior tracks whether the function reads or writes memory (ModRefInfo) and which locations it can access (argmem, inaccessiblemem and other). This patch changes it to track ModRef information per-location instead. To give two examples of why this is useful: * D117095 highlights a weakness of ModRef modelling in the presence of operand bundles. For a memcpy call with deopt operand bundle, we want to say that it can read any memory, but only write argument memory. This would allow them to be treated like any other calls. However, we currently can't express this and have to say that it can read or write any memory. * D127383 would ideally be modelled as a separate threadid location, where threadid Refs outside pre-split coroutines can be ignored (like other accesses to constant memory). The current representation does not allow modelling this precisely. The patch as implemented is intended to be NFC, but there are some obvious opportunities for improvements and simplification. To fully capitalize on this we would also want to change the way we represent memory attributes on functions, but that's a larger change, and I think it makes sense to separate out the FunctionModRefBehavior refactoring. Differential Revision: https://reviews.llvm.org/D130896
2022-08-28Use std::gcd (NFC)Kazu Hirata1-1/+2
To avoid changing semantics inadvertently, this patch casts arguments to uint64_t before calling std::gcd.
2022-08-25Revert "[CMake] Avoid `LLVM_BINARY_DIR` when other more specific variable ↵John Ericson2-2/+2
are better-suited" This reverts commit ad8c34bc3089d847a09bb740f7a58c96073e0959.
2022-08-24Exclude check-polly-unittests and check-polly-isl from check-allEli Friedman1-0/+2
The unittests are already included in check-polly, so check-all was running them twice. Running them twice causes a race on the output files, which led to intermittent failures on the reverse-iteration buildbot.
2022-08-24[CMake] Avoid `LLVM_BINARY_DIR` when other more specific variable are ↵John Ericson2-2/+2
better-suited A simple sed doing these substitutions: - `${LLVM_BINARY_DIR}/(\$\{CMAKE_CFG_INTDIR}/)?lib(${LLVM_LIBDIR_SUFFIX})?\>` -> `${LLVM_LIBRARY_DIR}` - `${LLVM_BINARY_DIR}/(\$\{CMAKE_CFG_INTDIR}/)?bin\>` -> `${LLVM_TOOLS_BINARY_DIR}` where `\>` means "word boundary". The only manual modifications were reverting changes in - `compiler-rt/cmake/Modules/CompilerRTUtils.cmake - `runtimes/CMakeLists.txt` because these were "entry points" where we wanted to tread carefully not not introduce a "loop" which would end with an undefined variable being expanded to nothing. This hopefully increases readability overall, and also decreases the usages of `LLVM_LIBDIR_SUFFIX`, preparing us for D130586. Reviewed By: sebastian-ne Differential Revision: https://reviews.llvm.org/D132316
2022-08-23[Polly] Don't use `llvm-config` anymore (in CMake sad path)John Ericson1-6/+3
If `LLVM_BUILD_MAIN_SRC_DIR` is not defined, just assume we are in regular monorepo layout. Non-standard (and not really supported) layouts can still be configured manually. Reviewed By: beanz Differential Revision: https://reviews.llvm.org/D132314
2022-08-18Revert "[cmake] Use `CMAKE_INSTALL_LIBDIR` too"John Ericson3-6/+6
This reverts commit f7a33090a91015836497c75f173775392ab0304d. Unfortunately this causes a number of failures that didn't show up in my local build.
2022-08-18[cmake] Use `CMAKE_INSTALL_LIBDIR` tooJohn Ericson3-6/+6
We held off on this before as `LLVM_LIBDIR_SUFFIX` conflicted with it. Now we return this. `LLVM_LIBDIR_SUFFIX` is kept as a deprecated way to set `CMAKE_INSTALL_LIBDIR`. The other `*_LIBDIR_SUFFIX` are just removed entirely. I imagine this is too potentially-breaking to make LLVM 15. That's fine. I have a more minimal version of this in the disto (NixOS) patches for LLVM 15 (like previous versions). This more expansive version I will test harder after the release is cut. Reviewed By: sebastian-ne, ldionne, #libc, #libc_abi Differential Revision: https://reviews.llvm.org/D130586
2022-08-17[test] Propagate HWASAN_OPTIONSVitaly Buka1-0/+2
2022-08-14[Polly] Remove the test case that depends on InstCombine and DeLICM.Roman Gareev1-79/+0
2022-08-09LLVM_FALLTHROUGH => [[fallthrough]]. NFCFangrui Song4-7/+7
2022-08-07[polly] Fixed a number of typos. NFCGabriel Ravier27-33/+33
I went over the output of the following mess of a command: `(ulimit -m 2000000; ulimit -v 2000000; git ls-files -z | parallel --xargs -0 cat | aspell list --mode=none --ignore-case | grep -E '^[A-Za-z][a-z]*$' | sort | uniq -c | sort -n | grep -vE '.{25}' | aspell pipe -W3 | grep : | cut -d' ' -f2 | less)` and proceeded to spend a few days looking at it to find probable typos and fixed a few hundred of them in all of the llvm project (note, the ones I found are not anywhere near all of them, but it seems like a good start). Reviewed By: inclyc Differential Revision: https://reviews.llvm.org/D131167
2022-08-07[Polly] Suppress the LLVM-IR output for pattern matching tests, if there is ↵Roman Gareev4-5/+5
no FileCheck-ing for it.
2022-08-07[Polly] Generalize the pattern matching to the case of tensor contractionsRoman Gareev21-21/+1734
The pattern matching optimization of Polly detects and optimizes dense general matrix-matrix multiplication. The generated code is close to high performance implementations of matrix-matrix multiplications, which are contained in manually tuned libraries. The described pattern matching optimization is a particular case of tensor contraction optimization, which was introduced in [1]. This patch generalizes the pattern matching to the case of tensor contractions using the form of data dependencies and memory accesses produced by tensor contractions [1]. Optimization of tensor contractions will be added in the next patch. Following the ideas introduced in [2], it will logically represent tensor contraction operands as matrix multiplication operands and use an approach for optimization of matrix-matrix multiplications. [1] - Gareev R., Grosser T., Kruse M. High-Performance Generalized Tensor Op­erations: A Compiler-Oriented Approach // ACM Transactions on Architec­ture and Code Optimization (TACO). 2018. Vol. 15, no. 3. P. 34:1–34:27. DOI: 10.1145/3235029. [2] - Matthews D. High-Performance Tensor Contraction without BLAS // SIAM Journal on Scientific Computing. 2018. Vol. 40, no. 1. P. C 1—C 24. DOI: 110.1137/16m108968x. Reviewed By: Meinersbur Differential Revision: https://reviews.llvm.org/D114336
2022-07-26[Polly] Insert !dbg metadata for emitted CallInsts.Michael Kruse5-15/+220
The IR Verifier requires that every call instruction to an inlineable function (among other things, its implementation must be visible in the translation unit) must also have !dbg metadata attached to it. When parallelizing, Polly emits calls to OpenMP runtime function out of thin air, or at least not directly derived from a bounded list of previous instruction. While we could search for instructions in the SCoP that has some debug info attached to it, there is no guarantee that we find any. Our solution is to generate a new DILocation that points to line 0 to represent optimized code. The OpenMP function implementation is usually not available in the user's translation unit, but can become visible in an LTO build. For the bug to appear, libomp must also be built with debug symbols. IMHO, the IR verifier rule is too strict. Runtime functions can also be inserted by other optimization passes, such as LoopIdiomRecognize. When inserting a call to e.g. memset, it uses the DebugLoc from a StoreInst from the unoptimized code. It is not required to have !dbg metadata attached either. Fixes #56692
2022-07-25Remove redundaunt virtual specifiers (NFC)Kazu Hirata10-27/+26
Identified with tidy-modernize-use-override.
2022-07-25[cmake] Support custom package install pathsJohn Ericson1-8/+15
Firstly, we we make an additional GNUInstallDirs-style variable. With NixOS, for example, this is crucial as we want those to go in `${dev}/lib/cmake` not `${out}/lib/cmake` as that would a cmake subdir of the "regular" libdir, which is installed even when no one needs to do any development. Secondly, we make *Config.cmake robust to absolute package install paths. We for NixOS will in fact be passing them absolute paths to make the `${dev}` vs `${out}` distinction mentioned above, and the GNUInstallDirs-style variables are suposed to support absolute paths in general so it's good practice besides the NixOS use-case. Thirdly, we make `${project}_INSTALL_PACKAGE_DIR` CACHE PATHs like other install dirs are. Reviewed By: sebastian-ne Differential Revision: https://reviews.llvm.org/D117973
2022-07-22Use any_of (NFC)Kazu Hirata1-2/+1
2022-07-20Use llvm::is_contained (NFC)Kazu Hirata1-2/+1
2022-07-15Use value instead of getValue (NFC)Kazu Hirata3-13/+13
2022-07-15[IR] Don't use blockaddresses as callbr argumentsNikita Popov1-1/+1
Following some recent discussions, this changes the representation of callbrs in IR. The current blockaddress arguments are replaced with `!` label constraints that refer directly to callbr indirect destinations: ; Before: %res = callbr i8* asm "", "=r,r,i"(i8* %x, i8* blockaddress(@test8, %foo)) to label %asm.fallthrough [label %foo] ; After: %res = callbr i8* asm "", "=r,r,!i"(i8* %x) to label %asm.fallthrough [label %foo] The benefit of this is that we can easily update the successors of a callbr, without having to worry about also updating blockaddress references. This should allow us to remove some limitations: * Allow unrolling/peeling/rotation of callbr, or any other clone-based optimizations (https://github.com/llvm/llvm-project/issues/41834) * Allow duplicate successors (https://github.com/llvm/llvm-project/issues/45248) This is just the IR representation change though, I will follow up with patches to remove limtations in various transformation passes that are no longer needed. Differential Revision: https://reviews.llvm.org/D129288
2022-07-13Use has_value instead of hasValue (NFC)Kazu Hirata1-1/+1
2022-06-29[Polly][MatMul] Abandon dependence analysis.Michael Kruse6-49/+63
The copy statements inserted by the matrix-multiplication optimization introduce new dependencies between the copy statements and other statements. As a result, the DependenceInfo must be recomputed. Not recomputing them caused IslAstInfo to deduce that some loops are parallel but cause race conditions when accessing the packed arrays. As a result, matrix-matrix multiplication currently cannot be parallelized. Also see discussion at https://reviews.llvm.org/D125202
2022-06-26Don't use Optional::hasValue (NFC)Kazu Hirata2-5/+5
This patch replaces x.hasValue() with x where x is contextually convertible to bool.
2022-06-25Revert "Don't use Optional::hasValue (NFC)"Kazu Hirata3-21/+22
This reverts commit aa8feeefd3ac6c78ee8f67bf033976fc7d68bc6d.
2022-06-25Don't use Optional::hasValue (NFC)Kazu Hirata3-22/+21
2022-06-22[Support] Change TrackingStatistic and NoopStatistic to use uint64_t instead ↵Mingming Liu2-3/+3
of unsigned. Binary size of `clang` is trivial; namely, numerical value doesn't change when measured in MiB, and `.data` section increases from 139Ki to 173 Ki. Differential Revision: https://reviews.llvm.org/D128070
2022-06-21[polly] #include <algorithm>Arthur Eubanks1-0/+1
For the usage of std::max in the header. Speculative fix for https://ci.chromium.org/ui/p/fuchsia/builders/toolchain.ci/clang-windows-x64/b8810806780048763729/overview reported in https://reviews.llvm.org/D125263.
2022-06-20Don't use Optional::getValue (NFC)Kazu Hirata1-3/+3
2022-06-19Use value_or instead of getValueOr (NFC)Kazu Hirata3-5/+5
2022-06-13[NFC][Alignment] Simplify code in JSONExporterGuillaume Chatelet1-5/+3
2022-06-13[NFC][polly] Removed dead codeGuillaume Chatelet1-15/+0
2022-06-07[NFC] Use predecessors to replace make_range.jacquesguan1-2/+2
Reviewed By: rriddle Differential Revision: https://reviews.llvm.org/D127085
2022-06-05Remove unneeded cl::ZeroOrMore for cl::opt/cl::list optionsFangrui Song13-39/+34
2022-06-05Remove unneeded cl::ZeroOrMore for cl::opt/cl::list optionsFangrui Song7-12/+9
2022-06-04Remove unneeded cl::ZeroOrMore for cl::list optionsFangrui Song1-5/+3
2022-06-04Remove unneeded cl::ZeroOrMore for cl::opt optionsFangrui Song18-115/+90
Similar to 557efc9a8b68628c2c944678c6471dac30ed9e8e. This commit handles options where cl::ZeroOrMore is more than one line below cl::opt.
2022-06-02[IR] Enable opaque pointers by defaultNikita Popov5-8/+8
This enabled opaque pointers by default in LLVM. The effect of this is twofold: * If IR that contains *neither* explicit ptr nor %T* types is passed to tools, we will now use opaque pointer mode, unless -opaque-pointers=0 has been explicitly passed. * Users of LLVM as a library will now default to opaque pointers. It is possible to opt-out by calling setOpaquePointers(false) on LLVMContext. A cmake option to toggle this default will not be provided. Frontends or other tools that want to (temporarily) keep using typed pointers should disable opaque pointers via LLVMContext. Differential Revision: https://reviews.llvm.org/D126689
2022-06-01[Polly] Fix -Wreorder-ctor. NFCFangrui Song1-1/+1
2022-06-01[Polly] Migrate -polly-mse to the new pass manager.Yang Keao16-325/+454
This patch implements the `MaximalStaticExpansion` and its printer in NPM. Reviewed By: Meinersbur Differential Revision: https://reviews.llvm.org/D125870
2022-06-01[ValueTracking] Enable -branch-on-poison-as-ub by defaultNikita Popov1-1/+1
Now that SimpleLoopUnswitch and other transforms no longer introduce branch on poison, enable the -branch-on-poison-as-ub option by default. The practical impact of this is mostly better flag preservation in SCEV, and some freeze instructions no longer being necessary. Differential Revision: https://reviews.llvm.org/D125299
2022-05-26[Polly][Test] Fix race condition while printing dot files.Michael Kruse2-4/+14
The tests dot-scops.ll and dot-scops-npm.ll both wrote to the same file scops.func.dot. If they are executed in parallel they will race for the file. Fix by renaming func to func_npm in dot-scops-npm.ll so this test writes dot scops.func_npm.dot. Long-term, we will probably pass a file name (prefix) to the printer pass such that we can use the guaranteed-unique LIT %t placeholder in tests.
2022-05-26[FileCheck] GetCheckTypeAbbreviation() to handle the misspelled case.Ivan Kosarev1-9/+9
Also fix directives not covered by D125604.
2022-05-26[FileCheck] Catch missspelled directives.Ivan Kosarev1-29/+29
Reviewed By: MaskRay Differential Revision: https://reviews.llvm.org/D125604
2022-05-19[lit] Fix setup of sanitizer environmentVitaly Buka1-4/+13
Not all options were propageted into tests. Reviewed By: ychen Differential Revision: https://reviews.llvm.org/D122869
2022-05-19[APInt] Remove all uses of zextOrSelf, sextOrSelf and truncOrSelfJay Foad1-1/+1
Most clients only used these methods because they wanted to be able to extend or truncate to the same bit width (which is a no-op). Now that the standard zext, sext and trunc allow this, there is no reason to use the OrSelf versions. The OrSelf versions additionally have the strange behaviour of allowing extending to a *smaller* width, or truncating to a *larger* width, which are also treated as no-ops. A small amount of client code relied on this (ConstantRange::castOp and MicrosoftCXXNameMangler::mangleNumber) and needed rewriting. Differential Revision: https://reviews.llvm.org/D125557
2022-05-17[Polly] Mark classes as final by default. NFC.Michael Kruse65-208/+205
This make is obivious that a class was not intended to be derived from. NPM analysis pass can unfortunately not marked as final because they are derived from a llvm::Checker<T> template internally by the NPM. Also normalize the use of classes/structs * NPM passes are structs * Legacy passes are classes * structs that have methods and are not a visitor pattern are classes * structs have public inheritance by default, remove "public" keyword * Use typedef'ed type instead of inline forward declaration
2022-05-16[polly] Remove 'using namespace llvm/polly' from ScopGraphPrinter.h header.Simon Pilgrim1-22/+25
As mentioned on D123678 this appears to be causing namespace resolution issues on some versions of gcc.
2022-05-09[polly] Fix type in function name. NFC.Michael Kruse1-3/+3
generated by cgit v1.1 at 2024-07-18 13:43:17 +0000