profile
viewpoint

tensorflow/mlir 1637

"Multi-Level Intermediate Representation" Compiler Infrastructure

joker-eph/mlir 0

"Multi-Level Intermediate Representation" Compiler Infrastructure

jpienaar/charted 0

Visualization toolkit for Dart language

jpienaar/chroma 0

A general purpose syntax highlighter in pure Go

jpienaar/delve 0

Delve is a debugger for the Go programming language.

jpienaar/hugo 0

The world’s fastest framework for building websites.

jpienaar/inc 0

an incremental approach to compiler construction

jpienaar/linguist 0

Language Savant. If your repository's language is being reported incorrectly, send us a pull request!

jpienaar/mlir 0

"Multi-Level Intermediate Representation" Compiler Infrastructure

CommitCommentEvent
CommitCommentEvent

push eventllvm/mlir-www

Jacques Pienaar

commit sha 14551af4f9e666c6188332236df844f5c57cfbdf

Reduce to 0.64.1

view details

push time in a day

push eventllvm/mlir-www

Jacques Pienaar

commit sha 02dd0bf6b7dcdab3179fd51a70bf6a28576f00df

Switch hugo version to fixed one again

view details

push time in a day

push eventllvm/llvm-project

Jacques Pienaar

commit sha fa7d04a0d3f40d1b9997775f24e9c6032011e15f

[mlir] Add short readme.txt to docs directory Summary: Refer folks to the main website and make it explicit that the rendered output is what is of interest and that the GitHub viewing experience may not match (even though we are trying to keep it as close as possible, the renderers do differ). Differential Revision: https://reviews.llvm.org/D74739

view details

push time in 4 days

push eventllvm/mlir-www

Jacques Pienaar

commit sha e1b1315cb49ea6d36f6e4aebd6b7c88c47f92753

Only exclude matching title, comments in code could also have same form but very unlikely exact same string

view details

push time in 4 days

push eventllvm/llvm-project

Jacques Pienaar

commit sha 1842fd50d2ff66209cee11268ddf4bee28c482f7

[mlir] Fix multiple titles We have one title in every doc which corresponds to `#`, in the some there are multiple and it is expected to be h1 headers (visual elements rather than organizational). Indent every nesting by one in all of the docs with multiple titles. Also fixing trailing whitespace.

view details

push time in 4 days

push eventllvm/llvm-project

Jacques Pienaar

commit sha 7baf2a434c8675752efeb185984ca01dbc03f7a4

[mlir] Start Shape dialect * Add basic skeleton for Shape dialect; * Add description of types and ops to be used; Differential Revision: https://reviews.llvm.org/D73944

view details

push time in 10 days

push eventllvm/mlir-www

Jacques Pienaar

commit sha aa541bb35e4563703b0d12305fcb31064bd8d008

Update openprojects.md

view details

push time in 15 days

push eventllvm/llvm-project

Jacques Pienaar

commit sha 2697e8bc1e63d48fca2094f136d8ec98ff020b83

[mlir] Update generic op ebnf to include region Summary: Optional regions are supported in the generic op print/parse form, update the docs to match. Differential Revision: https://reviews.llvm.org/D74061

view details

push time in 17 days

push eventllvm/llvm-project

Jacques Pienaar

commit sha 28cd54cdbba699707df39762f7f2095ed812f35e

[mlir] Fix clang 5 warning for missing braces

view details

push time in 18 days

push eventllvm/mlir-www

Jacques Pienaar

commit sha dcc0d9bdd8d7a0f3d2ccbe32af80965aa2edc0df

Update openprojects.md

view details

push time in 19 days

push eventllvm/mlir-www

Jacques Pienaar

commit sha 92f3897113bf471729b35022075d80c71917524b

Convert [TOC] to toc shortcode Prepend <p/> as that appears to render better (else too little vspace), the problem with render was due to bug that was resolved in newer hugo.

view details

push time in 20 days

push eventllvm/mlir-www

Jacques Pienaar

commit sha 09bfaa2fc60a5d40ccc801c1a707087fac417c1a

Remove old style indicator in codeblock

view details

Jacques Pienaar

commit sha 6e31f25bb4b1f43b4617e21b3bfc6380c1c36753

Add toc shortcode, only seems to work with blackfriday renderer though

view details

push time in 20 days

push eventllvm/llvm-project

Jacques Pienaar

commit sha c4b4c0c47c992e63fc0792d71bea82b6e6dcf0cd

[mlir] Expand shape functions in ShapeInference doc Summary: Start filling in some requirements for the shape function descriptions that will be used to derive shape computations. This requiement part may later be reworked to be part of the "context" section of shape dialect. Without examples this may be a bit too abstract but I hope not (given mappings to existing shape functions). Differential Revision: https://reviews.llvm.org/D73572

view details

push time in 20 days

push eventllvm/llvm-project

Jacques Pienaar

commit sha 1544cf2d7cecf61cf3a7d309b7a9fc92ef77c34a

[mlir] Fix errors in release & no-assert Seen on gcc 8, in release mode & assertions off warnings about logger, made all statements referencing logger inside LLVM_DEBUG blocks and ifdef a few variables only used in debug. This is mechanical fix to get CI green.

view details

push time in 21 days

pull request commenttensorflow/tensorflow

[MLIR] Swap target names MlirOptLib and MlirOptMain

(I could also do it if you prefer)

marbre

comment created time in 23 days

pull request commentcgo-conference/cgo2020

Update c4ml website url

Ping?

jpienaar

comment created time in 23 days

Pull request review commenttensorflow/tensorflow

[MLIR] Swap target names MlirOptLib and MlirOptMain

 cc_library(  # TODO(jpienaar): Update this. cc_library(-    name = "MlirOptLib",+    name = "MlirOptMain",

This one is in the lib directory while the other one is the main entry point, fix the cmake file instead.

marbre

comment created time in 23 days

pull request commentgoogle/iree

Modify mappings in MLIR_EXPLICIT_TARGET_MAPPING

Finally one should be generated from the other or foreign build rule used bazel side. For now updating cmake ones to match bazel ones would be a welcome contribution. There is more refactoring afoot in the cmake files but I think they would be independent of the renaming.

marbre

comment created time in a month

Pull request review commenttensorflow/tensorflow

[MLIR][XLA] Add HLO FuncOp to LHLO FuncOp legalization.

 class HloToLhloReturnConverter : public OpConversionPattern<xla_hlo::ReturnOp> { //   return //  } // }-struct HloLegalizeToLhlo : public FunctionPass<HloLegalizeToLhlo> {-  void runOnFunction() override {+struct HloLegalizeToLhlo : public ModulePass<HloLegalizeToLhlo> {+  void runOnModule() override {     OwningRewritePatternList patterns;-    ConversionTarget target(getContext());+    auto& context = getContext();+    ConversionTarget target(context);     target.addLegalDialect<xla_lhlo::XlaLhloDialect>();+    target.addLegalDialect<StandardOpsDialect>();+    target.addIllegalOp<mlir::ReturnOp>();+    target.addIllegalOp<mlir::TensorLoadOp>();+    target.addIllegalOp<mlir::TensorStoreOp>();+    target.addLegalOp<ModuleTerminatorOp>();+    target.addIllegalDialect<xla_hlo::XlaHloDialect>();+    target.addDynamicallyLegalOp<FuncOp>([&](FuncOp op) {+      bool result = true;+      for (auto& input : op.getType().getInputs()) {+        result &= input.isa<MemRefType>();+      }+      return result;+    });++    auto module = getModule();+    populateHLOToLHLOConversionPattern(module.getContext(), &patterns); -    auto func = getFunction();-    populateHLOToLHLOConversionPattern(func.getContext(), &patterns);-    if (failed(applyPartialConversion(func, target, patterns, nullptr))) {+    SmallVector<Operation*, 16> ops;+    for (auto& op : module.getBody()->getOperations()) {+      ops.push_back(&op);+    }+    if (failed(applyFullConversion(ops, target, patterns, nullptr))) {       signalPassFailure();     }   } }; +Type convertType(Type t) {

Function naming convention (unless overloaded from base class) should follow https://google.github.io/styleguide/cppguide.html#Function_Names

dfki-ehna

comment created time in a month

Pull request review commenttensorflow/tensorflow

[MLIR][XLA] Add HLO FuncOp to LHLO FuncOp legalization.

 struct HloToLHloReduceConverter       rewriter.replaceUsesOfBlockArgument(old_arg, new_arg);     }     // Add an argument for the result.-    entry_block.addArgument(+    auto resultArg = entry_block.addArgument(         entry_block.getArgument(original_arg_count).getType());     // Remove the old arguments.     for (int i = original_arg_count - 1; i >= 0; --i) {       entry_block.eraseArgument(i);     }-    // Insert terminator at the end.-    rewriter.setInsertionPointToEnd(&entry_block);-    rewriter.create<xla_lhlo::TerminatorOp>(loc);++    auto terminatorOp = entry_block.getTerminator();

The coding style here (excluding overloaded functions) follows lower snake case, so result_arg and terminator_op.

dfki-ehna

comment created time in a month

Pull request review commenttensorflow/tensorflow

[MLIR][XLA] Add HLO FuncOp to LHLO FuncOp legalization.

 struct HloToLHloReduceConverter       rewriter.replaceUsesOfBlockArgument(old_arg, new_arg);     }     // Add an argument for the result.-    entry_block.addArgument(+    auto resultArg = entry_block.addArgument(         entry_block.getArgument(original_arg_count).getType());     // Remove the old arguments.     for (int i = original_arg_count - 1; i >= 0; --i) {       entry_block.eraseArgument(i);     }-    // Insert terminator at the end.-    rewriter.setInsertionPointToEnd(&entry_block);-    rewriter.create<xla_lhlo::TerminatorOp>(loc);++    auto terminatorOp = entry_block.getTerminator();+    if (terminatorOp->getNumOperands() > 0) {+      auto operand = terminatorOp->getOperand(0);+      OpBuilder::InsertionGuard guard(rewriter);+      rewriter.setInsertionPoint(terminatorOp);+      rewriter.create<xla_lhlo::CopyOp>(loc, llvm::None, operand, resultArg);

Why is this only inserting a copy for one operand?

dfki-ehna

comment created time in a month

Pull request review commenttensorflow/tensorflow

[MLIR][XLA] Add HLO FuncOp to LHLO FuncOp legalization.

 class HloToLhloReturnConverter : public OpConversionPattern<xla_hlo::ReturnOp> { //   return //  } // }-struct HloLegalizeToLhlo : public FunctionPass<HloLegalizeToLhlo> {-  void runOnFunction() override {+struct HloLegalizeToLhlo : public ModulePass<HloLegalizeToLhlo> {+  void runOnModule() override {     OwningRewritePatternList patterns;-    ConversionTarget target(getContext());+    auto& context = getContext();+    ConversionTarget target(context);     target.addLegalDialect<xla_lhlo::XlaLhloDialect>();+    target.addLegalDialect<StandardOpsDialect>();+    target.addIllegalOp<mlir::ReturnOp>();+    target.addIllegalOp<mlir::TensorLoadOp>();+    target.addIllegalOp<mlir::TensorStoreOp>();+    target.addLegalOp<ModuleTerminatorOp>();+    target.addIllegalDialect<xla_hlo::XlaHloDialect>();+    target.addDynamicallyLegalOp<FuncOp>([&](FuncOp op) {+      bool result = true;+      for (auto& input : op.getType().getInputs()) {+        result &= input.isa<MemRefType>();

How about std::all_of ?

dfki-ehna

comment created time in a month

Pull request review commenttensorflow/tensorflow

[MLIR][XLA] Add HLO FuncOp to LHLO FuncOp legalization.

 class HloToLhloReturnConverter : public OpConversionPattern<xla_hlo::ReturnOp> { //   return //  } // }-struct HloLegalizeToLhlo : public FunctionPass<HloLegalizeToLhlo> {-  void runOnFunction() override {+struct HloLegalizeToLhlo : public ModulePass<HloLegalizeToLhlo> {+  void runOnModule() override {     OwningRewritePatternList patterns;-    ConversionTarget target(getContext());+    auto& context = getContext();+    ConversionTarget target(context);     target.addLegalDialect<xla_lhlo::XlaLhloDialect>();+    target.addLegalDialect<StandardOpsDialect>();+    target.addIllegalOp<mlir::ReturnOp>();+    target.addIllegalOp<mlir::TensorLoadOp>();+    target.addIllegalOp<mlir::TensorStoreOp>();+    target.addLegalOp<ModuleTerminatorOp>();+    target.addIllegalDialect<xla_hlo::XlaHloDialect>();+    target.addDynamicallyLegalOp<FuncOp>([&](FuncOp op) {+      bool result = true;+      for (auto& input : op.getType().getInputs()) {+        result &= input.isa<MemRefType>();+      }+      return result;+    });++    auto module = getModule();+    populateHLOToLHLOConversionPattern(module.getContext(), &patterns); -    auto func = getFunction();-    populateHLOToLHLOConversionPattern(func.getContext(), &patterns);-    if (failed(applyPartialConversion(func, target, patterns, nullptr))) {+    SmallVector<Operation*, 16> ops;+    for (auto& op : module.getBody()->getOperations()) {+      ops.push_back(&op);+    }+    if (failed(applyFullConversion(ops, target, patterns, nullptr))) {

What error do you get with applyFullConversion(module.getBody()->getOperations(), ... ?

dfki-ehna

comment created time in a month

push eventllvm/mlir-www

Jacques Pienaar

commit sha 47a1c8f93e563e4103751888ce27fe964ee18d48

Install doxygen before trying to use it

view details

push time in a month

push eventllvm/mlir-www

Jacques Pienaar

commit sha feb3cf9760e31af576a2b4d996aa0624db26cd43

Remove bool designator from cmake command line

view details

push time in a month

push eventllvm/mlir-www

Jacques Pienaar

commit sha 714f066f2e02255a31f71865b85dc7b3fc3f280d

Fix missing dash

view details

push time in a month

push eventllvm/llvm-project

Jacques Pienaar

commit sha e47b56139885e727c22ad212605a1e227755f799

[mlir] Revert MSVC specific part of whole_archive_link Revert the MSVC specific parts in whole_archive_link to previous form to potentially address https://bugs.llvm.org/show_bug.cgi?id=44660.

view details

push time in a month

push eventllvm/mlir-www

Jacques Pienaar

commit sha 84c04e417f44a6ee8405062d4898951c0b49f963

Add doxygen generation

view details

push time in a month

push eventllvm/llvm-project

Jacques Pienaar

commit sha e298e216501abf38b44e690d2b28fc788ffc96cf

[mlir] Bootstrap doxygen config Add basic doxygen config following clang and llvm example with minimal changes.

view details

push time in a month

PR opened cgo-conference/cgo2020

Update c4ml website url

For unknown reasons the https entry is returning an error, while if one goes to the http variant it works and ends up at this up at this updated https one. I've not dug into this at all, but updating so that link works.

+2 -2

0 comment

1 changed file

pr created time in a month

push eventjpienaar/cgo2020

Jacques Pienaar

commit sha d92fc2d22a2771b38fb14405a466d98d4dac03e2

Update c4ml website url For unknown reasons the https entry is returning an error, while if one goes to the http variant it works and ends up at this up at this updated https one. I've not dug into this at all, but updating so that link works.

view details

push time in a month

push eventllvm/llvm-project

Jacques Pienaar

commit sha 178562fb352d12b5235c63f75297b8a456b53a0f

[mlir] Enable specifying verify on OpInterface Summary: Add method in ODS to specify verification for operations implementing a OpInterface. Use this with infer type op interface to verify that the inferred type matches the return type and remove special case in TestPatterns. This could also have been achieved by using OpInterfaceMethod but verify seems pretty common and it is not an arbitrary method that just happened to be named verifyTrait, so having it be defined in special way seems appropriate/better documenting. Differential Revision: https://reviews.llvm.org/D73122

view details

push time in a month

push eventllvm/llvm-project

Jacques Pienaar

commit sha b70e4efb75bf1670fa335f3211c05b26b060ffde

[mlir] Generalize broadcastable trait operands Summary: Generalize broadcastable trait to variadic operands. Update the documentation that still talked about element type as part of broadcastable trait (that bug was already fixed). Also rename Broadcastable to ResultBroadcastableShape to be more explicit that the trait affects the result shape (it is possible for op to allow broadcastable operands but not have result shape that is broadcast compatible with operands). Doing some intermediate work to have getBroadcastedType take an optional elementType as input and use that if specified, instead of the common element type of type1 and type2 in this function. Differential Revision: https://reviews.llvm.org/D72559

view details

push time in a month

push eventllvm/mlir-www

Jacques Pienaar

commit sha 3bc9ff3a8ecb517d66e12c8fd0835a52f68759a3

Use hugo latest

view details

push time in a month

Pull request review commenttensorflow/tensorflow

[ROCm] Unit-test updates for the ROCm platform.

 TEST_F(ConstantFoldingTest, ConstantPushDownBiasAdd) { TEST_F(ConstantFoldingTest, MulConvPushDownTest_Conv2D_ScalarConst) {   for (string data_format : {          "NHWC",-#if GOOGLE_CUDA+#if GOOGLE_CUDA || TENSORFLOW_USE_ROCM

We can also address this in follow up, it might likely be there is not something that exactly first this (esp as Christian was happy with the change).

deven-amd

comment created time in a month

Pull request review commenttensorflow/tensorflow

[ROCm] Unit-test updates for the ROCm platform.

 glob_lit_tests(         ":test_utilities",     ],     driver = "@llvm-project//mlir:run_lit.sh",+    tags_override = {

That's reasonable yes.

deven-amd

comment created time in a month

push eventllvm/llvm-project

Jacques Pienaar

commit sha fa26a37d36991dd19276c2112ae03e962debe0e4

[mlir] Add shaped container component type interface Summary: * Add shaped container type interface which allows infering the shape, element type and attribute of shaped container type separately. Show usage by way of tensor type inference trait which combines the shape & element type in infering a tensor type; - All components need not be specified; - Attribute is added to allow for layout attribute that was previously discussed; * Expand the test driver to make it easier to test new creation instances (adding new operands or ops with attributes or regions would trigger build functions/type inference methods); - The verification part will be moved out of the test and to verify method instead of ops implementing the type inference interface in a follow up; * Add MLIRContext as arg to possible to create type for ops without arguments, region or location; * Also move out the section in OpDefinitions doc to separate ShapeInference doc where the shape function requirements can be captured; - Part of this would move to the shape dialect and/or shape dialect ops be included as subsection of this doc; * Update ODS's variable usage to match camelBack format for builder, state and arg variables; - I could have split this out, but I had to make some changes around these and the inconsistency bugged me :) Differential Revision: https://reviews.llvm.org/D72432

view details

push time in a month

Pull request review commenttensorflow/tensorflow

[ROCm] Unit-test updates for the ROCm platform.

 TEST_F(ConstantFoldingTest, ConstantPushDownBiasAdd) { TEST_F(ConstantFoldingTest, MulConvPushDownTest_Conv2D_ScalarConst) {   for (string data_format : {          "NHWC",-#if GOOGLE_CUDA+#if GOOGLE_CUDA || TENSORFLOW_USE_ROCM

@rmlarsen is there some joint declaration this could be guarded on? (E.g., TF_USE_GPU?) Or an attribute? (E.g., SUPPORTS_NCHW ? - I'm assuming there is not something that specific :) )

deven-amd

comment created time in a month

Pull request review commenttensorflow/tensorflow

[ROCm] Unit-test updates for the ROCm platform.

 package(licenses = ["notice"]) glob_lit_tests(     data = [":test_utilities"],     driver = "@llvm-project//mlir:run_lit.sh",+    tags_override = {

Do you perhaps have issues you could attach to these as to why they are disabled for rocm?

deven-amd

comment created time in a month

Pull request review commenttensorflow/tensorflow

[ROCm] Unit-test updates for the ROCm platform.

 glob_lit_tests(         ":test_utilities",     ],     driver = "@llvm-project//mlir:run_lit.sh",+    tags_override = {

Do you expect rocm to be used tflite side? if not, making it a default tag for all the tests might be better.

deven-amd

comment created time in a month

push eventllvm/mlir-www

Matthew Parkinson

commit sha da77547e6fac66d4b56a42693c2a88aa953e25e6

Typo in MLIR Getting Started guide (#6) Missing "S" from Windows instructions.

view details

push time in a month

PR merged llvm/mlir-www

Typo in MLIR Getting Started guide

Missing "S" from Windows instructions.

+1 -1

0 comment

1 changed file

mjp41

pr closed time in a month

pull request commenttensorflow/tensorflow

[XLA] Update the path for the LLVM FileCheck executable

@cheshire the other changes don't overlap with this, I expected that they would and might still be pending, but we do need to update this too, so LG.

deven-amd

comment created time in a month

pull request commenttensorflow/tensorflow

[XLA] Update the path for the LLVM FileCheck executable

I believe an internal change is in the works (look at my review queue internally) that may intersect with this one [AFK] and should resolved this too. I believe it is being submitted as we speak.

Not sure why it didn't trigger on the bots.

On Fri, Jan 10, 2020, 12:09 PM George Karpenkov notifications@github.com wrote:

@jpienaar https://github.com/jpienaar Do you think this commit is necessary? Apparently, the previous path has stopped working after d6ba353 https://github.com/tensorflow/tensorflow/commit/d6ba353dd974f3d883734e96653cdee7fe6abfbc . I wonder then how come our OSS bots are still passing.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/tensorflow/tensorflow/pull/35751?email_source=notifications&email_token=AAFMRTXXHH7NTCW6CQMUOB3Q5DIXHA5CNFSM4KFMRPNKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEIVCHGY#issuecomment-573186971, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAFMRTRSJR6ROTN66EGONNLQ5DIXHANCNFSM4KFMRPNA .

deven-amd

comment created time in a month

startedDoctorWkt/acwj

started time in a month

startedpapers-we-love/papers-we-love

started time in 2 months

push eventllvm/mlir-www

Jacques Pienaar

commit sha 85523e23f7fda2ace5f7b7f9e0d6441b3d428b21

Update talks with talks found on YouTube * Make all the tech talk descriptions follow the same format * Add upcomming talks section to allow highlighting upcoming talks

view details

push time in 2 months

push eventllvm/mlir-www

Jacques Pienaar

commit sha f4958c862d675ea300507bae0f41a3929e631698

Split out open design meeting presentation * Split out open design meeting * Add link to ODM

view details

push time in 2 months

push eventllvm/mlir-www

Jacques Pienaar

commit sha 81046b9be70384c6fd19205b73086992b1dc4a1b

Disable showing update and publish date until we sensible dates there

view details

push time in 2 months

push eventllvm/mlir-www

Jacques Pienaar

commit sha 9c6a12384669c662c4af2346f8450f72cb5d7d00

Avoid showing edit link for pages from llvm monorepo

view details

push time in 2 months

push eventllvm/mlir-www

Stella Laurenzo

commit sha 6338336b765a8211cf4ffc93b9211217a06c5d3f

Add IREE to users of MLIR page. (#4)

view details

push time in 2 months

PR merged llvm/mlir-www

Add IREE to users of MLIR page.
+4 -0

0 comment

1 changed file

stellaraccident

pr closed time in 2 months

push eventllvm/mlir-www

Jacques Pienaar

commit sha 1789251998249a41ae251f11cbce21aeb222f497

Remove else cases

view details

push time in 2 months

push eventllvm/mlir-www

Jacques Pienaar

commit sha 2017f93862a04807e1708e0e1b2ce7ec0c731804

Bump hugo versio and use MD render hook

view details

push time in 2 months

issue commentbep/portable-hugo-links

Example using anchors?

Sounds good, thanks I was uncertain if it was supported but some of my global configs were stopping it (I've only used Hugo for a couple of days, so all very new still :)). Very nice to have this feature.

jpienaar

comment created time in 2 months

issue openedbep/portable-hugo-links

Example using anchors?

I was trying use render-links example here with anchors, but that doesn't seem to work. For the regular page the relative URL works (yay), but not for those using anchors. Looking at what gets generated I see a nopPage for .relPermalink returned for the link with anchor.

I'm not sure if there may be some config that is wrong my side, or if relative links with anchors need to be specially handled.

created time in 2 months

push eventllvm/llvm-project

Jacques Pienaar

commit sha 8f1caf518f374b452fdff22dcd56bcc12c900338

[mlir] Only generate default for uncovered cases Have to explicitly check if all cases are covered instead.

view details

push time in 2 months

push eventllvm/llvm-project

Jacques Pienaar

commit sha f533fa3af5ef8e0805593f58eb15ddf307caf5fa

[mlir] Revert default case that was needed This one isn't always complete.

view details

push time in 2 months

push eventllvm/llvm-project

Jacques Pienaar

commit sha 98a88c1beec5beff03aa416803c14ae8800322ad

[mlir] Add parenthesis to avoid -Wparentheses

view details

push time in 2 months

push eventllvm/llvm-project

Jacques Pienaar

commit sha 3d83d8259c1b5c2f6a847158cf3c3482a4b032ee

[mlir] Remove redudant default cases These provide no value and trigger -Wcovered-switch-default.

view details

push time in 2 months

push eventllvm/mlir-www

Jacques Pienaar

commit sha 644dd0e4df8e1b2672bef23b5d27939c701ffbf5

Update link in _index to use relref too

view details

push time in 2 months

push eventllvm/mlir-www

Jacques Pienaar

commit sha d5c6e357a609bbab43129c8a875d71a83be1ef37

Update sed to change entire line rather

view details

push time in 2 months

push eventllvm/mlir-www

Jacques Pienaar

commit sha 46bb826a9bbcb456f5ab66e4077b8d1f6e406d3a

Update links to use relref Also use regex* as stopgap to generate relref of links from main page. * This will surely be fragile but would avoid all the current broken links as temporary measure.

view details

push time in 2 months

push eventllvm/llvm-project

Jacques Pienaar

commit sha 7544cb8807b2ecb2f5186ab18d0ffa865a493a47

[mlir][docs] Remove redundant path prefix ./ is not needed.

view details

push time in 2 months

push eventllvm/llvm-project

Jacques Pienaar

commit sha 430bba2a0f39b543e66119761d1526d037229936

[mlir] Make code blocks more consistent Use the same form specification for the same type of code.

view details

push time in 2 months

push eventllvm/mlir-www

pkanwar23

commit sha a0cd1eb2689221b5635f22efceb5202973f54d37

fixing links

view details

push time in 2 months

PR merged llvm/mlir-www

fixing links
+2 -2

0 comment

1 changed file

pkanwar23

pr closed time in 2 months

issue commenttensorflow/tensorflow

[MLIR] tf-opt can not run with -debug option

In InitMlir we parse the flags with the two different command libraries in use: if no -- is on the line, then all are passed to LLVM's cl, if -- then the part before -- is passed to LLVM and those after to TF's flags. Try

tf-opt -debug --

Please reopen if this doesn't resolve it.

qqsun8819

comment created time in 2 months

issue closedtensorflow/tensorflow

[MLIR] tf-opt can not run with -debug option

System information

  • Have I written custom code (as opposed to using a stock example script provided in TensorFlow):No
  • OS Platform and Distribution (e.g., Linux Ubuntu 16.04):Linux Ubuntu 18.04
  • Mobile device (e.g. iPhone 8, Pixel 2, Samsung Galaxy) if the issue happens on mobile device:No
  • TensorFlow installed from (source or binary):source
  • TensorFlow version (use command below): 2.0.0
  • Python version:3.7
  • Bazel version (if compiling from source):1.0.1
  • GCC/Compiler version (if compiling from source):6.4.0
  • CUDA/cuDNN version: 10
  • GPU model and memory: V100

Describe the current behavior in current tensorflow dialect (implemented in tensorflow/compiler/mlir), we use LLVM_DEBUG in various place to print some debug information during compilation , and if we run tf-opt , it's expected to print these informations with -debug option added, this is also the behavior with mlir-opt, but in current master branch , if I pass a -debug option to tf-opt ,an error occured

#bazel-bin/tensorflow/compiler/mlir/tf-opt -debug
tf-opt: Unknown command line argument '-debug'.  Try: 'bazel-bin/tensorflow/compiler/mlir/tf-opt --help'
tf-opt: Did you mean '--help'?

but if I run mlir-opt with -debug:

#bin/mlir-opt -debug
Args: bin/mlir-opt -debug

I've also compared the main function between tf-opt(https://github.com/tensorflow/tensorflow/blob/master/tensorflow/compiler/mlir/tf_mlir_opt_main.cc ) and mlir-opt (https://github.com/tensorflow/mlir/blob/master/tools/mlir-opt/mlir-opt.cpp ) and see no apparent difference between the two files

Describe the expected behavior

tf-opt can run with -debug so I can use the debug info to debug the compilation process

closed time in 2 months

qqsun8819

PR opened alecthomas/chroma

Update languages supported
+2 -2

0 comment

1 changed file

pr created time in 2 months

push eventjpienaar/chroma

Jacques Pienaar

commit sha 079cb85b19ee7dcec2437c05eb8fd0d6ffee231f

Update languages supported

view details

push time in 2 months

issue commentalecthomas/chroma

Updating ebnf syntax to match linguists?

Oops, and I was logged into the wrong account. Which would explain why I didn't get the notification either.

jpienaar

comment created time in 2 months

push eventllvm/mlir-www

Jacques Pienaar

commit sha cd2b4a509db4211251133ce447e058ce6b92bbff

Use slide menu instead

view details

push time in 2 months

push eventllvm/mlir-www

Jacques Pienaar

commit sha c528e0938094567307786fbda2ba1eb1c970b0ed

Expand gitignore to skip llvm_src and website/content/docs/

view details

push time in 2 months

push eventllvm/mlir-www

Jacques Pienaar

commit sha bc29269e3fb11ff6ec12db552f8e0e80ec541204

Skip directories where no md files found

view details

push time in 2 months

PR closed tensorflow/tensorflow

Reviewers
[mlir] cla: yes size:S

1.the type of op.getIndex() is unsigned, the value is always greater than or equal to 0 2.the function hasConstantoffsetSizesAndStrides is defined but not used

+1 -16

2 comments

2 changed files

leike666666

pr closed time in 2 months

pull request commenttensorflow/tensorflow

[mlir]

Please make a PR on https://github.com/tensorflow/mlir instead for these (third_party/mlir is a read-only copy of that repo here that will be removed).

leike666666

comment created time in 2 months

push eventjoker-eph/www-mlir

Jacques Pienaar

commit sha fce47cead3434e639b8688b07fae0ffd86bf1e20

Update config.toml

view details

push time in 2 months

issue openedalecthomas/chroma

Updating ebnf syntax to match linguists?

What problem does this feature solve?

Currently the syntax for EBNF in chroma differs from what GitHub linguist supports, this is to find out feasibility of getting these to match closer. This would allow using the same file, highlighted similarly when served from GitHub or Hugo.

What feature do you propose?

Update the EBNF grammer to match what GitHub does (https://github.com/Alhadis/language-grammars).

I'm not sure how many folks depends on the current EBNF grammar here.

created time in 2 months

PR opened gohugoio/hugo

deps: Update chroma

Updating chroma module to head to include some new language support added in chroma. Tried removing unused parts from go.sum but v0.7.0/go.mod gets re-added but seems to get added back when I run go test ./... and I'm not sure why.

Saw hugo does allow for packages to be referenced in between releases, but not sure if there is any other criteria.

+3 -2

0 comment

2 changed files

pr created time in 3 months

push eventjpienaar/hugo

Jacques Pienaar

commit sha 4fb912ba3bfc840c33099d0952ef8a8521c8e080

Update chroma to head Updating to include some new language support added. Saw hugo does allow for packages to be referenced in between releases, but not sure if there is any other criteria.

view details

push time in 3 months

push eventjpienaar/hugo

Jacques Pienaar

commit sha 9c7ee17b9abff189ba1af8e3ae3218129bb47b91

Update chroma to head Updating to include some new language support added. Saw hugo does allow for packages to be referenced in between releases, but not sure if there is any other criteria.

view details

push time in 3 months

push eventjpienaar/hugo

Jacques Pienaar

commit sha d38294272c74ac1221b7321f97eed3b744f872c3

Remove unneeded entry from go.sum

view details

push time in 3 months

create barnchjpienaar/hugo

branch : update_chroma

created branch time in 3 months

issue commenttensorflow/mlir

Simple lowering to LLVM

Hey, have you tried tensor_load/tensor_store ops here? That seems to do what you want.

doru1004

comment created time in 3 months

PR opened alecthomas/chroma

Add lexer for LLVM TableGen

Add a basic lexer for LLVM's TableGen (https://llvm.org/docs/TableGen/LangRef.html). Used an example from a tutorial (modulo verifying an additional template highlighting) for testdata and I just chose the mime type arbitrarily.

+171 -0

0 comment

3 changed files

pr created time in 3 months

delete branch jpienaar/chroma

delete branch : patch-1

delete time in 3 months

delete branch jpienaar/chroma

delete branch : mlir

delete time in 3 months

push eventjpienaar/chroma

jpienaar

commit sha 5a82e26836c81371561d0573b80228e5eb0a0db4

Update tablegen.go

view details

push time in 3 months

create barnchjpienaar/chroma

branch : tablegen

created branch time in 3 months

fork jpienaar/hugo

The world’s fastest framework for building websites.

https://gohugo.io

fork in 3 months

PR opened alecthomas/chroma

Avoid empty matches in MLIR lexer

Ensure at least one character is matched else lexing ends up making no forward progress.

+1 -1

0 comment

1 changed file

pr created time in 3 months

push eventjpienaar/chroma

jpienaar

commit sha 7245a309cc1b2d0f0b23151e4816cca2c1471ec6

Avoid empty matches in MLIR lexer Ensure at least one character is matched else lexing ends up making no forward progress.

view details

push time in 3 months

push eventjpienaar/chroma

Jacques Pienaar

commit sha a899a7f33563b21c8d874ce80b90814559528c05

Avoid empty matches in MLIR lexer Else lexing ends up making no forward progress.

view details

push time in 3 months

fork jpienaar/delve

Delve is a debugger for the Go programming language.

fork in 3 months

push eventjoker-eph/www-mlir

Jacques Pienaar

commit sha 69030dfe1716f24a9ea2f197f8dfffbbbe52d310

Add shortcode to include MD file in another MD file

view details

Jacques Pienaar

commit sha 9170103f3614627d1c782e2fb85037f6fa979377

Center image with text in header

view details

Jacques Pienaar

commit sha 621fabb0d02b492733705dbd12f5aa93a6802762

Add source as menubar item

view details

push time in 3 months

push eventjpienaar/www-mlir

Jacques Pienaar

commit sha 39a601675c0391c36577a7e73da7cd96f8e76fb3

Add content shortcode and align header image with text * Add content shortcode to enable importing one MD doc in another; * Align logo with text; * Move GitHub logo with link to code and make it text link in menu instead;

view details

push time in 3 months

more