[PATCH] D135067: [lit] RUN commands without stdin.

2022-10-04 Thread Christian Sigg via Phabricator via cfe-commits
csigg updated this revision to Diff 465096. csigg added a comment. Herald added a reviewer: jdoerfert. Herald added subscribers: cfe-commits, sstefan1. Herald added a project: clang. Fix clang tests that relied on empty input or accidentally did not provide the input in the RUN directive. Repos

[PATCH] D126158: [MLIR][GPU] Replace fdiv on fp16 with promoted (fp32) multiplication with reciprocal plus one (conditional) Newton iteration.

2022-06-05 Thread Christian Sigg via Phabricator via cfe-commits
csigg added a comment. This diff was recommited in 400fef081adbafc358858709861cdb14303de0e9 . Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D126158/new/ https://reviews.llvm.org/D

[PATCH] D126158: [MLIR][GPU] Replace fdiv on fp16 with promoted (fp32) multiplication with reciprocal plus one (conditional) Newton iteration.

2022-06-04 Thread Christian Sigg via Phabricator via cfe-commits
csigg added a comment. Thanks Mehdi for reverting. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D126158/new/ https://reviews.llvm.org/D126158 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https

[PATCH] D126158: [MLIR][GPU] Replace fdiv on fp16 with promoted (fp32) multiplication with reciprocal plus one (conditional) Newton iteration.

2022-06-03 Thread Christian Sigg via Phabricator via cfe-commits
This revision was automatically updated to reflect the committed changes. Closed by commit rGbcfc0a905101: [MLIR][GPU] Replace fdiv on fp16 with promoted (fp32) multiplication with… (authored by csigg). Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D1

[PATCH] D126158: [MLIR][GPU] Replace fdiv on fp16 with promoted (fp32) multiplication with reciprocal plus one (conditional) Newton iteration.

2022-06-03 Thread Christian Sigg via Phabricator via cfe-commits
csigg updated this revision to Diff 434142. csigg added a comment. Fix. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D126158/new/ https://reviews.llvm.org/D126158 Files: mlir/include/mlir/Dialect/LLVMIR/NVVMOps.td mlir/include/mlir/Dialect/LLV

[PATCH] D126158: [MLIR][GPU] Replace fdiv on fp16 with promoted (fp32) multiplication with reciprocal plus one (conditional) Newton iteration.

2022-06-02 Thread Christian Sigg via Phabricator via cfe-commits
csigg updated this revision to Diff 433772. csigg added a comment. Rename pass. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D126158/new/ https://reviews.llvm.org/D126158 Files: mlir/include/mlir/Dialect/LLVMIR/NVVMOps.td mlir/include/mlir/Dia

[PATCH] D126158: [MLIR][GPU] Replace fdiv on fp16 with promoted (fp32) multiplication with reciprocal plus one (conditional) Newton iteration.

2022-05-30 Thread Christian Sigg via Phabricator via cfe-commits
csigg updated this revision to Diff 432894. csigg added a comment. Fix. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D126158/new/ https://reviews.llvm.org/D126158 Files: mlir/include/mlir/Dialect/LLVMIR/NVVMOps.td mlir/include/mlir/Dialect/LLV

[PATCH] D126158: [MLIR][GPU] Replace fdiv on fp16 with promoted (fp32) multiplication with reciprocal plus one (conditional) Newton iteration.

2022-05-30 Thread Christian Sigg via Phabricator via cfe-commits
csigg updated this revision to Diff 432881. csigg added a comment. Rebase Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D126158/new/ https://reviews.llvm.org/D126158 Files: mlir/include/mlir/Dialect/LLVMIR/NVVMOps.td mlir/include/mlir/Dialect/L

[PATCH] D126158: [MLIR][GPU] Replace fdiv on fp16 with promoted (fp32) multiplication with reciprocal plus one (conditional) Newton iteration.

2022-05-30 Thread Christian Sigg via Phabricator via cfe-commits
csigg updated this revision to Diff 432880. csigg added a comment. Herald added a subscriber: mgorny. Make fdiv rewrite an NVVM transform pass instead. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D126158/new/ https://reviews.llvm.org/D126158 File

[PATCH] D126158: [MLIR][GPU] Replace fdiv on fp16 with promoted (fp32) multiplication with reciprocal plus one (conditional) Newton iteration.

2022-05-30 Thread Christian Sigg via Phabricator via cfe-commits
csigg updated this revision to Diff 432871. csigg added a comment. Rebase. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D126158/new/ https://reviews.llvm.org/D126158 Files: mlir/include/mlir/Dialect/LLVMIR/NVVMOps.td mlir/lib/Conversion/GPUToN

[PATCH] D126158: [MLIR][GPU] Replace fdiv on fp16 with promoted (fp32) multiplication with reciprocal plus one (conditional) Newton iteration.

2022-05-30 Thread Christian Sigg via Phabricator via cfe-commits
csigg added inline comments. Comment at: mlir/lib/Conversion/GPUToNVVM/LowerGpuOpsToNVVMOps.cpp:158 +// by the same divisor. +struct ExpandDivF16 : public ConvertOpToLLVMPattern { + using ConvertOpToLLVMPattern::ConvertOpToLLVMPattern; herhut wrote: > This patte

[PATCH] D126158: [MLIR][GPU] Replace fdiv on fp16 with promoted (fp32) multiplication with reciprocal plus one (conditional) Newton iteration.

2022-05-25 Thread Christian Sigg via Phabricator via cfe-commits
csigg updated this revision to Diff 432074. csigg added a comment. Rebase. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D126158/new/ https://reviews.llvm.org/D126158 Files: mlir/include/mlir/Dialect/LLVMIR/NVVMOps.td mlir/lib/Conversion/GPUToN

[PATCH] D126369: [LLVM] Add rcp.approx.ftz.f32 intrinsic

2022-05-25 Thread Christian Sigg via Phabricator via cfe-commits
This revision was landed with ongoing or failed builds. This revision was automatically updated to reflect the committed changes. Closed by commit rGc4bc416418a2: [LLVM] Add rcp.approx.ftz.f32 intrinsic (authored by csigg). Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https:

[PATCH] D126369: [LLVM] Add rcp.approx.ftz.f32 intrinsic

2022-05-25 Thread Christian Sigg via Phabricator via cfe-commits
csigg added a comment. Thanks Artem. I think I should be able to land it myself. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D126369/new/ https://reviews.llvm.org/D126369 ___ cfe-commits mailing list cf

[PATCH] D126158: [MLIR][GPU] Replace fdiv on fp16 with promoted (fp32) multiplication with reciprocal plus one (conditional) Newton iteration.

2022-05-25 Thread Christian Sigg via Phabricator via cfe-commits
csigg added a comment. In D126158#3534750 , @tra wrote: > I would suggest separating it into separate LLVM and MLIR patches. Thanks Artem. I separated out the LLVM changes in https://reviews.llvm.org/D126369. Repository: rG LLVM Github Monorepo CHA

[PATCH] D126369: [LLVM] Add rcp.approx.ftz.f32 intrinsic

2022-05-25 Thread Christian Sigg via Phabricator via cfe-commits
csigg created this revision. csigg added a reviewer: tra. Herald added subscribers: mattd, gchakrabarti, asavonic, sanjoy.google, hiraditya. Herald added a project: All. csigg requested review of this revision. Herald added subscribers: llvm-commits, cfe-commits, jdoerfert, jholewinski. Herald add

[PATCH] D126158: [MLIR][GPU] Replace fdiv on fp16 with promoted (fp32) multiplication with reciprocal plus one (conditional) Newton iteration.

2022-05-22 Thread Christian Sigg via Phabricator via cfe-commits
csigg created this revision. csigg added a reviewer: bkramer. Herald added subscribers: bzcheeseman, mattd, gchakrabarti, awarzynski, sdasgup3, asavonic, wenzhicui, wrengr, Chia-hungDuan, dcaballe, cota, teijeong, rdzhabarov, tatianashp, msifontes, jurahul, Kayjukh, grosul1, Joonsoo, liufengdb,

[PATCH] D75811: [CUDA] Choose default architecture based on CUDA installation

2020-03-14 Thread Christian Sigg via Phabricator via cfe-commits
csigg added a comment. > I've gone with the approach of trying the architectures in the most recent > non-deprecated order – sm_52, sm_30. I'm curious why you added sm_52 (I'm currently writing bazel rules for better CUDA support, and I'm using just sm_30 because that's been nvcc's default for

[PATCH] D75811: [CUDA] Choose default architecture based on CUDA installation

2020-03-14 Thread Christian Sigg via Phabricator via cfe-commits
csigg added a comment. > I'll be adding a `CUDA_ROOT` option to CMake that will be passed to clang as > `--cuda-path`. I'm not familiar with CMake and whether that option is picked up from an environment variable, but on Windows that environment variable that the CUDA installer sets is `CUDA_P

[PATCH] D72557: Add pretty printers for llvm::PointerIntPair and llvm::PointerUnion.

2020-01-20 Thread Christian Sigg via Phabricator via cfe-commits
csigg added a comment. Hi David. Thanks a lot for you review. I did another pass to switch to factory methods that catch exceptions without spewing common errors on the GDB console. Would you mind taking another quick peek? Thanks a lot for your patience! Repository: rG LLVM Github Monorepo

[PATCH] D72557: Add pretty printers for llvm::PointerIntPair and llvm::PointerUnion.

2020-01-20 Thread Christian Sigg via Phabricator via cfe-commits
csigg updated this revision to Diff 239160. csigg added a comment. Herald added a subscriber: wuzish. Fix base. I'm so terrible at this. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D72557/new/ https://reviews.llvm.org/D72557 Files: debuginfo-te