Page MenuHomePhabricator

JonChesterfield (Jon Chesterfield)
User

Projects

User does not belong to any projects.

User Details

User Since
Aug 11 2015, 1:08 PM (354 w, 2 d)

Recent Activity

Yesterday

JonChesterfield added a comment to D125060: [amdgpu][wip] Implement lds kernel id intrinsic.

Revision just uploaded annotates all kernels with an integer. That doesn't need to land with the intrinsic but it makes it much easier to drive runtime tests through this code so it's there for now.

Thu, May 26, 9:43 AM · Restricted Project, Restricted Project
JonChesterfield updated the diff for D125060: [amdgpu][wip] Implement lds kernel id intrinsic.
  • Rebase on main, append metadata to kernels in lower module lds pass
Thu, May 26, 9:33 AM · Restricted Project, Restricted Project

Mon, May 23

JonChesterfield added a comment to D125970: [amdgpu] Add amdgpu_kernel calling conv attribute to clang.

If it was adding a calling convention, sure - caution warranted. There's no llvm change here though, an existing CC is exposed to C++. No change to the type system either.

This is adding a user-facing calling convention to Clang and it changes the type system as a result. For example, lambda function pointer conversion operators sometimes are generated for each calling convention so that you can form a function pointer of the correct type (this might not be impacted by your change here); there's a specific number of bits for representing the enumeration of calling conventions and this uses one of those bits, etc.

Mon, May 23, 9:06 AM · Restricted Project, Restricted Project

Fri, May 20

JonChesterfield added a comment to D125970: [amdgpu] Add amdgpu_kernel calling conv attribute to clang.

If it was adding a calling convention, sure - caution warranted. There's no llvm change here though, an existing CC is exposed to C++. No change to the type system either.

Fri, May 20, 7:19 AM · Restricted Project, Restricted Project
JonChesterfield added a comment to D125970: [amdgpu] Add amdgpu_kernel calling conv attribute to clang.

A clangd buildbot (https://lab.llvm.org/buildbot/#/builders/131/builds/27770) failed on this with

Fri, May 20, 1:43 AM · Restricted Project, Restricted Project
JonChesterfield committed rG83c431fb9e72: [amdgpu] Add amdgpu_kernel calling conv attribute to clang (authored by JonChesterfield).
[amdgpu] Add amdgpu_kernel calling conv attribute to clang
Fri, May 20, 12:51 AM · Restricted Project, Restricted Project
JonChesterfield closed D125970: [amdgpu] Add amdgpu_kernel calling conv attribute to clang.
Fri, May 20, 12:51 AM · Restricted Project, Restricted Project

Thu, May 19

JonChesterfield added a comment to D125970: [amdgpu] Add amdgpu_kernel calling conv attribute to clang.

Unintentionally created this patch against an older version of main and it interacted badly with D124998 on the rebase. Rerunning tests now, and will leave this open for further comments for a little while. Thanks all

Thu, May 19, 3:09 PM · Restricted Project, Restricted Project
JonChesterfield updated the diff for D125970: [amdgpu] Add amdgpu_kernel calling conv attribute to clang.
  • Fix git merge misfires
Thu, May 19, 3:06 PM · Restricted Project, Restricted Project
JonChesterfield updated the diff for D125970: [amdgpu] Add amdgpu_kernel calling conv attribute to clang.
  • Fix git merge misfires
Thu, May 19, 3:04 PM · Restricted Project, Restricted Project
JonChesterfield updated the diff for D125970: [amdgpu] Add amdgpu_kernel calling conv attribute to clang.
  • Rebase on main
Thu, May 19, 3:03 PM · Restricted Project, Restricted Project
JonChesterfield added a comment to D125970: [amdgpu] Add amdgpu_kernel calling conv attribute to clang.

Thanks for accepting! I'm interested to learn more about how the calling conv works, e.g. if parts of it are implemented in clang and parts of it patched on the fly by opt, but that's downstream of easy access to writing C tests that use it.

Thu, May 19, 2:47 PM · Restricted Project, Restricted Project
JonChesterfield added a comment to D125970: [amdgpu] Add amdgpu_kernel calling conv attribute to clang.

Added a codegen test for arg passing. It establishes that most arguments are left alone, but structs passed by value are handled as an addrspace(4) byref. Letting opt -O2 run annotated some argument pointers as being in addrspace(1) which I think is wrong.

Thu, May 19, 2:06 PM · Restricted Project, Restricted Project
JonChesterfield updated the diff for D125970: [amdgpu] Add amdgpu_kernel calling conv attribute to clang.
  • Add O0 arg passing codegen test
Thu, May 19, 2:03 PM · Restricted Project, Restricted Project
JonChesterfield added a comment to D125970: [amdgpu] Add amdgpu_kernel calling conv attribute to clang.

OK, so that's a different thing. CUDA/HIP has a bunch of rules about implicitly tagging things with addrspace(1) at the call boundary. I don't think any of that magic should exist for C or C++, the developer gets to spell out the address space stuff they want explicitly, and if they want to link it against CUDA/HIP/OpenMP code they get to look up the rules. In particular, persuading clang to do the extra argument mangling stuff will get in the way of using this to create test cases.

Thu, May 19, 1:43 PM · Restricted Project, Restricted Project
JonChesterfield added a comment to D125970: [amdgpu] Add amdgpu_kernel calling conv attribute to clang.

need a codegen test to make sure amdgpu_kernel ABI is used in C/C++ for functions with this attribute

Thu, May 19, 1:06 PM · Restricted Project, Restricted Project
JonChesterfield requested review of D125970: [amdgpu] Add amdgpu_kernel calling conv attribute to clang.
Thu, May 19, 6:28 AM · Restricted Project, Restricted Project

Wed, May 18

JonChesterfield added a comment to D125252: [Libomptarget] Add basic support for dynamic shared memory on AMDGPU.

Looks about as expected, modulo Johannes' comment above. Could we do the getenv call once, instead of per-launch? Somewhere in the massive class constructor perhaps.

Wed, May 18, 7:40 AM · Restricted Project, Restricted Project

Thu, May 5

JonChesterfield requested review of D125060: [amdgpu][wip] Implement lds kernel id intrinsic.
Thu, May 5, 4:46 PM · Restricted Project, Restricted Project

Wed, May 4

JonChesterfield committed rGbc78c0995242: [amdgpu] Elide module lds allocation in kernels with no callees (authored by JonChesterfield).
[amdgpu] Elide module lds allocation in kernels with no callees
Wed, May 4, 2:42 PM · Restricted Project, Restricted Project
JonChesterfield closed D122091: [amdgpu] Elide module lds allocation in kernels with no callees.
Wed, May 4, 2:42 PM · Restricted Project, Restricted Project
JonChesterfield added a comment to D122091: [amdgpu] Elide module lds allocation in kernels with no callees.

Drop the caching of the value and pass in function instead of module. No functional change to previous diff, simpler code.

Wed, May 4, 11:07 AM · Restricted Project, Restricted Project
JonChesterfield updated the diff for D122091: [amdgpu] Elide module lds allocation in kernels with no callees.
  • Module->function, skip caching the attribute
Wed, May 4, 11:05 AM · Restricted Project, Restricted Project
JonChesterfield added inline comments to D122091: [amdgpu] Elide module lds allocation in kernels with no callees.
Wed, May 4, 10:28 AM · Restricted Project, Restricted Project

Sun, May 1

JonChesterfield added a comment to D122091: [amdgpu] Elide module lds allocation in kernels with no callees.

Didn't notice the hasFnAttribute request comment, change made. Anything else?

Sun, May 1, 2:16 PM · Restricted Project, Restricted Project
JonChesterfield updated the diff for D122091: [amdgpu] Elide module lds allocation in kernels with no callees.
  • rebase, and getFnAttr -> hasFnAttr
Sun, May 1, 2:13 PM · Restricted Project, Restricted Project

Apr 27 2022

JonChesterfield added a reviewer for D122091: [amdgpu] Elide module lds allocation in kernels with no callees: carlo.bertolli.
Apr 27 2022, 7:34 AM · Restricted Project, Restricted Project

Apr 19 2022

JonChesterfield accepted D124039: [OpenMP] Add better testing for the linker wrapper.

Nice! I like the strategy.

Apr 19 2022, 2:37 PM · Restricted Project, Restricted Project

Apr 18 2022

JonChesterfield accepted D122831: [OpenMP] Make the new offloading driver the default.

LGTM, thanks! If it fails CI we might want to submit the tests and the code change separately, but I'm fine with trying it as one block in the first instance.

Apr 18 2022, 10:07 AM · Restricted Project, Restricted Project, Restricted Project

Apr 14 2022

JonChesterfield accepted D123827: [libomptarget] [amdgpu] Hostcall offset check should consider implicit args.

Thanks! Good fix

Apr 14 2022, 5:30 PM · Restricted Project, Restricted Project

Apr 13 2022

JonChesterfield accepted D122987: [OpenMP] Make offloading sections have the SHF_EXCLUDE flag.

Thanks for the extra test lines!

Apr 13 2022, 7:40 AM · Restricted Project, Restricted Project
JonChesterfield added a comment to D122831: [OpenMP] Make the new offloading driver the default.

Editing the tests in place to check for new driver properties would mean we lose coverage for the old driver and get a lot of churn if we need to back out the change.

Apr 13 2022, 7:38 AM · Restricted Project, Restricted Project, Restricted Project
JonChesterfield accepted D122069: [Object] Add binary format for bundling offloading metadata.

Couple of nits above but basically I'm convinced. The gnarly part of binary formats is string tables and I'm delighted that part of MC was readily reusable. Wrapping the string table in different bytes to align with the elf format may still be a good idea but it's not an obvious correctness hazard.

Apr 13 2022, 7:27 AM · Restricted Project, Restricted Project, Restricted Project

Apr 6 2022

JonChesterfield accepted D123177: [libomptarget] Add device RTL to regression test dependencies..

Nice, thanks!

Apr 6 2022, 7:54 AM · Restricted Project, Restricted Project
JonChesterfield added a comment to D123177: [libomptarget] Add device RTL to regression test dependencies..

Extra dependency edge is great, thanks! Is LIBOMPTARGET_TESTED_PLUGINS new? It looks unused in this patch

Apr 6 2022, 12:33 AM · Restricted Project, Restricted Project

Apr 5 2022

JonChesterfield added a comment to D122987: [OpenMP] Make offloading sections have the SHF_EXCLUDE flag.

I think we need tests for coff and elf for the MC part, should be existing ones for the other flags. Ideal fit for the openmp sectiy, good find. Inclined to ignore the formatting warnings on old code.

Apr 5 2022, 7:53 AM · Restricted Project, Restricted Project

Apr 4 2022

JonChesterfield added a reviewer for D122091: [amdgpu] Elide module lds allocation in kernels with no callees: bcahoon.

Ron thought @bcahoon might be interested, adding as reviewer

Apr 4 2022, 10:58 AM · Restricted Project, Restricted Project

Mar 31 2022

JonChesterfield added a comment to D122091: [amdgpu] Elide module lds allocation in kernels with no callees.

@arsenm there's a bunch of stuff to do that builds on this. What further changes would you like?

Mar 31 2022, 7:10 AM · Restricted Project, Restricted Project

Mar 29 2022

JonChesterfield added reviewers for D122069: [Object] Add binary format for bundling offloading metadata: grokos, ABataev, ronlieb, tianshilei1992.

Added some reviewers. I'd much prefer this used an existing binary format, DIY is prone to errors and maintenance hassles down the road. Don't care as much about which format as about it being one with an existing, tested implementation and ideally existing inspection tools.

Mar 29 2022, 7:17 AM · Restricted Project, Restricted Project, Restricted Project

Mar 28 2022

JonChesterfield added a comment to D122091: [amdgpu] Elide module lds allocation in kernels with no callees.

What happens to the case where the entry point function makes call(s) to device functions, but none of them use lds?

Mar 28 2022, 6:51 AM · Restricted Project, Restricted Project

Mar 25 2022

JonChesterfield accepted D122504: [OpenMP] Make Ctor / Dtor functions have external visibility.

Nice, thanks. Wonder if we want protected visibility as well.

Mar 25 2022, 2:12 PM · Restricted Project, Restricted Project

Mar 24 2022

JonChesterfield added inline comments to D122091: [amdgpu] Elide module lds allocation in kernels with no callees.
Mar 24 2022, 3:59 AM · Restricted Project, Restricted Project

Mar 23 2022

JonChesterfield added a comment to D122352: [OpenMP] Do not create offloading entries for internal or hidden symbols.

Sounds good to me too, thanks!

Mar 23 2022, 3:00 PM · Restricted Project, Restricted Project
JonChesterfield added inline comments to D122091: [amdgpu] Elide module lds allocation in kernels with no callees.
Mar 23 2022, 9:36 AM · Restricted Project, Restricted Project
JonChesterfield updated the diff for D122091: [amdgpu] Elide module lds allocation in kernels with no callees.
  • Drop string bool parameter and unreliable test
Mar 23 2022, 9:36 AM · Restricted Project, Restricted Project
JonChesterfield updated the diff for D122091: [amdgpu] Elide module lds allocation in kernels with no callees.
  • invert sense of attribute
Mar 23 2022, 9:12 AM · Restricted Project, Restricted Project
JonChesterfield added a comment to D118493: Set rpath on openmp executables.

The rule that is broken is "standard RPATHs" Fedora installs libomp to /usr/lib64.

...

I think what we'll do in Fedora is just add -fno-openmp-implicit-rpath to the default CFLAGS so we can get the old behavior for official RPMs.
Mar 23 2022, 3:15 AM · Restricted Project, Restricted Project, Restricted Project

Mar 22 2022

JonChesterfield updated subscribers of D118493: Set rpath on openmp executables.

@estewart08 Fedora are rejecting some uses of rpath (and probably runpath). I can't find a corresponding page for redhat. This may become a problem for our aomp/rocm builds.

Mar 22 2022, 1:56 AM · Restricted Project, Restricted Project, Restricted Project
JonChesterfield added a comment to D118493: Set rpath on openmp executables.

It lets applications run with the libomp and libomptarget built with the toolchain. For users, they don't have to be root. For compiler devs, we test our work instead of whatever the distro had installed.

Mar 22 2022, 1:10 AM · Restricted Project, Restricted Project, Restricted Project

Mar 21 2022

JonChesterfield added a comment to D122091: [amdgpu] Elide module lds allocation in kernels with no callees.

The current behaviour is to allocate an instance of llvm.amdgcn.module.lds.t at address zero in every kernel in the IR module if the global llvm.amdgcn.module.lds is present. That instance only needs to be allocated in kernels which subsequently call a function which uses it.

Mar 21 2022, 11:02 AM · Restricted Project, Restricted Project
JonChesterfield added a comment to D122091: [amdgpu] Elide module lds allocation in kernels with no callees.

The other passes do know how much is allocated. I don't want to go digging around the IR looking for the llvm.donothing call, which iirc is removed before instruction selection

Mar 21 2022, 6:22 AM · Restricted Project, Restricted Project

Mar 20 2022

Herald added a project to D112717: [IR] Replace *all* uses of a constant expression by corresponding instruction: Restricted Project.
Mar 20 2022, 4:46 PM · Restricted Project, Restricted Project

Mar 19 2022

JonChesterfield updated subscribers of D122091: [amdgpu] Elide module lds allocation in kernels with no callees.

@ronlieb this is the patch I'd like to run through the internal CI

Mar 19 2022, 6:21 PM · Restricted Project, Restricted Project
JonChesterfield requested review of D122091: [amdgpu] Elide module lds allocation in kernels with no callees.
Mar 19 2022, 6:13 PM · Restricted Project, Restricted Project
JonChesterfield added a reverting change for rG98e474c1b321: [amdgpu][nfc] Pass function instead of module to allocateModuleLDSGlobal: rGbcbd4cf1f269: Revert "[amdgpu][nfc] Pass function instead of module to….
Mar 19 2022, 5:59 PM · Restricted Project
JonChesterfield committed rGbcbd4cf1f269: Revert "[amdgpu][nfc] Pass function instead of module to… (authored by JonChesterfield).
Revert "[amdgpu][nfc] Pass function instead of module to…
Mar 19 2022, 5:59 PM · Restricted Project
JonChesterfield committed rG98e474c1b321: [amdgpu][nfc] Pass function instead of module to allocateModuleLDSGlobal (authored by JonChesterfield).
[amdgpu][nfc] Pass function instead of module to allocateModuleLDSGlobal
Mar 19 2022, 9:42 AM · Restricted Project

Mar 17 2022

JonChesterfield accepted D121837: [OpenMP][FIX] Allow device constructors for AMD GPU.

Thanks!

Mar 17 2022, 9:59 AM · Restricted Project, Restricted Project

Mar 16 2022

JonChesterfield accepted D121837: [OpenMP][FIX] Allow device constructors for AMD GPU.

Nice, thanks!

Mar 16 2022, 12:53 PM · Restricted Project, Restricted Project

Mar 11 2022

Herald added a project to D94021: [OpenCL] Add clang extension for function pointers: Restricted Project.
Mar 11 2022, 5:36 PM · Restricted Project, Restricted Project
JonChesterfield updated the diff for D121499: [openmp] Generalise function instantiation macro from dlwrap.
  • apply refactor from D115980, removes typo-prone part
Mar 11 2022, 4:44 PM · Restricted Project, Restricted Project
Herald added a project to D115980: [libomptarget][amdgpu][nfc] Make API harder to misimplement: Restricted Project.

This can be picked up and simplified using D121499

Mar 11 2022, 4:29 PM · Restricted Project, Restricted Project
JonChesterfield added a comment to D121467: [Libomptarget] Create device math wrappers.

Given the header proposed in D121499, OpenMPMath.cpp can be replaced with:

Mar 11 2022, 4:16 PM · Restricted Project, Restricted Project
JonChesterfield requested review of D121499: [openmp] Generalise function instantiation macro from dlwrap.
Mar 11 2022, 3:57 PM · Restricted Project, Restricted Project
JonChesterfield committed rG75779435f3fc: [nfc][openmp] Swap arguments to remove noise from upcoming diff (authored by JonChesterfield).
[nfc][openmp] Swap arguments to remove noise from upcoming diff
Mar 11 2022, 3:09 PM · Restricted Project, Restricted Project
JonChesterfield added inline comments to D121468: [OpenMP] Add linking of OpenMP math library.
Mar 11 2022, 10:36 AM · Restricted Project, Restricted Project
JonChesterfield added a comment to D121467: [Libomptarget] Create device math wrappers.

Sounds good to me.

Mar 11 2022, 10:34 AM · Restricted Project, Restricted Project
JonChesterfield updated subscribers of D121276: [CMake] Include runtimes test suites in check-all.

@ronlieb this probably means our CI needs to do different things

Mar 11 2022, 9:21 AM · Restricted Project, Restricted Project, Restricted Project, Restricted Project
JonChesterfield added a comment to D121466: [OpenMP] Replace math headers with OpenMP wrapper calls.

We've got quite a lot of debt in this area and seem at risk of taking on more. Ideally the cuda and hip and openmp headers would be closer to a single header containing:

Mar 11 2022, 9:16 AM · Restricted Project, Restricted Project
JonChesterfield requested changes to D121467: [Libomptarget] Create device math wrappers.

OK, I think I follow.

Mar 11 2022, 9:11 AM · Restricted Project, Restricted Project

Mar 7 2022

JonChesterfield added a comment to D121057: [OpenMP][FIX] Ensure exclusive access to the HDTT map.

I'm ooo for a couple of days, @tianshilei1992 could you take a look?

Mar 7 2022, 5:40 PM · Restricted Project, Restricted Project
JonChesterfield added a comment to D121060: [OpenMP][NFCI] Use RAII lock guards in libomptarget where possible.

Found one error in the existing code but can believe there are more. We could split this patch into two, one replacing the DIY LockGuard with std::lock_guard and clang-format the thing, then a second replacing manual locking with uses of lock guard. Would be the same behaviour as this patch but make it easier to find the change in behaviour by inspection.

Mar 7 2022, 6:05 AM · Restricted Project, Restricted Project

Mar 5 2022

JonChesterfield added a comment to D121057: [OpenMP][FIX] Ensure exclusive access to the HDTT map.

Only comments on the helper type for now. I like the idea of a generic wrap-this-thing-in-mutex helper as a path to stepping on the remaining race conditions.

Mar 5 2022, 3:51 PM · Restricted Project, Restricted Project
JonChesterfield accepted D121060: [OpenMP][NFCI] Use RAII lock guards in libomptarget where possible.

Great, thanks!

Mar 5 2022, 3:38 PM · Restricted Project, Restricted Project
JonChesterfield added a comment to D121055: [OpenMP][CMake] Clean up the CMake variable `LIBOMPTARGET_LLVM_INCLUDE_DIRS`.

Nice, thanks!

Mar 5 2022, 3:35 PM · Restricted Project, Restricted Project

Mar 4 2022

JonChesterfield added inline comments to D120272: [CUDA] Add driver support for compiling CUDA with the new driver.
Mar 4 2022, 11:16 AM · Restricted Project, Restricted Project

Mar 2 2022

JonChesterfield added a comment to rG31efecfde988: Remove a top-level "using namespace" from TargetTransformInfoImpl.h.

Thanks! I get the opposite, building locally with clang sometimes breaks gcc buildbots.

Mar 2 2022, 6:53 AM · Restricted Project
JonChesterfield added a comment to rG31efecfde988: Remove a top-level "using namespace" from TargetTransformInfoImpl.h.

Broke a couple of buildbots, e.g. https://lab.llvm.org/buildbot/#/builders/193/builds/7799

Mar 2 2022, 6:46 AM · Restricted Project

Feb 23 2022

JonChesterfield added a comment to D120353: [OpenMP] Add option to make offloading mandatory.

Thanks! Seems a good thing to add to the offloading test runner, preferably in a separate change to avoid reverting this in case of unforeseen problems

Feb 23 2022, 10:13 AM · Restricted Project

Feb 21 2022

JonChesterfield accepted D120275: [OpenMP] Ensure offloading sections do not have SHF_ALLOC flag.

LG, thanks for including the test

Feb 21 2022, 12:12 PM · Restricted Project

Feb 16 2022

JonChesterfield reopened D119841: [OpenMP] Pass AMDGPU math libraries into the linker wrapper.

Thanks!

Feb 16 2022, 8:56 AM · Restricted Project
JonChesterfield added a comment to D119841: [OpenMP] Pass AMDGPU math libraries into the linker wrapper.

This is asserting somewhere in clang-linker-wrapper, e.g. https://lab.llvm.org/buildbot/#/builders/193/builds/6939

Feb 16 2022, 8:52 AM · Restricted Project
JonChesterfield added inline comments to D119841: [OpenMP] Pass AMDGPU math libraries into the linker wrapper.
Feb 16 2022, 6:20 AM · Restricted Project

Feb 15 2022

JonChesterfield added a comment to D119841: [OpenMP] Pass AMDGPU math libraries into the linker wrapper.

I'm not confident it's only attributes that we rely on mlink-builtin-bitcode patching up, that could be poor phrasing on my part.

Feb 15 2022, 1:18 PM · Restricted Project

Feb 14 2022

JonChesterfield added a comment to D119747: [Libomptarget][NFC] Remove constexpr to hide warnings.

Nvm, saw the commit message

Feb 14 2022, 1:59 PM · Restricted Project

Feb 12 2022

JonChesterfield accepted D119638: [OpenMP][NFC] Simplify identifying the device bitcode library.
Feb 12 2022, 10:35 AM · Restricted Project

Feb 11 2022

JonChesterfield added a comment to D119590: exclude openembedded distributions from setting rpath on openmp executables.

Cross compilers are a hazard here. I'd expect there to be a fairly long list of magic flags you need to pass to clang to get it to find the right libraries. Can you add fno-openmp-implicit-rpath to that list instead?

Feb 11 2022, 3:55 PM · Restricted Project, Restricted Project

Feb 10 2022

JonChesterfield accepted D119478: [Libomptarget][AMDGCN] add gfx90c target.

That's somewhat interesting, @arsenm the 4000 series APUs used to identify as gfx902 iirc - does the new name for existing hardware make sense to you?

Feb 10 2022, 1:30 PM · Restricted Project

Feb 8 2022

JonChesterfield accepted D119240: [OpenMP] Enable new driver tests for AMDGPU.

Thanks!

Feb 8 2022, 6:49 AM · Restricted Project

Feb 5 2022

JonChesterfield accepted D118988: [OpenMP] kmp_atomic_float10_max_min.c test should only be executed on x86 platform.

Looks reasonable to me. This is looking at 80 bit floating point which x86 can do (in some modes at least, this passes mlong-double-80 ). Long double elsewhere tends to be 64 or 128 bit.

Feb 5 2022, 2:20 PM · Restricted Project

Feb 4 2022

JonChesterfield accepted D119018: [OpenMP] Add -Bsymbolic to arguments for GNU linker.

Thanks!

Feb 4 2022, 10:21 AM · Restricted Project
JonChesterfield added a comment to D119018: [OpenMP] Add -Bsymbolic to arguments for GNU linker.

Can we drop an xfail in an existing test as part of this patch?

Feb 4 2022, 10:09 AM · Restricted Project

Feb 3 2022

JonChesterfield added inline comments to D118934: [OpenMP] Completely remove old device runtime.
Feb 3 2022, 11:58 AM · Restricted Project, Restricted Project
JonChesterfield updated the summary of D118934: [OpenMP] Completely remove old device runtime.
Feb 3 2022, 11:56 AM · Restricted Project, Restricted Project
JonChesterfield accepted D118934: [OpenMP] Completely remove old device runtime.

I've read this closely and can't think of anywhere else that needs to be patched. Thanks!

Feb 3 2022, 11:56 AM · Restricted Project, Restricted Project

Feb 1 2022

JonChesterfield committed rGf52927c122ed: Revert "[OpenMP][FIX] Explicit barriers in SPMD mode are not aligned" (authored by JonChesterfield).
Revert "[OpenMP][FIX] Explicit barriers in SPMD mode are not aligned"
Feb 1 2022, 6:57 AM
JonChesterfield added a reverting change for rG7b9844cc8dd0: [OpenMP][FIX] Explicit barriers in SPMD mode are not aligned: rGf52927c122ed: Revert "[OpenMP][FIX] Explicit barriers in SPMD mode are not aligned".
Feb 1 2022, 6:57 AM
JonChesterfield updated subscribers of rG7b9844cc8dd0: [OpenMP][FIX] Explicit barriers in SPMD mode are not aligned.

@ronlieb and I think reverting this is sufficient to get us a buildbot running again, going to try it. f52927c122edd7d237eb4adafd31eba93342923c

Feb 1 2022, 6:56 AM
JonChesterfield updated subscribers of rG8b7e99c41d61: [openmp] Disable tests that presently hang on CI.

@ronlieb @jdoerfert This was hoped to resolve the breakage from https://lab.llvm.org/buildbot/#/builders/193/builds/5772 which unfortunately has five changes in it,

Feb 1 2022, 5:37 AM
JonChesterfield committed rG8b7e99c41d61: [openmp] Disable tests that presently hang on CI (authored by JonChesterfield).
[openmp] Disable tests that presently hang on CI
Feb 1 2022, 5:01 AM