[all-commits] [llvm/llvm-project] d5038b: [libclc] Move __clc_ldexp to CLC library (#126078)

[Fraser Cormack via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: d5038b3774485d617e1300cf2f7b98c2460b9042
      https://github.com/llvm/llvm-project/commit/d5038b3774485d617e1300cf2f7b98c2460b9042
  Author: Fraser Cormack <fraser at codeplay.com>
  Date:   2025-02-26 (Wed, 26 Feb 2025)

  Changed paths:
    M libclc/CMakeLists.txt
    M libclc/amdgcn/lib/SOURCES
    R libclc/amdgcn/lib/math/ldexp.cl
    A libclc/clc/include/clc/math/clc_ldexp.h
    A libclc/clc/include/clc/math/clc_ldexp.inc
    A libclc/clc/lib/amdgcn/SOURCES
    A libclc/clc/lib/amdgcn/math/clc_ldexp_override.cl
    M libclc/clc/lib/generic/SOURCES
    A libclc/clc/lib/generic/math/clc_ldexp.cl
    M libclc/clspv/lib/SOURCES
    R libclc/generic/include/math/clc_ldexp.h
    M libclc/generic/lib/SOURCES
    R libclc/generic/lib/math/clc_ldexp.cl
    M libclc/generic/lib/math/ldexp.cl
    M libclc/generic/lib/math/ldexp.inc
    M libclc/spirv/lib/SOURCES

  Log Message:
  -----------
  [libclc] Move __clc_ldexp to CLC library (#126078)

This function was already conceptually in the CLC namespace - this just
formally moves it over.

Note however that this commit marks a change in how libclc functions may
be overridden by targets.

Until now we have been using a purely build-system-based approach where
targets could register identically-named files which took responsibility
for the implementation of the builtin in its entirety.

This system wasn't well equipped to deal with AMD's overriding of
__clc_ldexp for only a subset of types, and furthermore conditionally on
a pre-defined macro.

One option for handling this would be to require AMD to duplicate code
for the versions of __clc_ldexp it's *not* interested in overriding. We
could also make it easier for targets to re-define CLC functions through
macros or .inc files. Both of these have obvious downsides. We could
also keep AMD's overriding in the OpenCL layer and bypass CLC
altogether, but this has limited use.

We could use weak linkage on the "base" implementations of CLC
functions, and allow targets to opt-in to providing their own
implementations on a much finer granularity. This commit supports this
as a proof of concept; we could expand it to all CLC builtins if
accepted.

Note that the existing filename-based "claiming" approach is still in
effect, so targets have to name their overrides differently to have both
files compiled. This could also be refined.



To unsubscribe from these emails, change your notification settings at https://github.com/llvm/llvm-project/settings/notifications