tree 3252126e1346a22765abc8357fca81c61d7978ee
parent 369be311a7b8344699d12719a8fa557fe8934e46
author Joseph Huber <huberjn@outlook.com> 1742563552 -0500
committer GitHub <noreply@github.com> 1742563552 -0500
gpgsig -----BEGIN PGP SIGNATURE-----
 
 wsFcBAABCAAQBQJn3WjhCRC1aQ7uu5UhlAAAZ10QAIPUrAyKN7WnOpGaUlp8wSc4
 WCgFiG5BX/qEoV/mB3XGni5CuA0aCVYwuKjDM5nbHDu8ayliNl7oCoAVpEOeJcIp
 1N8fd/eqjWkTMry8tvJHuc4vf+y0xGMIa50nl7AwcFcqDsSng2dijHIZHo7t/rQC
 mbTRYC3PH2R9VT6DrxJLAgm6w4gyjUz79F/ANBMxarIzI1mOKbgHXWOEJw/ya5Th
 OroqvReFShDPV6GEuTfvdrR4PR8T6j+zg/P1xw5l4XdMd62oWH6a7F+gOzJ4NUqV
 h6t6ytwGkPrTUitP3SnQO+EYggwaSkdTNfdmjvS6eKEMlTiapW3jgkoGEEZyg3KE
 Qo0OmVsGnAUuz8TYhLx+b1D3dscmDC9WVMQuqtvXL9GH47UZgbJst4qMwkFpAeOS
 8FwX/nuptWi+Kn7lH2V5OufEh2EYTeOGkcbv5vL9Isg1IOpaEmx0VZWlLsSASnSp
 ygW2/xSxoYX/KjS5XD7C+el17qhWAvIPjEQ2kNevWTzk0u6YL5+cK/pChXHIqZWF
 ILlqeVyoddGeyTMBk8XLcuhq4Th0UdPb4DZ3f0HLqb4X96UeEfj7sRuwyI2KyD42
 Iqp1pNu372nvH2FHcoEzTwHFLBwIq3y8Te24X0gk2w+NwZNKLwKilSf8x4/pvn4t
 Y94hS0qfZ9p26EAFOM6K
 =gu1t
 -----END PGP SIGNATURE-----
 

[Clang] Permit `-Xarch_` to be used with `--offload-arch` (#131884)

Summary:
The `--offload-arch` option is very complicated, but roughly behaves as
the `-march` option for several compilations at once. This creates
problems when we try to combine multiple separate architectures into
one, as happens with SYCL, OpenMP, and HIP w/ SPIR-V.

The existing solution used by OpenMP is the `-Xopenmp-target` option,
this lets you select which `--offload-arch` options go to which
toolchain. This patch permits `-Xarch_` to be used in the same way.

There are concerns about whether or not this falls into the `-Xarch_`
umbrella because it changes the driver behaviour, but I think this is
the
easiest way to handle this problem. The existing solution seems to be
prefixing things and adding more magic handling into `--offload-arch`.
Like SYCL is doing `nvidia_gpu_sm_89` instead of just `-Xarch_nvptx64
--offload-arch=sm_89`.

The only reason this is more complicated than just doing `-Xarch_sm_89
-march=...` is because we need to know to create multiple jobs for each
architecture.