add patches to fix PyTorch-1.12.1 w/ foss/2021b + CUDA v11.5.2 on POWER

[Flamefire]

(created using eb --new-pr)

Same as #18490 but for the CUDA version (single EC changed due to time it takes for testing the CUDA version)

[Flamefire]

Test report by @Flamefire
SUCCESS
Build succeeded for 1 out of 1 (1 easyconfigs in total)
taurusml20 - Linux RHEL 7.6, POWER, 8335-GTX (power9le), 6 x NVIDIA Tesla V100-SXM2-32GB, 440.64.00, Python 2.7.5
See https://gist.github.com/Flamefire/62b26f0d25529ca1b359ca275b03b05e for a full test report.

[Flamefire]

Test report by @Flamefire
FAILED SUCCESS
Build succeeded for 0 out of 1 (1 easyconfigs in total)
taurusi8018 - Linux CentOS Linux 7.9.2009, x86_64, AMD EPYC 7352 24-Core Processor (zen2), 8 x NVIDIA NVIDIA A100-SXM4-40GB, 470.57.02, Python 2.7.5
See https://gist.github.com/Flamefire/785fea163a066ee7621194c5cba657f9 for a full test report.

I traced the failure in test_cpp_extensions_jit to a driver issue: We have CUDA 11.4 as the driver but this uses 11.5.
Installing CUDAcompat/11.6 resolves the issue. I hence consider the run as a success.

[Flamefire]

@boegel Remaining issue found and resolved. So this is good to go from my side.

[Flamefire]

Test report by @Flamefire
SUCCESS
Build succeeded for 1 out of 1 (1 easyconfigs in total)
taurusi8012 - Linux CentOS Linux 7.9.2009, x86_64, AMD EPYC 7352 24-Core Processor (zen2), 8 x NVIDIA NVIDIA A100-SXM4-40GB, 470.57.02, Python 2.7.5
See https://gist.github.com/Flamefire/b1ba6a20f88f1b2251f1bc548dcde428 for a full test report.

[branfosj]

Test report by @branfosj
Using easyblocks from PR(s) easybuilders/easybuild-easyblocks#2983
SUCCESS
Build succeeded for 1 out of 1 (1 easyconfigs in total)
bear-pg0208u09a - Linux RHEL 8.6, x86_64, Intel(R) Xeon(R) Platinum 8360Y CPU @ 2.40GHz (icelake), 1 x NVIDIA NVIDIA A100-SXM4-40GB, 520.61.05, Python 3.6.8
See https://gist.github.com/branfosj/71dadbe33659eb03cdea532eed45dbe0 for a full test report.

[branfosj]

Going in, thanks @Flamefire!