[IPEX deprecation] Add Benchmarks job to the PyTorch pin update validation workflow

[kwasd]

The remainder of the:

• [IPEX deprecation] Process to update upstream Pytorch pin and automation around it #1933

blocked by:

• [Benchmarks] Deprecate import intel_extension_for_pytorch in benchmarks including decoupling XeTLA build which require ipex.h #1911
• [FA][Upstream PT] XPU out of memory raised by FA kernel with upstream pytorch #2042
• [Benchmarks][Upstream PyTorch 2.5] Triton and XeTLA softmax performance degrades in comparison with torch 2.1 / ipex 2.1 test proxies #2106

See more details in the issue comments and conversation in the linked PR's:

• Build benchmarks with upstream pytorch #1905
• Remove import intel_extension_for_pytorch from benchmarks #2147