tests: benchdnn: inputs: matmul: test bmn clipping

[hidefromkgb]

This PR is intended to test if clipping B, M, N helps reduce the overall Matmul LLM CI time.
Addresses MFDNN-14287.

[hidefromkgb]

make test
set test_scope=NIGHTLY
disable test_device_cpu
disable benchdnn_all
enable benchdnn_matmul
enable arch_gpu_xe-hpc
enable arch_gpu_xe-hpg-atsm
enable arch_gpu_xe-hpg-dg2
enable arch_gpu_xe-lp
enable arch_gpu_xe-lpg
enable arch_gpu_xe-lpg+
enable arch_gpu_xe2-hpg-bmg
enable arch_gpu_xe2-lpg
enable arch_gpu_xe3-lpg

[kealan-barbieri]

would it be possible to maintain a complete set for optional use outside nightly runs?

[hidefromkgb]

would it be possible to maintain a complete set for optional use outside nightly runs?

option_set_fwks_llm_gpu is not an independent set, as it's generated from KPIs/gpu/common/LLM/*.log files using gen_benchdnn.sh (see onednn-perf-report) — which is why I don't believe it's worth keeping the original set.