[CI][Benchmarks] update to latest compute-benchmarks, add syclpreview

[pbalcer]

This patch updates to the latest revision of compute-benchmarks and updates the scripts to match.

There two important changes:

• the addition of syclpreview SubmitKernel benchmarks
• introduction of "eventless" variants of SubmitKernel for L0 and UR. These will now be grouped together for easy comparison.

[pbalcer]

running api_overhead_benchmark_sycl SubmitKernel out of order not using events, iteration 0... 
api_overhead_benchmark_sycl SubmitKernel out of order not using events complete (api_overhead_benchmark_sycl SubmitKernel out of order not using events: 20.494 μs).
running api_overhead_benchmark_sycl SubmitKernel out of order, iteration 0... 
api_overhead_benchmark_sycl SubmitKernel out of order complete (api_overhead_benchmark_sycl SubmitKernel out of order: 20.209 μs).
running api_overhead_benchmark_sycl SubmitKernel out of order with measure completion not using events, iteration 0... 
api_overhead_benchmark_sycl SubmitKernel out of order with measure completion not using events complete (api_overhead_benchmark_sycl SubmitKernel out of order with measure completion not using events: 23.158 μs).
running api_overhead_benchmark_sycl SubmitKernel out of order with measure completion, iteration 0... 
api_overhead_benchmark_sycl SubmitKernel out of order with measure completion complete (api_overhead_benchmark_sycl SubmitKernel out of order with measure completion: 23.147 μs).
running api_overhead_benchmark_sycl SubmitKernel in order not using events, iteration 0... 
api_overhead_benchmark_sycl SubmitKernel in order not using events complete (api_overhead_benchmark_sycl SubmitKernel in order not using events: 18.429 μs).
running api_overhead_benchmark_sycl SubmitKernel in order, iteration 0... 
api_overhead_benchmark_sycl SubmitKernel in order complete (api_overhead_benchmark_sycl SubmitKernel in order: 20.597 μs).
running api_overhead_benchmark_sycl SubmitKernel in order with measure completion not using events, iteration 0... 
api_overhead_benchmark_sycl SubmitKernel in order with measure completion not using events complete (api_overhead_benchmark_sycl SubmitKernel in order with measure completion not using events: 22.462 μs).
running api_overhead_benchmark_sycl SubmitKernel in order with measure completion, iteration 0... 
api_overhead_benchmark_sycl SubmitKernel in order with measure completion complete (api_overhead_benchmark_sycl SubmitKernel in order with measure completion: 23.572 μs).
running api_overhead_benchmark_l0 SubmitKernel out of order not using events, iteration 0... 
api_overhead_benchmark_l0 SubmitKernel out of order not using events complete (api_overhead_benchmark_l0 SubmitKernel out of order not using events: 11.019 μs).
running api_overhead_benchmark_l0 SubmitKernel out of order, iteration 0... 
api_overhead_benchmark_l0 SubmitKernel out of order complete (api_overhead_benchmark_l0 SubmitKernel out of order: 12.327 μs).
running api_overhead_benchmark_l0 SubmitKernel out of order with measure completion not using events, iteration 0... 
api_overhead_benchmark_l0 SubmitKernel out of order with measure completion not using events complete (api_overhead_benchmark_l0 SubmitKernel out of order with measure completion not using events: 14.565 μs).
running api_overhead_benchmark_l0 SubmitKernel out of order with measure completion, iteration 0... 
api_overhead_benchmark_l0 SubmitKernel out of order with measure completion complete (api_overhead_benchmark_l0 SubmitKernel out of order with measure completion: 16.541 μs).
running api_overhead_benchmark_l0 SubmitKernel in order not using events, iteration 0... 
api_overhead_benchmark_l0 SubmitKernel in order not using events complete (api_overhead_benchmark_l0 SubmitKernel in order not using events: 11.727 μs).
running api_overhead_benchmark_l0 SubmitKernel in order, iteration 0... 
api_overhead_benchmark_l0 SubmitKernel in order complete (api_overhead_benchmark_l0 SubmitKernel in order: 13.336 μs).
running api_overhead_benchmark_l0 SubmitKernel in order with measure completion not using events, iteration 0... 
api_overhead_benchmark_l0 SubmitKernel in order with measure completion not using events complete (api_overhead_benchmark_l0 SubmitKernel in order with measure completion not using events: 17.521 μs).
running api_overhead_benchmark_l0 SubmitKernel in order with measure completion, iteration 0... 
api_overhead_benchmark_l0 SubmitKernel in order with measure completion complete (api_overhead_benchmark_l0 SubmitKernel in order with measure completion: 18.010 μs).
running api_overhead_benchmark_ur SubmitKernel out of order not using events, iteration 0... 
api_overhead_benchmark_ur SubmitKernel out of order not using events complete (api_overhead_benchmark_ur SubmitKernel out of order not using events CPU count: 88240.000 instr).
api_overhead_benchmark_ur SubmitKernel out of order not using events complete (api_overhead_benchmark_ur SubmitKernel out of order not using events: 12.991 μs).
running api_overhead_benchmark_ur SubmitKernel out of order, iteration 0... 
api_overhead_benchmark_ur SubmitKernel out of order complete (api_overhead_benchmark_ur SubmitKernel out of order CPU count: 93190.000 instr).
api_overhead_benchmark_ur SubmitKernel out of order complete (api_overhead_benchmark_ur SubmitKernel out of order: 13.537 μs).
running api_overhead_benchmark_ur SubmitKernel out of order with measure completion not using events, iteration 0... 
api_overhead_benchmark_ur SubmitKernel out of order with measure completion not using events complete (api_overhead_benchmark_ur SubmitKernel out of order with measure completion not using events CPU count: 90535.000 instr).
api_overhead_benchmark_ur SubmitKernel out of order with measure completion not using events complete (api_overhead_benchmark_ur SubmitKernel out of order with measure completion not using events: 18.063 μs).
running api_overhead_benchmark_ur SubmitKernel out of order with measure completion, iteration 0... 
api_overhead_benchmark_ur SubmitKernel out of order with measure completion complete (api_overhead_benchmark_ur SubmitKernel out of order with measure completion CPU count: 95274.000 instr).
api_overhead_benchmark_ur SubmitKernel out of order with measure completion complete (api_overhead_benchmark_ur SubmitKernel out of order with measure completion: 18.130 μs).
running api_overhead_benchmark_ur SubmitKernel in order not using events, iteration 0... 
api_overhead_benchmark_ur SubmitKernel in order not using events complete (api_overhead_benchmark_ur SubmitKernel in order not using events CPU count: 88240.000 instr).
api_overhead_benchmark_ur SubmitKernel in order not using events complete (api_overhead_benchmark_ur SubmitKernel in order not using events: 13.120 μs).
running api_overhead_benchmark_ur SubmitKernel in order, iteration 0... 
api_overhead_benchmark_ur SubmitKernel in order complete (api_overhead_benchmark_ur SubmitKernel in order CPU count: 93190.000 instr).
api_overhead_benchmark_ur SubmitKernel in order complete (api_overhead_benchmark_ur SubmitKernel in order: 13.503 μs).
running api_overhead_benchmark_ur SubmitKernel in order with measure completion not using events, iteration 0... 
api_overhead_benchmark_ur SubmitKernel in order with measure completion not using events complete (api_overhead_benchmark_ur SubmitKernel in order with measure completion not using events CPU count: 90535.000 instr).
api_overhead_benchmark_ur SubmitKernel in order with measure completion not using events complete (api_overhead_benchmark_ur SubmitKernel in order with measure completion not using events: 18.120 μs).
running api_overhead_benchmark_ur SubmitKernel in order with measure completion, iteration 0... 
api_overhead_benchmark_ur SubmitKernel in order with measure completion complete (api_overhead_benchmark_ur SubmitKernel in order with measure completion CPU count: 95274.000 instr).
api_overhead_benchmark_ur SubmitKernel in order with measure completion complete (api_overhead_benchmark_ur SubmitKernel in order with measure completion: 17.366 μs).
running api_overhead_benchmark_syclpreview SubmitKernel out of order not using events, iteration 0... 
api_overhead_benchmark_syclpreview SubmitKernel out of order not using events complete (api_overhead_benchmark_syclpreview SubmitKernel out of order not using events: 19.775 μs).
running api_overhead_benchmark_syclpreview SubmitKernel out of order, iteration 0... 
api_overhead_benchmark_syclpreview SubmitKernel out of order complete (api_overhead_benchmark_syclpreview SubmitKernel out of order: 19.861 μs).
running api_overhead_benchmark_syclpreview SubmitKernel out of order with measure completion not using events, iteration 0... 
api_overhead_benchmark_syclpreview SubmitKernel out of order with measure completion not using events complete (api_overhead_benchmark_syclpreview SubmitKernel out of order with measure completion not using events: 22.879 μs).
running api_overhead_benchmark_syclpreview SubmitKernel out of order with measure completion, iteration 0... 
api_overhead_benchmark_syclpreview SubmitKernel out of order with measure completion complete (api_overhead_benchmark_syclpreview SubmitKernel out of order with measure completion: 22.387 μs).
running api_overhead_benchmark_syclpreview SubmitKernel in order not using events, iteration 0... 
api_overhead_benchmark_syclpreview SubmitKernel in order not using events complete (api_overhead_benchmark_syclpreview SubmitKernel in order not using events: 18.364 μs).
running api_overhead_benchmark_syclpreview SubmitKernel in order, iteration 0... 
api_overhead_benchmark_syclpreview SubmitKernel in order complete (api_overhead_benchmark_syclpreview SubmitKernel in order: 20.167 μs).
running api_overhead_benchmark_syclpreview SubmitKernel in order with measure completion not using events, iteration 0... 
api_overhead_benchmark_syclpreview SubmitKernel in order with measure completion not using events complete (api_overhead_benchmark_syclpreview SubmitKernel in order with measure completion not using events: 21.439 μs).
running api_overhead_benchmark_syclpreview SubmitKernel in order with measure completion, iteration 0... 
api_overhead_benchmark_syclpreview SubmitKernel in order with measure completion complete (api_overhead_benchmark_syclpreview SubmitKernel in order with measure completion: 23.325 μs).

[pbalcer]

@intel/llvm-gatekeepers please merge, the container build failure is unrelated (#18024 (comment)).