Avoid creating a lot of queries for equal issue mode

[pgmpablo157321]

fix #2007

[github-actions]

MLCommons CLA bot All contributors have signed the MLCommons CLA ✍️ ✅

[arjunsuresh]

Thank you @pgmpablo157321 for the fix.

Can you please answer this question on effective_samples_per_query? For the benchmarks with equal_issue_mode, the effective_samples_per_query must be a multiple of the dataset size right? But in the previously submitted logs this is not the case. May be loadgen is not populating this value correctly?

https://github.com/mlcommons/inference_results_v4.1/blob/30d19b768aed8d9894307050039ccecdfae8b479/closed/NVIDIA/results/H200-SXM-141GBx8_TRT_MaxQ/mixtral-8x7b/Offline/performance/run_1/mlperf_log_detail.txt#L35