MLPerf Inference v3.0 closed/Dell r7515_q4_pro-qaic-v1.8.3.7-aic100 — 4x Qualcomm Cloud AI 100 Pro, bert-99/99.9 Offline+Server raw logs

Fields

id

sources/bench-mlperf-v3-0-dell-r7515-4xqaic-ai100-bert

URL

https://github.com/mlcommons/inference_results_v3.0/tree/main/closed/Dell/results/r7515_q4_pro-qaic-v1.8.3.7-aic100

Publisher

MLCommons (Dell submission)

Kind

benchmark

Published

2023-08-02

Accessed

2026-10-14

notes

MLPerf Inference v3.0 closed submission on Dell r7515 with 4x Qualcomm Cloud AI 100 Pro (q4_pro config, Apps SDK v1.8.3.7, aic100 backend). bert-99 Offline 3059.8 samples/s VALID; bert-99 Server 2724.39 completed samples/s; bert-99.9 Offline 1487.12; bert-99.9 Server 1372.9. BERT-large SQuAD, INT8. Values read verbatim from performance/run_1/mlperf_log_summary.txt. The last round in which Qualcomm submitted Cloud AI 100 (base, not Ultra) results; v4.0 has only CTuning/glow entries.

Notes

MLPerf Inference v3.0 closed submission on Dell r7515 with 4x Qualcomm Cloud AI 100 Pro (q4_pro config, Apps SDK v1.8.3.7, aic100 backend). bert-99 Offline 3059.8 samples/s VALID; bert-99 Server 2724.39 completed samples/s; bert-99.9 Offline 1487.12; bert-99.9 Server 1372.9. BERT-large SQuAD, INT8. Values read verbatim from performance/run_1/mlperf_log_summary.txt. The last round in which Qualcomm submitted Cloud AI 100 (base, not Ultra) results; v4.0 has only CTuning/glow entries.

Cross-references

This record references no other records.