استخدام GPU Perf Engineer
شرح موقعیت شغلی
1) HPL/STREAM/OSU/NCCL/RCCL benchmark suite. Per-platform ground truth fingerprints (S3). Ph
2: AMD MI350X RCCL baselines + FP4/FP6 profiling. Ph
3: NVL72 topology benchmarks, GB300 HGX HBM3e profiling; TE (Transformer Engine) throughput tracking.
AI Eng consulted on RAG retrieval schema
2) GPU topology CRD schema: NVLink gen, NVSwitch presence, AMD Infinity Fabric type. Ph 2: GB300 HGX NVLink counters + AMD MI350X ROCm/HIP stack. Ph 3: NVL72 NVSwitch affinity, topology-aware placement. Ph
4: NVL72 rack-scale logical unit (72-GPU), AMD MI400 Helios compatibility tracking
2: AMD MI350X RCCL baselines + FP4/FP6 profiling. Ph
3: NVL72 topology benchmarks, GB300 HGX HBM3e profiling; TE (Transformer Engine) throughput tracking.
AI Eng consulted on RAG retrieval schema
2) GPU topology CRD schema: NVLink gen, NVSwitch presence, AMD Infinity Fabric type. Ph 2: GB300 HGX NVLink counters + AMD MI350X ROCm/HIP stack. Ph 3: NVL72 NVSwitch affinity, topology-aware placement. Ph
4: NVL72 rack-scale logical unit (72-GPU), AMD MI400 Helios compatibility tracking
مهارتهای مورد نیاز
- GPU
حداقل سابقه کار
- مهم نیست
جنسیت
- مهم نیست
وضعیت نظام وظیفه
- مهم نیست