You can do it and may be ok for single user with idle waiting times, but performance/throughput will be roughly halved (closer to 2/3) and free context will be more limited with 8xH200 vs 16xH100 (assuming decent interconnect). Depending a bit on usecase and workload 16xH100 (or 16xB200) may be a better config for cost optimization. Often there is a huge economy of scale with such large mixture of expert models so that it would even be cheaper to use 96 GPU instead of just 8 or 16. The reasons are complicatet and involve better prefill cache, less memory transfer per node.
alt Hacker News
You can do it and may be ok for single user with idle waiting times, but performance/throughput will be roughly halved (closer to 2/3) and free context will be more limited with 8xH200 vs 16xH100 (assuming decent interconnect). Depending a bit on usecase and workload 16xH100 (or 16xB200) may be a better config for cost optimization. Often there is a huge economy of scale with such large mixture of expert models so that it would even be cheaper to use 96 GPU instead of just 8 or 16. The reasons are complicatet and involve better prefill cache, less memory transfer per node.