[HN Gopher] Gaudi 2 - An AI Accelerator Card with 96GB of HBM2E ...
___________________________________________________________________
Gaudi 2 - An AI Accelerator Card with 96GB of HBM2E Memory
Author : peter_d_sherman
Score : 20 points
Date : 2023-09-21 13:04 UTC (9 hours ago)
(HTM) web link (habana.ai)
(TXT) w3m dump (habana.ai)
| akreal wrote:
| I was curious about energy efficiency and took two samples from
| the linked MLPerf GPT3 results. H100 seems about three times more
| efficient than Gaudi2. 256 Gaudi2 600W TDP: 256 *
| (442.578 / 60) * 0.6 = ~1133 kWh 512 H100 700W TDP: 512 *
| (64.264 / 60) * 0.7 = ~384 kWh
| sm_1024 wrote:
| I'm guessing H100 has 2x host energy overhead for connecting
| those GPUs? That might offset some of the perf/W benefits of
| nvidia's offering.
| mdaniel wrote:
| I have to say, I was expecting the PyTorge migration guide
| <https://docs.habana.ai/en/latest/PyTorch/PyTorch_Model_Porti...>
| to be filled with a lot more silliness while trying to break the
| CUDA stranglehold so I wish them all the best
| Reubend wrote:
| It's great that their advertised performance is competitive, but
| I don't see any mention of price on this website, and I wasn't
| able to easily find any prices in their "Acess/Buy" page either.
|
| So until I hear otherwise, I'm going to assume that this doesn't
| provide a better price to perf ratio than existing commercial
| setups.
___________________________________________________________________
(page generated 2023-09-21 23:03 UTC)