The bar graph seems a little whacky. It groups the TPU (which can only do FP16) ... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

		gok on Feb 23, 2018 \| parent \| context \| favorite \| on: Benchmarking Google’s new TPUv2 The bar graph seems a little whacky. It groups the TPU (which can only do FP16) with the FP32 results from the GPUs, then puts the FP16 GPU results off to the side even though that's much closer to what the TPU is doing. Impressive results regardless though; quite a bit faster than V100 than the paper specs would suggest.

ekelsen on Feb 23, 2018 | [–]

It also seems like the price comparison should compare with the fp16 numbers on both platforms, not the fp32 numbers.

elmarhaussmann on Feb 23, 2018 | [–]

Author here.

Good point, I agree that the FP16 GPU results should be closer or grouped with the TPU results. We'll try to update accordingly.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact