3 ms·
Will there be a customer grade TPU in the near future? Or won't they be able to be as price/performance efficient compared to (nvidia) GPUs?
by zapnuk 7y ago
Will there be a customer grade TPU in the near future?
Or won't they be able to be as price/performance efficient compared to (nvidia) GPUs?
- onion2k 7y agoWill there be a customer grade TPU in the near future? Would the nVidia Jetson count?
- alexhutcheson 7y agoJetson is for inference, not training.
- The_rationalist 7y agoARM socs already have NPU, but I guess they doesn't count.
- rwmj 7y agoKendryte K210 maybe? It's cheap as chips (pun intended!) I think I got mine for £40 including shipping. https://kendryte.com/ https://kendryte.com/ Note this is only for inference. For training you'll have to use a GPGPU or one of the chips in this article.
- Q6T46nT668w6i3m 7y agoApple’s neural engine in their A-series of SoCs.
- lern_too_spel 7y agoThis article is about hardware for training neural networks, not the inference chips that are in most phones today.
- JoeDaDude 7y agoDo the Google Coral products count as consumer grade? https://coral.withgoogle.com/docs/accelerator/datasheet/ https://coral.withgoogle.com/docs/accelerator/datasheet/
- lern_too_spel 7y agoNo, those are designed for inference.
- dgacmu 7y agoConsumer turing cards are about the closest you get right now. They're pretty reasonable bang for buck for training. They have tensor cores - not quite as many as Ulta, but the entire chip runs at a higher clock rate and the price/performance is better if you don't mind losing a gig or so of RAM and some memory bandwidth.