What are the ways to make models work on AMD, Intel, etc faster? Do AMD, Intel, Tenstorrent, Qualcomm (due to their Modular.ai acquisition) combine forces to take on CUDA?
DeepSeek drops CUDA kernels for fast inference, if users care about having day-0 fast inference, you'll have to play catch up every time.
Also, I've heard many western labs use open source CUDA kernels (usually from Chinese labs), IMO it makes little sense to move away from CUDA right now unless you're a big player that can afford a patchy ecosystem with internal development (e.g. Anthropic with JAX/TPU).
Precisely. But that might change. Once Tenstorrent releases a high VRAM card that only does linear algebra then the price might genuinely start to make sense.
I don't think it needs higher VRAM, the whole point is horizontally scaling.
Most models have moved to MoE and expert routing which is done over infiniband, so tensor parallelism where Tenstorrent was really weak at isn't used much, I can see it as a viable alternative where latency isn't a concern but cost is.
Also even if they wanted to increase the capacity on their GDDR, they wouldn't be able to since they are limited by the GPU memory channels.
reply