6 ms·
I think calling VLIW "an adandoned design" is somewhat of an exaggeration, such architectures are pretty common for embedded audio processing.
by owlbite 8mo ago
I think calling VLIW "an adandoned design" is somewhat of an exaggeration, such architectures are pretty common for embedded audio processing.
- mike_hearn 8mo agoSure. I did mention DSPs. But how many people write code for DSPs?
- matt_d 8mo agoWorth adding on that note: From JAX to VLIW: Tracing a Computation Through the TPU Compiler Stack, https://patricktoulme.substack.com/p/from-jax-to-vliw-tracing-a-computation https://patricktoulme.substack.com/p/from-jax-to-vliw-tracin... Google’s Training Chips Revealed: TPUv2 and TPUv3, HotChips 2020, https://hc32.hotchips.org/assets/program/conference/day2/HotChips2020_ML_Training_Google_Norrie_Patil.v01.pdf https://hc32.hotchips.org/assets/program/conference/day2/Hot... Ten Lessons From Three Generations Shaped Google’s TPUv4i, ISCA 2021, https://gwern.net/doc/ai/scaling/hardware/2021-jouppi.pdf https://gwern.net/doc/ai/scaling/hardware/2021-jouppi.pdf
- mike_hearn 8mo agoThanks, that JAX writeup was interesting.
- HarHarVeryFunny 8mo agox86-64 SSE and AVX are also SIMD
- vel0city 8mo agoSIMD and VLIW are somewhat similar but very different in the end.
- HarHarVeryFunny 8mo agoTrue. The ISA in this Anthropic machine is actually both, VLIW and SIMD, and both are relevant to the problem.