←back to thread

1-Bit AI Infrastructure

(arxiv.org)
157 points galeos | 5 comments | 15 Nov 24 14:28 UTC | HN request time: 0.206s | source
1. js8 ◴[20 Nov 24 12:19 UTC] No.42193249[source]▶
It's technically not 1-bit, but 2-bit.

Anyway, I wonder if there is some HW support in modern CPUs/GPUs for linear algebra (like matrix multiplication) over Z_2^n ? I think it would be useful for SAT solving.

replies(4): >>42193552 #>>42193565 #>>42195167 #>>42197753 #
2. almostgotcaught ◴[20 Nov 24 13:00 UTC] No.42193552[source]▶
not CPU/GPU but on FPGA finite field arithmetic is a thing; plenty of stuff like this around https://ieeexplore.ieee.org/document/4392002
3. scarmig ◴[20 Nov 24 13:01 UTC] No.42193565[source]▶
There's carry-less multiplication (https://en.m.wikipedia.org/wiki/CLMUL_instruction_set), introduced by Intel in 2010.
4. JKCalhoun ◴[20 Nov 24 16:02 UTC] No.42195167[source]▶
Or, technically, 1.58 bit. ;-)
5. meindnoch ◴[20 Nov 24 20:23 UTC] No.42197753[source]▶
https://en.m.wikipedia.org/wiki/CLMUL_instruction_set