undefined | Better HN

0 pointsjoefourier18d ago0 comments

That depends what kind of ASIC you’re talking about. Cerebras can run models like GLM 4.7 with 355B parameters.

0 comments

Cerebras just uses SRAM instead of DRAM. An ASIC would instead hardwire the neural network.

Surely it's more of a spectrum? From a CPU, to a TPU, to a chip that hardwires softmax attention but lets you store arbitrary weights, to one that hardwires the weights directly.

j / k navigate · click thread line to collapse

0 pointsjoefourier18d ago0 comments

That depends what kind of ASIC you’re talking about. Cerebras can run models like GLM 4.7 with 355B parameters.

0 comments

cubefox17d ago

Cerebras just uses SRAM instead of DRAM. An ASIC would instead hardwire the neural network.

joefourierOP17d ago

Surely it's more of a spectrum? From a CPU, to a TPU, to a chip that hardwires softmax attention but lets you store arbitrary weights, to one that hardwires the weights directly.

j / k navigate · click thread line to collapse