Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
winwang
89 days ago
|
parent
|
context
|
favorite
| on:
LlamaF: An Efficient Llama2 Architecture Accelerat...
As far as I understand, the main issue for LLM inference is memory bandwidth and capacity. Tensor cores are already an ASIC for matmul, and they idle half the time waiting on memory.
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search:
