Y
Hacker News
new
|
ask
|
show
|
jobs
by
taneq
2 days ago
You might be conflating throughput with latency. 100k tok/s is very different to 1 tok/10us.
1 comments
ssivark
1 day ago
When doing auto regressive inference, how often do you do a CUDA kernel call? What is the main bottleneck at the throughputs you're operating?
link