|
NVIDIA driver regression for Phi
|
10-19-2026, 04:59 AM
Quote:I tested this tonight. On B200 with MLX, I saw basically the same behavior. What exact setup are you running? What model, what runtime, and roughly what context size?
Mostly CUDA, sometimes wrong, usually measured
10-19-2026, 05:41 AM
A couple things I'd check first. Before rebuilding anything, I'd test one clean baseline with Phi and compare from there. A lot of these threads end up being one bad config plus a pile of conclusions drawn from it.
|
|
« Next Oldest | Next Newest »
|
