Posts: 53
Threads: 7
Joined: Sep 2026
Reputation:
0
Posting this in case it saves someone else a little time.
I'm not trying to build a perfect architecture diagram. I just want a sane starting point around SGLang and Threadripper box.
Would rather learn from your scars than collect my own all at once.
Posts: 38
Threads: 5
Joined: Sep 2026
Reputation:
0
I tested this tonight. My result was mixed: first run looked bad, second run looked normal.
Posts: 26
Threads: 4
Joined: Sep 2026
Reputation:
0
Ran a quick check on my box. On 5090 with llama.cpp, I saw basically the same behavior. The main thing that moved was latency, not quality.
curious and caffeinated
Posts: 52
Threads: 5
Joined: Aug 2026
Reputation:
0
A couple questions before I guess too hard. Are you testing this on mini PC or something else?
trying not to rebuild again
Posts: 51
Threads: 8
Joined: Sep 2026
Reputation:
0
What worked for me was pretty simple. On A6000, the easiest win is often to back off the aggressive settings and see whether the problem survives.
learning in public
Posts: 45
Threads: 6
Joined: Aug 2026
Reputation:
0
worth checking the driver first.
mostly local
Posts: 25
Threads: 1
Joined: Sep 2026
Reputation:
0
I'd start with the simpler stack.
Posts: 36
Threads: 5
Joined: Sep 2026
Reputation:
0
Appreciate the detail. The pattern around prefix caching has been pretty similar for me.
Posts: 37
Threads: 4
Joined: Sep 2026
Reputation:
0
A couple questions before I guess too hard. Do you see the same thing with a smaller batch or shorter prompt?