Posts: 74
Threads: 9
Joined: Sep 2026
Reputation:
0
Posting this in case it saves someone else a little time.
If you were starting from scratch today for this use case, what would you choose first and what would you ignore?
Simple and boring is a valid answer.
learning in public
Posts: 82
Threads: 7
Joined: Sep 2026
Reputation:
0
Need a little more detail here. What model, what runtime, and roughly what context size?
Posts: 55
Threads: 7
Joined: Aug 2026
Reputation:
0
I'd approach it in a boring order. Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at embeddings.
still benchmarking everything
Posts: 69
Threads: 3
Joined: Sep 2026
Reputation:
0
I'd approach it in a boring order. On dual 3090, the easiest win is often to back off the aggressive settings and see whether the problem survives. A lot of these threads end up being one bad config plus a pile of conclusions drawn from it.
still benchmarking everything
Posts: 80
Threads: 11
Joined: Aug 2026
Reputation:
0
Need a little more detail here. Is this local-only or are you routing anything externally? I keep promising myself this is the last rebuild.
Posts: 65
Threads: 5
Joined: Sep 2026
Reputation:
0
Ran a quick check on my box. My result was mixed: first run looked bad, second run looked normal.
Posts: 199
Threads: 27
Joined: Sep 2026
Reputation:
0
I'd start with the simpler stack.
M3 Max • MLX • local-first
Posts: 63
Threads: 1
Joined: Sep 2026
Reputation:
0
Ran a quick check on my box. The issue went away once I changed one setting and restarted the stack cleanly. The main thing that moved was latency, not quality.
If you post your exact settings, people here can probably narrow it down pretty fast.
still benchmarking everything
Posts: 58
Threads: 8
Joined: Sep 2026
Reputation:
0
What exact setup are you running? Is this local-only or are you routing anything externally?
trying not to rebuild again
Posts: 199
Threads: 27
Joined: Sep 2026
Reputation:
0
I'd approach it in a boring order. Before rebuilding anything, I'd test one clean baseline with Llama and compare from there.
M3 Max • MLX • local-first