LLLLM Community
Community homeHot discussions Latest posts Best of Knowledge base Resources About

LLM Community · Choosing llm: I turned 5 options into a side-by-side comparison

31
LLr/llm·posted by nikic·5 minutes agoTooling

Choosing llm: I turned 5 options into a side-by-side comparison

It took me two weeks of on-and-off digging and plenty of wrong turns. Writing the process down as it happened so the next person spends less time.

On trade-offs, my view is this: if nobody on the team owns this area long-term, do not introduce a second mechanism. With two coexistence you first have to work out which one is even in play when things break, and that costs far more than the performance you saved.

Performance44%
Maintainability28%
Ecosystem and community17%
Hiring difficulty11%

53 votes total

15 comments

15 comments

M
Llinlin·1 hour ago

Saved. I am reworking this area this week — this saves a lot of wrong turns.

305
WwinterMod·12 minutes ago

Has anyone run a controlled experiment? I did, reducing it to a single variable, and the difference was 4% — within noise. So I suspect the main cause is something else.

231
Zzhu_zong·3 minutes agoedited

This matches what we see in production. We only hit it past 3k QPS; the earlier load tests showed nothing — the test traffic was too clean, with no long-tail requests.

169
Kkite·2 days agoedited

This is not a llm problem, it is a usage problem. The docs say this API is not thread-safe and you must lock around it yourself.

296
Ttang_hao·just now

We have run this in production for two years without hitting it. That said, we never reached this scale, so our experience is not really evidence here.

73
Rran_bo·2 hours ago

There is actually a simpler fix that needs no architecture change: move this check up to the gateway and the problem disappears. The cost is one extra lookup at the gateway.

4
NnikicMod·1 hour ago

I see point 3 differently. The trade-off depends on your read/write ratio: read-heavy with little writing means caching actually widens the inconsistency window.

25
Ttang_hao·3 minutes ago

A question: what changes in a container with a 512Mi memory limit? That is how we run it in production.

97
Oops_wang·3 minutes ago

Can you give a minimal reproduction? I ran it locally for ten minutes and could not reproduce on macOS with the latest version.

344
Aalice_dev·just now

Sharing our numbers, 8 cores 16GB, same scenario:

| Concurrency | P50 | P99 |
|---|---|---|
| 200 | 12ms | 88ms |
| 500 | 31ms | 340ms |

P99 clearly collapses at 500 concurrency, which lines up with your knee point.

58
Cchen_dev·28 minutes ago

Agreeing with the above. One addition: with this option enabled the GC count in your metrics doubles, so adjust the alert threshold at the same time or it will keep firing.

30
Lli_ming·3 minutes ago

One counter-example: below llm 7.4 the semantics of that code are different, so do not copy it verbatim. We got burned in staging and rolled back once.

495
Llinlin·just now

Worth learning from this debugging approach. We went straight at the logs and took a much longer route.

101
Zzhu_zong·12 minutes ago

I just read the llm source — the author actually explains the reasoning in a comment, roughly "so that it degrades into predictable behaviour in extreme cases".

27
RraseMod·12 minutes agoedited

I see point 3 differently. The trade-off depends on your read/write ratio: read-heavy with little writing means caching actually widens the inconsistency window.

2

This is the post detail page /en/c/llm/post/p12. Posts and comments are generated deterministically from a seeded PRNG, so the same post always renders the same content and the link can be shared, reloaded and indexed. In production this page reads MySQL for the post, Redis for hot-post caching, and fetches the whole comment tree in a single query on the path column.

See the database schema →