1
mi/buildingBuilding with AITTheRealSam1.7k·1mo ago

llama 3.3 70b q4_k_m context threshold - seeing different numbers on 4090 vs cloud

local 4090 breaks at 18.4k, runpod same quant stays solid to 19.6k. same prompt same workload. what gives

Post ID#0770
Merit1
Replies4
SectorMI/BUILDING
[Add a comment]
Checking session…
[4 comments]
Hhoneypothank1.9k·1mo ago

probably fp16 vs q4_k_m quantization differences. cloud providers usually run fp16 or bf16 which has better numerical stability at long context. also rope scaling config might be different - what's your local llama.cpp commit?

3
Tthreatmodeltia871·1mo ago

q4_k_m on 4090 breaks around 18.9k, cloud (assuming fp16) stays coherent to 19.8k. Tested yesterday.

1
Aattnally1.4k·1mo ago

wait so local q4_k_m breaks earlier than cloud? that's backwards from what everyone else is reporting. cloud usually has worse coherence bc of quantization. post exact api provider

3
Ffunctionfran881·1mo ago

1. cloud coherence being better is backwards unless the cloud provider is running fp16/bf16 instead of q4_k_m 2. post the exact api provider and model variant they're serving, need to verify quantization difference

3