I think the title is clear enough, not too specific about the brand if the pricing is good enough.
I have a Corsair Vengeance LPX kit already and something matching would be great but largely a non-issue.
1 Like
Bought Patriot Viper kits (128GB total), okish price.
Woah buddy! What are you doing with all these rams man
Local LLMs, for now GPT OSS 120B and will test GLM Air 4.5. Qwen3 Next is also on the radar once llama.cpp adds support for it.
I do have it paired with a 7900XT (20GB) to speed up stuff.
1 Like
Which processor have you paired it up with? What’s the TPS and prompt processing time you’re getting with OSS 120B?
5900X for now due to budget constraints, bought it used on the forum here.
TPS varies from 13-15 tokens/sec. I haven’t used llama-bench to get the exact TPS/PP speed, just testing on LM Studio currently.
I’ll move to Linux soon as soon as my SSD arrives, will have better/exact numbers then.
1 Like