Yes I have to cap it because of electric wiring

I have ran some strata tests will post later. Also did Maya tests. Multiple agents test, etc

Hmm for me it's when tasks a long for example with /goal. A lot of tool calling can happen and I think a better quant will be able to do the job in significantly less tool calls than a smaller quant. I also stress test quants against tools i use often to see to decide

Please compare quants you’re running vs seeing when comparing speeds

Does it work with strata or exlama?

rockstar marketing is already putting "one of the biggest on tv" on the box

Yes, I have some tests with Maya and strata as well. Will post soon

Yeah it runs way faster on strata just showing results for a bunch of stuff for some followers

Ohhh yes I am running bpw 6.05 which I think is about q8

What quant size?

leaning into local ai and capping the m6 at 32gb is very apple

one prompt, one file, a full asteroids game. running locally on my 8x 3090 rig qwen 3.8 flash-next on exllamav3 (6.05bpw + MTP draft) 124 tok/s 0.44s to first token ~1,270W while generating (8 cards, each capped at 160W) @Alibaba_Qwen @turboderp_

😂

Sheeeeesh

What’re you going to mine?

Lmaooo this is wild hahaha

glad i bought this rig one piece at a time. pricing the whole thing at today's prices would be a very different conversation

I haven’t finished the new cyberpunk season 2 yet but I’m on episode 8 So far I like the first season better

haven't measured it yet. tensor parallel testing is next and i'll share numbers