Subsea Engineer, Trading Indicator Dev and AI-Memory Layer Specialist | @echo_locate_ai | @rainbowtrends
NFA
Joined May 2019
- Tweets37.8K
- Following1.7K
- Followers7.6K
- Likes47.9K
Coworker gave me a scrap computer that was going to the landfill. I threw in some spare parts and a pair of 5060TI's and I'm getting over 65 tok/s full context using Strata and Qwen Next Flash IQ3_S. This is nuts.
DDR4 64gb, 2x5060TIs 16gb, 1tb NVME and 5600x
tradingview.com/blog/en/mark…
Update from TradingView just dropped.
IDK how you get this many different AIs to work together, but its pretty cool to see it work.
Back home, got DeepSeek-V4.1-Flash-1017 on a pair of 3090s in a 2014 DDR4 workstation doing 25 tok/s using 1M context. I can't stop laughing and I know Nikos is only going to make this thing faster.
ISTA IQ2_XXS Quant using 1M context
My 2014 HP Z840 with a pair of 3090s is doing 10 tok/s on DDR4 first attempt with no tuning. Absolutely insane what @coldniko and ISTA are doing.
I'm not really sure what all those alarms were coming out of my server after trying to add 2 more sticks of ram to max out all the slots, but after a few rotations of sticks it finally booted.
My 2014 HP Z840 with a pair of 3090s is doing 10 tok/s on DDR4 first attempt with no tuning. Absolutely insane what @coldniko and ISTA are doing.
Thanks to all that came out, heading to the RT Discord for the rest of the night to hang out with the crew.
We are live: youtube.com/live/8glrl79MgYc…
We are live: youtube.com/live/8glrl79MgYc…
Weekly Whale Watch - LIVE xitter.cf/i/broadcasts/1lKQRWbXo…
