Artificial Intelligence thread

[PopularScience]

A technological miracle! Tsinghua breaks through the computing power problem of large models. RTX 4090 can run the full version of DeepSeek single-handedly

Please, Log in or Register to view URLs content!

 

[9dashline]

A technological miracle! Tsinghua breaks through the computing power problem of large models. RTX 4090 can run the full version of DeepSeek single-handedly

Please, Log in or Register to view URLs content!

This legit? whats the catch, mich slpwer tps?

 

[OptimusLion]

Tsinghua University's open source project breaks through the bottleneck of large model computing power: RTX 4090 single card achieves full-blooded operation of DeepSeek-R1

It is very difficult for ordinary users to run the real full-blooded version of DeepSeek-R1 on ordinary hardware, and the cost of renting a server is stressful even for developers

Please, Log in or Register to view URLs content!

 

[dingyibvs]

This legit? whats the catch, mich slpwer tps?

Some have shown that you can run the full version on a $2000 rig with a lot of system RAM, it's just pretty slow. Looks like this rig also has a lot of system RAM, so the question is how much does a single 4090 speed up the process?

 

[tphuang]

you just need to have enough VRAM to hold the model to run it. inference speed is another story.

 

[pbd456]

VRAM is on chip like cache? Or more like hbm?

 

[Proteus]

Some have shown that you can run the full version on a $2000 rig with a lot of system RAM, it's just pretty slow. Looks like this rig also has a lot of system RAM, so the question is how much does a single 4090 speed up the process?

Please, Log in or Register to view URLs content!

The post right above you have the answer.
This the stream video test on 14 RTX 3090s.