@User_M31i
iAccount based inIndia!
About this account
- Account based in
- India
- Connected via
- India Android App
! X says this location may be affected by a proxy or VPN.
Account-level information from X, not a live location or the device used for a specific post.
(CUDA Kernels, C,C++, python, cryptology.)
Singularity.
Joined July 2019
- Tweets5.9K
- Following5.3K
- Followers727
- Likes29.7K
User M31 retweeted
“Every difficulty in life presents us with an opportunity to turn inward and to invoke our own inner resources. The trails we endure can and should introduce us to our strengths.”
― Epictetus
Make precision alignment of your models and precisely fine tuning, on your GPUs and sell the weights. Business model??
Why??
Because for doing precision alignment and fine tuning, you need massive compute, you cannot done this in you laptop or PC, or small server, you need massive data centers,
So, all these big tech firms are doing, that finding the best, highly precision, aligned and fine tuned, weights, for the future tech, so they build framework and blueprint for the future that is coming.
GPUs are mostly for heavy graphics uses, but we can use them for AI and other heavy works where we need high computation. That is why we need to do parallelisation of GPUs, so we can extent their compute and made them more useful as according to our need. That is why we need CUDA for NVIDIA GPUs, and OpenCL for other company GPUs available in the market. So where we can use CUDA to make our GPU more useful and effective. 👇
•AI
•Image rendering.
•3D simulation.
•Weather prediction.
•Atom simulation.