By using this site, you agree to our Privacy Policy and our Terms of Use. Close
JEMC said:
Bofferbrauer2 said:

Apparently, SRAM, so basically all the cache on CPUs and GPUs, almost didn't shrink at all compared to 5nm. I think some AMD engineers saw the writing on the wall early that they moved it out of the chip and into it's own die both on the X3D CPUs and the RDNA3 GPUs.

**slides**

https://fuse.wikichip.org/news/7343/iedm-2022-did-we-just-witness-the-death-of-sram/

https://www.xda-developers.com/how-death-sram-affect-future-pc/

Oh, they did:

https://www.youtube.com/watch?v=8XBFpjM6EIY

Skip to around the 3 minutes mark.

Yeah, it already started getting apparent at 7nm, but nowhere near as obvious or bad as the "jump" to 3nm. AMD has already put them into their own individual chiplets, Intel is about to do the same with their tiles. Now the question is where NVidia is on the technology.