Bondrewd
Veteran
He's still the only Weibo tourist who filters his shit before posting so kudos to him I guess.I noticed that Greymon's accuracy is slowly reducing
Looks like somebody clipped his sources
Last edited:
He's still the only Weibo tourist who filters his shit before posting so kudos to him I guess.I noticed that Greymon's accuracy is slowly reducing
Looks like somebody clipped his sources
But is there any truth to what he posts, then?He's still the only Weibo tourist who filters his shit before posting so kudos to him I guess.
Some yes some no!But is there any truth to what he posts, then?
IIRC you mentioned that it will not be MCM?that's bs.
That's N33.IIRC you mentioned that it will not be MCM?
That's N33.
He's BSing the TO date for N31.
not too familiar with AMD naming conventions, but Is 33 the 'big chip' and the 31 the mainstream chip?So it has not yet taped out, or was it taped out earlier? Because some say that N33 will be the only RDNA3 part for 2022, while some say that also highend should come next year. But, if N31 taped out, the chance to see it in action next year rises a lot.
The opposite.but Is 33 the 'big chip' and the 31 the mainstream chip?
It can and probably will use, even if GDDR6 densities of that time force it to use clamshell designs.If 256bit N21 can have 32GB
https://videocardz.com/press-releas...o-v620-with-navi-21-gpu-and-32gb-gddr6-memory
why shouldn't N33 have 128bit and 16GB?
Clamshells are very not practical in mainstream stuff.why shouldn't N33 have 128bit and 16GB?
Uhhhh ohhhhh.It can and probably will use
It really is a single GPU with a somewhat funky topology.I really think with the newer fabric AMD will achieved unified cache with dual/quad gpus
Uhhhhh no only the biggest bars go!As other suggest, AMD will dip into RT wholly, when gamers demand it...
Nope.The IFC on Navi 31 GPUs will be split between two chips since it is an MCM design so it's still 256 MB per chip
Not a 'possibility' the thing is made like that.but there's also a possibility of getting stacking technology on the next-gen chips
bingo.Aren't the rumored active interposers / bridges (incorporating Infinity Cache) for Navi 21/22 already a 3D Infinity Cache solution?
Which I/O pins?I think the idea is to bond (more than two) cache + I/O chips below the two compute dies in Navi31/32, not above. For the simple reason that you need to connect those I/O pins to the PCB.