AMD RDNA3 Specifications Discussion Thread

If true from a Performance per Area point of view N33 is nothing short of amazing, getting N22 top bin performance (and more) from a die smaller than N23, on the same process, is really great.
Yea raw GM204 energies from the thing.
Would cost the same $329 too haha.
getting N22 top bin performance (and more) from a die smaller than N23, on the same process
Yuo forgot power but yea (mostly not fucked RDNA3 is vewy nice).
 
If true from a Performance per Area point of view N33 is nothing short of amazing, getting N22 top bin performance (and more) from a die smaller than N23, on the same process, is really great.
>=1.4x perf/w, perf/area and transistor on a similar node is engineering wizardry, Maxwell comparison is very apt and am excited to see what it can do in an APU. It'll be approaching 2x perf/w of Navi 10 as well
 
On the SIMDs, is there any idea here on how often it will dual issue ? I mean vaguely of course. I assume it's most of the time or it would hardly seem worth it.
There’s a recent Mesa commit that puts it all together. Most useful dual issue instruction is probably fmac, which is a 2-operand version of fma that’s commonly used for matrix multiplies. Dual-issue dot2c I could see being useful too.

In short, if your shader is fp math heavy then VOPD will help a lot. Integer math, load/store, no help at all.
 
Anyway, the 7600 should be a compelling package. $349ish for a 50%(+) performance bump over the 6600xt should be great for sales. The return of the better than console settings "cheap" card!
This would be pretty much ideal for me. (this is N33 right?)
 
Anyway, the 7600 should be a compelling package. $349ish for a 50%(+) performance bump over the 6600xt should be great for sales. The return of the better than console settings "cheap" card!

I've advised a few people who don't immediately need a new GPU to hold off for N33. Just have to hope AMD doesn't pull an Nvidia with the pricing on this one. At <=$350 this could be the GPU savior.
 
Real scores faster than 4080 in vulcan but slower in opengl.. open gl are very accurate in gaming results
 
Real scores faster than 4080 in vulcan but slower in opengl.. open gl are very accurate in gaming results
The second benchmark is for OpenCL.
 
Real scores faster than 4080 in vulcan but slower in opengl.. open gl are very accurate in gaming results

That link has Vulkan and OpenCL. Only one of those is relevant for gaming.

Regards,
SB
 
Back
Top