Nvidia Gets SALTY

[Bofferbrauer2]

Pemalite said:

Bofferbrauer2 said:

Seriously Perm?

If it was just a die Shrink plus increased clock speeds and more bandwith, then it wouldn't have outperformed the old Vega 64 by such a large margin. Radeon VII has some architectural improvements, because the numbers don't add up otherwise. It's certainly not the Bandwith, as many of these games listed weren't limited in that domain to begin with.

Vega 7's clockspeed increased by 16.4%. - That means the Render Output Pipelines (A big limiter on Vega), Texture Mapping Units operate that much faster. - The ROPS is a big one as AMD always seems to be ROP starved.

Compute increased by 9.5% over Vega 64.

Bandwidth increased by a whopping 112.27%.

The 25% or so increases is more or less from increases in clockrate and that bandwidth boost... Vega was never compute bound to begin with.

So yes, seriously. Vega 7 is just like RX 590, few enhancements... Bulk of the gains due to clock increases thanks to the smaller fabrication process opening up extra headroom.

Is there some secret sauce hidden somewhere? Possibly, but we don't know at this point in time... And it is best to leave such speculation until Anandtech has done a thorough analysis on the hardware when it releases.

Just to add in something I wasn't sure at that point, Vega 10 FP64 can only can reach 1/16th of FP32 while Vega 20 can do 1/2, so the chip must have changed quite a bit under the hood. Double Precision is also something that cost quite a lot of energy, hence why it got cut down  in modern GPUs both by AMD and NVidia. NVidia nowadays mostly uses 1/32th, and all RTX cards, including the Quadro, do so. The original Titan had 1/4, while the Volta-based Titan V has 1/2 like the Vega 20, making the latter the probable target if the FP64 capability of the Radeon VII hasn't been cut down.