AMD Vega 10, Vega 11, Vega 12 and Vega 20 Rumors and Discussion

AlexFromAU said:
I'm surprised this hasn't been posted here yet, basically there was a 3DMark 11 Performance benchmark test ran on 4th July, and it's PCI-ID is 687F:C1 which means that this is Vega.

See here: http://www.3dmark.com/3dm11/11592003/

If you compare it, you can see that driver name is shown under Graphics Card section: http://www.3dmark.com/compare/3dm11/11592003/3dm11/12256169

And it is yet again slower than the GTX 1080.
Click to expand...

Does a normal GTX 1080 run at 1870 Mhz core and 1450 Mhz memory?

Also 32GB versus 8GB memory. Maximum Turbo Clock at 4.9Ghz versus 4.0Ghz
 
Maybe, the TBR rasterizer depends on the use of primitive shaders.

To quote the article: "The new programmable geometry pipeline on Vega will offer up to 2x the peak throughput per clock compared to previous generations by utilizing a new “primitive shader.” This new shader combines the functions of vertex and geometry shader and, as AMD told it to me, “with the right knowledge” you can discard game based primitives at an incredible rate. This right knowledge though is the crucial component – it is something that has to be coded for directly and isn’t something that AMD or Vega will be able to do behind the scenes."

https://www.pcper.com/reviews/Graph...w-Redesigned-Memory-Architecture/Primitive-Sh
 
Hope to see FS results soon rather than more 2011 720p benchmarks, does ~15% better than the 1080FE here, so hopefully it can extend it further on FS benches.
 
Clock for clock the Vega FE performs worse than the Fury X apparently

vega-v-furyx-doom-4k.png


Article:
http://www.gamersnexus.net/guides/2977-vega-fe-vs-fury-x-at-same-clocks-ipc

Video:
 
I haven't seen the video yet, but that Doom result shows the same IPC as Fiji within a rather small margin of error, not lower.
 
Geeforcer said:
How did they equalize the memory bandwidth?
Click to expand...
They didn't, the Fury X retains the slightly higher bandwidth, which likely accounts for the slight advantage to FuryX. So normalizing the bandwidth and the results, likely they're basically identical performance at the same clocks.
 
So there's still hope for gaming drivers next month that give it a nudge further, if AMD can get it to 20% over a 1080FE it'd be a decent buy. Power consumption is still a worry though.
 
Clukos said:
It trails the Fury X in several tests
Click to expand...

It's practically the same scores between one card and another, save for that AOTS benchmark, for which they state the following:

IMPORTANT NOTE: We have uncovered some curious behavior with Ashes of the Singularity in the process of benchmarking Vega: FE for our next feature test. Our FPS numbers steadily increased pass-to-pass on Vega, growing at one point in a couple percent with no changes between runs (other than more runs). We are unclear if this is some sort of caching going on that is Vega-specific or if it’s just how the game works overall; we’ll have to run non-Vega cards through to see if the same performance behavior emerges. This could impact our above results, assuming “more tests = better performance.” For now, we took the first three tests from each card and averaged them, as those numbers were the most consistent with one another (+/- 1FPS).
Click to expand...

Sure there are some differences, but Vega also has less raw memory bandwidth so it could be coming just that.
When the 32CU Tonga was compared to 32CU Polaris 10 there was a clear advantage from GCN3 to GCN4. But here we see almost zero difference between Fiji's GCN3 and Vega's "GCN5"?


It sure looks like it's just running Fiji's driver path in some kind of fallback/compatibility mode.
And it also doesn't seem to me that Vega's architectural advantages are broken in the hardware, because if we look at the specviewperf scores with normalized clocks:

k29HxXB.png



Here we see 15% to 670% boosts at the same clocks, but in gaming this translates to nothing?
I don't buy it.
 
ToTTenTranz said:
Here we see 15% to 670% boosts at the same clocks, but in gaming this translates to nothing?
I don't buy it.
Click to expand...
Architecture changes for packed math, like dual rate fp16. None of those hardware additions affect gaming (yet) which is all fp32.
 
Malo said:
Architecture changes for packed math, like dual rate fp16. None of those hardware additions affect gaming (yet) which is all fp32.
Click to expand...
Dual rate FP16 would never result in 6.8x more performance. And according to the author, this particular snx-02 test measures vertex performance.
Besides, is FP16 being used in 3D art software like 3DSMax and Maya, both showing close to 30% boosts on same clocks? And 3D CAD design like solidworks and Catia showing 50-100% boosts?
 
ToTTenTranz said:
It's practically the same scores between one card and another, save for that AOTS benchmark, for which they state the following:



Sure there are some differences, but Vega also has less raw memory bandwidth so it could be coming just that.
When the 32CU Tonga was compared to 32CU Polaris 10 there was a clear advantage from GCN3 to GCN4. But here we see almost zero difference between Fiji's GCN3 and Vega's "GCN5"?


It sure looks like it's just running Fiji's driver path in some kind of fallback/compatibility mode.
And it also doesn't seem to me that Vega's architectural advantages are broken in the hardware, because if we look at the specviewperf scores with normalized clocks:

k29HxXB.png



Here we see 15% to 670% boosts at the same clocks, but in gaming this translates to nothing?
I don't buy it.
Click to expand...

"Pro-ish" vs "Normal" drivers... Plus only a 1/4 memory, which someone once told me was super-important in pro apps.
 
Geeforcer said:
"Pro-ish" vs "Normal" drivers... Plus only a 1/4 memory, which someone once told me was super-important in pro apps.
Click to expand...

Whoever told you that was right and it probably referred to many professional user's real-world environments, but not for all benchmarks in specviewperf 12.
For example, the Maya 04 benchmark sees only a small boost between Pro and gaming graphics cards:

wVxsJWF.png


Yet in this benchmark, Vega FE sees a 40% improvement over Fury X, clock-for-clock.
 
ToTTenTranz said:
Whoever told you that was right and it probably referred to many professional user's real-world environments, but not for all benchmarks in specviewperf 12.
For example, the Maya 04 benchmark sees only a small boost between Pro and gaming graphics cards:

wVxsJWF.png


Yet in this benchmark, Vega FE sees a 40% improvement over Fury X, clock-for-clock.
Click to expand...


It's the drivers in other cases. In snx-02, which you specifically brought up, the mighty W5100 outperforming Fury 2.5x.
 
You must log in or register to reply here.
Back
Top