Rx Vega hits SiSoft Sandra, 35% faster than gtx 1080.

[Coaxialgamer]

Quote

Some new benchmarks have surfaced over the last week which show an unreleased AMD Vega-based GPU being put through its paces. The compute-intensive SiSoft Sandra test has it comfortably ahead of the GTX 1080.

Quote

The most recent benchmark to appear is the SiSoft Sandra general purpose compute test which seems to be using a later revision of the Vega card than we’ve previously seen in the Ashes of the Singularity or CompuBench results. The new code, 687F:C3, uses the same essential device ID but what looks like a later revision code, potentially making these latest results more in line with the final release. The previous results have come from a card with the device ID of 687F:C1.

The overall general purpose compute power of what we expect to be the flagship AMD Radeon RX Vega card comes in at 2756.69 Mpix/sec, while the GTX 1080 can be found with scores of 2,050.72 Mpix/sec. That’s a lead just shy of 35% for the AMD card when it comes to the computational side of the equation. Running the same test on our GTX 1080 Ti though has the latest Nvidia card running at 3,451.05 Mpix/sec, or some 25% quicker again than the RX Vega

Well this is nice. Hopefully Vega is up to snuff. Because lately i haven't felt very good about it. 

 

https://www.pcgamesn.com/amd/amd-rx-vega-sisoft-benchmarks

Original source(german) :https://www.computerbase.de/2017-03/amd-vega-4.096-shader/

 

Just realized the image i linked was so blurry it's basically useless. 

Here's the sisoft link :http://ranker.sisoftware.net/show_run.php?q=c2ffcdf4d2b3d2efdbe9dcebdaebcdbf82b294f194a999bfccf1c9&l=en

[Starelementpoke]

1 minute ago, Coaxialgamer said:

Well this is nice. Hopefully Vega is up to snuff. Because recently i haven't felt very good about it. 

 

https://www.pcgamesn.com/amd/amd-rx-vega-sisoft-benchmarks

Salt and all that. But neat.

[EPPHO]

1 minute ago, Coaxialgamer said:

Well this is nice. Hopefully Vega is up to snuff. Because recently i haven't felt very good about it. 

 

https://www.pcgamesn.com/amd/amd-rx-vega-sisoft-benchmarks

Oh shit bois let's hope the price will be decent

[rhavimarques]

is there a chance its another dual-gpu card?

[Techstorm970]

Finally!  An article that doesn't compare TFLOPS! 

 

[CostcoSamples]

Wow, that should put it squarely into GTX 1080 Ti competition!!

[79wjd]

20 minutes ago, Coaxialgamer said:

 

You'd think with all that Vega power, we could get a few more pixels. 

[silberdrachi]

If benchmarks are starting to "leak" what time frame are we looking at on these things?

 

Thinking about upgrading my monitors to one of those dell 34in ultrawides and that would need a new gpu. Sadly used 1080s on ebay havent dropped a lot in price from the TI buyers. And not sure id trust them anyway

[Techstorm970]

Just now, silberdrachi said:

If benchmarks are starting to "leak" what time frame are we looking at on these things?

RX Vega will probably be released at Computex in May.

[Coaxialgamer]

19 minutes ago, rhavimarques said:

is there a chance its another dual-gpu card?

Well sisoft only detects one gpu thread, so extremely unlikely. 

[Coaxialgamer]

18 minutes ago, CostcoSamples said:

Wow, that should put it squarely into GTX 1080 Ti competition!!

Which sadly isn't anywhere near enough for it to be competitive, considering vega 10 is around 530mm², compared to gp 102 at 471mm², on a less dense node. Plus, the 1080ti isn't even a fully active chip. 

Unless the final retail chip is faster ofc. 

[Coaxialgamer]

18 minutes ago, djdwosk97 said:

You'd think with all that Vega power, we could get a few more pixels. 

Just realized the pic provided in the article is cr*p. I replaced it and put the sisoft link below the article. 

[ivan134]

26 minutes ago, silberdrachi said:

If benchmarks are starting to "leak" what time frame are we looking at on these things?

 

Thinking about upgrading my monitors to one of those dell 34in ultrawides and that would need a new gpu. Sadly used 1080s on ebay havent dropped a lot in price from the TI buyers. And not sure id trust them anyway

May, which is when Prey 2 drops.

[Prysin]

@App4that

yo, cry. now.

[Valentyn]

32 minutes ago, techstorm970 said:

RX Vega will probably be released at Computex in May.

Sounds about right, then give or take another month or two before people can actually get something in their hands.

Well depending on if it's another RX 480 fiasco launch.

[Mr.Meerkat]

17 minutes ago, Coaxialgamer said:

Which sadly isn't anywhere near enough for it to be competitive, considering vega 10 is around 530mm², compared to gp 102 at 471mm², on a less dense node. Plus, the 1080ti isn't even a fully active chip. 

Unless the final retail chip is faster ofc. 

?? I think he's talking performance wise...

 

AMD has always had more "processors" in the GPUs for the same performance so the die being larger is no surprise to anyone (not mentioning how HBM may be a contributing factor) although 512 extra cores doesn't sound like it would be enough to make up the difference if we were to look at what happen previously  

[jasonc_01]

42 minutes ago, CostcoSamples said:

Wow, that should put it squarely into GTX 1080 Ti competition!!

The 1080ti is still 25% ahead of the this new Vega test in Sandra

[CostcoSamples]

1 minute ago, jasonc_01 said:

The 1080ti is still 25% ahead of the this new Vega test in Sandra

oh...  =(

[Valentyn]

2 minutes ago, jasonc_01 said:

The 1080ti is still 25% ahead of the this new Vega test in Sandra

 

Has no one noticed the "vega" GPU on that page is running at 344Mhz, and on an FX 970 motherboard?

[Liltrekkie]

Oh cool, another modern heater! 

 

Just kidding haha.

 

Seriously though, I wonder what affect HBM has on die size, I also wonder what these revisions mean, clock speed boosts? Enabling more feature? Figuring out power requirements or usage?? 

 

I feel like vega 10 could be similar to the 1080TI or just a few percent under it. Even if Vega was 3-8% "slower" if it was priced right, it would be a very good buy. 

[Coaxialgamer]

7 minutes ago, Mr.Meerkat said:

?? I think he's talking performance wise...

 

AMD has always had more "processors" in the GPUs for the same performance so the die being larger is no surprise to anyone (not mentioning how HBM may be a contributing factor) although 512 extra cores doesn't sound like it would be enough to make up the difference if we were to look at what happen previously  

That's only really been true since maxwell and pascal  which were able to do so with it's vastly higher clocks ( and even then there are exceptions). Better ipc and extra cores increase die size. Clock speeds do not. 

Look at the the rv770 vs gt200 ( 256mm² vs 571mm²), cypress vs gf100 (334mm² vs 529mm²) or even hawaii vs gk110 ( 438mm² vs 561mm²) as a recent example. I could even argue that fiji ( 596 mm²) was more dense than gm200( 601mm²) because it had 900 million more transistors in a die size which is slightly smaller.  In all those cases, amd was able to squeeze the same amount or more power into smaller spaces ( though in the case of rv770, it had a half node advantage). 

 

Hbm2 isn't really a factor either, as the memory controllers are smaller than gddr5(x) controllers, and the memory chips aren't being counted. 

 

You do have to keep in mind vega is poised to increase clocks and ipc over gcn, so we can't compare them. 

 

My point was that if can't engineer a chip that's faster than the titan xp despite being considerably larger and on a more advanced node, it's going to be bad. Especially considering vega will be competing against volta until navi comes out, and there are no node shrinks in sight. 

[Coaxialgamer]

8 minutes ago, Liltrekkie said:

Oh cool, another modern heater! 

 

Just kidding haha.

 

Seriously though, I wonder what affect HBM has on die size, I also wonder what these revisions mean, clock speed boosts? Enabling more feature? Figuring out power requirements or usage?? 

 

I feel like vega 10 could be similar to the 1080TI or just a few percent under it. Even if Vega was 3-8% "slower" if it was priced right, it would be a very good buy. 

Hbm memory chips are not counted when it comes to gpu die size, and the memory controllers are smaller than  traditional gddr controllers. 

[TheRandomness]

8 minutes ago, Liltrekkie said:

Oh cool, another modern heater! 

 

Just kidding haha.

 

Seriously though, I wonder what affect HBM has on die size, I also wonder what these revisions mean, clock speed boosts? Enabling more feature? Figuring out power requirements or usage?? 

 

I feel like vega 10 could be similar to the 1080TI or just a few percent under it. Even if Vega was 3-8% "slower" if it was priced right, it would be a very good buy. 

HBM stacks are separate to the GPU itself.

[Coaxialgamer]

10 minutes ago, Valentyn said:

 

Has no one noticed the "vega" GPU on that page is running at 344Mhz, and on an FX 970 motherboard?

Considering it also reports it to have 16kb of cache, i would argue that reporting isn't perfect. The article also points that out. 

[jasonc_01]

11 minutes ago, Valentyn said:

 

Has no one noticed the "vega" GPU on that page is running at 344Mhz, and on an FX 970 motherboard?

Its being misread by sandra