AMD Radeon VII Benchmarks Seemingly Leak: Stiff Competition for RTX 2080

[Guest]

The AMD Radeon VII's 3DMark and Final Fantasy XV alleged benchmark results have been leaked.

AMD Radeon VII Benchmarks Seemingly Leak: Stiff Competition for RTX 2080 : Read more

 

[redgarl]

Well, I must say I am surprised. However, I didn't know the following graphs were the FFXV benchmarks.

I wonder why even publishing those right now.

 

[mjbn1977]

why are you only showing the Firestrike benchmarks? There are also Timespy results available, and in those the Radeon VII is quite a bit behind the RTX 2080. The Radeon only got 8700 points in that one. The RTX 2080 gets around 10500 in Time spy.

 

[Dantte]

why are you only showing the Firestrike benchmarks? There are also Timespy results available, and in those the Radeon VII is quite a bit behind the RTX 2080. The Radeon only got 8700 points in that one. The RTX 2080 gets around 10500 in Time spy.

Shhhhh. Just buy it!

 

[spentshells]

Nice, I think I should get one of those.

 

[jimmysmitty]

Hate "leaks". Mainly because they are all just a crap shoot.

Here is what I found. I found the regular preset and beyond the fact that each of the nVidia run ones are by different users they also each are running different driver versions:

25.21.14.1694 Regular 416.94 (November)
25.21.14.1771 Extreme 417.71 (January)
25.21.14.1634 Ultra 416.34 (October)

That alone invalidates this. They are all different systems with different motherboards, SSDs and RAM. That could affect any number of performance metrics. Then there is the OS. All are 10 but are they on the same patches?

Then last but not least in Fire Strike, the one I found, the actual software version they ran is different with the "Vega VII" one being newer than the RTX 2080 benchmark. That can change performance scores alone let alone anything else I mentioned.

There are also no details on the GPU itself such as what the clock speed was when ran. It just shows 0MHz. Was it stock? Overclocked? With none of this information it should be tossed out as nothing.

Why is Toms posting a rumor with no research into it?

 

[jh953zm]

Nice results, but I have the RTX 280 Ti, so AMD will have to come out with a new graphics card that scores higher than my card at the same or less power draw before I switch.

 

[jh953zm]

I meant 2080 Ti

 

[drivinfast247]

Nice results, but I have the RTX 280 Ti, so AMD will have to come out with a new graphics card that scores higher than my card at the same or less power draw before I switch.

I meant 2080 Ti

That's cool! As the Radeon VII costs half as much it's not supposed to be in competition with the 2080ti.

 

[nadhir.brighet]

Final Fantasy shouldn't be used at all. It's nice to have some competition, hopefully it'll have some impact on the prices.

 

[alextheblue]

Why is Toms posting a rumor with no research into it?

Look at the author. They still are posting bad info and never corrected the last article. For example:

By disabling or gimping the Radeon VII's FP64 performance, AMD effectively creates a huge gap between the two models. Despite all this, the Radeon VII should be capable of pumping out 862 GFLOPs of FP64 performance and stay ahead of the Nvidia Titan RTX (509 GFLOPs) and GeForce RTX 2080 Ti (420 GFLOPs).

"You keep using that word, I do not think it means what you think it means."

Also your FP64 numbers are wrong. Radeon VII is 1:8 for FP64 (vs 1:2 for the MI50), which is double what you are reporting here and in the last article. It's also double the ratio of previous Vega Radeons. If you don't believe me, reach out to AMD for clarification. Side note, Nvidia gimps their FP64 even more on consumer and prosumer cards like their Titan RTX. So... yeah.

https://twitter.com/RyanSmithAT/status/1085680805802733568?s=20

 

[Kenneth Barker]

For all you first time AMD fans here. You are about to get a rude introduction into how Radeon launches go. The last 4 years have been the same sad and disappointing pattern. We will see soon enough.

When Vega launched way back, they had to push the savings found in freesync monitors to help justify the card during press release. VII will be slower than the 2080 in MOST titles. It will cost more and consume more power. That has been the going average for Radeon high end for a good long while now. I would be willing to put money on this coming true and being the case with VII.

AMD pushes the hype train really, really hard. Often failing to deliver. You can certainly appreciate their cards for what they are, but it does nothing but make you look bad when the same things happen time and time again.

 

[mjbn1977]

I agree. They would have been better of in keeping the expectations right and say, hey....we have a card that can beat the RTX 2070....for the fanboys, since it will cost $200 more.

anyway....did anyone see leaked R VII benchmarks from 3Dmark Port Royale? ;-))))

 

[konjiki7]

What a crap article.... To many variables between software versions, drivers and pc configs. The FFXV benchmark of all things.... SMH Maybe Tom's is in desperate need of clicks?

 

[urbanman2004]

Definitely taking these results w/ a fist full of salt

 

[mjbn1977]

I don't think so. It is just the typical chatter right before release of a new videocard. Reviewers start reviewing cards and that's why you start benchmark results popping up in online databases of benchmark tools. I think its not wrong to mention it, especially since all those leaked benchmarks were pretty correct with the RTX 2080 and 2080Ti, and had the ballpark number pretty much spot on. I don't see why that should be different this time around....

 

[shrapnel_indie]

I don't think so. It is just the typical chatter right before release of a new videocard. Reviewers start reviewing cards and that's why you start benchmark results popping up in online databases of benchmark tools. I think its not wrong to mention it, especially since all those leaked benchmarks were pretty correct with the RTX 2080 and 2080Ti, and had the ballpark number pretty much spot on. I don't see why that should be different this time around....

Agreed it's chatter... and a leak.

The Radeon numbers may be close or bang-on.... or not... that I don't think is the real issue. I think the real issue is the comparison(s) made against the RTX-2080 itself. All benchmark suites from the RTX are not consistent in who ran them, the hardware configuration, and the software/drivers/OS configuration. Because of this, it's hard to compare properly.

 

[mjbn1977]

well, I ran multiple benchmarks with my RTX2080 over the last view months, with different nVidia Drivers, different Windows Versions, etc.....the benchmark results usually only fluctuate less than 1% at factory oc. even if I use different drivers or have CPU overclocked or not. By the way, my RTX 2080 reaches 2300 points more in Timespy compared to the leaked Radeon VII results. And even my Firestrike results are higher than the leaked one from the Radeon. But it Firestrike the difference is indeed not much.

 

[johnrob]

I will happily wait for some legit benchmarks.

Rumors are great and all, but I'll likely have my next graphics card for 3 years so I'll make sure it's the one I want.