Radeon RX Vega 56 Reportedly Beating GTX 1070 by 20%
Click here to post a comment for Radeon RX Vega 56 Reportedly Beating GTX 1070 by 20% on our message forum
sunnyp_343
Having better TFLOPS not means that it give better FPS..GG
fantaskarsef
Well they tested games... now I'd only be interested in 1080p graphs (still the most common gaming resolution if steam hardware survey is anything to go by). Also, was BF1 run in dx11 or dx12? Doom in OpenGL or Vulkan?
But it does look good, yes. The price seems to be pretty reasonable for that performance, too. I guess it will grab sales from the 1070. If only it would have been around a year earlier, I'm fairly sure it would have killed the 1070 indeed, now people already spent their money over the last months I'm afraid.
Jeffs
No one is saying anything about Tflops. In the article there is a little list that compares the amount of FPS on Vega 56 to a GTX1070...
I'm curious for the benchmarks we will see when the cards launch. If these rumours are true then the vega56 could be a good price/performance contendor.
chronek
I wonder if lastest vega driver fix was related with that, if was then eth miners helped amd find performance problem (that one when more than 2GB of memory is used). It is sad that they will release vega in bundles just to make it unprofitable for miners to buy.
I wonder if water cooled vega will beat 1080 ti about the same margin, I wonder how 4x0/5x0 cards performance will increase after they will release that driver.
AlmondMan
Nice... hope it's true 🙂 I could use an upgrade after I went to 3440x1440.
Stormyandcold
When we look at HH's numbers though for 2k resolution vs these leaked numbers, then, it's not so clear;
•Battlefield 1: 95.4FPS (GTX 1070: 72.2FPS) (Guru3D GTX1070 80fps) Vega 56 still better.
•DOOM: 101.2FPS (GTX 1070: 84.6FPS) (Guru3D GTX1070 113fps) GTX1070 better?
Waiting for your review HH.
Kohlendioxidus
Looks like I was right, the same happend with Ryzen aswell and after launch everibody was like :3eyes:. This number put RX56 almost in the GTX 1080 teritory actually -> fact that needs to be confirmed, but just imagine what the full Vega can achieve. I think they will even kick GTX 1080 Ti ass aswell! In this case I will really consider buying a RX 64 Vega as it is at least double the power of my 290x 🤓!
Nurmi
New 2019 must click leak Ayy you wont believe the results Anunnaki confirmed
-YouTurd
Ryu5uzaku
Kind of hard to compare two different reviews which is the issue of course. No idea where they have gotten their numbers from in the game if so.
But I am sure we will get an awesome review from HH. Until then I am taking all with grain of salt.
@Kohlendioxidus
If it is "only" 1080 level it's already pretty much double your card. 1080 ti more then double.
TheDeeGee
60 Watts more to make it 20% faster... hmmm i dunno.
Also 1070 is from AMDs 400 Series... so ofcourse a new Series will be faster.
Fact is AMD is still struggling with TDP/Performance Ratio.
__hollywood|meo
kyle @[H] commented that he doesnt put much stock in tweaktowns #s, but then again, with their beef from years ago...no way to tell if its professional skepticism or just a long memory :P
prices already leaked. 56 will be $400, & due to mining, 1070s are currently basically impossible to find at that price point. eth is still floundering, however...cards from both companies should be easier to find """"soon""""
The Commenter
If it's 20% faster than a 1070 that means it's as fast as a 1080.
BReal85
Sorry, but given these news, you are crying about the TDP/performance ratio, which means Vega56 seems 20% faster but having 40% more TDP (actually reference 1070 was consuming more power than its TDp, so 40% might not be true at all) compared to the 1070. Really? 😀
What is a hundred times more imporant is that leak just a customer luring trick or is that really true.
Ricepudding
Far as i know the mining craze has ended again for the most part, currency have dropped massively again. tons of rx480/580s back on ebay 😀
Still these results are not really impressive, woo a new card has beaten an old card, this happens every year... just AMD cant seem to keep up in the GPU market right now, with Volta round the corner it is quite worrying for them, if they even have a plan to go against that at all?
fl2015
Those numbers sound a bit suspect tbh especially doom as I can get more frames @ 1440p with a 6 year old processor. Also an extra 8 CU's would put the vega 64 well ahead of the 1080, something in their own slides they sort of admit is not the case:
http://i.imgur.com/OZIulX1.jpg
What the slides do show is higher minimum fps, that might make these results make more sense if the "source" was told to calculate average fps with a lower percentage minimum than most reviewers tend to use. That might make Vega look a smoother experience but not faster.
I'd wait for an actual review or at least a proper leak before boarding the AMD hype train.
Redemption80
Even the AMD slides have the 1080 slightly ahead in Doom when it comes to certain minimums, so do not see the 1070 being that far behind Vega 56.
This leak just looks dodgy, and actually a bit lazy.
fantaskarsef
Full Vega = Vega64, which AMD themselves compared to a 1080 @4K, which historically seemed to be AMD's territory. Look at their own slides when they paper launched the cards. And sorry, the 1080Ti is out of AMD's current reach, don't see either the Vega 56 nor the Vega 64 beating that card, or AMD would have stickered it onto their own presentations already.
Also, we're talking about stock clocks, and we know that currently Nvidia seems to OC better than AMD might, we will have to see if AMD's stock air coolers are that great for a high TDP card like Vega.
It will come down to real reviews across more games and resolutions (like Hilbert's reviews). We've seen 4 games, which isn't all that comprehensive, and we don't know all the real settings used (even the OP's linked post doesn't give all the info).
gianluca
That VEGA is a reference card and its TDP is lower than the one of my RX480 Nitro+ OC (225W on paper), offering a lot more powaaa. In regular usage and in game (and even in benchmarks in my personal experience) you'll never see the maximum TDP. And the customer can always undervolt the card, keeping the same performance, consuming less and keeping the card way silent and cool.
AMD tends to give his products more voltage than what is needed. Don't know why, but it's like that since a decade at least. The owners of AMD cpus and gpus know this very well. I wouldn't worry about the TDP, at all. 😉
user1
Wouldn't be suprising, the 1070 is heavily cut down from the 1080, 33% less cuda cores active, the rx vega 56 is only cutdown by 14% from the vega 64,
Assuming similar performance between the 1080 and rx vega 64 that would put vega 56 ahead the 1070 even if vega 64 is slightly slower than the 1080.
Given the default clocks of vega 56 it will likely have alot of headroom as well if you can feed and cool it.
inb4 1700mhz+ with water cooling.
fantaskarsef
Pretty spot on. Yet, if we're talking about water cooling, compare apples to apples and run both a 1070 and a 1080 water cooled against Vega 56 and 64 respectively.
Not sure they'll win lots in comparison if every chip is OCed to their maximum under water.
Don't get me wrong, the Vega56 still looks like a great card, better bang for the buck than the Vega64 to me personally, but don't get in heads over heals and compare an overclocked Vega56 to a 1080Ti, that's a comparison AMD can't win. Even Vega64 is compared to the 1080 by AMD themselves.
What's yet to see is how the board partner's cards will OC and how water cooled Vega64 performs, these are the real questions, and probably the best selling cards of AMD's lineup anyway.
