This got me thinking: What is the techpress doing? The amount of "ignore Vulkan" and "choose Gameworks" for benchmarks is scary to me. I just do not get what is happening right now! Is Nvidia now the supreme evil and invading our brains or just average ignorance (I did not say stupidity!) of the people who should advise to their best knowledge? Especially the "ignore Vulkan" is giving me headaches as that is playing in M$ hands with their DX12 only on Win10 policy. This board of promoters has to mean something!
There is more... In my overview I have noticed some interesting stuff as well...
If you don't want to watch - Techpowerup have changed their graphs a few days after the FU edition 1060 review... And what do you know, the performance difference have increased from 6/7% to 9/10% in 1440/1080p respectively.
Very nice break down of the numbers and performance of both cards. Thank you. I do think Vulkan is game changer that will enable AMD to stay strong in years to come their performance and foresight to develop this API further with their own products in mind was smart.
It just makes no sense right now to purchase nvidia regardless of whether it is priced equally or not due to the lack of optimizations and support for newer API's.
Thank you... I have not seen by now, I think this is the first time, a review site to change it's benchmark numbers after release... I mean comon... I loved their graphs, but I guess now I have to use another website for graphs... On top of that the ignorance of other websites, testing with HBAO+, an Nvidia feature, Project Cars, etc... It's horrible. If you don't know 100% what are you looking at, there are many things that are misleading... PS: I follow ComputerBase.de since back in the ATi 3000 series... Still remember seeing interesting numbers, different from most other places there...
clueless on whether it warrants it or not ? Looks like it is handled until windows pushes it harder. Like their app store ? Which i nuked ha ha suckers.............................. lol
There is one game with final support, two in beta, and one that is listed as upcoming. Hard to benchmark something that is not prevalent yet.
Also a pathetic list of supported games, most of which being in beta, "support coming soon", and even a couple that are game breaking on systems so bad that in one instance it was removed.
Did they ever respond? Sometimes reviewers test with a driver, put out the video, and then come to find a better driver was released on day 1 or a couple days after they had tested but still before launch. There is probably more to this story...
To who? To an unknown channel with 10 subscribers and a video wit less than 60 views? Nobody cares about that, and nobody cares to investigate.
The day of the review was the launch day. So that is not an argument. They literally came back and changed the numbers. I would like to see atleast mention somewhere in the review "We retested so and so because of issues with the drivers.". Other websites are doing things like this, putting disclaimers, driver issue gave us 10fps instead of the expected 59... We will update the review with newer driver... Not TPU...
I would like to remind you, that is exactly the same argument, that was used back ini the day for DX12... 2 years later...
Yes, but the list will grow in just about an year. From that point forward, the list will only grow. So ignoring the growing list, because it's small today...
Look, i understand the argument. I, for example, am not playin any DX12 or Vulkan games. Still, if i buy a GPU today, i would like to know it will support the new technologies at the moment, because i will not change the GPU next year. Still rocking 270X... My only regret is, that if i was to buy the 260X i would have had Freesync and not needed to upgrade, before i buy new monitor. Now i need to change my GPU... So i rather make sure it will last me a bit.
I wonder how many reviewers changed the RX 270X from being 10% weaker to 18% stronger than the GTX 760, a shame, I thought of buying the GTX 760 originally.
Exactly the reason i loved TechPowerUp... They literally redid every test on every GPU for every new review. 3 reviews down the line, everything is retested with the newer drivers and is more accurate. Now i can't trust them to be accurate, after i have caught them changing benchmarks with no disclaimers or whatever.
Fair point, they should 100% put a disclaimer somewhere if this was the case. But, I do understand that they may have initially released the page with older drivers than those available on release. We're heard reviewers say they have gotten multiple pre-release drivers from Nvidia, so maybe they did the tests 4 days before the launch, but then the afternoon of the day prior to launch, they get the release day driver and that driver makes the difference. Again, they should 100% mark that as the case, but I can see why they might release the original page and then update it when they can do thorough test. Again, no contest that they should have marked it somewhere if this is the case, but there is probably a reason that they were updated other than them being Nvidia shills...
The problem I see is that by showing anything right now, when both DX12 and Vulcan are in their infancy, the results in a year may be drastically different. I mean when any reviewer tests DX11 games, you know that those tests are going to be not only representative of that game by the results as a whole will indicate which card will win in the majority of DX11 games. If the 390 beats the 970 in 8 games and the 970 wins in 2, you know the 390 is probably a better card, and the 970 wins in some outliers.
For example, look at this Legit Reviews Article's DX12 tests: http://www.legitreviews.com/nvidia-evga-geforce-gtx-1060-video-card-review_184301 The RX 480 loses or ties to the Gtx 1060 SC in every DX12 benchmark they did, although the stock Gtx 1060 did lose in a couple instances, and these were all stock clocks of these three cards. Obviously as it sits, if the cards had been overclocked to their potential, the Gtx 1060 would actually beat the RX 480 in every DX12 test as it sits right now. The thing is, they only have two benchmarks, and one of them is 3D Mark Timespy, which was barely just released. You can't draw a definitive conclusion going forward from that, there isn't a large enough sample size or any maturity in the drivers from either camp. Yes the 1060 won today, but in a month when the drivers are just a tad bit better, the situation may be completely revered. There just isn't enough data to assume anything really, I mean there are five DX12 games with somewhat final support and only one Vulcan game with final support.