Gadgetory

have your favorite tech reviewers and also the press been getting sent golden sample r-tx and GTS gravis cards well today we're gonna be exploring this question more in depth where we have a retail sample for an RT x 2060 this is the cheapest card I could find in Australia on a sale on eBay I got it for a little over 500 Aussie dollars which would be roughly close to around the retail price in the USA and then besides that the RT X 2070 here which is a review sample sent from Nvidia we have an RT X 27 T retail Galax ample which is currently the cheapest in Australia as well coming a little bit under 700 Aussie dollars or at least that's what I got it for then this one here was the card that we used in the original comparison against the RT X 2070 versus the RT X 2060 if you guys haven't seen that video I'll put the link up here for you where a lot of you guys said in the comments that this RT X 2060 looked like it was underperforming and so I went back and double-checked the numbers and Metro Exodus definitely was standing out as giving odd results but some people in the comments were saying that this game was giving out odd results to begin with but I then sent Galax an email and said look this card could possibly have some weird results so they decided to send us out another r-tx 2060 for this video for the comparisons so let's put all these cards on the test bench and then come back and see if the press have been getting sent those golden samples so I finished up testing all yesterday at night and I was benchmarking in double-checking results and this gave me the answers that I really needed to know so I'll be dropping the yes man happy used parts Cape hunt act and I'm gonna be putting on the serious face new parts investigation for you guys and the answers I really needed to get came from these two graphics cards right here the galaxy x models ones a black edition ones a white edition they're the exact same card at the core exact same coolers one just got a wide aesthetic and they even share the same v bios so when I put this card in after this black card it actually recognized it as the exact same graphics card and even applied the same overclocks that I left when I took this out and this is really interesting because as we go through some of the results here we can see that this white variant here is actually performing pretty close to this gigabyte edition here which is a retail sample one that comes in around MSRP and some of the results were really really crazy especially Far Cry 5 where we look at this Galax Black Edition here clearly coming out well in front of even the overclocked gigabyte retail sample and then we look at the power consumption results and we can see that they were in favor of this black edition here and so moving through some of the other results here and I tested these four games of 1440p as you guys have been requesting we saw this black edition here just give the retail sample of beat down and its own white edition variant a beat down as well apex legends was a similar story I decided to start testing at water treatments since I believe this game is now hard locked at 144 FPS so I tested again 1440p was a great resolution to test that to make sure we weren't hitting that cap and we can see in tandem that the r-tx 2070 samples I have here actually are performing very close to one another there's not too much of a discrepancy here but these three are TX 2060 samples clearly have some variants going on Resident Evil doesn't really show this a whole lot and the synthetics benchmarks like 3d mark firestrike don't show this either in fact it has the retail sample gigabyte card pulling out ahead of these two galaxy cards and moving on to time spy extreme showed similar things but then we moved back to the games there is a clear difference so what exactly is going on here and this is where we're going to try and answer some questions because this is the worst I've ever seen it in the history of benchmarking on my channel and benchmarking in my life starting off first you guys are probably gonna want answers as to why different reviewers are getting different results and I believe it has to do with GPU boost 4.0 versus the previous editions of GPU boost that was released with ten series seven series and six series cards and in that manner this time around it's worse than it's ever been because the GPU boost 4.0 isn't just detecting temperatures and also voltages it's now detecting the quality of the silicon on board and boosting that gravis card in games according to that silicon itself so in other words yes it's very possible that the press and your favorite tech reviewers are getting sent out golden samples and that is affecting their benchmarks and sometimes it can be to the tune of 10 or more percent which is quite a big sum to skew out results and also looking back at my 20 60 vs 2070 video the conclusion and the results don't seem too far-fetched to what you could expect if you go out and buy a retail sample of both these cards so in that case other reviewers could have got a really good 2060 in and then got a really lackluster 2070 and that would show that the 2060 especially when it's overclocked is catching the 2070 which again as I said I haven't seen that around the studio here but further proof of this and things that weird me out is these cards here perform pretty much identical in the simulated benchmarks like fire strike and time spy extreme but when we look at the power consumption results they were pretty much the same as well and the only indicator that I had to show that these cards were differing in the gaming benchmarks was the schools themselves but also the core clocks that averaged out in MSI Afterburner this black edition here was clearly overclocking better on the core clocks then this white edition here but from going on forward from here in every review I do with these graphics cards especially the main one in focus I'm gonna be showing the core clocks averaged out and also the core clocks averaged out when they're overclocked those are the two main factors that will affect performance you guys get an indication of how well this card is performing compared to other reviewers cards and then you can check their results and cross-reference and say hey he got a bad 2060 or he got a really good 2060 and that's affecting the results so that's likely where you're seeing a lot of these variances come from and especially I've seen it myself in the last couple of months numbers have started being all over the place as hopefully this gives you guys some more answers but when we compare it to a CPU review I think it's even worse than the press getting sent out ring is for CPUs and the reason being in those cpu reviews you can at least see that every CPU that gets sent out as long as you turn off that multi-core enhancement is going to run at the same clock speeds and so for ninety nine hundred K you can say okay this guy's getting four point seven gigahertz out of the box but his chip is overclocking higher and we can see that with the 5.1 gigahertz numbers and that's why it's crucial now to I guess with GPU reviews now include the core clocks on what they're averaging out to when that said person is testing the games and so if you guys are scratching your heads I'm gonna throw another curveball at you and this has got to do with why I test on the same day in the same ambient temperatures with the same driver even though I don't test as many games as other reviewers or I don't test as many graphics cards I always make sure you guys are getting these same apples-to-apples results on that particular day although particular two days across the same drivers and if we don't do that then we could add even more variants into the equation for instance if we're testing on a four one seven driver or four one a driver and then testing that against another card which is running four one nine drivers and the ambient temperatures were twenty degrees on the four one nine driver card versus the four one seven we could see some results that just balloon out of control and now I know a lot of the colleagues on youtube like gamers next is hardware on box J's to sense all those guys do their best to give you guys apples to apples results but unfortunately firm the results i've seen here today the silicon lottery is going to affect GPU benchmark numbers that you guys are seeing so the best thing to do from here would be to hopefully test more retail samples as that's going to be very indicative of what you guys are going to get in terms of performance but on top of that I'd definitely go check out people like a dog TV who are analyzed all the different reviews out there so you guys can then get a mean from the average of reviews and you guys I hope you enjoyed today's video if you did - be sure to hit that like button also let us know in the comment section below what you guys think of GPU boost 4.0 and now the silicon lottery in my opinion affecting reviews this is going to be an interesting topic and I'm actually going to make this a two-part series we're in part two I'm going to be testing out ambient temperatures and fan speeds and how they can also affect numbers that you see in reviews and in case of the r-tx 2070 samples that I had here they perform pretty much around the same ballpark so there's nothing to worry about in terms of variants of the r-tx 2070 numbers you guys have been seeing here on the channel and this RT x 2070 was a retail sample - so that's good to see with Galax and speaking of Galax themselves the one thing I will say is it's great to know that they're not sending me out cherry-picks samples as reference with this white card here this came in before this black card here so if they had a need to send me out a golden sample or they were sending our golden samples they would have never sent me out this in the first place so I'm glad Galax in this case KF a - and Europe are doing a great job of being an honest company and I get that vibe when I go to their events and collaborate with them anyway guys I'll catch you in another tech video very soon peace out for now bye you