Hi, this is Wayne again with a topic “Reviewing Nvidia GPUs Isn’t Worth It”.
Nvidia launched the RTX 470 super last week on the 17th and will be launching the 480 super next week on January 31st. There was also a 4070 TI super. I think I can’t remember which other super there was it doesn’t matter. The point is, we haven’t done an LTT review of either of them um, because we didn’t feel like dealing with nvidia’s six embargos six for this launch for all for each of the three cards and remember Guys these are not new gpus. These are a thing that we already had, but like slightly recom jiggered six and they had six embargos. Is that three per card that is two per card? There were three cards: oh yeah, so there’s an unboxing embargo for each card review embargo and then a review embargo. Oh wait did I say six.
I meant nine embargos so or maybe it’s not nine, because at least one of them has a separate review embargo for the founders Edition and for the a cards. I can’t I can’t remember, but basically I saw the schedule and I said: okay number one fck off Nvidia just launched them Just Launch them we’ll be happy to review them, we’ll put them all together. Number two. I just I’m sorry, but I just I just think it’s ridiculous for our viewers to have to dig through all the different videos. You know you want nine videos yeah no well, because every one of the like following videos is just going to be all the graphs. You just looked at with one more data point on it, so we made the decision when we saw this to do one review video, what a Terry move of the new super lineup.
Thank you um. It wasn’t just my decision. We talked about it as a team.
We just looked at it and we went. This is absurd. Maybe you know if we have time we’ll do a couple of them on short circuit, but no we’re going to do a review of super.
You guys can just watch one video you’ll know everything you need to know about super the super 4,000 yeah exactly sounds cooler, yeah, no, just just ridiculous. So instead you guys got the. You know how we’re parallelizing our GPU benchmarking, video, which is going to be super cool, super cool, get it uh because – and you won’t necessarily see it on the super review here. But in future we are going to be able to, hopefully massively increase the number of fresh results that we can have on our Benchmark, charts um because, as you guys know, we retest as long as there isn’t a new driver. What was that? Did you hear that yeah uh, I opened a video on YouTube and I guess it was still piped through cool okay. The point is um: okay, okay, we’re muted, no, the the the video yeah yeah, okay, uh.
The point is that we are so we’re able to provide fresh testing data and we’ll be able to provide so much more so as long as they’re um, unless as long as there, unless there hasn’t been a driver update at all no OS updates no driver updates. We retest and we’ll be able to do that with a much much wider pipe with the three CPUs that we have found are within. I think it ended up being like 0.25 % of each other um in productivity.
It doesn’t matter it’s in the video speaking of hav memorized. You should watch that video too cuz you’re talking about GPU testing, but the CPU testing video was sick. No, no! No! That’S for GPU testing! Oh it’s the same thing, yeah! No! No! No! No! No! No! No, it’s not the same thing.
No, I meant you’re we’re talking about the same video. We were talking about the same video and a lot of people misunderstood that um we are validating CPUs that are the intro, explains it pretty. Well, it’s just a long video, nearly identical to each other, so that when we are doing GPU testing, we can test three gpus at a time yeah, instead of one GPU at a time, standardization bench has a Delta of what was it 8 % or something? I don’t remember what the percentage it was as much as 8 %, and I think that was in factorio or something like that. If it has that much of a Delta, then all your results are screwed. So you you need similar platforms for standardization and um, so so a lot of people really confused by that video, like I remember, seeing one bit of feedback on it that we should um. If we, oh man, I forget what it was, but it was basically like.
No, you should just rerun The Benchmark more times and I’m like no, no, no, we already. We already do that. This is that that’s not what that is not what we’re talking about at all.
We already do multiple runs and then uh the outliers go away and then the median or meor or whatever we have smart people who do that stuff, um yeah, really cool video. That kind of thing doesn’t get a lot of views, but what it is it’s foundational. So when the lab’s website is fully launched when we start to populate it with articles, this is what we’re going to be able to kind of Point people to and go okay yeah.
This is the rigger that is going on behind the scenes so that, when you guys just see a graph on the lab’s website, you’ll know it’s good to go and in a way, I’m actually, I’m really looking forward to the lab’s website. Beity error, like even in that video after all the after all the QC there were like, were, I think, two bars that were oh cuz. You can update it. Mislabeled one of them was 1 % and one of them uh said one. It said 1 % when it was supposed to be averag and it average when it supposed to be 1 %, but when everything’s coming straight out of the results database and the graphs are just automatically generated, we don’t have to worry about human input, creating graphical assets. That’ be nice, it’s going to be so.
There could still be ISS isues with anyways. Whatever nothing is infallible. Yeah, but it’ll be nice, but it’ll there will be less human Intervention, which is and with the site at least even if there are still issues uh. We can update it um a lot of people asked – and this is a question – that’s coming up here. Why not just force the clock speed. Do you do that? Do you run your computer like that? No, oh, it’s actually, a surprisingly unrealistic thing.
In current year, yep 100 % so uh. No, we we think it’s worth the work of making sure that we are running the system, the way that you would run the system. So that’s why we that’s why, when it was time to get over 10 CPUs, we went for a modern top tier chip, because we want to see what the maximum difference that we can see in these gpus is assuming that we can reduce the other system bottlenecks.
As much as we can – and we spent the time testing it to make sure that we have consistency, we’re going to have to continue to revalidate and with every new platform, we’re going to have to do this process again, but you guys are worth it. You guys are worth it these things take time. Man like, I think I think a lot of people are are looking at the slow progress in the lab going like oh yeah, you know like.
Why is it so slow? You know why why were there M Engineering takes time, and it’s just like guys, because there’s always going to be little bit of learning as we go, and these things take time. Yeah, it’s going to keep getting better, though that I can promise you it’s going to keep getting better hell. Yeah, it was a great video. It was fun, doesn’t get a lot of views that kind of thing, but is what it is.
Unfortunately, that’s okay, we’re just going to do it because we care and at the end of the day, if those kinds of videos just don’t get views, then that’s fine, we’ll have the website, but we still want everything we do to have this Foundation. That’S what we are building towards, that’s the that’s! The vision from our chief optometrist .