Yes, he compared a x2 RX480 against a single 1070 and Maxwell Titan X. Really shady, specially when he didn't have on-screen-monitoring for this side-by-side footages. Also, he touched on the subject of "smoothness" but never measured it when he had previously made a video on how to use PresentMon which is a software that works on DX12 and Vulkan APIs.
Really, I thought he was showing how 2x RX 480's and DX12 helped Ryzen much more then the 7700K?
2 Likes
Oh I now see where he was speculating on GPU at the end. I think he was mostly babbling and that was not scripted or the point of the video. I mostly tuned that part out. Again I would never use one review as definitive info but as a direction for further investigation.
1 Like
Well, tbh I found his video interesting since he pointed out that reducing the GPU bottleneck (actually, I'm not sure how 1 1070 performs vs 2 rx480), or using a GPU that makes better use of DX12 does reduce the gap between the 1800X and the 7700k. Which is quite in contrast to other reviews that tried to reduce the GPU bottleneck in other ways.
One thing is absolutely true regardless of what you think about the guy: There has not been enough testing between CPUs running AMD cards. Dual 480s are a thing, Furys are still a thing (and not that far away from 1070s, just sayin'). But nearly every review with gaming benchmarks runs team green exclusively.
Stuff like that is why I still like to watch his videos.
4 Likes
That's partly AMD's fault, it's their job to provide review samples of these cards. They know they don't have the mindshare so it's up to them to make that bridge with the reviewers. They've done a good job so far with the Ryzen though, not so much on the GPUs.
BS! Think about how many RX 480 videos there were.
1 Like
AdoredTV didn't have a FuryX or 390x or a 290x, he would have to get those out of his pocket, this is the problem with many small review sites.
I'm not talking about him, I mean all the other reviewers.
Like how Hardware Unboxed specifically said in one of his videos recently that he decided to test with the 1060 instead of the 480 because he was also testing with the 1070 and 1080. He definitely has access to AMD cards (I'm assuming at least one Fury card and likely several 480s), but didn't bother testing with them. It seems like a reasonable thing to do. To want to see how things work with the AMD gpus. Once you get the picture with the higher end NV cards, there isn't much of a reason to test with the 1060. That data point can be basically extrapolated from the higher end NV cards. Scale it down some and there you go. But we haven't been seeing how well it performs with AMD cards.
I think that there have been a lot of trends recently when it comes to testing Ryzen, from testing various resolutions, to testing with cores disabled, to testing with various speed memory, etc. And I think that this will be a trend soon enough, and it might all be because of Adored that it happens. And that is something that I am thankful for. It was something that was going largely unnoticed by the rest of the media and it is definitely something that I will be interested in seeing. That said, I hadn't really been that concerned about it because we are definitely going to see how well Ryzen performs with AMD gpus once Vega comes out. People are basically going to be forced (at least for this release, and I assume with most gpu releases from here on out) to have an Intel system and an AMD system. Preferably a 7700k and an 1800x since those are the parts that perform best in games from both sides.
3 Likes
I think the basic takeout is that DX11 vs DX12 results for the same CPU depend heavily on the GPU used. This has a few corollaries:
- Any comparison between DX11 and DX12 depends on both the CPU and the GPU used, since all 4 combinations gave different numbers
- Any comparison between CPUs in terms of relative DX12 vss DX11 performance depends on the GPU used
- Any comparison between GPUs under either DX11 and DX12, and the difference between the two, will depend on the CPU used
This would mean that tests based on a single hardware configuration (up to the tested component) convey limited information on what to expect from the parts tested. In this example, an Nvidia GPU owner would draw misleading conclusions about CPU performance in DX12 from a review using an RX480 CF, and vice versa. The owner of an i7 CPU would draw misleading conclusions from a GPU review under a Ryzen system. And so on.
More generally, I think this series of videos adds to the more general topic of how many limitations gaming benchmarks have as a way to learn about CPUs. Although you could say all of them are just different faces of the same limitation, namely, that many things are being tested at the same time: the CPU, the GPU, the API, the game itself (let's not forget this was all for one game: repeating all these tests for a different game will probably deliver yet more different numbers).
If a gamer sees a test for a game he plays, using a very similar hardware configuration to his own, he can obtain useful information about a particular upgrade. As a general test of the capabilities of a CPU, gaming benchmarks are not very informative.
3 Likes
I think the bulk of the really in-depth analysis will largely remain on the smaller channels like MindBlank Tech and AdoredTV. The rest will go with the run-of-the-mill Max/Avg/Minimum just because of how time consuming it is to do just those 3 let alone do a extensive analysis for every game like AdoredTV did just for Tomb Raider alone (100+hrs).
1 Like
I don't think any review site should be complaining about to much options to create content.... It is their job.
2 Likes
That would be ideal since I'm sick of hearing "it feels smoother on [x]" with no graphs and numbers to back it up.
I value that actually more than numbers. Because it means someone actually played games on that system and not hit a button and some time later copied numbers into a spreadsheet.
If that's that case why even do benchmarks at all? Might as well say, "This renders more frames than this, and I swear it's smoother. Well, that's it folks, thanks for watching!".
I don't like to take things at face value.
Yup, I would be fine with that from someone who really plays games and knows how the thing compares. Like what Jay did. Build it, use it, play games on it and then tell me about the experience. That is what I care about.
But you do. Doesn't matter if reported in english or in numbers, you have to be able to trust the person giving you the information.
Look, Steve from Gamers Nexus has said multiple times that he basically doesn't play games anymore and that his own system is kind of old. I don't think the rest of the staff is any different. Same goes for a lot of those review sites and I am not questioning their numbers or integrity here, I do believe that they all do the best they can. But most of them simply don't play games anymore, they benchmark, fully automated if they can. They don't know if that thing in that level of the game is still making a little stutter or not because that isn't part of the benchmark. So I understand to some extent that you want your numbers but don't think that those are all there is. Benchmarks can only paint a very small part of the picture.
4 Likes
Nah, not me. If I want to get the value for my money it's gonna take more than just anecdotal evidence or pretty words.
This asshole gets it. Yeah typical of Scottish/Irish/English humour. American don't get it at all and jump on the defensive like you had just shit on the deceleration of independence. It is so easy to get a rise out of them.
We have an orange snowflake for a president, trust me, we know jokes. In the case for AdoredTV's "joke" it was out of character which made for an out of context "joke". As for Bitwit, he does skits oh his channel (even his sponsored ads) and sticks phallic objects on his forehead, there's context to the ridiculous things he says.