Gaming Question for our [H] Readers in Regards to Video Card Reviews

B

Brent_Justice

Moderator
Joined
Apr 17, 2000
Messages
17,755
We are currently evaluating changing up our gaming suite, keep it updated and current.

Can you please tell me if Skyrim performance in our evaluations is still important to you? Considering it is a DX9 game, been out a while, all card max it out at 2560 with 8X AA+FXAA, but we know it is popular. When you read our reviews, does Skyrim impact your buying decision and if it disappeared from our reviews, would it be a bad thing?

Also the same goes for Deus Ex, another popular game, DX11 with some Tessellation, but again, if this disappeared from our reviews, would it matter, and does it impact you.

Following that, about adding games, would you find benefit in us adding The Witcher 2 Enhanced Edition? Though this game is a year old, it's a 2011 game, would including it now, a year later, mean something to you when reading our performance results?

Finally, a couple new games coming out in May, any interest for Nexiuz and Max Payne 3?

Thanks for your feedback!
 
I am fully in favor of ditching Skyrim; popular though it may be, you are right, it takes nothing to run.
I would not miss DXHR in the reviews either.

Witcher 2, I was actually thinking about that the other day, wondering why it isn't in the lineup. It would be perfect for GPU evals with the sheer brute force needed to run the game. Ask me, I'll tell you it should have been in the batch since launch. My 6950/2gb/unlock won't get anywhere near 60 frames on ultra at 1920x1080. I'm sure even cutting edge single-gpu solutions can't max it properly. Maybe in dual-gpu mode, but even then, I'm still talking 1080p, and your evals do awfully high resolutions. This game is the perfect real life video bench, especially the prologue scenes, which could easily be replayed exactly the same way for consistent testing.

I don't care about Max Payne, and I don't know what that Nex thing is, but if they push the hardware envelope, why not.
 
Last edited:
Skyrim is something I play but good performance is to be expected on newer cards so I guess that could go. Maybe include it in lower-end reviews? Max Payne 3 is supposed to have enhanced PC graphics (DX 11...etc) so I guess we'll see on the 29th. Never heard of Nexuiz. I don't play Deus-ex nor the Witcher so those metrics have little context for me personally.
 
I wouldn't want to see Skyrim and DE:HR go. While they're not pushing the envelope in graphics tech, they are relevant because many people play them so they definitely influence buying decisions. I wouldn't mind seeing a couple of additional games, crysis 2 is a mainstay of many reviewers simply because it is still played quite a bit and is very demanding (perhaps the most demanding game out there aside from metro 2033). Same goes for metro 2033, but not many people play that anymore so that could factor as an argument against it. I think witcher 2 : EE is a great addition. Witcher 2 is honestly one of the best looking games to be released in the past year, I love it and would love to see it added to the benchmark suite.

Basically, I understand that testing games takes a ton of time but a balance between popular /new titles and titles that represent cutting edge technology would be great. This is why I love witcher 2, its a popular title and is somewhat cutting edge as well. I'd also make an argument for crysis 2 and the upcoming diablo 3 / max payne 3.

Regarding nexuiz, hard to say! I don't think it will be a very popular game, although if its a cutting edge game graphically that would make a very strong argument for adding it. I just got in the beta,i'll have to check it out sometime this week.

Just my .02 :) Thanks for listening
 
Skyrim is still a game I (and I'm sure plenty of other people) play fairly often. However as you've stated the vast majority of GPUs have no issue at all keeping the game at the 60fps cap. I see no reason to keep Skyrim (or Deus Ex for that matter) in your reviews. That being said, if there was a way to (somewhat easily) implement a benchmark of a modded Skyrim I would be all for that. Benchmarks of games running with heavy mods are the one benchmark I always seem to have trouble finding, I understand this could be because of all the variables involved when using mods...

As far as The Witcher 2 EE, I think that would be a great idea. There are still many a card out there that are brought to their knees by the Ubersampling in The Witcher 2.
 
Last edited:
Would it be possible to do a Witcher 2 review on a limited scale such as just 7970's and 680's so that we may see how you would approach benching it? I loved the game, but if the cards of today can easily handle the game minus ubersampling I'm not sure how interested I would be. Might be best as a one time deal.

On the flip side, I greatly respect the support the developers have been showing. Spreading interest for the title via being a performance review staple might not be a bad thing.
 
Skyrim obviously needs to go - besides it running great on most decent cards, every decently informed gamer and/or buyer of a powerful graphic card(s) has modrf the hell out of the graphics - heck the steam workshop practically begs the average joe to install enhancements
 
I still play Deux Ex from time to time, it's a decent DX11 game to benchmark, but I wouldn't mind you moving it out of the review rotation for a newer game.
 
I think that Skyrim would be better suited to cpu reviews, personally. Dues ex isn't exactly a looker either and runs well enough on most cards. I wouldn't miss it if it were to go. One game that I would like to see is Serious Sam 3, that does a number on most high end cards.
 
Skyrim would make a great benchmark fully modded with texture packs, but I see no way this could be standardised. Right now I am interested in how well a fully modded skyrim runs on different graphics cards, but it seems like this an impossible benchmark situation with the mods constantly changing and everyone using different ones.

However.... I dont know if a fully modded Skyrim would scale the same as vanilla. If the percentage of increase is about the same (i.e. card A is 30% faster than card B at stock and on a fully modded Skyrim) then I would like to see it stay. I doubt it would be that simple though. Modding adds a lot of textures and some mods add a lot of effects so my guess would modding wouldn't follow a linear scale.

And of course [H]s testing method doesn't lend itself to that type of situation either, except for the apples to apples comparison.
 
Max Payne 3 had some interesting system requirement. IF, and this is a big IF, that game turns out to be a demanding game, then of course it would be great for [H] to use it in future benchmark. I'm not familiar with Nexiuz but if its graphically demanding, then I think it would be a nice addition. Same goes for Witcher 2

I agree that its good for the games lineup to have some relevance to the current interest of the readers, but it should also be the major factor when making the decision of which card to get. Games like Skyrim can easily run on any modern graphic card that it doesn't matter if you have a HD 7970, HD7870, GTX 680 or GTX 580, there's no any difference in the gameplay experience. So how the card performs in these games is not going to be a factor that helps us decide which card to get because we know either one is fine here, and therefore we look to other more demanding games to decide.

Just my opinion anyway. And thanks [H] for listening to the readers opinion.
 
relevant games for me include skyrim and shogun 2 because of eyefinity performance, deus ex is great and relevant but too easy on system req's.
 
Do more 1920x1080 benches with the high end cards^^ To me, it's hard to gauge how a card will run at 1080p when all used are 2560x1560 benches. I know you're doing highest possible settings, but I hate to have to go to toms or anandtech for their benches sometimes just to see 1080 because they don't do real world testing :( On topic though, The Witcher 2 EE would be a good one
 
IMO keep skyrim..

for me its very important for reviews to use games that "most" people play. Whether its relevant or not, its a popular game.

I mean when i look at a review I immediately look for games i play to see how it would be different for how I play.

look at the review posted here linking to a site that failed to include BF3 as part of their review suite.. Although BF3 is important to reviews part of it is because that is what people are playing now.



If you do decide to get rid of skyrim, maybe think about just adding a small single page section on "How this game plays for popular games".. that may go right along with your game play experience vs benchmarks. You get the gameplay in some of hte more demanding games but a quick look at how some of hte most popular games perform with the card too.
 
I will add that Crysis 2 with the DX11 and High-Res packs might be a good choice. Still pushes my system in a lot of ways at 1080p.
 
MavericK96 said:
I will add that Crysis 2 with the DX11 and High-Res packs might be a good choice. Still pushes my system in a lot of ways at 1080p.
Click to expand...

The problem is Crysis 2 with DX11 is just horrible inefficient. Remember how it uses extreme levels of tessellation on random objects for no reason?

Witcher 2 would be a great addition, still surprised you didn't add it to the line-up last year. It has the highest resolution textures I've seen in a game so far and looks far better than most DX11 titles.

Max Payne 3 might be a good as well depending on how well DX11 is implemented, since I don't imagine the DX9 mode will push systems much, considering it's a console port.
 
I would recommend Alan Wake to be added in the testbed. The PC version released in Feb 2012 with fantastic graphics and brought significant improvements in visual effects, image quality over the Xbox 360 version. Its really demanding when maxed out and has a state of the art engine with nice features

http://www.pcgameshardware.com/aid,...mpared-to-360-looks-great-at-60-fps/Download/

And its a console port done right unlike other games which have crude patches for DX11 features and are mostly an afterthought. I am not entirely convinced of Batman Arkham City's DX11 implementation. Game studios should look at DICE and FrostBite 2 used in BF3 as the benchmark to design a game engine holistically with maximum rewards to the user experience.

Nexuiz , Max Payne 3 might be good candidates but unless they really push the graphical envelope in features / image quality you can hold off.

Another suggestion is when benching the ultra high end cards (USD 500 or above) the choice of games should be restricted to the most demanding games. Games which do more than 60+ fps maxed out at 2560 x 1600 are not really demanding. BF3, Alan Wake, Crysis 2 (DX11 with high res textures), Witcher 2, Shogun 2: Total War are some of the most graphically demanding games released in the last 12 months. Also if its possible can the testbed include atleast 8 games. 4-5 is too less
 
Last edited:
AceCR42 said:
The problem is Crysis 2 with DX11 is just horrible inefficient. Remember how it uses extreme levels of tessellation on random objects for no reason?
Click to expand...

I disagree completely. Crysis 2 is very relevant in benchmarking tessellation performance, i'm not sure I agree that it uses extreme levels of tess on random objects for no reason. Tessellation was added after the game was developed, so it was an all or nothing approach.
 
I'm in favor in Skyrim being kept on as that game and BF3 are some of the top listed games over in General Hardware for what people want to play on the new/old PC we're helping them build/upgrade. With that said, if you did actual multi-player GPU reviews of BF3 and if that requires you to drop Skyrim, I have no problems with that. I'd rather have multi-player GPU/CPU reviews of BF3 than a single player BF3 + Skyrim GPU/CPU reviews.

If you guys haven't thought about it yet, a one-shot review of Diablo 3 with different GPUs would be nice to see. That's the other top listed game I've seen.

It would not matter to me if Deux Ex disappeared. It doesn't show up too often in the aforementioned lists. Witcher 2 EE would be interesting to see as I do see that game show up a bit in General Hardware. Nexiuz (the CryEngine 3 version) looks good but I doubt it's gonna be that popular. As for Max Payne, I really don't see that being all that popular among PC gamers either. But I could be wrong.

EDIT: Something else I thought of: More VRAM usage reporting with any game. At the moment, there are some really well priced 1GB and 1.2GB Nvidia cards and AMD doesn't have that many cards to match the price points of those Nvidia cards. However I do have concerns about their future ability to play games due to their somewhat limited VRAM.
 
Last edited:
AceCR42 said:
The problem is Crysis 2 with DX11 is just horrible inefficient. Remember how it uses extreme levels of tessellation on random objects for no reason?

Witcher 2 would be a great addition, still surprised you didn't add it to the line-up last year. It has the highest resolution textures I've seen in a game so far and looks far better than most DX11 titles.

Max Payne 3 might be a good as well depending on how well DX11 is implemented, since I don't imagine the DX9 mode will push systems much, considering it's a console port.
Click to expand...

Crysis 2 can be benched with extreme for Objects and Ultra for the rest to be a reliable indicator of GPU performance. Techreport found that use of tesselation in Crysis 2 Ultra mode was excessive and inefficient.

http://techreport.com/articles.x/21404

So their benches keep objects at extreme, rest of the settings at ultra

http://techreport.com/articles.x/22653/11
 
I agree with Danny. Its not so much about how taxing the game is as much as it is Skyrim and BF3 are the games that 95% of upgraders and builders ask "will this card play...". And as somebody that's on a few other forums answering these questions, I'd much rather be linking reviews by [H] than Tom's or Anand because they've got the games prospective buyers are wanting to play. [H] has the best way of reviewing cards by not just concentrating on whose graph on the benchmark table is bigger but by concentrating on the highest playable settings. Don't get me wrong, I want taxing games. I don't need to see a pair of overclocked GTX680's getting 3000 fps in Call Of Duty but I would like to see how it handles Crysis 2, Batman AC and even Crysis 1.

I also think it should be expanded a little to include a few more games because Metro, Batman AC and Crysis 2 are more important to me than Skyrim and BF3 but The Witcher could be more important to somebody else. So I think its good to have a good mix of different games out there so buyers can judge how well a card will work with the games they'll play the most because its not always the case the if a card pwns in one game that it'll pwn in a different one.
 
Last edited:
raghu78 said:
Crysis 2 can be benched with extreme for Objects and Ultra for the rest to be a reliable indicator of GPU performance. Techreport found that use of tesselation in Crysis 2 Ultra mode was excessive and inefficient.

http://techreport.com/articles.x/21404

So their benches keep objects at extreme, rest of the settings at ultra

http://techreport.com/articles.x/22653/11
Click to expand...

Thanks for those links. I had heard some vague stuff about Tessellation in Crysis 2 being messed up, but didn't hear a solution. I'll have to test Object detail at Extreme and see if I get better performance.
 
Keep Skyrim, but bench it with fullscreen SGSSAA (not TrSSAA/AAA). Generally, I would like to see more SGSSAA benchmarks to have a counter pice to the trend of cheap blurry post-AA filters. Skyrim is perfect for that. And the latest cards really have the power to use it in many games.

BF3 and Crysis 2 are good because their engines will be used in the not too distant future with big selling games.

I agree with raghu78 that more games are needed. 4-5 are too little. 6-8 would be better.
 
I would love if you guys could add Metro 2033 again, that is still one of, if not the most, demanding game to max out.
 
I'd prefer demanding games in the reviews. And by that I don't mean demanding as in you can't do 9999x AA/MSAA/FXAA/ETC in it.

Personally I'd also be very careful about the game picked for a specific review though. Some games favor one side or the other to a very large degree, and since you guys usually run with few games in your reviews having a one-sided mix could greatly skew the overall impressions of a card. That said, you guys have usually handled that part pretty fairly.
 
I think The Witcher 2 and Metro 2033 would be relevant additions, and The Witcher 2 is a fantastic title, one of the best games I have ever played as well as being both gorgeous to look at and graphically demanding.

Personally, in terms of review strategy, I think you should pick 3-4 core games that are very GPU intensive regardless of age/popularity, and then do 2 titles or so that are new and relevant, if not necessarily intensive. This would be a good balance to strike.
 
xIronCrossx said:
I would love if you guys could add Metro 2033 again, that is still one of, if not the most, demanding game to max out.
Click to expand...

As a consumer I think most people would like to see 6-8 games spanning a mix of the most demanding games (metro 2033 like you mentioned fits this) and the most relevant games that people are actually playing. Also testing for both single screen and surround, which you guys have been doing, I like that. I also like that H does unique things in their reviews (highest playable settings, framerate analysis not just by the numbers, minimum framerates) so I like all that stuff but 1-2 more games would be killer. Although I understand that game testing takes a long time..
 
Rizen said:
I think The Witcher 2 and Metro 2033 would be relevant additions, and The Witcher 2 is a fantastic title, one of the best games I have ever played as well as being both gorgeous to look at and graphically demanding.
.
Click to expand...

Yep can't go wrong with Witcher 2. It really is a fantastic looking game.
 
Personally I'd like to see ArmA2 OA, but that might be a better one for CPU reviews. Still GPU intensive as well, though.
 
I think Skyrim and DX:HR need to be kept. While for high end card reviews I don't think its nessacarry, midrange/low end cars could still show nice performance differences in them.

I would suggest adding Diablo 3 when it releases. The engine at first does not seem very demanding but I would be interested to see the full game in action. I assume if its anything like the first two, the later levels when your screen is covered with enemies and spell effects could push videocards pretty hard.
 
MavericK96 said:
Personally I'd like to see ArmA2 OA, but that might be a better one for CPU reviews. Still GPU intensive as well, though.
Click to expand...

Yea last time they asked this question, they tried to integrate Arma2 and it was just too CPU bound to really showcase the GPU differences.
 
The problem is that almost no games push video cards anymore. Battlefield 3 is the only one and even then many cards are pretty good with it until you hit the higher resolution multimonitor setups.
 
H-street said:
Yea last time they asked this question, they tried to integrate Arma2 and it was just too CPU bound to really showcase the GPU differences.
Click to expand...

Yeah, I figured that might be the case. I'd like to see it used in CPU tests, though. :)

demingo said:
I would suggest adding Diablo 3 when it releases. The engine at first does not seem very demanding but I would be interested to see the full game in action. I assume if its anything like the first two, the later levels when your screen is covered with enemies and spell effects could push videocards pretty hard.
Click to expand...

I really don't think Diablo 3 is going to push a system in any significant way.
 
Dan_D said:
The problem is that almost no games push video cards anymore. Battlefield 3 is the only one and even then many cards are pretty good with it until you hit the higher resolution multimonitor setups.
Click to expand...

BF3 is certainly not the only game that pushes GPU's. My main thing is that more than 4 games would be ideal for giving people an idea of whats good for their purchase, using a mix of popular games and games that really push the limits of GPUs' - close to 7-8 would be a sweet spot. Just my opinion of course anyway ;) Witcher 2 with ubersampling, crysis 2, metro 2033 all push hardware to the limits.
 
Last edited:
xoleras said:
BF3 is certainly not the only game that pushes GPU's. My main thing is that more than 4 games would be ideal for giving people an idea of whats good for their purchase, close to 7-8 would be a sweet spot. Witcher 2 with ubersampling, crysis 2, metro 2033 all push hardware to the limits.
Click to expand...

I'll give you Metro 2033. I forgot about that one. The Witcher 2, maybe. Crysis 2, not from what I remember. I only had issues with the demo when I went into multimonitor territory. The same as I did with BF3. Otherwise, smooth as hell. And that was a couple video cards ago.
 
I never played Skyrim, so I'm not versed, but it sounds like nothing is phased by it anyway.
Deus Ex could exit.
Metro 2033 is a tough game on GPUs, I always recommend it for people that want to see what their computer is capable of.
If you want to keep a DX9 game, I'd use Hard Reset, that game has alot of bling.:D

Witcher 2EE sounds like a good choice.

It's just really difficult to find a game currently that challenges the GPUs.

I second a graph of VRam useage and maybe one of GPU useage too.
 
xoleras said:
BF3 is certainly not the only game that pushes GPU's. My main thing is that more than 4 games would be ideal for giving people an idea of whats good for their purchase, using a mix of popular games and games that really push the limits of GPUs' - close to 7-8 would be a sweet spot. Just my opinion of course anyway ;) Witcher 2 with ubersampling, crysis 2, metro 2033 all push hardware to the limits.
Click to expand...

I would love to see more games too actually. I was just reluctant to suggest it as I have no idea the amount of workload that goes into each benchmark given [H]'s method of benchmarking. But if you guys could squeeze in an additional game or two, that would be great.;)

Personally I would wait for [H]'s review before making a purchase, however late it may be. But I can't speak for everyone of course.
 
I could see Skyrim going away but not sure about DX:HR. It would be quite nice to add at least the Witcher 2 to the suite as it is current and punishes hardware.
 
You must log in or register to reply here.
Back
Top