Designing Our Next GPU Test Suite
by Ryan Smith on March 14, 2010 12:00 AM EST- Posted in
- Ryan's Ramblings
In keeping with our desire to refresh our GPU test suite periodically, we’re going to be redoing our GPU test suite to rotate in some more modern games, along with rotating in some DirectX11 games capable of taking advantage of this generation of GPU’s full capabilities. And while we already have a pretty solid idea of what we’re going to run, we wanted to throw out this question anyhow and see what responses we get.
What games would you like to see in our next GPU test suite, and why?
What we’d like to see is whether our choices line up with what our readers would like to see. We can’t promise that we’ll act on any specific responses, but we have our eyes and ears open to well-reasoned suggestions. So let us know what you think by commenting below.
Facebook
Twitter
RSS
240 Comments
View All Comments
alphacheez - Sunday, March 14, 2010 - link
I'm not sure this exactly falls into the category of the GPU test suite, but I'd like to see what games can be played well on lower end/older/integrated graphics.It's not very interesting to see that something (like integrated graphics) gets 5 FPS in Crysis Warhead; it's more interesting to see what games can be played, maybe Half-life 2, Quake 4, or even back to Doom 3 (or games from that sort of vintage) to give an idea of the age of games that give a good experience. This would especially be interesting to include in reviews of Ion-style notebooks (which are often CPU limited), CULV notebooks (which tend to be GPU limited) and other low performance/integrated systems.
therealnickdanger - Monday, March 15, 2010 - link
+1+1
+1
As a gamer who's not too interested in the latest and greatest eye-candy, I really like the idea of legacy benchmarks for mainstream and low-end graphics cards/IGPs. One thing that drives me insane when shopping around for a new laptop or netbook is trying to find information on "what games can it play?". In all sincerity, the only PC games I play are Halo, WoW, Unreal Tournament (original and 2004), and some occassional Source-based games (CS, DoD).
dgschrei - Sunday, March 14, 2010 - link
I would like to see Bad Company 2 in this lineup for two reasons.1: At this moment it is probably the best selling game on PC so there will be a large user base. This will give a lot of people a good comparison point since they already know what fps their card can do in this game.
2: It's currently the only popular DX11 enabled game. They only use it for soft shadows, but still.
I also disagree about the no synthetic benchmarks argument. They are important, because they allow us to see, which one of the 2 GPU companies has currently the better architecture.
Just take Unigine. Although the reported power of the GTX480 is not so far away from a HD5870 it seems to annihilate it there because Nvidia focused on tesselation power. We wouldn't be able to see this in a simple game bench where anything from a certain shader to lazy programming to the CPU can limit the fps.
They should however not factor into the final rating.
shadow2200 - Sunday, March 14, 2010 - link
What's really needed is software that stresses the maximum amount of cores in a CPU,supports all the features within DX11,and is stressfull enough to put even multi-GPU setups to the test,and no software exists that can do all that...At least not yet unfortunately.(maybe the next version of 3Dmark supporting DX11?)Though as a special section,one thing that can make up for the lack of the above,at least in terms of being GPU demanding,is running benchmarks using 3 displays,as a single 30" display costs a fortune still(1000~1500$),and users can easily get 3,24 inch LCD's for much less than that(1/2 as much),and it will be much harder to run those 3 at 5760*1200 resolutions,than it will running a single 30" LCD at 2560*1600.
Direct comparisons between brands in the above scenario should only be allowed when all GPU makers have a single card that can output to 3 displays though,though obvously it would be a moot point in multi GPU shootouts(2 HD 5870's versus 2 Fermi cards for instance),as both can do it from what i've seen.
grv - Monday, March 15, 2010 - link
furmarkHillBeast - Monday, March 15, 2010 - link
Too right. Furmark is a brilliant test. Even if they don't use it to bench it, they should at least run a system for say a few hours running benchmarks to get a rought idea of how reliable these cards are and how hot they will get under worst circumstances.Also they need to test it INSIDE a case so we can know how hot it will be in real life rather than on a bench.
Ryan Smith - Wednesday, March 17, 2010 - link
We do exactly that. Our load temperature tests are done inside a full Thermaltake Spedo case.kanabalize - Sunday, March 14, 2010 - link
Often we will see only FPS types of games in the benchmarks.. although i feel they are important as many GPU intensive games are made for that genre, let there be some other type of game genre like RPG, MMORPG, RTS, Simulation and etc...Also, since graphics card today comes with many other features like physics engine and GPGPU functionality, It would be good if these features are taken into consideration when doing benchmarks as GPU are not used only for gaming nowadays.
HillBeast - Monday, March 15, 2010 - link
The main thing is that while MMORPGs, RTS and the like are popular, they aren't intense enough in my opinion and you reach a point where it won't be hard enough on the GPU for it to be accurate.HotFoot - Tuesday, March 16, 2010 - link
But that should illustrate a point - that depending on your needs, a top-of-the-line 5970 may be no better than your current 8800GTS.Just like it was kind of silly to suggest the i7 980 was the best CPU yet for playing WoW because it pushed the frame rates from some rate too high for the human eye to pick up to an even higher rate.