|
|
| Cluster Shout Box Link » /matrix/cluster.asp/40 |   GX-Brandon (285) Aug 29, 2007 - 12:37 pm
| I don't think it will. I think the reason there's so much doom and gloom right now is because the consoles are at the beginning of a new life cycle, but as time passes on the PC will just pull further away from consoles and the segment will grow as gaming in general grows. The PC market may not grow as quickly but it will always be there.
What does disappoint me a little though is hearing companies like id talking about dumbing down their idtech 5 engine for consoles. The old id wouldn't do something like that imo. Flag this | Edit this post |
| Cluster Shout Box Link » /matrix/cluster.asp/40 | GX-Brandon (285) Aug 29, 2007 - 12:24 pm
| BTW: An update on BioShock testing. The 2900 XT is scaling TREMENDOUSLY under 32-bit now that I'm forcing DX9.
I'm still having problems getting FRAPS to work with the GeForce cards. Everytime I start FRAPS the game crashes. Flag this | Edit this post |
| Cluster Shout Box Link » /matrix/cluster.asp/40 | GX-Brandon (285) Aug 29, 2007 - 12:22 pm
| Zane: The OS reports a lower figure but it does recognize all your RAM. I'm surprised Microsoft didn't fix this with Vista, WinXP had the same problem.
With so much system RAM, I doubt you'll see any benefit from Readyboost. You might as well take that 1GB flash drive and use it for conventional storage. Flag this | Edit this post |
| Cluster Shout Box Link » /matrix/cluster.asp/40 | GX-Brandon (285) Aug 29, 2007 - 12:17 pm
| | I don't think so. Waaay back in the days we the editors at FS would challenge other websites and even companies to matches of Quake/Counter-Strike. That was a loooong time ago though. Flag this | Edit this post |
| Cluster Shout Box Link » /matrix/cluster.asp/40 | GX-Brandon (285) Aug 29, 2007 - 12:16 pm
| I've successfully swapped out the motherboard without getting the BSOD in WinXP before, but it's a crapshoot. I always recommend a fresh format before re-installing the OS although I do realize it's a bit inconvenient.
Your RAID question depends on the motherboard you are upgrading to. If you're upgrading for example from an old nForce mobo w/RAID to a Bearlake mobo with RAID it's not going to work. Flag this | Edit this post |
| News Link » /news/newsarticle.asp?searchid=17093 | GX-Brandon (285) Aug 24, 2007 - 10:37 pm
| | No matter what I do, I cannot get my DX9 water to look that bad. My best guess is that's not using high quality graphics settings, or perhaps that video comes from the BioShock demo and it doesn't have the water of the final game, because my water definitely doesn't look that bad. Flag this | Edit this post |
| News Link » /news/newsarticle.asp?searchid=17093 | GX-Brandon (285) Aug 24, 2007 - 08:08 am
| | I wouldn't say you need DX10 to get nice smoke. Really it all depends on how the developer chooses to implement it. I know Crytek is using DX10 for effects like motion blur, and the geometry shader has been mentioned in the past for use with things like better vegetation and fur. Flag this | Edit this post |
| News Link » /news/newsarticle.asp?searchid=16699 | GX-Brandon (285) Jul 27, 2007 - 09:44 am
| Yeah, you will get no argument from me there, I agree that the 1950 Pro and 7900 GS are amazing values, especially for the vast bulk of today's games.
It's the games of tomorrow that may show promise for GPUs like the 8600 GTS. Game devs can use DX10 shaders to improve performance (similar to SM 2.0b and SM3.0) or to enhance graphics. If they chose DX10 to improve performance, there's a possibility that the 8600 GTS could outperform the 1950 Pro/7900 GS, etc, so it really just comes down to how game devs end up using DX10. Flag this | Edit this post |
| News Link » /news/newsarticle.asp?searchid=16699 | GX-Brandon (285) Jul 27, 2007 - 09:16 am
| | Remember that Editors Choice is not about value, that's what the Bull's Eye award is for. Editors Choice is about being the best, and while I admit it's terribly overpriced and isnt a value at all, the card delivered among the top 1 or 2 in performance and cooling, only falling last in the price category. Flag this | Edit this post |
| News Link » /news/newsarticle.asp?searchid=16519 | GX-Brandon (285) Jul 17, 2007 - 08:49 am
| This is coming straight from the game devs at Gas Powered:
"First off, because of the multi-core nature of modern machines and the multiple threads in SupCom, you CANNOT just use FPS as a measure of overall system performance. If you have a fast GPU and slow CPU, you could still get a good framerate. And likewise, if you have a chump GPU and a fast CPU, your framerate might be lowish while you still get a reasonable game experience.
Repeatability is tricky because we go to great lengths to do things differently based on how the machine is performing. If your machine is running at 20fps vs. 40fps, we only render half as many frames per tick.
So a run of the same 10 minutes will involve a lot more work on a faster machine than on a slower machine. We do this because we want to give the player the best experience their machine can supply.
One thing you can do is run the game single threaded and with a pretend framerate. If you do that, the work it does is 100% repeatable, but you are no longer making use of the multiple cores."
Because of the challenges of properly pushing single vs dual vs quad cores in Supcom, that's probably why Anand chose to test replay playback performance -- that's the only way you can properly push all 3 configs w/ 100% repeatability.
Again guys, I've been discussing this topic with GPG since Oct of last year, I really want to make SupCom into a good CPU benchmark, but right now it's just not there yet. I also know that AMD and Intel are discussing it with them as well. Flag this | Edit this post |
| News Link » /news/newsarticle.asp?searchid=16519 | GX-Brandon (285) Jul 16, 2007 - 02:28 pm
| Yeah, that's not gameplay testing though. Unfortunately all he's doing is testing how quickly the CPU can play back a replay.
What I've been lobbying for is for Gas Powered to modify the SupCom Mark test so it runs more like the CPU test in 3DMark 06. Or at least modify the replay system so that I can use FRAPS with it to accurately reflect gameplay performance. Intel and AMD have also been lobbying for modifications to the test, basically right now it's more indicative of multiplayer performance than single-player. Flag this | Edit this post |
| News Link » /news/newsarticle.asp?searchid=16519 | GX-Brandon (285) Jul 16, 2007 - 02:22 pm
| Rainbow Six Vegas is one title I've been toying with the idea of including for months since it's using UE3, or at least an earlier variant than the one we'll see later this year in UT3 or BiA, it has some limitations of its own though that I also don't like.
I was really hoping that STALKER and more recently Lost Planet would show better results with quad-core, but all my testing so far hasn't shown that potential. Flag this | Edit this post |
| News Link » /news/newsarticle.asp?searchid=16519 | GX-Brandon (285) Jul 16, 2007 - 12:23 pm
| » Supreme Commander testing
I'd love to be able to test with SupCom, I've been speaking with Gas Powered since Oct of last year (before it was released) about incorporating a performance test into the game. Unfortunately, the way it's implemented now there's no way for me to reliably test CPU performance with the game.
We all know the game takes advantage of multi-core, but unfortunately the built-in performance test doesn't test this. This also doesn't work with playing back replays with FRAPS: neither method runs multithreaded.
The only way to test CPUs properly is to actually play the game with FRAPS running in the background. But the problem with this method is repeatability. Also, the game devs told me the game is constantly scaling itself based on your current performance, there's no way to turn this feature off. So basically, you cant test 2 different CPUs the same way each time.
Bottom line is for right now there's no way to test actual gameplay performance. Flag this | Edit this post |
| News Link » /news/newsarticle.asp?searchid=16439 | GX-Brandon (285) Jul 11, 2007 - 04:20 pm
| » Oblivion settings
If you guys recall my Oblivion performance articles from a LONG time ago, I'm running the high-end Oblivion.ini, not the one from my mainstream performance article. The only thing I turn off is shadows on grass, everything else is cranked up high, but not quite maximum. I do this because I still compare previous gen cards (like the GeForce 7800) in my articles, and if I were to crank everything up to max the 7800s would just chug more in foliage than they already do now. Obviously I didn't include previous gen cards in this particular article, but normally I do try and compare the latest to the previous gen as you can see in the Radeon HD 2600 article.
It's never a good idea to compare numbers though, just because you can get totally different numbers based on the area you are testing -- just look at the difference in frame rate between the mountains area and foliage area. It's huge!
The mountains area is probably not intensive enough with today's high-end cards and it's possible that I should just drop tests from that area altogether. It's something that I may have to consider, it would definitely save me time! Flag this | Edit this post |
| News Link » /news/newsarticle.asp?searchid=16439 | GX-Brandon (285) Jul 10, 2007 - 09:25 pm
| | As I mentioned on the system setup page, we're using custom drivers provided directly to us by the driver teams at AMD and NVIDIA -- these aren't public drivers. We've been told though that the components in these drivers will ultimately be released to the public, but I can't give you an exact driver version to be on the lookout for. Flag this | Edit this post |
| News Link » /news/newsarticle.asp?searchid=15541 | GX-Brandon (285) May 17, 2007 - 07:32 am
| | Actually I was referring to all the DX10 games that will be coming out in the next 6 months. UT3, Hellgate London, Crysis, Lost Planet, EVE, DX10 patches for Company of Heroes, Supreme Commander, FS X, Call of Juarez, etc. Flag this | Edit this post |
|
