hrm to include UT2003 or not in my 9700 Pro review?

[ben6]

That is the question. I realized in the middle of finishing out my review yesterday that the UT 2003 demo came out yesterday. As , honestly, I don't play UT, I'm wondering if it's worth it to include it in review just for the sake of having it in the review. As it is now I think my review is too long for most people's tastes. What's your opinion?

 

[Jazz]

u know u want to id say put it loads of people wanna know how how good it performs on a Ut2003 so id say put it in

 

[dksuiko]

Why not?

 

[Johnny Rotten]

That is the question. I realized in the middle of finishing out my review yesterday that the UT 2003 demo came out yesterday. As , honestly, I don't play UT, I'm wondering if it's worth it to include it in review just for the sake of having it in the review. As it is now I think my review is too long for most people's tastes. What's your opinion?

I definitely would. The unreal engine is going to be a defacto next gen benchmark.

 

[Pete]

Yes. People will be upgrading their PC's for games like UT2K3 and D]|[.

 

[Dave Baumann]

Ben, the only thing I would caution you on is making sure the demo is consistent in both your own benchmarks and possibly with other sites. If you read the Benchmarking UT2003 thread in the games forum then it seems there may be a few inconsistencies with the bot matches. So, all I'm saying is don't rush it to include it and make sure you know what the demo is doing.

 

[Matt Burris]

Dave:

I've noticed it too, but I'm not too positive if what I saw was different in each benchmark. It just feels like it's different sometimes...

 

[Dave Baumann]

I've noticed it too, but I'm not too positive if what I saw was different in each benchmark. It just feels like it's different sometimes...

I'll admit, I've not had time read through that thread completely and understand it yet, but Daniel made some reference to the bot match being dependant on screen ratio, so when benchmarking through the resolutions it may be that 1280x1024 is different to the rest of the resolutions because its a different aspect ratio - you might need to benchmark using 1280x960 to keep consistency within the UT benchmark. That does present problems for cards that do not natively support this resolution though (such as most of the Radeon's, bar 9700).

 

[demalion]

They all support it, it is simply that ATi does not make it available. Easiest way to add it is to use something like Rage3D Tweak...I'd expect PowerStrip also has the ability. Another option is adding it to the install file for the drivers (as per Omega/Plutonium drivers, etc).

No, I don't know why they don't include it, and I usually don't think about it much primarily because adding it back on driver updates is habitual now.

 

[noko]

Yes PowerStrip does have the ability to get Radeons and beyond at 1280x960 resolutions.

 

[Bambers]

Wierd thing is the drivers specifically prohibit 1280x960. Other resolutions are just not added but 1280x960 is the only res put in the RestrictedModes entry in the registry and I've found powerstrip is unable to add this res unless all the restricedmodes entries are deleted.

 

[borzwazie]

what kind of machine are you doing this on? So far, it looks like the game is really cpu-limited at least up to 2.0 GHz. The only thing that separates the men from the boys (so to speak) in this benchmark is the addition of AF and FSAA.

 

[pascal]

Flyby are OK, but botmatches are problematic and totally CPU limited.

What about some nice high quality FSAA screenshots?

 

[Nagorak]

Wouldn't it be best to benchmark from a recorded demo? Does the demo have demo recording abilities (that sounds kind of weird, hehe)?

 

[cm]

AFAIK not even the game will support demo recording/playback. The will add that with the first patch.

Lo guys, btw

 

[Reverend]

Wouldn't it be best to benchmark from a recorded demo? Does the demo have demo recording abilities (that sounds kind of weird, hehe)?

My email exchanges regarding this with Sweeney on 24 August :

2) Ability to record demos and benchmark it - possible? How? Same procedure (demorec -> timedemo 1 -> demoplay) ?

You can do that, but demo playback adds various kinds of overhead that might affect the demo. What we've done for the official benchmarks is to place camera movement paths in the existing levels, and have the demo code do a flyby of the level along the path. This gives the most consistent
performance results. In this case, we're measuring rendering performance
(both the CPU and the GPU aspects of it), but not things like AI and game
code, which we don't attempt to benchmark because we can't guarantee
consistency between runs.

Add to this Daniel Vogel's remarks.

 

[WaltC]

Wierd thing is the drivers specifically prohibit 1280x960. Other resolutions are just not added but 1280x960 is the only res put in the RestrictedModes entry in the registry and I've found powerstrip is unable to add this res unless all the restricedmodes entries are deleted.

Don't know that I checked prior to the 2.3 Catalysts, but 1280x960 is definitely a resolution selectable right from the WinXP SP1 display properties, with my 2.3 Catalyst-powered R9700. Or, are you talking about within the UT2 demo?