does 3dmark really mean nothing?

[0ldman]

Do a search here... it is absurd, but it has been posted more times than I can count.

High poly count, high shader usage, etc... never mind the fact D3 is a low poly game, openGL, custom shaders, etc... the comparison was made.

 

[Scali]

Are you saying that the 3DM05 engine is in any way representative of the performance we'll see in the 3D game engines of 2005 (which is about 17 days away) or even 2006? I doubt we'll see any 2005 games running at 22fps average on a 6800U in 1024x768 with no AA.

You are overlooking the fact that 3DMark has only one detail setting: the highest. Games generally have various detail levels, so even if there will be games that run at 22 fps averae on a 6800U at the HIGHEST detail level, they probably won't default to that, so most people/benchmarks will get higher average framerates.

As a benchmark it seems pretty useless because it doesn't tell us how well our hardware will perform in actual games - and isn't that the whole *point* of a benchmark? All this benchmark really tells us is how well the latest hardware runs 3DMark05 and how well ATi and NV have optimized their drivers for it (which they both do, bigtime).

First of all, the point of a benchmark is never to tell how the hardware will perform in actual games, obviously. Only those actual games will give you an exact answer (each game only for itself ofcourse). The point of a benchmark is to measure performance on a given set of operations.
In the strictest sense you are right in saying 3DMark only tells us how well the latest hardware runs 3DMark... (which is obviously also true for any real game benchmark).
However, in the broader sense, 3DMark tells you how well the latest hardware runs scenes with an average polycount of the scenes in 3DMark, and with the average shader use and effects of the scenes in 3DMark. If you understand that, then you can understand what 3DMark can and cannot tell us about performance in future games. If you don't understand that, sell your PC and get an XBox or something.

 

[Scali]

Considering I can play Doom3 nicely and still have yet to see that damn proxycon demo right...

Funny enough Battle Of Proxycon actually ran nearly twice as fast as Doom3 (timedemo1) on my box, in 640x480. In 1024x768 they were pretty well-matched though.

But to me, the most important thing that 3DMark03 showed was how well videocards got away with shadowvolumes. My brother had a Radeon 8500 at the time, and it was just horrible in 3DMark03. Newer cards clearly handled shadowvolumes lots better, and we saw the same in Doom3 ofcourse.
Without 3DMark03 you wouldn't have much of an idea of what the difference between an 8500 and say a 9500 would be, especially when Doom3 is not out yet. With 3DMark03 it became painfully obvious that an 8500 was not going to cut it when shadowvolumes arrived, while the new generation, even the low-end cards, had no problems.

Problem with such uses of 3DMark is that it requires insight from the user, which apparently is missing with some people here. They apparently think there's nothing more to graphics than a framerate, and if the framerate of benchmark A doesn't correspond to the framerate of benchmark B, then benchmark A must be useless (for some reason they never say benchmark B must be useless...). What they should be saying is: Benchmark A performs a different set of operations from benchmark B.

 

[0ldman]

I have a 4200, and even the 8500 plays Doom 3 well.

Sure, I'd love a faster card, actually get to see the heat effects, but I'm not buying a new card when the one I have occasionally drops to 25fps, 90% of the time its capped at 60fps.

 

[Scali]

I have a 4200, and even the 8500 plays Doom 3 well.

Sure, I'd love a faster card, actually get to see the heat effects, but I'm not buying a new card when the one I have occasionally drops to 25fps, 90% of the time its capped at 60fps.

Note what I said about detail levels in games, compared to 3DMark.
And the 8500 is considerably worse than your 4200 I'm sure.
The 8500 could only play at 640x480 or 800x600 at low or medium detail, it would drop below 15 fps quite quickly actually.
Also, afaik Doom3 uses a single-pass shading approach while 3DMark03 uses a two-pass approach for ps < 1.4.

Also, I'm not sure what you mean to say with your "buying a new card"-comment... 3DMark should never be a reason to buy new hardware. It can just point out weaknesses in your current system. Whether you feel that you need to do something about these weaknesses or not, is up to you. I didn't go out and buy a new card either when my 9600Pro scored only 1350 in 3DMark05. On the other hand, I'm not going to rant on a forum that I can still play Half-Life 2 at high quality with all effects on, while I get a low score and bad framerate in 3DMark05. That's not the point.

The point is that I know that my system is particularly weak at high polycounts, and this is a problem that may degrade the experience in future games. Just like the 8500 degraded the experience in Doom3. My brother has since upgraded to a GeForce 6800LE, and the Doom3 experience is much better now. Now where does 3DMark fit in? Simple... while the GeForce cards are excellent cards for Doom3, they haven't always been that strong in Direct3D titles. And one of those titles was Half-Life 2, which my brother was looking forward to. Radeon would have been a sure bet for Half-Life 2, but was not so great for Doom3. Thanks to 3DMark's scores (and some other popular games), we could conclude that a 6800LE was about as fast as 9800Pro (about the same price back then) in most Direct3D scenarios. So it would be a good card for Half-Life 2 aswell. And indeed this turned out to be true.

 

[CleanSlate]

Sounds like you ened a complete upgrade 0ldman, do like I did grab a 21 inch CRT flatscreen trinitron refurb with a 2 year replacement warranty for ~240. That'll uncap that 60fps to 85 all the way through 16by12 so you'll have a real reason for getting up and buying a nice shiney new 6600gt later on.

~Adam

 

[0ldman]

Sounds like you ened a complete upgrade 0ldman, do like I did grab a 21 inch CRT flatscreen trinitron refurb with a 2 year replacement warranty for ~240. That'll uncap that 60fps to 85 all the way through 16by12 so you'll have a real reason for getting up and buying a nice shiney new 6600gt later on.

~Adam

I'm shooting for a 6600GT VIVO one day, but Doom3 is capped at 60fps, as that is my primary game right now, and I have no trouble running it.

I can play any game I want @ 1024x768 on my monitor with the exception of Doom3, which runs at 800x600.

90% of my gaming is on my TV, which my card will only output to 800x600 anyway. Warcraft III and Doom3 are the only games that even stress my system, and they run fine. Everything else runs well over 100fps if I turn off Vsync with AF and AA cranked.

This is my TVbox, the only real gamer in the house, Tbred 2100+ @ 2074MHz, 1GB of PC2700 @ 180MHz 2-3-2-11, Ti4200 VIVO @ 290/600, Shuttle AN35N Ultra 400... the weak spot I see is the integrated sound, which I just haven't found a card I'm happy with yet.

Considering I am still in the sweet spot of recommended systems for all games coming out today, I see no reason to upgrade. When I finally get an HDTV, I'll upgrade. This system will not be able to record HD at realtime.