Jump to content
NEurope
Sign in to follow this  
Minlack

3DMark06

Recommended Posts

Whats your specs? I'm gunna wait untill the whole thing cools down and download it nice and fast.

Share this post


Link to post
Share on other sites

well, i get barely 1k on '05, so i think i'm gonna wait until i get a new graphics card... (my bottle neck)

Share this post


Link to post
Share on other sites

Dude 4279 is fucking insane..

 

http://anandtech.com/video/showdoc.aspx?i=2675&p=3

 

Unless you have some ridiculously expensive system, you're either lying or didn't run at the default settings, which of course you can't use for comparison.

 

My best scores on the previous ones:

 

3DMark05: 2542

3DMark03: 6355

3DMark2001: 18717

 

woot woot

Share this post


Link to post
Share on other sites

AMD Athlon 64 X2 2264 MHz

1024 MB

NVIDIA GeForce 7800 GTX

 

Default settings. Does the link not work? Cos my specs are on that page.

Share this post


Link to post
Share on other sites

cba with 06, only just got 05 and it runs really badly (1005 :(). im away to put some more ram in but my laptop needs repairing first, its an acer so i should get it back from repair in time for 3DMark07 (if im lucky)

Share this post


Link to post
Share on other sites

3DMark06: 748

 

woot woot!

 

3DMark06 is rubbish, hell 05 was rubbish. You can't run the tests smoothly even with the highest end kit available, which is fucking pointless. 3DMark2001 is last great ones since it represents relevant use and ran smoothly on decent machines of the time.

 

Plus 06 has a shit demo (I fucking hate Proxycon ever since it came in 03, it's so FUCKING BORING, and 06's is almost the same as 05's!!) and is restricted in features and it's not free to upgrade (I bought 05, only fuck knows why).

 

Here's an example of what I mean:

 

http://service.futuremark.com/compare?3dm06=12410

 

That is the #1 score for 06 so far.

 

Yet it couldn't achieve even 1 frame per second in the CPU test, and all the other tests are just about what you'd expect games to run at (30-60fps) but look at the ridiculous system he has to have to achieve that!

 

This benchmark belongs in a fantasy world. No, "future proof" is not a good argument, since they'll bring out another 3DMark in a year or two years anyway so that argument is fucking redundant. They should scale the benchmark to look okay on midrange systems and smooth on high end systems, but they've scaled it so it so ridiculously, especially with this stupid SM3.0 requirement to do some tests, they should only be optional because in real games they're rarely used or don't make much difference, especially not as rift-like as the test here would show.

 

They need to drop back and make more balanced benchmarks. I remember the days of watching 2000 and 2001 demos and thinking they were pretty, and they ran smooth, even on my system for the time (and it wasn't a particularly uber system or anything) and they've just fucked the curve entirely.

 

Not a good move, futuremark.

Share this post


Link to post
Share on other sites

RoadKill: (Talking about 3DMark05 now:) The tests runs smoothly for me when it tests with the graphics card. The ones that doesn't run smoothly is like that for everyone, because it tests how the CPU alone can render the shit. It doesn't use the graphics card at those tests.

 

In 3DMark05 I got:

 

3DMark Score: 6004 3DMarks

CPU Score: 5194 CPUMarks

 

I'm gonna get the new version now and test the shit there. =)

Share this post


Link to post
Share on other sites
RoadKill: (Talking about 3DMark05 now:) The tests runs smoothly for me when it tests with the graphics card. The ones that doesn't run smoothly is like that for everyone, because it tests how the CPU alone can render the shit. It doesn't use the graphics card at those tests.

I know exactly what the CPU tests are but that doesn't mean they couldn't scale them better.

 

Oh by the way, here's my results in the ORB:

 

2001: http://service.futuremark.com/compare?2k1=8472460

03: http://service.futuremark.com/compare?2k3=3786171

05: http://service.futuremark.com/compare?3dm05=662292

06: http://service.futuremark.com/compare?3dm06=13608

Share this post


Link to post
Share on other sites

And these were my results in 3DMark06. =)

 

3DMark Score 2148 3DMarks

SM 2.0 Score 1058 Marks

SM 3.0 Score N/A

CPU Score 1625 Marks

 

Comparison URL: http://service.futuremark.com/compare?3dm06=14312

 

Minlack: And wow, the graphics cards make alot of difference. You got like double my score and the only big difference is your graphics card, that is one of the best out there. And my card (ATI Radeon X850XT) was the best one like three months ago. =P But my card's core clock is faster then yours, but your memory clock is over 1 GHz, mine is only like 500 MHz, I guess how fast the memory is makes alot of difference, it has the same amount of memory as mine, 256 MB. Your AMD64 X2 4400 is just some MHz faster than mine.

Share this post


Link to post
Share on other sites
3DMark06: 748

 

woot woot!

 

3DMark06 is rubbish, hell 05 was rubbish. You can't run the tests smoothly even with the highest end kit available, which is fucking pointless. 3DMark2001 is last great ones since it represents relevant use and ran smoothly on decent machines of the time.

 

Plus 06 has a shit demo (I fucking hate Proxycon ever since it came in 03, it's so FUCKING BORING, and 06's is almost the same as 05's!!) and is restricted in features and it's not free to upgrade (I bought 05, only fuck knows why).

 

Here's an example of what I mean:

 

http://service.futuremark.com/compare?3dm06=12410

 

That is the #1 score for 06 so far.

 

Yet it couldn't achieve even 1 frame per second in the CPU test, and all the other tests are just about what you'd expect games to run at (30-60fps) but look at the ridiculous system he has to have to achieve that!

 

This benchmark belongs in a fantasy world. No, "future proof" is not a good argument, since they'll bring out another 3DMark in a year or two years anyway so that argument is fucking redundant. They should scale the benchmark to look okay on midrange systems and smooth on high end systems, but they've scaled it so it so ridiculously, especially with this stupid SM3.0 requirement to do some tests, they should only be optional because in real games they're rarely used or don't make much difference, especially not as rift-like as the test here would show.

 

They need to drop back and make more balanced benchmarks. I remember the days of watching 2000 and 2001 demos and thinking they were pretty, and they ran smooth, even on my system for the time (and it wasn't a particularly uber system or anything) and they've just fucked the curve entirely.

 

Not a good move, futuremark.

 

Think of the benchmark as what could be done in PC games now or near future. It's not a reason to leave a feature out of a benchmarking program just because current games don't support it. While I do agree with you that the series went downhill from 3dmark05. I expected that at least the "game demo" would run smoothly as it did in 3dmark2001 with a relatively slow system. But no, it was some places even slower than the benchmark itself.

 

And using synthetic benchmarks to measure real game performance isn't the best way to go. Because some games are optimised for nvidia cards and some to the ati ones.

Share this post


Link to post
Share on other sites
Think of the benchmark as what could be done in PC games now or near future. It's not a reason to leave a feature out of a benchmarking program just because current games don't support it. While I do agree with you that the series went downhill from 3dmark05. I expected that at least the "game demo" would run smoothly as it did in 3dmark2001 with a relatively slow system. But no, it was some places even slower than the benchmark itself.

 

And using synthetic benchmarks to measure real game performance isn't the best way to go. Because some games are optimised for nvidia cards and some to the ati ones.

 

The problem is it isn't representative of future PC games, because another version of 3DMark will come out in a years or so time anyway so it doesn't really mean shit. It should run on and benchmark comparably machines today, not just a few uber high end machines (Because you really need Shader Model 3 to compete, and that even excludes such things as X850 XTs, which are by no means slow).

 

It's recycled as well, I'm fed up with seeing Proxycon yet again, even with almost identical scripting to the 05 version, yet they have the cheek to make you pay for it again even if you bought 05? Obscene.

 

Edit: My machine (9800 Pro, 3Ghz P4, 1GB RAM) came in the top 9% of all systems scanned on the Futuremark game advisor, yet I have absolutely no hope of scoring a decent score in 06, or even 05 for that matter.

 

This is a benchmark simply and only for high end cards that support Shader Model 3 and is basically useless outside of this restrictive boundary.

 

This is my complaint. They're not making a good benchmark anymore, they're just closing in on the high end, perhaps 1% of computers that have a tight knit 3DMark circle jerk.

Share this post


Link to post
Share on other sites

Well the release cycle used to be about two years(first 01 then 03 then 05, leaving the 01SE out bacause it was just an update) , so then it made sense to have the most current graphic gimmicks and such implemented.

 

But for some reason the cycle is now only a year, so maybe they hadn't got the time to make something new? Just tweaked the proxycon a little.

 

Well lets hope that the next one will be better. Can't even be arsed to download the 06 because it would probably look like a slide show on my 9700 Pro.

 

And for the reasons it requires sm3. It would also be unfair benchmark for the high end cards if it wouldn't have the support for it. And the X850 XT certainly isn't slow(but it isn't ultra high end either), the lack of sm3.0 support is one reason not buy it.

Share this post


Link to post
Share on other sites
And for the reasons it requires sm3. It would also be unfair benchmark for the high end cards if it wouldn't have the support for it. And the X850 XT certainly isn't slow(but it isn't ultra high end either), the lack of sm3.0 support is one reason not buy it.

 

But what happens if someone had invested in X850 XT crossfire? That's far from slow, and it lacks SM3.0 support but in real terms that's no huge deal.

 

This benchmarks makes it out to be the most important thing ever.

Share this post


Link to post
Share on other sites
But what happens if someone had invested in X850 XT crossfire? That's far from slow, and it lacks SM3.0 support but in real terms that's no huge deal.

 

This benchmarks makes it out to be the most important thing ever.

 

I would be just happy to own two x850 xt crossfires, playing new games etc, not worrying about my benchmarking score and lack of SM3.0 support :)

 

But the benchmark probably gives it a bit too much weight in final score if it affects the score that badly.

Share this post


Link to post
Share on other sites

i got 2340

 

my specs are as follows

 

Athlon Xp 3200+ @ 2.21GHZ

1.5gb pc3200 ram

x850xt pe

Share this post


Link to post
Share on other sites

The company was better when it was called MadOnion.

 

Try finding 3dmark 2000. I think there was a 2000 version. It has been awhile.

Share this post


Link to post
Share on other sites
The company was better when it was called MadOnion.

 

Try finding 3dmark 2000. I think there was a 2000 version. It has been awhile.

 

Yeah there was 2000 and it was badass, the music in the demo was totally sweet.

 

I still always type "madonion.com" when I go to the site, and always move the 3DMark shortcuts into the madonion.com start menu folder. MadOnion forever!

Share this post


Link to post
Share on other sites
Sign in to follow this  

×