Update: 3DMark 06 has been released, get it from FutureMark (http://www.futuremark.com). It weighs in at 577MB. This thread is for all the discussion of 3DMark06.



Here's an interview with the head developer of the next 3DMark for DirectX9.0:
http://www.beyond3d.com/interviews/fm05/p2/

It's still under development, but here's a screenshot:
http://img486.imageshack.us/img486/1078/next3dmark3gj.th.jpg (http://img486.imageshack.us/my.php?image=next3dmark3gj.jpg)

It will incorporate SM3.0, have complex HDR effects and hopefully be a decent tech demo and reasonable benchmark, as have all the previous 3DMark series.

There's also a nice historical look at the previous 3DMarks on this page (http://www.beyond3d.com/interviews/fm05/p2/index.php?p=06) of the article.

If there's one thing that's great about 3DMark IMHO, it's that it's made it easier for people to benchmark and compare performance, and that's not an easy thing to do in the PC world.

For me the best thing about the 3DMark series is that i can use the online comparison to see how people are performing on almost an identical spec to me. Im not at all concerned about the people on uber-systems and how they score because thats irrelevant to me and my system. I want to see how the next guy on the same processor, memory and GPU as me is scoring and if hes scoring better why, thats the true beauty of the 3d marks.

From what i remember the scores vary even between identical systems which just goes to show how even the smallest difference in a system or the tiny little bottleneck you never realised you had can affect the whole systems performance.

I have absolutely no intrest in the benchmark in terms of getting as high a score as possible, we all see the "nerds" who tweak every nipple on their pc to just get that ultimate score but the more savvy pc users realise that they are completely burning out their system for that one score then they revert back to a more stable setting for everyday use.

id get probly a 1000 score on that :x

That's stunning. I can only get 3300 on 3DMark05. Looks I'll need a brand new PC to run that at all.

The new 3DMark06 might be close, there is a teaser video on the main page looking similar to the pic posted above by Koroush and it says 'Unleashed Soon'.

http://www.futuremark.com/

Cheers

Apo

Interesting, thanks :) It may well be out shortly given it's already 2006 - I'm guessing early-mid February just as the next-gen cards are announced. That way everyone with current-gen stuff will run it, feel inadequate and be ready to upgrade all over again :)

Interesting, thanks :) It may well be out shortly given it's already 2006 - I'm guessing early-mid February just as the next-gen cards are announced. That way everyone with current-gen stuff will run it, feel inadequate and be ready to upgrade all over again :)

the beauty of business


the graphics look awesome but it just looks like a souped up 3D mark 2005

I am really looking forward to the first Direct X 10 version of 3D mark, thats something that I'm looking forward to

"The next 3DMark will require 256MB to ensure that there is no unnecessary swapping going on (System RAM <-> Video RAM). According to our mission of producing forward looking benchmarks to show performance of new hardware, we wanted to take a leap forward, and let the artists have more freedom to create visually stunning content. This doesn’t mean that the next 3DMark won’t run on 128MB hardware, but the performance hit can be quite severe. Several games are already limited by the amount of VRAM (at very high detail settings, which is comparable to the detail level in the next 3DMark), so I don’t see any problems with our decision. It is a known fact that VRAM has an impact on performance if the game is using a load of huge textures & data. Enable AA & AF on top of that and you need 512MB to have acceptable fps. The way we see it is that 256MB is nowadays the standard/default amount of VRAM for relatively high-end cards. 128MB was still a pretty decent amount 1-2 years ago, but things have changed pretty rapidly since. I wouldn’t be surprised if 512MB would be seen as a new standard/default in less than 2 years. Bring on 1GB cards!"

Well I don't even think about buying new card and I look forward for some great slideshow with my 128MB 6600GT:D

Well I don't even think about buying new card and I look forward for some great slideshow with my 128MB 6600GT:D
I bet we will all be looking at a slideshow, and personally I'm not so sure that it will look too nice on my system (http://forums.tweakguides.com/showthread.php?t=1116) either.

I really hate that developers can just start thinking that everyone will be able to get the new "standard" card as soon as it is avaliable. Becasue right now, I know that many people are using cards around the 9800pro level and it's just not fair to them... The engines need to be properly optimized, e.g.) just like the Unreal Engine was.

3Dmark really has nothing to do with game developers. As history has shown, it takes roughly 6-12 months before the features in the latest 3DMark even remotely work their way properly into a game. So I really wouldn't worry too much about it.

I wish I could even run 3D Mark 2005, just look at my system, I can barely run 3D mark 2003.

Heh, it reminds me of when I bought my 5900XT, so I decided I'd give myself a visual treat and see how 3DMark05 performed. I think I had like, 1FPS in the demo where the dragon thing jumps over a flying ship. I was devastated.

That's why I don't care about my 3DMark score as far as all current games run fine for me at decent settings. I think that 3DMark score is important when you are choosing new card to buy and not after you have actually bought it.

Just got this link in the mail, thought it might be of interest to some.

http://www.gamespot.com/promos/2006/3dmark06/index.html

Merged into existing topic on 3dMark06,
Sharpy

Oh boy...I can see how horribly aged my graphics card is getting....thanks Red Skull....thanks......:(

Of course I'll download it an run it anyway..... :D Comes out tomorrow....

Edit:
Ignore mine since the 2 posts above have been moved in here. Had trouble connecting to the forum last night at all so i posted in this existing thread and did not see the other one before it was moved here.

.....
Just read at gamespot that 3dMark06 will be out today. Hope it is no exclusive dl, since there is nothing new on futuremark.com by now. Maybe gets updated later since it is 5:26am in europe now, Wednesday 18th.

http://www.gamespot.com/promos/2006/3dmark06/index.html?tag=psbox_3dmark_gsb

Cheers

Apo

:D I am pretty sure it will be posted later on. I mean give the guys a chance to wake up bud! ;) :p

Good find on this BTW but I am with most of the guys here....my rig is barely able to run '05 I am sure '06 is gonna really stress it! :eek:

Can be found among other places at:

http://downloads.guru3d.com/download.php?det=1297#download

There is also a torrent link on that page.

Wow cool. I havent tried any 3dmark on my new computer (see sig). Im downloading this - probably to get a sucky score and think my computer sucks, lol. Nah.. its still 3dmark and nothing else :)

3dmark 06... according to bit-tech, they were running crossfire x1800XTs and still were getting less than 30fps on 1280x1024

can't wait to see how poorly my x800xl will perform....

/* me brabbing popcorn to watch slideshows! */

Shoot, 3dmark05 just spanked my x800xl for the first time yesterday, I finally downloaded it. Im not in a rush to download 06.

Haha... fps was around 20 all the time... grr. Anyways, looks like it was played at 1280X1024. Got a score of 4400 marks. Maybe id get better if i do a test after a reboot... but meh...:P

Here is a pick with cpu score too: http://img523.imageshack.us/img523/4254/3dmark069ra.jpg

It out and available for download from the futuremark site (http://www.futuremark.com/products/3dmark06) There is a basic free version as well as advanced and profesional versions.

Key 3DMark06 Advanced and Professional Features

* DirectX® 9 3D game performance benchmark
* Includes separate graphics card and CPU tests
* Produces workloads that simulate next generation 3D gaming requirements
* Advanced 3D game engine supports HDR Rendering with SM3.0 Shaders, in addition to SM2.0 Shaders
* Includes two HDR/SM3.0 game tests and two SM2.0 game tests
* CPU performance testing via AI and physics workloads for both single core systems and multi-threaded, multi-core and multiple processor environments
* 3D graphics feature tests include Fill Rate, Pixel Shader, Vertex Shader, SM3.0 tests, and Batch Size tests

P.S. It's 580MB in size

merged with main 3Dmark06 thread,
Sharpy

Yeah just try to get a connection with the site. They must be getting mobbed. 580mb :(

It's not a problem if you have a fileplanet subscription or bittorrent it is quite readly available but expect dial-up flashbacks due to server loads (for a while i was getting 3k/s and less :eek: on DSL) Don't need it just yet will wait a few days to let the rush pass.

Oh yeah Im waiting no doubt. I havent installed a bittorent utility yet on the new machine otherwise I might just grab it now. I can knock down a 580mb file fairly fast with the broadband but Im guessing it would be a nightmare right now.

Of course I just downloaded 3dmark05 yesterday on the new machine, figures :)

OK, catalyst 6.1 is released, anyone else that wont be surpriced if most ati-owners see a few 100s of points more in 3dmark06? :)

Update:
Nvm that, i only saw a 2p increase lol. Now im bored with 3dmark06... it only makes me feel bad about my computer... lol :)

Specs:

P4 2.4b OC'ed to 2.6ghz
Corsair XMS 1024mb
ATI 9800pro 128mb (6.1 Catalysts)

Total 3DMark06 score: 773
SM 2.0 score: 374
CPU score: 730

Haha, I wasn't expecting much of a score, but I think this is what 3DMark05 should have been.

If you haven't tried out the newcomer of 3D benchmarking yet I higly recommend you download the basic version of 3D Mark 2006.

Anyway, it has come to my attention that you might run into issues if you're using a motherboard based on the VIA chipset.
If your PC automatically reboots when trying to launch 3D Mark 2006 I suggest you try this:
1. First go to system properties by simple right-clicking "This Computer" and choose properties.
2. Go to "Advance" tab and click the "Start and recovery" button
3. Uncheck option "Automated reeboot" and hit OK, press yes on the next window that pops up.
Now run 3D Mark again and await the bluescreen.
If it says the problem is related to something called "DRIVER_IRQL_NOT_LESS_OR_EQUAL" take a look at the bottom of the screen.
Here it says what caused the problem, if you read VIdePort.sys the problem is caused by a driver called VIA IDE Accelerator Driver. A driver released by VIA for enchaning IDE-disk performance (I Think Koroush mentioned it in the latestTWC).
Uninstall this, restart and run the 3D Mark again.
(PS you might want to change the "Automated reeboot" option back)

Merged with main 3DMark06 thread

catalyst 5.13

Total 3dmark06: 1730
SM2.0 score: 881
cpu score: 947

catalyst 6.1
Total 3dmark06: 1734
SM2.0 score: 883
cpu score: 947

No difference whatsoever.

Yay! Personally i'm impressed :D. I think having dual core helped a bunch.

3DMark06 Score 4211 3DMarks
SM 2.0 Score 1745 Marks
SM 3.0 Score 1716 Marks
CPU Score 1465 Marks

1 - Return to Proxycon 13.6 FPS
2 - Firefly Forest 15.4 FPS
CPU Tests
1 - Red Valley 0.5 FPS
HDR Tests
1 - Canyon Flight (SM 3.0) 14.8 FPS
2 - Deep Freeze (SM 3.0) 19.5 FPS


When I did a search and compare I saw that someone had almost identical specs (same cpu speed and graphics card speed) and they got 6238 marks Could this be because they were using a SLI setup?

My results @ stock speeds, stock settings, Basic Edition:

Score: 4257

SM2.0 Score: 1710

HDR/SM3.0 Score: 1710

CPU Score: 1666

X2 4400+, 7800GTX, 2GB RAM. Looks like I've got a "balanced" system ;)

As for 3DMark06, I'm disappointed the basic edition just rehashes the same main 3D tests from 3DMark05. This is the first time 3DMark has done this - usually it's great to see brand new stuff for each 3Dmark release.

The only thing which really impresses me is the Deep Freeze benchmark, that looks incredible and ran quite well on my system.

It IS a bit disappointing isn't it to see the same tests?

My final score: 3927 3D Marks

SM2.0 Score: 1700
HDR/SM3.0 Score: 1695
CPU Score: 1103

I'm very happy with my score, considering some of the other scores you guys are getting. I thought mine would be a lot lower

It's a bit scary how the CPU tests run at like 0 FPS. Why is that?

mine

527

haha yeah go my awesome 2600+ and 9600XT !!!

Khoroush,

My new rig will have the same CPU and same amount of RAM as you do (check my sig). The difference in the GPU is that I am going for 2 7800 GTX 512MBs in SLI.

I dont know how much more points that will score... I hope it crosses the 5k barrier...

But the way it looks like... it squeezes the hell out of a kick *** rig even. The screenshot that really impressed me was the Canyon Flight test in the gorge with the sunlight shining on the water... HDR at its best... absolutely amazing.

Yeah the 2 HDR tests look awesome i think. I especially like the snow one, mainly cause it is new heh, but also cause it looks so impressive

I got the BSOD when trying to run the program. This is what came up...

IO system verification error in pcibus.sys WDM Driver Error 224
pcibus.sys+36e at B4B6F36E

:(

Just tried 3DMark05 and got the samething..

That's a shame (I haven't tried to run it yet)

What's your system specs and are you running beta drivers or official releases?

I got 1583 with catalyst 5.13... Friend got with X850 XT and almost same other specs over 2000.. I though I could have had at least 1800...

Have a question for y'all... I can't run the program because it defaults the test rez to 1280x1024, and it won't let me change it since it's the "basic" edition. The most my monitor can handle is 1024x768. Any command line vars I could use to fix?

Command line options are available only in the Professional Edition. I guess you need to get an Advanced Edition license which allows u to change the benchmark options and then run the test at 1024 x 768. But sadly u cannot compare scores run at resolutions lower than the default.

I think it might be time I considered upgrading my PC...

I got 199 - yes One hundred and ninety-nine (I haven't missed a digit off!)

Noo... I Got 1925!!!!! :(.

3700+ @ 2477Hz
EVGA 7800GT CO 473/1200
EVGA NF41
1x1gig Mushkin XP4000

3DMARK06
3DMark Score....3829 3DMarks
SM 2.0 Score....1693 Marks
SM 3.0 Score....1711 Marks
CPU Score........977 Marks

3DMARK05= about 8000

The theory is correct from what i've read in the futuremark forums and that is people that don't get the best score start considering building or upgrading right away. Like Karoush said this stuff usually isn't implemented for many month's so give it another half or full year then you can start upgrading and during that time your waiting you can save up even more money so you get a even more kickin system like a dual core processor maybe even save up for sli rig :D.

Haste makes waste because then next year when 3dMark 07 comes out you'll just have to do this all over again lol

3dmark is all crap I dont care, but I'm really looking forward to upgrading this summer with a dx10 video card.

There is absolutely no point to upgrading for 3DMark06. It is specifically designed to be super-intensive.

On my rig, which is a fairly decent one, including a 7800GTX - most of the benchmark runs at ~10FPS. When I know for a fact that most games like FEAR, COD2 etc. run at 1280x1024 with 2xAA, 4xAF, everything on maximum at 30-60FPS easily.

What's even funnier is the CPU test for me is also absolutely dog slow. I get 0.5-1FPS the whole time, yet I have an X2 4400+. Does that mean I should upgrade to an FX-60? Of course not.

The program is firstly designed to increase hardware sales, secondly to stress test your system, and thirdly as a benchmark. It is no longer a decent tech demo (since they made the free tests identical to the last version almost), and it is no indication whatsoever of game performance, not even upcoming games.

I am going to get more than 10FPS at 1280x1024 in Elder Scrolls: Oblivion for example - I guarantee it. So please, don't upgrade for 3DMark06, because even if you get an FX-60 and two 512MB 7800GTX's in SLI you will probably see a slideshow.

Mine's overdue a graphical upgrade anyway - I wouldn't do it just to get better results in 3dMark. I think I got less than 500 on the 05 version and that looked lame when I ran it. My GPU is three years agos technology and it's beginning to creep in some games now. I think it's almost time to pass the GeForce 5700 over the missus' rig!

well its not 3dmark thats got me mad, its my 30fps in militia with my x850 and pentum d 830 and gig of ram...

What's even funnier is the CPU test for me is also absolutely dog slow. I get 0.5-1FPS the whole time, yet I have an X2 4400+. Does that mean I should upgrade to an FX-60? Of course not.


Wow, that's a relief. I thought It was just my old, but still reliable, AMD Athlon 64 3200+ that was getting 0-1 fps average on the CPU tests it kinda made me disappointed about it.

I too am VERY disappointed. Antartica looks fantastic but everything else is just a rehash of the old tests with some new special effects. Granted these special effects make it look a lot better, like the soft shadows and HDR, what gets me most is the upping of the resolution. They obviously only upped the resolution, like you said, to make people think they need to buy better hardware.

When if you put things back to 1024x768 if you had a pro version I'm sure you'd see similar results as 3DMark05.

They will argue no doubt that 1280x1024 is now the "standard" gaming resolution. Which I agree, it is a fairly standard gaming resolution.

However to combine strenuous stress-test-like 3D graphics with a 1280x1024 resolution, and lock it in at that res, they are only forcing everyone to think their systems are inadequate. It achieves nothing else. It doesn't benchmark the system properly, it doesn't scale well (as the CPU results show), and it is certainly not indicative of any game now, nor in the next few months.

Sure, it may be vision of games in 6-12 months, but who knows what hardware will be out by then, and how many games will adopt all of these effects. If previous 3DMarks are anything to go by, it took ages for all these effects to work their way into actual games.

I see it likes this. 3Dmark2003 is more demanding than Quake4. What does that tell me? Well... 3 years later after a 3dmark-release.. we will have games that demand that much, but then the graphics are much nicer and STILL perform better.

Let alone all the effects in a single game. I have yet to see a game that combines HDR with soft shadows. FEAR's soft shadows are very inefficient, not to mention ugly. I'm not sure of a single game that uses sub surface scattering either.

i got 2766 which i think is freaky nice. who else has a 6800GS i wanna compare scores

I actually like the fact that 1280x1024 is the base res. That is the res that I want to play at. It is also good to distinguish the higher performance cards, who only really shine at higher resolutions. I've been saying for ages that the res needed to be raised and I'm happy to see they've done it.

Rehashing the old demos was a shame though, but let's not forget the fact that ultimately this is a tool, not a toy. The demos get the job done, although they may not be as interesting to watch.

I scored 1112 with stock speeds on my GFX card.

Devlopers are getting worse and worse with the patches, hot fixes, security updates, rushed items. It would've been better if they released the lastest version next year with new content and maybe some real benchmarks. Fear is probably once of the most intensive games graphic wise out on the market and my fps ranges from 30-70 in the game.

Oh and anotherthing when that CPU test started after about 30 seconds my cpu fans rpm shot up. This is the first time they've done that too, that really did freakin stress it.

The reason I didn't upgrade this year is because dx10 isnt supporting dx9 cards at all, but there might be a emulation or something of dx9, which will probably take a lot of resources.A dx10 card is a must for me since I dont want to be outdated, and I think it was dumb to upgrade to the higher end graphics cards this passed summer(IMO).

not trying to argue or anything, but this only really needed to be an update from a certain perspective. one complaint about 3dmark 05 was that it is cpu limited already (geforce 7800 gtx 512 sli and fx60 or something). because it was cpu limited already, it sort of ceased to be a great benchmarking tool because soon all equipment would get virtually the same score.

this sort of suggested that a new version that was not cpu limited was needed. so they rehashed it for new equipment. also this probably wasn't planned as a full new benchmark like previously. they usually release only every 2 years (3dmark 2001, 2003, 2005 => 2006 is odd).

as far as i could tell, the best equipment could only run 3dmark 05 pretty well initially too, so its kinda similar.

it may not be perfect, but i'm impressed even though i can't run it. it looks better, uses pixel shaders 3.0, HDR, and multi-core. i'm impressed.

http://img68.imageshack.us/img68/2573/3dmarkscore8fu.jpg

there's my score, At first I was like oh s*** I'm going to have to upgrade, but then I saw what most people were getting, and my PC still runs everything well, so I'm going to keep what I have for a while.

(3dmark 2001, 2003, 2005 => 2006 is odd)

Actually 2006 is even ;)

By both forcing you to buy the full version of 3DMark to get decent new demos,

I don't see how Futuremark is "forcing" you to buy the full version, access to the demos that give you a 3DMark score is free. If the "decent new demos" you refer to are the 3D graphics feature tests (such as the Fill Rate, Pixel Shader, Vertex Shader, SM3.0 tests, and Batch Size tests), then I think it's perfectly understandable that you have to pay for the extra features. Wasn't it the same way with '03 and '05?

I mean that with the Basic Edition you pretty much get rehashed 3DMark05 demos. People will feel they have to pay to buy the full version to get whizz-bang new demos. But yes, the full version no doubt doesn't have anything phenomenal either.

In the past 3DMark always has good new demos for each release in the free version. This tradition has been broken for obvious reasons - to firmly encourage people to buy the full version. That's fine from a commercial point of view, but not really something I am going to encourage in a commonly-used benchmark.

Oh, sorry I misunderstood. Yeah, I was a little disappointed that they rehashed some of the demos. I personally think they just wanted to get a fully-fledged DX9 benchmark out (with SM 3.0, HDR, etc.) before Vista comes out, without spending time getting whole new artwork for each demo. Plus it's kind of interesting seeing the graphical difference in the demos between '05 and '06, and seeing how far DX9 has come.

Just to clear this up, you don't get new demos with the full version. All you get is the options to change the settings they are viewed at. Just as with all the other 3DMarks.

My score is 4759 with X1800xt 512mb (722 core 900 ram) 1GB Ram with an 3800+ x2 at 2.4. For sure 3dmark have something wrong with the ATI Cards. I spent enought money all ready for sure no upgrade for 2 years

in 3dmark05, my score was ~5200 and I can still play most new games at 1280x1024 with minimum effects. So I think a score of 4000+ on 3dmark06 is quite awesome.

i think they would be dumb to not find ways to get you to buy the product. I mean its only like 20$. Big whoop. They are a company and they SHOULD be trying to sell their product. It just makes sense. Companies cant afford to make bigger newer better products without money. Maybe previous versions didnt make enough revenue so they wanted to boost sales... or perhaps they are just greedy. But either way, if they can get more money, then there is no one to blame but the customers for giving them the money.

If there is a "huge problem" with the way futuremark does things, dont buy their products (kinda like the whole BF2 thing). But you have to be prepared to have less if you arent going to buy them.
________________________________

edit: went and downloaded it myself.
Total: 2284
SM2:964
HDR/SM3.0: 873

Wow. Im weak

I updated it 901 on cpu, 2288 total

I wont even bother looking into 3dmark2006. My system is the king of 3dmark 2003 and makes a decent stab at 2005 but Id just feel ill if I ran the new one.

Anyway, as has im sure been said a hundred times in this thread, who cares if your pc baulks at this benchmark. If your machine plays what you want it to play well now, and if you think it will make a good fist of the games you want in the near future, thats all that matters.

My modest Intel/AGP rig has no troubles with the FEAR's and QUAKE4's of this world and Im sure it'll will run Oblivion respectably as well (with alittle compromise on the AA/AF front etc.) Im prepared to do that, so its all good....... to use an Americanism! :D

599 3DMarks was my score (man that's terrible). I can see how the slideshow induces a state of upgrade envy. :) Although, from what I've seen, even with the best system on the planet will you get anything near a decent score so I would really ignore it. 3DMark is a good test for "Future" video cards, FutureMark (get it? :) ) that will come out in the next year or so....

The systems which score really well are overclocked systems with two 7800GTXs in SLI (scoring around 7000 points). Not really any indication of a system which will do well in future games, since by the time such games come out 2 7800GTX's should be beaten by 1 next-gen card, not to mention hardware DX10 support.

3DMark06 isn't terrible, don't get me wrong. But I don't like where it's going and I feel sorry for the people who are getting upset over their nice systems being "inadequate" when this is not necessarily the case. As I say on the front page, wait for a game which truly stresses your system, since Futuremark is not necessarily an indicator of the "Future", and the "Marks" it assigns are seemingly not balanced at all, and do not scale well.

Out of curiosity I went back to the earlier versions and tried them. Here are the results:

3DMark 03 - 5827
3DMark 05 - 2796
3DMark 06 - 599

I find it interesting that between 03 and 05, a 2 year period, the performance expectation has dropped about 2x from 5827 to 2796. But from 05 to 06, a 1 year period, I have a score that has dropped nearly 5x from 2796 to 599.

That will be due to the 256Mb memory requirement and the strain of 1280x1024 on an older card card. Also you do not have SM3.0 support. In previous tests your card has met the requirements therefore scored better.

This is what I got:

3d marks score: 1320
SM 2.0 score: 652
CPU score: 901

Athlon 64 3300+ 2.4Ghz, Ati x700pro 256mb oc'd 450/500, 512mb pc3200

man i feel better hanging around you guys than when I'm reading that monthly edition of maximumpc! lol

3dmark05 was 6052 with cat 5.13
3dmark06 is 2049 with cat 6.1
(sm2 - 1052 / cpu - 1048)

p4 3.6 925x Ex
800mhz fsb
2g DDR2 533 PC24200
ati x800xt pci-e

Something I have noticed is that Intel cpus perform better than AMD cpus on in 3dmark06.

I am of course talking of course of the single core cpus. that might attributed to intel's hyperthreading technology. My cpu is a 3800+ and scored 947. troyboy30's cpu is a 3.6 and outperformed me.

Surely, 3dmark06 has been optimized to run better on dual core cpus and it is probably a good thing. we can now truely see how a pentium cpu with hyperthreading can perform even without having dual cores....

Imagine if games were designed to work with dual cpus. AMD would probably not even be selling single cores to gamers anymore.

Just downloaded 3D Mark 06 and installed it. When I click on the "Run 3DMark" button or on the "Demo..." button, the screen blacks out and the monitor light changes colour fro mgreen to orange. Generally, this means that either there is no signal or an unsupported resolution was attempted. Repeatedly pressing ESC brings it back to the desktop, so it is not like the system froze or anything.

OK, my monitor is an old ViewSonic that *does* support 1280x1024, but only at 60Hz. Does 3DMark 06 try any other refresh rate, like 75Hz or something? If so, is there any way in which I can force it to use either 1280x1024 at 60Hz or lower the screen resolution to 1024x768?

My rig is in the sig, I am using the official 81.98 drivers and there is no overclock on any system component. 3DMark itself reported that my DirectX is up-to-date. Any suggestions?

I've noticed that a 6600gt and a x1600xt beat the living daylights out of a x800xt, which is because of the shader model 3, I think 3dmark06 is too retarded and is a over rated benchmarking product now.

Please don't attempt to swear in posts, even if it's asterisked out. I've changed your "bad word" to the word "daylights" as it fitted. Cheers, deltd

I've noticed that a 6600gt and a x1600xt beat the living daylights out of a x800xt, which is because of the shader model 3, I think 3dmark06 is too retarded and is a over rated benchmarking product now.

Please don't attempt to swear in posts, even if it's asterisked out. I've changed your "bad word" to the word "daylights" as it fitted. Cheers, deltd

Correct. 3dmark isnt only about performance, but graphics too. I, however, do not like the fact tht 6600GT gets more points than the X800XT, that is a card, at the very least, twice as fast. Anyways... as so many other say - only real games are the real benchmarks for gamers.

Danny - Regarding the differences between the 6600GT and the X800XT in terms of supported Shader Model versions etc. Seems that 3DMark06 uses SM3.0 to get that performance boost and for those cards, Nvidia supported it earlier than ATI. While I do agree it is not good to send the message to an X800XT owner that your video card is "obsolete" (due to a lower score). I do agree that baselining the standards level is a good thing. This means that to be competitive in the market and to make better games possible, a certain minimum level of technology should be the norm at any given time. In this case, 3DMark06 is saying that SM3.0 should be a requirement so if you ARE going to buy a new card in the near future (within 6 months to a year), you better make sure it has SM3.0 support. Just one example, but I think that high profile synthetic benchmarks like this really help to ensure companies like ATI and Nvidia keep offering the best they can to consumers (even if the rate that new GPU technology is going is a little quick in my book). :)

indeed. if you were benchmarking cars, for getting from place to place, a clunker might score almost as high as a brand new ferrari. But when you throw in tests for the special features or looks, the ferrari will always win... The X800 series does a marvelous job rendering pictures, but if you want the latest features in SM3 (which is clearly become industry standard now) you will need a newer ATI card or equal nVidia card.

AMD Zone (http://www.amdzone.com/modules.php?op=modload&name=Sections&file=index&req=viewarticle&artid=229&page=3) have a comparision article for 3DMark, and come to a similar conclusion to that which I ranted about earlier on the front page:3DMark 2006 is certainly good for eye candy, but we would not recommend one use it as the main reason for choosing hardware. We always must recommend real world application performance numbers to synthetic tests any day of the week. Among other things it appears to have 'Intel Optimizations' which give illogical results in the CPU test.

I read a review on Guru3d.com of 3DMark2006 and sadly they said it was really good. From the review...

Don't you just love what the new updated '06 has become?

*sigh*. I agree with you Koroush in that its pretty disappointing, even as a benchmark tool as opposed to real world game performance.

I usually agree with most of the reviews done on Guru3d.com done by the admin there, Hillbert, but not this one

Hm, I'm new to the benchmarking scene, so... I'm not quite sure what to make of all this. My PC is not top of the line by any standard, but it still did a good job with FEAR, CoD2, etc., even before I reformatted and optimized. I could never play with all the eye-candy maxed out, but it looked great and ran fine. I don't know how my pre-optimization score would have compared to my current setup, but here are my scores, for what they are worth:

3DMark Score: 1994
SM2.0 Score: 1072
CPU Score: 762

It was pretty much a slideshow the whole time--I think the max FPS I got was in the "Return to Proxycon" scene, topping out at a whopping 15 and mostly in the single-digits thereafter ;) The CPU test was disgusting, but from what I gather it was pretty bad all around. I found one other person with the same CPU/Graphics card and beat him by 23 points overall. Of course, I did have .5 gig more RAM than he did.

Burr, that sucked. My results are horrible, lol.

3DMark Score: 2583
SM2.0 Score: 1091
HDR/SM3.0 Score: 988
CPU Score: 997

I got that on a:
CPU: AMD Athlon 64 3200+ CPU @ 2.55 GHz
Motherboard: EPoX 9NPA+ Ultra
RAM: GeIL Value 1GB(2x512MB) 184-Pin DDR SDRAM @ 275 MHz [2.5 8-4-4]
Graphics: eVGA GeForce 6800 GS [Core: 479 MHz Memory: 1.05 GHz]
Hard Drive: Maxtor DiamondMax 10 200GB w/16MB Buffer

It just occured to me how rare it is to see anyone showcasing an Intel CPU in their signature anywhere.

It just occured to me how rare it is to see anyone showcasing an Intel CPU in their signature anywhere.
Haha, that's because we all know better, lol.

It just occured to me how rare it is to see anyone showcasing an Intel CPU in their signature anywhere.

There is nothing wrong with intel's. I'm perfectly content with my Pentium D. Might just be me but I think I made the right choice. :)

At the moment most people are choosing AMD because of its generally better performance in gaming, but please let's not turn this into a fanboyish affair.

Just interesting that Futuremark would optimize for Intel and not AMD where the Gaming market clearly puts all their money.

eh, where the gaming sector is putting their money right NOW... maybe 2006 should be amd optimized... but its hard to switch around who you are going to optimize with all the time, they probably have a contract or something... and with the 65nm stuff getting ready for actually making something of itself, mayhaps intel can still make a comeback... amd isnt even to 65nm and intel already is prepped for 45nm... at this rate they can put their memory controller onchip too and still have more room than AMD..... just imagine that

Imagine me still buying AMD.

At the moment most people are choosing AMD because of its generally better performance in gaming, but please let's not turn this into a fanboyish affair.
Of course, i'm no fanboy. My loyalties are with whoever has the best stuff at the time. And right now that would be AMD.

eh, where the gaming sector is putting their money right NOW... maybe 2006 should be amd optimized... but its hard to switch around who you are going to optimize with all the time, they probably have a contract or something... and with the 65nm stuff getting ready for actually making something of itself, mayhaps intel can still make a comeback... amd isnt even to 65nm and intel already is prepped for 45nm... at this rate they can put their memory controller onchip too and still have more room than AMD..... just imagine that

hold on man! Here (http://news.com.com/IBM+to+help+AMD+on+future+chips/2100-1006_3-979718.html) is a link from a 2003 article talking about AMD and IBM pairing up to make 65nm and 45nm. They weren't too far for the estimate that AMD will begin launching 65nm chips in mid-2006; the article said late 2005.

In another article from 2003, here (http://www.infoworld.com/article/03/12/09/HNamdnanometer_1.html), it says AMD is eyeballing 45nm process for early 2007, backing up the previous article.

Now these articles are outdated for sure and I'm pretty sure 45nm won't happen until 2008.

The heat is on! I can't wait to see an Intel chip with a memory controller onboard, so we can now see what the true potential of an Intel chip is.

I didn't want to spark a war or get off topic lol i'm sure Intel just paid futuremark a bunch of money to get optimizations for their line of products.

Yes AMD is the "gamers" cpu but I don't think that had to do anything with intel getting the optimizations. Like I said its about money and clearly futuremark is out for a profit with this lastest release and these two companies will fight for their processor to be the "better" proformer even if its only by a couple marks.

For example the slogan "Runs best on intel with HT tech. Or you have to use Zboard keyboad templates when you play battlefield 2 or you'll suck at it" :p doesn't mean that it is true.

Finally downloaded it:

4010 3DMarks
SM2: 1534
HDR: 1617
CPU: 1789

This is with my overclocked system so I'm not sure what I would've scored at stock speeds. A lot less probably.

Finally downloaded it:

4010 3DMarks
SM2: 1534
HDR: 1617
CPU: 1789

This is with my overclocked system so I'm not sure what I would've scored at stock speeds. A lot less probably.

wow. thats about what I got in 3dmark05! (4400). Run it again in the summer and see how well you do ;)

Finally downloaded it:

4010 3DMarks
SM2: 1534
HDR: 1617
CPU: 1789

This is with my overclocked system so I'm not sure what I would've scored at stock speeds. A lot less probably.

4010 Is a good score. So far the higest is 10386(not by me :cool: )

You can see a good comparison to other rigs here http://www.xtremesystems.org/forums/showthread.php?t=88168

They also have the records for overclocking and bandwidth on that site, every time I look at the overclock forum I drool on my self a little...http://justjustin.nsync.nu/emoticonsforjjb/drool.gif

There is an interview with two Futuremark guys here (http://www.pcinpact.com/articles/d/63/1.htm), answering a few pertinent questions that have been on peoples minds.

PCi : Why did you choose to use the same three 3DMark05's scenes (although we appreciate they have been improved), rather than creating new ones for the 3DMark06 version ?

TS : Rapid advanced in raw 3D processing power (SLI and Crossfire systems) enabled us to revisit those scenes and raise those to totally different level. We thought that it will be very good way to show in concrete way what a difference you will get with better hardware.

Well here are my scores with everything set to default - see sig for specs

7055 3DMarks

SM2.0 Score - 3884
HDR/SM3.0 Score - 3978
CPU Score - 1086

3DMark06 must have optimisations for Dual-Core processors, because even the 3800+ gets a better CPU score than mine.

Well, might have to go a Dual-Core Opteron in the near furture...

Yes, it has alot of dualcore optimizations.

http://www.xbitlabs.com/articles/video/display/3dmark06_8.html

The FX57 is getting beaten by most dualcores. I would however not worry. I mean we are talking about a FX57. The most highend single core cpu (?). I dont think you need to go with a new cpu for at least a year. But this is just me :)

Yes, it has alot of dualcore optimizations.

http://www.xbitlabs.com/articles/video/display/3dmark06_8.html

The FX57 is getting beaten by most dualcores. I would however not worry. I mean we are talking about a FX57. The most highend single core cpu (?). I dont think you need to go with a new cpu for at least a year. But this is just me :)

a year? holy sweet mama! we still have sometime to go before single cores are out....a little bit more than a year I think.

And again, while 3DMark may have dual core optimizations, we will have to wait and see just how much games implement these sort of optimizations. As yet a dual core CPU does not see any major improvement in most recent games. I certainly wouldn't despair about having to upgrade an FX-57 just yet :)

Here is a run at 3DMark06 with my system overclocked in the system specs and my X1900XT @650/1550

3DMark06
http://img347.imageshack.us/img347/3981/3dmark0665015500cx.th.jpg (http://img347.imageshack.us/my.php?image=3dmark0665015500cx.jpg)

a year? holy sweet mama! we still have sometime to go before single cores are out....a little bit more than a year I think.

I said at least a year. And with that i mean that the FX57 will be a top-cpu for that long. Afaik - there arent really coming too many new cpus from AMD, i know the AMDX2 5000+ and FX62 are to be released, but i dont think two dualcore cpus will kill of a CPU as the FX57.

Dyanmic, here we see how much dualcore gives to the points. I mean, im on a 2.2ghz (AMDX2 4400+, no overclocking) and a X1800XT (factory overclock, 700/1.6), and my points are almost like yours. If you had a X2 overclocked to 2.5, i bet youd get much more points if you had a dualcore.

But as Koro said, no need to upgrade because of 3dmark.

Yeah, cause what Quake4/COD2 are suppose to support dualcore and look at the stats, not so good. The cod2 patch for dualcore had no performance gain, but they are probably going to be games out to support it this year. A nice article here if you want to read it....

http://www.xbitlabs.com/articles/video/display/3dmark06_7.html

Additional Comment:
I just finished putting my HS back on my X1900XT and i can clearly state that my memory is 1.1ns. I've also cleaned off the thermal compound that was there from factory and put AS5. I set my fan to max with OVERCLOCKER.exe, 4950RPM and benchmarked with 3DMark06 with clocks 689/797 set with OVERDRIVE and didn't break 60c underload.

3DMark06

http://img99.imageshack.us/img99/6291/3dmark05x1900xt6897975ca.th.jpg (http://img99.imageshack.us/my.php?image=3dmark05x1900xt6897975ca.jpg)

A nice perfomance boost wirth the new ati 6.2 drivers. From 4759 3dmarks to 5270. I overclock also my AMD to 2500 and my ram to 250(500)

Indeed, I like what ATi have done to their drivers. From being plain bad, to simply amazing. I mean, when i bought my card, its was it vs the 7800GTX512. Today, reading reviews on my card against X1900XT and 7800GTX (with more recent drivers) shows my card FAR ahead of the 7800GTX512, and very close to X1900XT. Good job ATi!

Im hoping to break 5k in 3dmark06 and 10k in 3dmark05 (see current scores in sig), all this on stock settings without any manual overclock. ^_^

Man, you guys are making me jealous of your ATI cards now....now im second guessing my decision.. :D

Im Amazing!

CPU Scores:
Athlon 64 3000+ (venice) @ 2.4ghz : 900 CPUmarks
Opteron 64 165 (denmark) @ 1.8ghz : 1300 CPUmarks
Opteron 64 165 (denmark) @ 2.6ghz : 1950 CPUmarks

total:
2200 with the 3000+ (with the GPU overclocked too)
2300 with the 165 overclocked, 1900 with it not


(Ran it again, this time with GPU overclocked and the 165 overclocked)
2653 Total
1047 SM2
920 SM3
1921 CPU

In the sig your opteron is at 2.6, tried benchmarking with those settings or do you have it at 2.4 for stabillity? :) Anyways, very nice overklocking ;)

I gotta read up on how to do everything ^_^ One over at driverheaven had his AMDX2 4400+ @ 2.75 ! Thats almost an FX62 (in terms of ghz). Id be happy if i got mine up to 4800+ or, if i really have luck - to FX60 specs :D

ooops! yeah, it was at 2.6 when i benchmarked. Im gonna get an aftermarket cooler and push myself even farther i think ^_^

Here are my 3DMark06 figures:
7903
3572 sm2.0
3343 sm3.0/hdr
2128 cpu

My 2 GTX are clocked at 480 core/1.25 ram (not an oc), CPU at 2.83GHz (~200MHz oc).

Nvidia driver version 81.98 (non WHQL version from December).

Ok, some data for the 84.20 beta driver under 3mark06:
Total 3Dmarks 8128
SM2.0 - 3671
HDR/SM3.0 - 3485
cpu - 2143

Any increase is a good increase. Not seen any image degradation or games fail to run under these drivers yet.

Hi, I'm new here but anyway with stock clocks........wait, let's start with my PC specs:
ATI x800 Pro 256mb,
1.5 GB RAM 2800,
Audigy 2 ZS Sound Card,
Pentium 4 2.8ghz,
Win XP Home Edition SP2
P4V8X-X Motherboard.

With stock clocks, I got 1514 score.

With overclocked (560 core, 540 memory), I got 1759 score.

This was with using 3DMark06.

Is there something wrong with this picture?

Hi, I'm new here but anyway with stock clocks........wait, let's start with my PC specs: ATI x800 Pro 256mb, 1.5 GB RAM 2800, Audigy 2 ZS Sound Card, Pentium 4 2.8ghz, Win XP Home Edition SP2.

With stock clocks, I got 1514 score.

With overclocked (560 core, 540 memory), I got 1759 score.

This was with using 3DMark06.

Is there something wrong with this picture?

nope you are within specs. When my system was at stock speeds, my score was

1734.The performance of the x800pro is a tiny bit less than the x800xl, but the score difference between your system and mine is mostly related to the cpu

Yes but that's not a bad cpu, with those specs I would have expected to get a little bit better performance....what the hell am I doing wrong?

Welcome Captian Kaminsky,
Please double check that you are posting in the correct thread in the future. Ive moved the post related to your question and cleaned up the previous thread.

Have a good day sir.

Yes but that's not a bad cpu, with those specs I would have expected to get a little bit better performance....what the hell am I doing wrong?

Absolutly nothing. Your videocard doesnt support SM3.0, so not all test are run and thus you get a low score (you dont get any SM3.0 score to add up). You should try 3dmark05 instead. Expect a score around ~5000 with your system.

Grrr....ok I'll try it out. Thanks for the help.

Finally gave it a try!



Total Score: 4373

SM2.0 Score: 2101

HDR/SM3.0 Score: 2066

CPU Score: 917 (yeah kept me back a little bit)

Total score: 5795
SM2.0 Score: 2781
HDR/SM3.0 Score: 2868
CPU Score: 1150

Total Score 5403
SM2 2174
HDR/SM3 2407
CPU 1638

Total Score 4433
SM2 2184
HDR/SM3 2158
CPU 870

AMD X2 4400 Dual Core Stock (2.2GHz)
MSI K8N Diamond Plus
Dual MSI GeForce 7900 GTs 256MB in SLI
2GB Corsair Twinx 4400 Extreme Overclocking
620W Enermax Powersupply

From an upgrade from a P4 3.2 @ 3.15 w/HT, AGP 6800 Ultra 256MB, 2GB corsair twinx 4400 ram, asus P4C800-E Deluxe to this, it was a pretty big jump. Especially since i had my 566MHz Celeron, 866MHz P3, 3.2GHz P4.

3DMark 01: 25464
3DMark 03: 18453
3DMark 05: 10963
3DMark 06: 6039
>>>SM2.0: 2432
>>>HDR/SM3.0: 2804
>>>CPU: 1679

Overall,
I was VERY happy with this system upgrade. I didn't upgrade based on 3DMark 06 or 05. My Asus motherboard crapped so i just sent it back, passed it to my mom and i built a new computer. I encourage everyone to upgrade to SLI. You don't need the latest and greatest videocards for it eather. Dual 6800 GTs/Ultras, 7800GTs. I picked 7900 GTs because just one outperformed the 7800GTX for less the price. I don't recommend the 7900GTX unless you want to go extreme.

I was just reading this week, i know it's old news but im excited to see that nvidia released it's new quad GPU setups. A basic alienware computer with Quad SLI is $5,999. YIKES! That will be my next upgrade in 2-3 years :)
http://www.nvidia.com/object/IO_30228.html

World’s fastest class claim based on testing Quad SLI against the leading competitor, the ATI Radeon X1900 XTX Crossfire. Performance tested on Monolith’s F.E.A.R game run at 2560x1600 resolution with soft shadows enabled. NVIDIA Quad SLI scores 47 frames per second versus 26 frames per second for ATI Radeon X1900XTX Crossfire. Benchmarks were run on an Athlon FX-57 system with 1GB RAM and Windows XP SP2. Performance tested on Doom3 game run at 2560x1600 with 4xAA/8xAF. NVIDIA Quad SLI scores 108.9 frames per second versus 84 frames per second for ATI Radeon X1900XTX Crossfire. Benchmarks were run on an Athlon FX-57 system with 1GB RAM and Windows XP SP2. Performance tested on Chronicles of Riddick game run at 2560x1600 with 4xAA/8xAF. NVIDIA Quad SLI scores 82.9 frames per second versus 53.52 frames per second for ATI Radeon X1900XTX Crossfire. Benchmarks were run on an Athlon FX-57 system with 1GB RAM and Windows XP SP2.

World’s fastest class claim based on testing Quad SLI against the leading competitor, the ATI Radeon X1900 XTX Crossfire. Performance tested on Monolith’s F.E.A.R game run at 2560x1600 resolution with soft shadows enabled. NVIDIA Quad SLI scores 47 frames per second versus 26 frames per second for ATI Radeon X1900XTX Crossfire. Benchmarks were run on an Athlon FX-57 system with 1GB RAM and Windows XP SP2. Performance tested on Doom3 game run at 2560x1600 with 4xAA/8xAF. NVIDIA Quad SLI scores 108.9 frames per second versus 84 frames per second for ATI Radeon X1900XTX Crossfire. Benchmarks were run on an Athlon FX-57 system with 1GB RAM and Windows XP SP2. Performance tested on Chronicles of Riddick game run at 2560x1600 with 4xAA/8xAF. NVIDIA Quad SLI scores 82.9 frames per second versus 53.52 frames per second for ATI Radeon X1900XTX Crossfire. Benchmarks were run on an Athlon FX-57 system with 1GB RAM and Windows XP SP2.

Not a fair comparision. FOUR cores vs TWO cores. I don't want to be tagged with fanboyism, but if you go percentage wise, TWO Ati cores are faster than FOUR of Nvidia's.

Clearly, the Quad core system was limited by the cpu and ram. This means that before these quad video cores, we need better cpus and faster ram.

Sure it's coming with AM2 and the Intel Conroe.....I have my doubts. Currently, AM2 only improves the ram bandwidth by about 12%(rough estimate). Not enough to see the true potential of a Quad core system.

So if there isn't the hardware for people to run the hardware to it's full potential, like how you mentioned the CPU bottleneck, why would they release these quad SLI cards? Is it safe to say it's completely not worth getting a quad SLI system within the next year? I remember reading soomewhere how quad core and multicore will be released within the next couple of years or so. Correct me if im wrong, ATIs crossfire seems to have more performance then nvidias SLI.

Correct me if i'm wrong here too; I guess this world is all about profits and getting stuff out as quick as possible instead of taking the time to do things right...

the quad core system was probably introduced for bragging rights....It is clearly not for the average user.

The avid gamer still can get by with only one GPU. It wouldn't make sense that you would HAVE to buy 2 videocards just for gaming. That's just stupid.

So that's the way I see it. While it's always fun to here about companies going over the top with their products, it doesn't serve me as a customer...in fact it's just money wasted that will make my next videocard purchase more expensive because money isn't spent on developping new more powerful cores, it's spent making 4 cores working together, something I'll never own.

Sorry for my ranting....I've been sleepless for days because of exams and thankfully they end tomorrow.

For the 3DMark05 with SLIAA-6835
For the 3DMark05 with SLIMultiGPU @ High Quality-9042
For the 3DMark05 with SLIMultiGPU @ Quality-9559
For the 3DMark06 with SLIMultiGPU @ High Quality-6274
For the 3DMark06 with SLIMultiGPU @ Quality-6398
Not too shabby.

I try not to get hung on the 3Dmark scores. But it is interesting to see how youe computer stands up to them. I can play any game on the market now and im sure for a couple more years i will be able to at full detail. I don't plan on upgrading again until DX10 comes out and when CPUs can push quad SLI to it's limits. I am looking at an Antec P180 case for a new case. Right now im using a modded P160 case. Took my side cover to a shop that my father works at and has a biohazard symbol machines out real nice and threw some pexiglass in with a 15" green cold cathode, replaced the eyesoreing blue superbright LEDs with superbright green ones to go with the theme. My old cathode is crapping out and i just ordered new ones so i think i may wait until after the next LAN party to get a new quieter case that will keep things nice and cool. Anyone have any suggestions for a case with the looks of the P180 but also keeps things nice and cool on the inside?

Something is obviously wrong here...I play BF2, FEAR, and Oblivion all the time but when I ran 3Dmark06 right now FPS was constantly under 10

Something is obviously wrong here...I play BF2, FEAR, and Oblivion all the time but when I ran 3Dmark06 right now FPS was constantly under 10

nothing wrong, just that 3dmark06 is doing what it's supposed to, standing up to recent video cards until more powerful ones are available.

weird, on oblivion i have the graphic settings pretty high with HDR and 3.0 pixel shading enabled...pretty much what the new 3dmark is testing right?

That's right, but it is being alot more demanding with more complex shaders and probably a higher resolution than you're gaming at in Oblivion (1280x1024). 8xAF and a minimum 256MB video RAM requirement (it will work with less but will be slow) is also in effect.

I overclocked my GPU last night to 408/1.01 but when I check the system details on 3dmark it says it's 351/1.0...351 was the core frequency before i overclocked. Is the program confused? Because my forceware still says 408/1.01

Futuremark benchies give false readings on clockspeeds extremely often. Nothing to worry about there.

My 7800GT has just been sold and it benched 3707 (http://service.futuremark.com/compare?3dm06=269097). I wonder what my 19XT will do next week. :)

X1900 in crossfire, both cards at stock with master x1900xt @ 625/1450, slave x1900xtx @ 650/1550.

Default benchmark settings:

Total score : 6914

SM2.0 Score : 3560
HDR/SM3.0 Score : 4286
CPU Score : 1035

Run with 4XAA and 8xAF (AA looks lovely with the HDR in the benchmark):

Total score : 6252

SM2.0 Score : 3286
HDR/SM3.0 : 3384
CPU score : 1034

Both runs done with a single core Amd 3500 @2.67

ok, here are my results :(

3DMark score: 2225
_______________________

SM2.0 score: 890

HDR/SM3.0 score: 838

CPU score: 1072

This seems like my graphics card is clearly my "bottleneck" but I thought my 6800 would have performed better than this??? I've gone through all of the tweaks realted to this area.

Any thoughts anyone?

First run with my new x1900xt nets me 4752 (http://service.futuremark.com/compare?3dm06=278154).

4125 with a lot of glitches and errors during the "deep freeze" portion of the test with my 3-day old 7900 GT card. Hopefully when I receive the RMA this week or next, it'll score higher.

Correct me if I'm wrong, but I believe the remarkable new forceware driver that's just out fixes that?

New score in sig!

Correct me if I'm wrong, but I believe the remarkable new forceware driver that's just out fixes that?


I wish, the RMA that's on it's way with the modified board fixes the issues. :)

Some numbers from my recent upgrade to 7900GTX SLI:

3DMark06 - 3DMarks 9472, SM2.0 4461, HDR/SM3 4372, CPU 2098.

Pleased with that.

So I found something odd recently.

I built a pretty nice system this last December. About a month after I built it, the new 3D Mark came out and I ran it. Got a 3802.

So in the last few months I've gone through the full Windows Tweakguide, as well as the Nvidia Tweakguide, learned a lot, done a lot of optimizations. Yay!

So I was suprised when I just ran 3D Mark and got a 3693 - A LOWER score. Obviously, it's entirely possible that in my optimizing I complete screwed something up, but I'm not sure how - I didn't do any of the really dangerous stuff.

Looking at the details, both my SM 3.0 and CPU scores are higher - but my SM 2.0 dropped from 1552 to 1360, dragging the total down. Anyone know if this is the latest official NVIDIA drivers being more tuned to SM 3.0? Or just the fact that my system is 6 months older then the first time I tested? Sorry if this comes too close to 'asking for help' - I'm willing to do plenty of tests to investigate this, just don't know where to start, or if it's something simple/natural that is okay.

Also, there is someone with my exact same system (they actually have slightly LESS RAM) that has a score over 6000 - nearly double mine. I figure they must be overclocked, but still - nearly double? Is something wrong with my system?

Well, it would certainly help if you post your system specs. No amount of overclocking will give a nearly double score in any benchmark. At most, you may expect a 20-30% increase, if you are seriously OCing your system with proper cooling.

Regarding why you got a lower score after tweaking, some times it is really difficult to identify exactly which tweak will cause this problem although there are some usual suspects. But that can be identified only with your system specs.

I get a 3d06 score (basic test) @ stock of 3680 - not too bad for my setup I guess.

Thanks for the help - yeah, that kind of suprised me - even before the tweaking, with the better score, his score was a good 40% greater than mine.

Unfortunately I was very dumb and did a fair amount of tweaking at one time, so finding the 'issue' will be damn hard - I figure I might try doing a super-far-back system restore (if I have such a restore point) and seeing if it gets a higher score.

Anyway, my specs:
Processor:AMD Athlon 64 X2 Dual Core Processor 4200+
Graphics: BFG GeForce 7800 GT 256 MB - I think BFG overclocks them a bit, for what it's worth.
Case (if this matters) XG Dragon. 500 Watt power supply which seems to be pretty reliable. Cooling is 1 80mm fan in the front, 1 80mm fan in the back, anda side air duct. I have an extra fan I can stick in if I need to but haven't as I enjoy the quietness of the case.
RAM: CORSAIR XMS 1GB (2 x 512MB) 184-Pin DDR SDRAM Unbuffered DDR 400 (PC 3200)
Hard Drive: Western Digital Cavier SE 320 GB 7200 RPM SATA150
OS: Windows XP Pro
Sound Card: Soundblaster X-Fi
Monitor: Dell 2005FPW

So yeah, not top of the line but not a bad system methinks, was hoping it would last a few years (with the exception of RAM, which I fully expect to upgrade during the computer's lifetime) - but already I get only 16-17 FPS in Oblivion's forested areas, and that's AFTER using the tweakguide and turning down some settings. So maybe my tweaking is behind it but, as noted, my score was low compared to others BEFORE that.

Idea: I saw no SLI indicators in the PC Mark Database, and when I compared my system some test titled "Dual GeForce 7800s" or something was near the top of the list. Is it possible its comparing my system to SLI systems? If so, that would explain the score difference.

Surely 16-17 FPS in Oblivion with your system specs indicate something wrong somewhere, unless you are running at a very high res like 1600x1200 with 4xAA or probably higher. Why, I get that FPS with my system running Oblivion at 1024x768 no AA everything Medium or lower. Not as pretty as I would like but it runs.

Is this only in Oblivion, or are you seeing this low FPS in most other games as well? Did you recently switch video cards, especially from ATI to nVidia? Have you tried re-installing the entire setup, as per the TGTC? I mean starting from the OS.

It could be the nvidiadrivers. I know that 3dmark06 scores differ from driver to driver. E.g the highest i´ve ever got (see sig) was with Catalyst 6.3. The 6.5 drivers are ~50points "slower" than the 6.3s.

And the latest games use shadermodel3.0 so that would mean that your computer is now faster than before in the latest games, where the performance is really needed.

Fortunately, the computer runs every other game like butter - Oblivion is the first to give me real slowdown.

I AM running at 1680x1050, with HDR on, but no AA (can't have them both at the same time) and no vsync (as much as the tearing annoys me, with vsync it can become unplayable). Yeah, when I got this monitor I worried about running it at really high resolutions but my friend who had this monitor said the performance difference was negligible...and as skeptical as I was, I found he was right; if I set it to 1024x768 the performance increase is barely noticable, if I remember correctly.

No, I haven't switched any hardware since getting the computer.

Yeah, I'll probably try re-installing the OS one of these days if I have time, if only for the experience and to follow Tweakguide's recommended setup - school just got out so never any time to deal with this before. That wipes out your hard drive data though, doesn't it? I can back up all the important stuff on a DVD, I suppose...

Fortunately, the computer runs every other game like butter - Oblivion is the first to give me real slowdown.

I AM running at 1680x1050, with HDR on, but no AA (can't have them both at the same time) and no vsync (as much as the tearing annoys me, with vsync it can become unplayable). Yeah, when I got this monitor I worried about running it at really high resolutions but my friend who had this monitor said the performance difference was negligible...and as skeptical as I was, I found he was right; if I set it to 1024x768 the performance increase is barely noticable, if I remember correctly.

No, I haven't switched any hardware since getting the computer.

Yeah, I'll probably try re-installing the OS one of these days if I have time, if only for the experience and to follow Tweakguide's recommended setup - school just got out so never any time to deal with this before. That wipes out your hard drive data though, doesn't it? I can back up all the important stuff on a DVD, I suppose...

Only GPU's with 512mb RAM or more handle high resolutions with less impact on performance as far as I know. If you have a widescreen with native of 1680x1050 Ideally you'd need a 512 card

Repair reinstall will allow you to keep your data on HDD, but if you can do a clean reinstall thats better

Really? Well damn, cause I can't afford one of those (or anything else added to this computer) - this was the first time I built my own computer and I mostly relied on the advice from my techy friend, and his system runs fine with this. One thing I hadn't considered, however, is that he's deaf, and therefore runs everything without sound, which I imagine frees up some system resources for graphics.

Oh well, guess I'll just have to get a second card and go SLI one of these days, though I suppose the problem with that is that you don't get the features of the new GPUs.

It also just occured to me that if I did a fresh install, I'd have to go through the tweakguide ALL OVER AGAIN - it took me a long time to do it the first time and I don't look forward to repeating the process. So I guess I'll just hope that Oblivion is the only offender.

One thing I did find when tweaking is that my graphics card and sound card are on the same IRQ (I think that's what it's called) - dunno entirely how that happened. Apparently this can (maybe?) be fixed by moving it from one PCI port to another, but I'm fairly nervous about doing that - I had to get help installing it last time due to this one weird headphone-port-plug. How much of an effect on performance does it have when the graphics and sound card share the same IRQ?

In my opionion you shouldn´t worry too much about loosing a few points in 3dmark (expesially since both your cpu-score and sm3.0-score are now higher).

The reason you only notice "lag" in Oblivion is because Oblivion is such a demanding game. A 7800GT is simply not enough to run the game at max settings. You need to go through the oblivion tweakguide and lower a few things that might lower the graphics only a little bit while improving performance (like changing shadow filtering to low or applying the grass = 120 tweak - much more of these in the tweakguide).

http://www.hardforum.com/showthread.php?t=1035693

That is a topic with 7800GT users and what settings they use ingame (Oblivion). As you can see, some are running in 1024X768 and bloom instead of HDR etc. One was even running in 800X600. So i think your computer is just fine as it is now :)

Good points Danny. I've never paid much attention to 3dmark scores. Its an artificial benchmark that doesnt mean all that much. To me, much better tests are running benchmarks of particular games that i play, so i can see the real difference in fps and so forth

4010 3DMarks
SM2: 1534
HDR: 1617
CPU: 1789


I just ran 3DMark06 again after 4 months after posting the above; there's been practically no change to my system in terms of overclocking, only different graphic drivers. Here's the score I had today.

4405 3DMarks
SM2: 1750
HDR: 1745
CPU: 1848

That's quite a jump!

I ran it earlier today and got:


8808 3dmarks.

SM2.0 score : 3795

HDR/SM3.0 score : 4430

CPU score : 1944

However, if i run it say 5 or 6 times in one afternoon the results do vary, not by much, but vary nonetheless.... I also get better CPU scores in XP X64 but lower overall, SM2 and HDR.... Weird.

It is, afterall, only a synthetic bench test.

eligy thats a nice CPU score u got

I just tried this test and I fear my score isn't very good.

3829 3D marks

SM2.0 score : 1751

HDR/SM3 score : 1742

CPU score : 902

CHeers... THe thing with 3dmark is, as Koroush said at the start of the thread somewhere, its very much a '6 months or so from now' thing. Every time I watch it run and see my framerates running at 1/3 of what they are in games I think 'what on earth is this doing to my PC cos I can't see whats making it use so much power'.

Zorrofox, I think that's a pretty respectable score matey, especially on a single 7800. Is it a 256 or 512 MB card?

It's just a 256Mb model. Someone else I read had the same CPU and only a GT version of my card and they got a much higher CPU score. I just wondered why. I've only just installed nTune 5.0 last night. I ran the Coarse tuning option and things have definitely improved but I wonder if there's much point in running the Fine option. It takes 3 hours! I don't know if I can do without my PC for that long :D

He had overclocked his CPU pretty much and thus got a much higher cpu-score. :)

You're talking about me? It's cause my CPU is a dual-core, whereas zorrofox has a single-core 3800 (at 2.4ghz), so I have a 100mhz advantage.

I see. Thanks for the heads-up. I think I'll run the ntune fie tune function this avo while the footie's on. Don't suppose I'll be using the PC much.

Additional Comment:
Aye, it was you! Sorry, I didn't note your name before I started to post.

That clears things up then. I know this isn't really the thread for it but do you think I'll see much of a difference after running the ntune fine tune function?

You're talking about me? It's cause my CPU is a dual-core, whereas zorrofox has a single-core 3800 (at 2.4ghz), so I have a 100mhz advantage.

..not only the 100mhz, but a dual core scores alot higher than a evenly clocked single in 3DMark06

3DMark06 - 3DMarks 9472, SM2.0 4461, HDR/SM3 4372, CPU 2098.

sweet...lucifer
that definetly screams over my 1500 score :P

haha, 1578 without an overclock on my X800 Pro. 3Dmark just makes me feel bad, but finally got to installing it - VF700 cooler coming for my card, going to overclock and look at the improvement. Man, if I had $4000, what would I do with it...

In the 4850 range, normally. I hit 5200+ with my video card overclocked once, but i didn't want to take any chances and clocked it back down.

X2 4400, 7900GT, 2GB RAM

CPU @ 2.75Ghz and GPU 605/1600
3DMARK06 Scores = 9048
http://img128.imageshack.us/img128/5817/rock7li.th.jpg (http://img128.imageshack.us/my.php?image=rock7li.jpg)

Gamign performance is a ripper. I can use 16AF in Oblivion now with 1600*1200.

Bought the card sometine ago(2 months) for $835 from MSY after selling my 7800GTX SLI for $743.

CPU @ 2.75Ghz and GPU 605/1600
3DMARK06 Scores = 9048
http://img128.imageshack.us/img128/5817/rock7li.th.jpg (http://img128.imageshack.us/my.php?image=rock7li.jpg)

Gamign performance is a ripper. I can use 16AF in Oblivion now with 1600*1200.

Bought the card sometine ago(2 months) for $835 from MSY after selling my 7800GTX SLI for $743.

Do you find much performance gain over the SLI 7800GTX's? Were they standard cards or the 512MB xxx editions? I seriously considered the GX2 but was recommended that financially it wasn't worthwhile as there would not be much of a performance boost. Having said that, since you managed to get a damn good price for the old cards I reckon you did the right thing anyways.

Do you find much performance gain over the SLI 7800GTX's? Were they standard cards or the 512MB xxx editions? I seriously considered the GX2 but was recommended that financially it wasn't worthwhile as there would not be much of a performance boost. Having said that, since you managed to get a damn good price for the old cards I reckon you did the right thing anyways.
I think you should stick with what you have. They should perform similiar to 7950GX2.

My experience with 7950GX2 vs 7800GTX SLI(256Meg), hmmmm honestly, not much of a difference!! Huge difference when I went from single 7800GTX to SLI!!!!!!!

Although I can now use 16 AF, I still don't find much improvement in image quality which is normally observed across other games when using AF. I guess Oblivion already looks so good when run at 1600*1200.

Perhaps there might be difference at resolutions higher than 1600*1200???

Cheers
Sandeep

Hmm I run oblivion in 1920x1440 and its fine with 8xAF, I dont bother with 16x as while its pretty if you just look at the image, and great for screenshots, I don't notice the slightest difference in quality during actual playing, i never stand still that long in a game.

my 7800's are the 512 MB higher clock frequency models, so I guess you are correct, performance should be similar, certainly the guys at overclockers.co.uk and a few other stores I asked said I'd be as well waiting for the 8xxx series cards now. Still, yours will be more power efficient and a touch cooler I expect.

CHeers... THe thing with 3dmark is, as Koroush said at the start of the thread somewhere, its very much a '6 months or so from now' thing. Every time I watch it run and see my framerates running at 1/3 of what they are in games I think 'what on earth is this doing to my PC cos I can't see whats making it use so much power'.


The reason 3dmark06 brings cards to it's knees is because the benchmarks make it seem like there is 30 people under 1 guy, so that really pushes your graphics card.

Results on on my new XPS M1710 (specs in sig):

4643 3DMarks

SM 2.0 Score: 1920

HDR/SM3.0 Score: 1887

CPU Score: 1633

This is with old drivers and with all the standard bloatware on the machine (although I am not so sure the latter makes much difference here), so perhaps the scores might improve once I get a chance to tweak, update and optimise.

I did notice that the CPU tests were a slideshow at about 1 FPS.

not too bad STR, I have the same card and my score is 5300
although my CPU score is only 1000

Nice. Do you know how I could get up there? I guess maybe that indicates the difference between nVidia's mobile and desktop 7900GTX. I assume you aren't overclocking yours. Maybe a driver update will give me a boost. I have been considering getting the ones from laptopvideo2go rather than the ones on the Dell site.

Though, I am quite happy with the score. Most reviews I have seen for the laptop at similar configurations have it at around that. Of course, the real tests come with the intensive games and I haven't tried any out yet but the score seems to indicate I am in good stead.

I would use the latest available on nvidias site.

http://www.nvidia.com/object/winxp_2k_mobile_84.63.html

But then again, i dont know if dell/laptopvideo2go has newer drivers.

6236 3Dmarks
SM2.0: 2766
HDR/SM3.0: 2601
CPU: 1783

this is with CPU @ 2.4 [normally runs at 2.6 and i'll try it up there later] with my my new 7800GT SLi setup. not bad considering i paid considerably less for my two cards than i would have for a 7900GTX and am getting better performance.


6358 3Dmarks
SM2.0: 2774
HDR/SM3.0: 2624
CPU: 1917

second one is with the cpu back up at 2.6ghz. doesnt really make a huge difference to the overall score.

10520 marks

SM2.0 4458
HDR/SM3.0 4888
CPU 2690

http://service.futuremark.com/compare?3dm06=489640

Deeko thats a killer CPU score u got, feel the power of Conroe :)

Deeko thats a killer CPU score u got, feel the power of Conroe :)

Yeah Im not finished yet by a long shot, just had it a few days, I got to 3ghz without even upping the Cpus vcore, Im pushing for as close to 4ghz as I can get when my new Ram arrives. Im just using a 1Gb stick of average Corsair XMS DDR667, waiting on some G.Skill DDR800 low latency to give me more performance and aid me in overclocking it more. Ill post back in the coming week with my new results.

Edit: Forgot to say, cards are at stock speeds ,cpu load temp is 40C with two instances of P95 running.

Updated results with the 64x2 at 2.9GHZ :)

http://service.futuremark.com/orb/projectdetails.jsp?projectType=14&projectId=452453

3DMark Score 9104 3DMarks

SM 2.0 Score 4187 Marks

SM 3.0 Score 4175 Marks

CPU Score 2104 Marks

very nice, but previos conroe score is simply stunning

Heh heh aye, conroe's are awesome..... I'm having a hard job holding off buying one just now.... Still, am giving it til the end of the year when there's a greater range of boards and chips to choose from....

Figured this would be the place to post this, since the other forums really don't fit with this kind of post.

I downloaded 3DMark06 (unregistered) to try and see what kind of results my new laptop could produce. It's a Dell Inspiron 9400 with a Core Duo T2600 CPU, a 256MB nVidia GeForce Go 7900 GS GPU, and 1GB RAM. I ran the tests to see how well it could do.

...it fell flat on its face.

The first test, Return to Proxyon (sp?), was pretty bad. The FPS ranged from ~15-20 FPS to 8 FPS, depending on what's on screen. The Firefly Forest test was a bit more stable, but still hung around ~13-15 FPS. The test after that was just a slideshow, and I cancelled it right there.

This just boggles my mind. This is the same machine that can produce over 100 FPS on the CSS Stress Test (1280X1024, DX9.0, everything set to max settings, including AA and AF), produce 60+ FPS for F.E.A.R (1024X768, everything to highest except Corpse Detail set to Low to remove bodies faster, and Shadows set to Medium and not Soft), produce 40+ FPS for CoH Demo (1280X1024, everything set to High except AA off), and other games with similar results (Oblivion, BF2142 Beta and DoW). Yet this benchmarking program, of which I expected that this machine would be able to handle, is brought to its knees easily by it.

What is going on here? How is it that during actual gameplay the machine handles perfectly, but fails terribly with 3DMark06?

Merged with the main 3dMark 2006 thread. Please search for existing threads on a topic before posting - SortYa

The "slideshow" you are refering to is normal. It doesn't test the videocard but the cpu. Let 3dmark run the full tests and post the score you get in the forums and compare with what other users got.

Well, it did say it was a CPU test, but I don't understand why it's so slow. You'd think that with a T2600 processor, it could handle that test...it's not the highest-end CPU out there now, but it's not at the point of being obsolete...

the way 3Dmark uses it's engine doesnt correspond to normal game engines or how they perform. i'm not sure of exactly how it works, but i understand that alot more is going on than is actually rendered, for example each bot that is displayed may act as 3 or four when running the calculations.
The CPU tests always run like a slideshow, even the most powerfull cpus out cant render that kind of scene in real time the way gpu's do. my 3Dmark06 score around 6300 with a cpu score just shy of 2000 and the cpu test always run at 1-2 fps.
And as Ontarian said, what you're experiencing is perfectly normal. just let it run it's course and see what your score is.

Hmmm...well, I thought at first that it was a problem, but I guess it really isn't. Might have to rerun it to see what happens.

Still not to happy about the FPS from the previous tests, though... :(

Updated Score:

3Dmark06 : 10822

SM2 Score : 4469

SM3 Score : 4902

Cpu Score : 3024

http://service.futuremark.com/compare?3dm06=495070

Still overclocking the Cpu so hopefully will break 11K soon.

Ugh...my score is utter CRAP.

3DMark06: 3526
SM2.0: 1370
HDR/SM3.0: 1367
CPU Score: 1707

What a bunch of BS. Guess this new laptop isn't so great, after all...

That is a very good score for a laptop. I don't know why you are complaining so much. 3dmark is a SYNTHETIC benchmark, meaning that it TRIES to mimic real world results.

Trust me on this, here are my scores (from a long time ago)

Total 3dmark06: 1734
SM2.0 score: 883
cpu score: 947

ontarian_frog is right. 3dmark06 doesn´t mean much anyways. Its a good way of measuring diffrent drivers/hardware on your computer - and maybe compare with others.

To run the (gpu-)tests decent you need the highest crossfire and SLI cards. The CPU tests will always run 1-2fps... no CPU can do better.

What 3dmark06 is good at is to make people feel bad about their computers - and this has just been proven :)

Ive just overclocked my CPU to 2.5ghz and ran prime95 for 24h without any errors. Ill try to run 3dmark again soon and will see if the overclock had any effect on performance or not.

/EDIT;
Here are the scores;
3dmark06 score: 4941 (grr give me 5k)

SM2.0: 2008
HDR/SM3.0: 1963
CPU score: 1927

http://img86.imageshack.us/img86/6291/dkc7.th.jpg (http://img86.imageshack.us/my.php?image=dkc7.jpg)

SM2.0: 2240
HDR/SM3.0: 2494
CPU score: 2666

This is the best score I was able to get with my system so far with the noob friendly OC profiles in the BIOS. Funny thing is, Speedstep works even when I overclocked conroe to 3.0GHz. It idles at 2.0GHz and hits it back to 3.0GHz under load. Thats just great for daily use. With my Zalman CNPS9500, OC'ing from 2.4 to 3.0GHz only increases temps to about 3 degrees higher :eek: From 20C to 23C... Conroe is an overclocking beast...

It's kind of funny, because with 3DMark05, I get double the score of 6355. Which is weird, because the two Return to Proxycon demos they have look the same...

There is a lot more detail in the Proxycon test with much better lighting used. Also 3dmark06 uses HDR and Shader model 3.0 in some tests while 3dmark05 uses only the older shader model 2.0 and no HDR in all its tests. HDR and SM3 will place a massive load on the graphics card than what SM2 will. Just wait until SM4 is released with DX10, I wonder just how badly HDR/SM4 tests in a future 3dmark will bring the first wave of DX10 cards to their knees :)

Yeah, lighting does look a bit better in 06 than it does in 05, now that I look at the two. Ah well, one can't really judge the power of their computer through these benchmark programs, it seems...

[content removed]

You have already created a new thread for the exact same problem here (http://forums.tweakguides.com/showthread.php?t=4451). Kindly avoid posting the same thing in multiple threads. ~Stryker

http://img301.imageshack.us/img301/6769/3dm6926va6.th.jpg (http://img301.imageshack.us/my.php?image=3dm6926va6.jpg)

The cpu score is less than 100 points higher now with my new X2 4400+ clocked at 2730mhz or so, however it's made a huge difference to the overall score. i havnt changed any gfx settings since i last ran it and scored 6300 or so. just swapped out the 2.6ghz 3800+ with a 2.7ghz 4400+.

I also noticed a significant jump in my 3dmark06 score when I had an Amd platform. I was running a Venice 3500 overclocked to 2.7ghz with 512kb cache a while back and I scored about 6600. I then upgraded to an amd 3700 San Diego core with 1meg cache and overclocked it to 2.8ghz, the extra cache and 100mhz bagged me about 700 points. Im certain it was the extra cache in the 3700 that helps to power a dual card SLI/Crossfire system better than the 512kb cache Amd Cpus.

i'm not sure of the differences between the Toledo and Manchestor cores other than the extra 512Mb cache. but it would seem that the toledo's definately outperform the manchestors. Maybe it is just the bigger cache, as the cpu score increase isnt much to write home about

i was struggling to get the 4400 past 2.8ghz without a big voltage boost, not realising Cool'n Quiet was on. disabling seem so have made it much more stable at lower voltages so there maybe some more scores soon

Additional Comment:
http://img147.imageshack.us/img147/8994/7000sn3.th.jpg (http://img147.imageshack.us/my.php?image=7000sn3.jpg)
Made it to 7000. it just took a few more mhz and the ram tightened up to 2-3-2-5. i could push it further but the cpu temps are getting up towards 50C and i'd rather keep them down. there's no real need to push it too far so early. i'm pleased with 7000. :D

3DMark Score 2098 3DMarks
SM2.0 Score 1059
HDR/SM3.0 Score 709
CPU Score 652


Why am i so low! I have a 3.0 ghz pc and a 6800gt OC video card, i am still way to low, don't you guys think?

to be honest, your score sounds about right for the hardware you have. that 6800 will really struggle with the SM3/HDR tests. Even nvidia 7 series cards struggle with that test. 3Dmark06 is really tough and essentially is designed to put the very latest [2006's] hardware through the most rigorous test possible. if you're getting decent framerates in the games you play i wouldnt worry about your score so much.

Thanks, that does help me feel a little better.

I just always have this paranoid fear that i am missing a simple setting that will boost my computer to wear it really should be...

I had ran 3dmark basic when i first built my system, wasnt to happy with score but faced it and moved on. When putting my system together i rushed it cause i wanted to get it up and listen to it purr lol.

Well after a couple of months i got sick of the cables just hanging in there all over the place..plus i figured probably wasnt too good for airflow...so i pretty much tore the whole thing down and got the cables all neatly tied in the corners keeping main area "cable clear". looks very professional now :).

Well i noticed my temps didnt change a whole lot except for northbridge droped ~7C, and cpu droped about same. Got the urge to run 3dmark basic again and WOW!!! my scores jumped alot. the one that changed was SM3.0/HDR it was 1882 before the change and it went up to 3442..allmost doubled. I am clueless to what made them jump so high...i know i had power hooked up to both cards, sli was enabled...i dont know im just stumped.

If anyone might have some suggestions as to what did this change let me know.

thanks Crimson

Merged into existing 3DMark thread - don't create new threads for existing topics

Additional Comment:
Update: when i ran that test after i messed with the cables i ran it on stock frequencies for my video card, just to see how high i could get it i oc'd my vid card from 475 core to 522 and from 680 memory to 774 and my score jumped from ~6700 i had before i messed with cables (had same oc) to 8129.
I dont understand...it makes absolutely no sense how my scores jumped so high after messing with my cables..but hey im loving it btw i idle at 50 with those frequencies and maxed at 58 after running 3dmark 3x consecutively.

Crimson
p.s. sorry about that mixup Koroush

Not happy with a 6k break? I only get on 1.7k 3dmark06. I guess you can never please some people.

Update for 3dmark06:

11015 (http://service.futuremark.com/compare?3dm06=530881)

SM 2.0 Score 4511 Marks

SM 3.0 Score 4974 Marks

CPU Score 3143 Marks

Is this the average, "1 - Return to Proxycon 35.022 FPS?"

If so, wow.

yea, i believe that is an average. 11000 is an incredibly high score. Deeko's rig, with Raptors, would be my dream machine :cool:

Not happy with a 6k break? I only get on 1.7k 3dmark06. I guess you can never please some people.

I wasnt happy with my score becuase i knew my system could score better then that using the orb on futuremark's web site. But the whole point of my post was not about how i wasnt happy with my 6k break it was about why my score jumped from ~6k to ~8k after just messing with my wires. Speaking of which i did some more tests and it wasnt just a freak thing..i still get 8k and i dont know why. ill post more if i find out but anyone else is welcome if they have some suggestions :)

Crimson

This score makes me cry. :(

688 3DMarks

SM2.0 Score: 332
HDR/SM3.0 Score: 162
CPU Score: 779

[Edit] A newer score. (After following some more advice from the TGTC)
717 3DMarks

SM2.0 Score: 350
HDR/SM3.0 Score: 167
CPU Score: 768

Check this (http://www.nordichardware.se/forum/viewtopic.php?topic=148299&forum=1) guy out! It's in swedish but just watch the pic's. 8800gtx's in sli:eek:

This score makes me cry. :(

688 3DMarks

SM2.0 Score: 332
HDR/SM3.0 Score: 162
CPU Score: 779

[Edit] A newer score. (After following some more advice from the TGTC)
717 3DMarks

SM2.0 Score: 350
HDR/SM3.0 Score: 167
CPU Score: 768

My Hewlett-Packard Pavilion a819n, the last retail PC I'll probably ever buy, just finished a run of 3DMark06. Here's the score:

316 3DMarks :eek:

SM2.0 Score: 147
HDR/SM3.0 Score: N/A :D
CPU Score: 732

A few specs: ASUS "Goldfish" mATX motherboard; Intel 2.8GHz Pentium 4 (Prescott core); 2.0GHz DDR400 RAM; 200GB Maxtor SATA HDD; nVidia GeForce 6200 PCI video card (256MB).

Pretty sad machine, even with the addition of a PCI video card. :p

Crossed the 10k barrier with my gtx :)

http://img362.imageshack.us/img362/6597/3dpx2.th.jpg (http://img362.imageshack.us/my.php?image=3dpx2.jpg)

My specs got me somewhere in the 4000/5000 range, I'm not sure though..

I'd like to run a test now, but i got to get up early tomorrow..I'll post it tomorrow evening.

I've just scored a pretty nice 6,757; without trying to shut anything down. I bet with a little tweaking I could break 7K... maybe a little OC'ing ;)

Very nice Sharpy,
I was thinking of getting one of the 8800gtx's because I dont know if I can hold out long enough for the new ati cards.
You have a great looking build on your hands..:thumbup:

My score is so sad. :(

2384
SM2.0 Score: 1006
HDR/SM3.0 Score: 903
CPU Score: 947

I really dont understand why some of you guys think your score is so bad. Programs are updated to stress newer released hardware and you can only expect so much if you run the newer program. Hardware can only do what it was built to do. :thumbup:

Cheers Ron i need to do a bit more fiddling on my overclock tho cuz my ram speed is only equiv to that of pc3200 with this 3ghz overclock on the e6600.

Im a newbie when it comes to this and 3g is my highest im lookin on stock cooling, ran 8 hrs of orthos today with no temps over 60C so i will have to read up on getting the ram back up to speed

3893 3Dmarks...

isn't that a bit low for my specs? probably not..I don't really care though, as long as my games run smooth..:)

3893 3Dmarks...

isn't that a bit low for my specs? probably not..I don't really care though, as long as my games run smooth..:)
Single core processors get killed in this. Im kinda curious, what did you score for CPU? My processor is clocked same as yours and I got 947.

My CPU scored 925..Don't know why it's lower though..Maybe because I'm not much of an overclocker and I've got some cheap RAM.

Nothing wrong with your CPU score, I have a 3400+ (754) and it scores 972. But its clocked at almost 2.6ghz. 3d mark 06 is catered to Dual core CPU's more.

Additional Comment:
My current score:

Total 3D marks: 5278

SM2.0 Score: 2654

HDR/SM3 Score: 2669

CPU score: 972

Total 3D Marks : 5260

SM2.0 Score : 2622
HDR/SM3 Score : 2746
CPU score : 946

This was with 91.47 (custom one's bundled with the card i have) .. i expect a slight increase with 92.91 that i'm using now , but am too lazy to run the whole things over again .. but damm it is so partial to dual core CPU's .. i've seen similar config. PC's with dual core 4200+ at stock hitting around ~5800 easily ..

I installed the 6.12 cats and came up with this: 4345

SM 2.0=1929
HDR/SM 3.0=1982
CPU= 1066

Are the 79xxgt series Nvidia cards really that much faster than my x1800xt?

stock speeds

3DMark Score 2927 3DMarks

SM 2.0 Score 1306 Marks

SM 3.0 Score 1187 Marks

CPU Score 871 Marks

I installed the 6.12 cats and came up with this: 4345

SM 2.0=1929
HDR/SM 3.0=1982
CPU= 1066

Are the 79xxgt series Nvidia cards really that much faster than my x1800xt?

Not by that much in 'real world' gaming. In SLI the difference is more pronouced though.

Are the 79xxgt series Nvidia cards really that much faster than my x1800xt?

I don´t know why, but 3dmark06 seems to dislike the X1800 series of cards.

http://www23.tomshardware.com/graphics.html?modelx=33&model1=521&model2=551&chart=206

Using that link, we see that in most games, the X1800XT PE (the XTs clocked at 695/1.6 - like mine) are faster than the OC'd 7900GTs (didnt find a regular one, so i compared OC against OC). In games like Titan Quest and Oblivion (outdoor), the X1800XT is almost twice as fast (1280X1024 settings).

I think this is why you should not rely so much on 3dmark-scores when buying a new GPU. It´s not always the full truth.

I myself don´t use 3dmark to compare my systems performance with other computers, but instead on how diffrent drivers/system settings (or maybe OCs) on my own computer affect my system performance. But then again, even OCs that gives 4-500 points increases in 3dmark06 - give 0fps increases in e.g. FEAR/Oblivion.

My total 3Dmark 06' Scores
6291

SM 2.0 2513
SM 3.0 2817
CPU 1911

My score using the 93.71 TweaksRUs tweaked drivers:
10156

SM 2.0 Score-4477
SM 3.0 Score-4260
CPU Score-2907 :thumbup:

Link (http://service.futuremark.com/compare?3dm06=703211)

Main Test Results
3DMark Score 9563 3DMarks
SM 2.0 Score 4296 Marks
SM 3.0 Score 4584 Marks
CPU Score 2144 Marks

Windows XP heavily tweaked, all hardware stock clocks (for now).

Danny : I agree with you there bro. A system that is highly ocd and tweaked for benchmarks, will post good number yet usually isn't ideal for playing actual games. I've seen some really odd stuff in regards to sli and 3dmark.

3DMark Score: 9702
SM 2.0 Score: 4460
SM 3.0 Score: 4595
CPU Score: 2145

97.92 drivers; nothing special done with them - the nvidia control panel doesn't work correctly for me in windows x64.

Results page:

http://service.futuremark.com/orb/resultanalyzer.jsp?projectType=14&XLID=0&UID=8013237

New card : 8800GTX

3dmark06 : 11552

Sm2.0 : 5085

sm3.0 : 5162

Cpu : 2941

http://img111.imageshack.us/my.php?image=3dmark06on8800gtx11525sk9.jpg

Wow you guys have good scores!

I just ran 3dmark06 and here is my meager score:

3dmark06 : 2418

SM2.0 : 1014

SM3.0 : 928

CPU : 944

poopy.

Ran the benchmark with my HD2900 again (after all, i got the full version with this card :))

Nothing was overclocked, everything stock and driver settings on max IQ (without AF/AA/vsync forced);

no AA/AF: 8002.
8xAA/16xAF (forced from 3dmark06): 5248.

My X1800 TOP did 5004 without AA/AF so surely theres an improvement over my last card :)

10078 on default settings with the system in my sig.

Did everyone else get like 1 FPS on the CPU test?

Did everyone else get like 1 FPS on the CPU test?

Yes, I even got 0 FPS with my crappy Pentium 4.

I currently get a score of 9909 with the specs in my sig and running under Vista. I really want to break 10k so a little more overclocking/tweaking needed.

Have you tried raising the PCI-E reference clock in the BIOS yet? Most people reach their highest overclocks with a setting of 110-115 instead of the default setting of 100. Just don't go too far.

Here are my scores, The first is at stock settings - 540/1296/792
The second score is with the over clock at - 621/1458/902


http://img245.imageshack.us/img245/1449/3dmarkscore1stockcc8.th.jpg (http://img245.imageshack.us/my.php?image=3dmarkscore1stockcc8.jpg) http://img245.imageshack.us/img245/1539/3dmarkscore2jpgoc621anddx9.th.jpg (http://img245.imageshack.us/my.php?image=3dmarkscore2jpgoc621anddx9.jpg)

9053, not bad I think, but I atill need to do a bit more tweaking :)

Ran the benchmark with my HD2900 again (after all, i got
no AA/AF: 8002.
8xAA/16xAF (forced from 3dmark06): 5248.


Any ideas as to why i only get about 7500points? my specs are pretty much the same as yours. I cant get to the 8000 point mark even when i overclock my card. Oc´ed i get about 7700 points.
Are you running the test with the default 1280*1024?

3DMark06 build 1.1.0 default test at 3.72GHz: 3dmarks-11569/sm2-5206/sm3-5052/cpu-2978.

Default clock (2.93GHz) results are 11121/5158/5003/2609.

My 8800GTX is not overclocked.

just posting some results i have:
specs(below in my sig)
CPU at 3.2Ghz : 6301 on 158.22NGO drivers at 1280x1024 no aa/af
this is the highest i've reached so far but i haven't tested with the new 162.18 whql's. will report back soon with more results.

Ps: my gpu clock is normal as there are many manufacturers giving out pre-oc'd cards with warranty at around the same clock speed but i have a friend who's running twin 86GTS's volt modded and running on 900mhz core. now that is the ultimate overclock!

Any ideas as to why i only get about 7500points? my specs are pretty much the same as yours. I cant get to the 8000 point mark even when i overclock my card. Oc´ed i get about 7700 points.
Are you running the test with the default 1280*1024?

Yes 1280X1024. Are you on Vista 64bit aswell? The 64bit might help out on performance. I know e.g. in Stalker that my card had much better performance in Vista64 than in XP32.

Also my CPU has a 2mb L2 cache, wich yours doesnt. This might also affect the results a bit (but im not sure).

By the way, when i run my CPU overclocked to 2.475mhz (RAM at 225mhz 2.5-3-3-7-T1) i get a score closer to 9000 (no AA/AF). Do you get 7700 with CPU overclocked to 2.6?

13315

Was my score, judging by other ppls scores with a similar setup seems like my system is running as it should,

These new Cat drivers sure made a differance to my system hope ATI can keep it up with further releases.

Spec is listed in "My Current Machine Topic"

Thanks in advance

Thats a sweet score :thumbup:. Top setup.

Ok guys, see if you can help me out. I have been doing testing for a few days now on my system. I started at stock settings, and have moved up in 10mhz increments on my core clock speed of my GPU (8800gtx). I noticed an increase until i got to 650. At this point, my numbers started lowering or staying the same.

I figured this was because of a wall i was hitting so i accepted that. I lowered the mhz and retested, and again, found my score to be lower.

I have a few questions, why would my numbers be lowering at all if i was raising the clock speed, is it because i was not raising the memory speed (which i was planning to do later with stock core speeds)?

Lastly, and this is my biggest issue, what would cause my CPU score to lower continually? I have rebooted in between tests, and run in the same enviroment each time, yet my CPU score gets lower as my Core speed gets higher.

Forgot to update my score here. Updated score : 13333

http://service.futuremark.com/compare?3dm06=2606056

That run was on XP, Ive yet to install and try 3dmark06 on Vista x64, might just wait until 3dmark07 :)

As said in my current machine topic got a extra 400mhz out of my E6600. downcloked my RAM slighty and got a new score of

14847

Thanks in advance

Can you post up your free valid Orb link Whiteowl so I can compare my score to yours with the breakdown, cheers :)