Post your 3dmark05 results

[Viper96720]

Got an awesome 1659

 

[user1234]

Argument: "I have dual raptors in raid-0 and my sequential read speed is 120 GB/s so my system is better than yours !!!! haha"
Fact: "Your Celeron/R7500/256MB system will never be better than my P4-EE/6800U/1GB system, no matter what storage subsystem you use. pwned"

Got it ?

 

[HumblePie]

user1234 pwn this...

What do the majority of people on these forums or that use benchmarks do? They UPGRADE. They don't buy WHOLE systems at once usually. They buy the PARTS they NEED for what they want. Hell, that's why when you look at the left hand side of your screen you see tons of sub categories. Video, sound, hard drives, blah blah blah. If I'm trying to create a entire system from scratch OR trying to upgrade a single component I would RATHER use a test that specifically targets THAT ITEM for performance. The more specific the test the easier it is to evaluate exactly what I want.

Think of it this way. If I'm into weight lifting and I'm looking to make my BICEPS bigger amd I going to get a rowing machine or a dumbell? They both work the bicep right? Which will target specifically the bicep though and get the results that I want if all I want to up is my bicep? DUH the dumbell you dumbell.

If I'm looking to get the BEST video card out on the market and I am comfortable with the rest of my system for now, do you think using 31 mark 01 or 3d mark 05 would be the better benchmark to use? DUH 2005. Why? because 2001 is influenced more by total system performance then graphic card performance while 2005 is the opposite. Too many other benchmarks are the same way.

Truth be told, if I had a specific GAME in mind, and who the hell builds a computer system just to play one game only and never anything else, I would use that game benchmark as well.

If I'm building a system from scratch I would like to have a specific benchmark that tests and stresses each individual part of a computer seperately and tries to minimize any outside variables that would cause flunctuations in performance. I DO NOT use graphical benchmarks to stress test my CPU. That's what Sandra, CPUmark, prime95 and others are for. Same thing with memory.

Here's the trick. If you buy all the best SUB components, the whole is going to be greater then the sum. You can only be as strong as your weakest link. Get it? This is why this test does EXACTLY what many GAMERS have been asking for. Well almost exactly. The other part of the test they ould have added was an OpenGL portion to the test. Beyond that, I expect graphical benchmarks to stress graphic cards and NOT WHOLE SYSTEMS.

And for your example of a raptor... look I can take 4 raptors or 4 high performance SCSI drives on a cheap a$$ CPU/MOBO/Memory system and still do things like Video Toaster. Something your P4 10,000 Ghz, 500 GB of memory or whatever asanine system you want to use will NEVER be able to do. Why? Because video toaster is an application that REQUIRES a certain transfer rate for your hard drive configuration. If your hard drives suck you can't do video toaster. Got it? This means that celeron with 4 raptors or scsi's will be BETTER then your system using video toaster.

Same thing with this example. With a crappy system and the right video card I can play some games better then someone else with a great system but a crappy video card. The vice versa holds true.

So in your own words.. [pwned]

 

[jiffylube1024]

Originally posted by: user1234
Argument: "I have dual raptors in raid-0 and my sequential read speed is 120 GB/s so my system is better than yours !!!! haha"
Fact: "Your Celeron/R7500/256MB system will never be better than my P4-EE/6800U/1GB system, no matter what storage subsystem you use. pwned"

Got it ?

Well actually, the dual Raptors will be better than the P4-EE/6800U/1GB at file copying and other things which depend heavily on HD sequential read speeds, which is exactly what your first theoretical argument presents.

Similarly, 3dmark 05 gives you an indication of pure PS 2.0+ shader performance, which is an estimation of game performance in the future. By all current trends, over the next couple of years, games will rely much more heavily on PS performance, so knowing which cards have strong PS performance is extremely relevant to future-minded consumers.


Nobody is saying it is the be-all, end-all test for games, because as we all know real-life games are CPU limited as well, not to mention faster GPU's require more and more powerful CPU's to keep them fed with data.

As a matter of fact, a test which isolates the GPU entirely like 3dmark 2k5 I would argue is more relevant than one which depends on the CPU as well, because this way you can really focus on which GPU is more powerful, and then pick accordingly when you build a new system. You don't have to worry that in this test they used a crap CPU and in another one they used a bleeding-edge CPU and got different results. The GPU's just line up in terms of performance in ultra-heavy PS 2.0+ usage.

Don't bring in stupid analogies like dumbass comparing a Celeron/R7500/256MB RAM system with dual Raptors (which, again, probably would rock as a fileserver too ) to a "top of the line" system. It reduces the credibilty of your argument to nill, as does the use of the word "pwned" at the end.

 

[user1234]

I agree that to choose components for an upgrade you need to look at component specific benchmarks like this one. But most people don't benchmark component themselves before choosing the right one, they just read reviews and look at comparative benchmark articles in professional hardware sites. In the end, after people choose their components, they care about the real world performance they will get in their favorite apps and games, so benchmarks which give a good estimate of overall real world performance are the most useful in comparing systems with many different components. This is also what people want to see after performing an upgrade to evaluate how effective tthe upgrade was in the context of their overall system. But 3dcrap is not such a benchmark, it's only a component benchmark masquarading itself as a gaming benchmark.

You need to understand that my point is that 3dm5 should not be used as a predicor of overall gaming performance, but many people still believe that, unfortunately for them.

Bottom line (admitted even by 3dm5 fanboys): 3DMark05 is not a gaming benbhmark

[3dm5 fanboys pwned]

 

[iwantanewcomputer]

2424 -> 2678 once i overclock my 9800 pro from 378/338 to 430/380...now time to overclock the cpu

 

[Travis6586]

I got 4,292. Not too shabby I guess. But all the tests were really choppy. Gonna be amazing when there's a video card out there that can run those tests like it ain't nothin.

http://service.futuremark.com/compare?3dm05=48929

 

[HumblePie]

user1234.. what are you smoking? I don't read reviews and look at graphs and benchmarks when I'm looking to buy something for "overall" performance. There's WAY to many variables to make a decision based off general and overall benchamarks. Sorry, when I upgrade or buy something I want to know exactly how that single peice of hard ware performs independantly of anything else. The good reviews and benchmarks do this, the bad ones don't.


Take this case in point. A few months ago, there was an review with benchmarks on anandtech comparing an A64 processor against an Intel processor in GAMING and REAL WORLD performance. This was considered the WORST review ever posted on anandtech. Why? because the guy was screwed up majorly. He used the BEST P4/motherboard combo and used only the best A64 CPU and a medicore motherboard as far as performance. To top it off. He used the BEST memory available for the P4 setup with awsome 2-2-2 timings. For the A64 it was like 3-4-4 timings. It was a horrible comparison. Every test showed the p4 barely beating or tieing, but never losing to the a64. We all know that this is completely FALSE. They also used really horrible benchmarks like Winstone and content creation for "real world" performance. BLAH!

If you want to show a real comparison between an A64 and a pentium you make everyting as equal as you can get it.Get the best a64 and the best p4. Get the best performing motherboard available for each cpu by TESTING THEM ALL. Use the exact same memroy. Same everything else. If anything is slightly different it will throw off msot tests. Why? because MOST tests are too damn dependant on too many variables. The you use test that stress the CPU only as much as it can without stressing the rest of the system. This means stuff like using Pi and what not. Don't use winstone.

Do you get it? 3D mark tests the graphics card based almost SOLELY of Dx9 and PS2+ shaders. If a card can't don't those then it will look horrible in that benchmark. Which is what all the 5xxx series nvidia cards suck at. If I want to play games that I know use DX9 and tons of PS2+ shaders... *cough half life 2*cough* then I would love to have a benchmark that stresses those.

 

[jim1976]

For what it's worth .... 5349

BTW I saw the highest score with a 6800U was somewhere near 6850. I looked at his details and said that ran it with SM2.0?? Typo??

 

[FuFighterStan]

I got 5160 or so... Athlon XP @ 2.33 ghz with my 6800 GT at 400/1100 using the 66.70 drivers. My friend had similar large increase with those drivers, going from about 4500 on 61.77 to 5300 using the 66.70.

 

[gururu]

Originally posted by: FuFighterStan
I got 5160 or so... Athlon XP @ 2.33 ghz with my 6800 GT at 400/1100 using the 66.70 drivers. My friend had similar large increase with those drivers, going from about 4500 on 61.77 to 5300 using the 66.70.

over 5000 seems pretty darned good. is the bench pretty smooth with that kind of score? half the bench is played at 1fps on my system. even the graphics seem kind of bland/bleached. pics from other sites show vibrant shiny images.

 

[CaiNaM]

Originally posted by: DAPUNISHER
CaiNaM,

There's good logic in what your stating, I'm just disenchanted and "acting out" Having hit a hotdeal of $299.99 for a BFG GT without the assistance of 3DM in the evaluation and decision process to purchase, leaves me convinced that I do not need their bench for any reason whatsoever other than to see the impressive eye candy. Yes, it's a preview of things to come when it first shows up, but for me that is where it's value ends. Particularly since by the time a number of games *especially good ones* that load the card as heavy as '05 does, hit the shelves, I'll likely need a new card anyways. Thus I've concluded most of it's benefits are dubious at best, at least in my case I will concede it has its uses, but I still feel it's not deserving of the significance many attach to it.

i will certainly agree with you there

honestly, when d3 came out and nv40 rules that, i still said both the ati and nv products are good.. with 3md showing and advantage towards ati, i still believe both products are good.

regardless of whether you're winning or losing at whatever benchmark, it's obvious these card perform at a high lvl n today's game regardless of whose flag you're waving

 

[jim1976]

Originally posted by: gururu

Originally posted by: FuFighterStan
I got 5160 or so... Athlon XP @ 2.33 ghz with my 6800 GT at 400/1100 using the 66.70 drivers. My friend had similar large increase with those drivers, going from about 4500 on 61.77 to 5300 using the 66.70.

over 5000 seems pretty darned good. is the bench pretty smooth with that kind of score? half the bench is played at 1fps on my system. even the graphics seem kind of bland/bleached. pics from other sites show vibrant shiny images.

Well all you have to do is check the framerates. I cannot say it's smooth. Just smoother than previous gen cards I guess. This is a bench made for next gen gpus in order to ran flawlesly. Pics are shiny and vibrant in my system. Nowhere near bland/bleached.

 

[user1234]

3DMark05 score = 2335

Radeon 9700 Pro @ 350/325 MHz (core/mem)
Athlon XP @ 11x210.5MHz (FSB) = 2315 MHz
2x256MB (dual channel) PC3500 3-5-5-9 @ 2x210MHz = 420 MHz DDR
Shuttle AN35N Ultra Mobo (nForce2 400 Ultra chipset)

BTW, this system runs Doom 3 pretty well on medium quality @1024x768.
Also, I use the Audigy 2 ZS which increases performance by about 10% in many games because of it's hardware sound accelaration, but this stupid benchmark doesn't capture that like many other things it ignores.

 

[gururu]

Originally posted by: user1234
3DMark05 score = 2335

Radeon 9700 Pro @ 350/325 MHz (core/mem)
Athlon XP @ 11x210.5MHz (FSB) = 2315 MHz
2x256MB (dual channel) PC3500 3-5-5-9 @ 2x210MHz = 420 MHz DDR
Shuttle AN35N Ultra Mobo (nForce2 400 Ultra chipset)

BTW, this system runs Doom 3 pretty well on medium quality @1024x768.

I even pwned some 9800 pro/xt systems...

thats a great score for a pro. I got around 2190 with my 9700pro @ 340/315.

 

[user1234]

thanks. I used the latest official drivers off ATI's website 6.14.10.6476, and running Windows XP Pro w/SP2. That's about it (btw, this is just my lan party/htpc rig). What's the specs for your system ?

btw, I saw some guy that posted here with a 9700 pro running at close to 400MHz core. I wonder if mine could do that, it seems an awful lot above stock. I will try it sometime. Last time I tried to find the max core I stopped at 350 and have no problems.

 

[Zebo]

GT Sucks Wildly Overclcoked 448/1200 maybe it's my mobile....A64 down?? maybe culprit...

Anyway I'm really having a case of "buyers regret" when I see these X800pro flashed to x800XTPE getting 1500 more pts for the same damn price.

 

[HumblePie]

Also, I use the Audigy 2 ZS which increases performance by about 10% in many games because of it's hardware sound accelaration, but this stupid benchmark doesn't capture that like many other things it ignores.

AGAIN your perponderous of ignorance astounds me user1234....

Having a sound card only speeds up SOUND DECODING. Notice in the bench there is NO SOUND. They do this because they want to show you the performance of your GRAPHICS CARD and not your sound card. If you want a benchmark for sound go find something else and not 3D mark you doofus. The audigy doesn't "magically" give you more frames per second from some never-never land. If you have no sound in the benchmark it doesn't matter how fast your sound card is.

Use the right tool for the RIGHT job. Grr....

 

[lordtyranus]

Originally posted by: Zebo
GT Sucks Wildly Overclcoked 448/1200 maybe it's my mobile....A64 down?? maybe culprit...

Anyway I'm really having a case of "buyers regret" when I see these X800pro flashed to x800XTPE getting 1500 more pts for the same damn price.

The athlon 64's are performing very poorly. My Barton XP and 412/1140 GT got 5333 pts.

Anyway, the 16 pipe x800s really pulled away bigtime. even without AA/AF and 3dc used. Nvidia had their proprietary DST shadowing used as well; remove that and they drop 500 pts.

 

[Zebo]

6670 Driver update helped almost 5800

 

[Uhtrinity]

6318

 

[Zebo]

Originally posted by: apoppin

Originally posted by: DAPUNISHER
Weird that the URL died so fast, I captured it though so here is the project the guy posted text

Can't read it.

Do you think it is for "real"? Techreport only got 5771 with the same GPU (and 3DMark probably IS gpu dependent)
:roll:

We know nVidia's SLI score is verifyable.

edit: Doom iii is the WORST (major-release) game i can remember (for YEARS)

Have you heard of overclocking? Or is that not fair? About as fair as using $1200 in cards against $600

 

[user1234]

Originally posted by: HumblePie

Also, I use the Audigy 2 ZS which increases performance by about 10% in many games because of it's hardware sound accelaration, but this stupid benchmark doesn't capture that like many other things it ignores.

AGAIN your perponderous of ignorance astounds me user1234....

Having a sound card only speeds up SOUND DECODING. Notice in the bench there is NO SOUND. They do this because they want to show you the performance of your GRAPHICS CARD and not your sound card. If you want a benchmark for sound go find something else and not 3D mark you doofus. The audigy doesn't "magically" give you more frames per second from some never-never land. If you have no sound in the benchmark it doesn't matter how fast your sound card is.

Use the right tool for the RIGHT job. Grr....

hmm.... what shall we do with you ? you just can't seem to understand what I'm saying, can you ? The comment about the sound card is just one more illustration that 3dm5 doesn't correlate to actual gaming performance, because real world gaming performance depends on your sound card hardware. So what is 3dm5 good for really in the end ? Iguess, it's good for comparing two or more video cards when all other components are the exactly same. But how many people test multiple video cards ? Most people read comparative reviews of video cards before purchasing one. And what's the point of running this benchmark on your system after you upgrade, and comparing it with some other schmoe somewhere on the internet ? What conclusion can you draw from the fact that your score is higher (or lower) by x percent ? The other guy's system may very well play most game faster than yours even if your score is higher, and in any case the percentage difference doesn't correlate to real world gaming performance. Ergo, this is quite a useless benchmark. And mostly painful to watch too.

For example - X800XT get 20% higher score than 6800U at stock speeds. How many games do you know where X800XT produces 20% faster frame rate than the 6800U on the exact same machine ? I know of many games where 6800U is faster, and sometimes 25% or more faster. So again, what's the relevance of 3dm5 score here ?

[pwned] "if you had half my intelligence, you'd be the king of your village"

 

[Zebo]

Originally posted by: Uhtrinity
6318

Nice setup. I love to see value systems hanging with the big boys. Well, relative value

What mem are you using and what card was able to flash?Are you Oced?

I am so getting a x800pro vivo. Between the new benchmarks with new cat drivers showing total domination, not just this bench either, Nvidia driver WHQL failures and cheats and inferior AA/AF it really looks like ati is the card to buy.

 

[user1234]

Originally posted by: Zebo

I am so getting a x800pro vivo. Between the new benchmarks with new cat drivers showing total domination, not just this bench either, Nvidia driver WHQL failures and cheats and inferior AA/AF it really looks like ati is the card to buy.

BS-crap, and especially surprising to hear from a heavy Nvidia user like yourself. Nvidia is dominating in sales in all the medium and high market segments, and its 6x00 cards are recommended over ATI by just about every computer magazine and website. And this 3dcrack rigged horrid app can't even put a small dent in that. Of course driver optimizations can change things, but I we will have to wait for real benchmarks with the new drivers to see if there is a significant change. At least for Nvidia, the new beta driver does improve scores a little even in more reliable benchmarks (I noticed about 1% improvement).

Maybe you're blinded by the excitement of getting a new card, but please let's stick to the facts - it is currently widely accepted that in the new generation of video cards, nVidia GPUs are on top. I'm sure you read some of these reviews. Sometime, if you try to jump the gun, you just get shot.