Printable View
After looking at this thread, I suppose you have to be a programmer to be excited about 3DMark06, lol, because only a programmer would see and understand the differences and what's going on. I don't know, but there just seems to be a lot of more negative people at XtremeSystems then the other forums I frequent. Unfortunitly, it's not justified in this case.
Maybe if you hate the benchmark, you should just stop benchmarking. Better yet, get a compiler and make your own.
Who cares if the result is .5 fps or 500 fps. It is a comparison test to see what relative gains in peformance are possible. The fact that they finally recognize the CPU as a key factor in the gaming benchmark is a step in the right direction. And, hampering cards with no SM 3.0 support only makes sense. Why give a better score to a card that does not support current feature sets? I think 06 will reflect relative gaming performance better than 05.
I've been touting 3DMWank as just that, a wank. What's a wank?, a wank's not the real thing, and that's what 3DM is. It's just a wank for an e-pene.Quote:
Originally Posted by Cybercat
Seriously though 3Dm is sooooo unrepresentative of real world in game preformance I'm surprised anyone takes the pos seriously. Especially as a valid indicator regarding how graphics cards/cpu's should preform in current and up coming games.
Why not call it 3DM05 rehashed to 06.
Why not compile a benchmark program from the source gode of various engines such as D3,HL2 etc.
some more detailed benching
http://review.gzeasy.com/graphics/20...cs_4657_6.html
don't get me wrong, I've always liked the 3DMark for at least one reason or another. '05 was a great middle-ground test that did a good job indicating shader and vertex performance, before it was discovered to be CPU-limited. I call it middle-ground, because unlike so many game engines out there it didn't swing towards one company or another.Quote:
Originally Posted by DudeMiester
'06 just seems like a minor upgrade, and simply below the standards of Futuremark, who has always wowed us with brand new tests that offered quality we've never seen before. Plus, it really doesn't seem like the performance justifies those improvements.
um, where is the proof? Btw, I don't think any of it looks much better than hl2 lost coast which I run at 1024x768 2x aa 4x af with a oc'ed x800p and I dont drop below ~30fps. What I hate is that futuremark has always reffered to 3dmark as "The Gamer's Benchmark" or something to that effect. HahQuote:
Originally Posted by RAMMAN
just read the old threads about it in the news section.
I wonder what is going on with the photos that we got of the dock city next to the water. My theory is that this was just a minor upgrade. I have reason to believe they are putting the enphasis on 3dmark07 which will be dx10 and realeased later this year.
Well imho, if you want real world game performance, then bench with games! lol! I see no reason why 3DMark06 should be indicitive of gaming at all, despite what they may advertise. Personally, my only concern is if it looks good visually.
Anyways, if you look closely you will see that 3DMark06 has many more polygons then HL2, resulting in more fine detail and smoothness. Also, when you see it in motion, you will realise the dynamic nature of the lighting and HDR, which in HL2 is largly static. Also, HL2 doesn't have advanced blooms (like the star bloom seen in the images), and not only that, it has a narrower dynamic range then the floating point format that 06 uses. Valve used an integer based format for speed, and because you can get MSAA with it on older cards. Also, Valve only uses HDR materials on the most visible surfaces, whereas 3DMark06 will use it on everything. Again, if you look at the sea-monster, that reddish glow will probably be dynamic and flowing (it's probably a procedural effect), something you would never see in HL2. Also, HL2 has much less detail in water ripples, nor did it have deformable water surfaces, which it appears that 3DMark06 does have (see the large waves in the water on which smaller details are overlaid). All in all, 3DMark06 has many more small details and subtle/atmospheric effects, however you must look closely to see them!
I hope you take a second glance, and compare them closely side by side.
freecableguy should bench his dual 275 system.
Yes. I want to see the 275 system + two 512 GTX all over this.Quote:
Originally Posted by STEvil
To The Top!
Hold on a minute guys, don't forget what this part of the forum is called .. it's not called XtremeSystems Graphical Benchmark as a Representation of Games Appreciation Team :D
3dmark is provided by Futuremark for us to gain points,be the best and then taunt some poor sucker who is 1 point less :p:
Seriously though I think we need to get hands on before passing judgement. 03 got a right bashing when it came out but now I quite like the test because it is both gpu and video memory bandwidth limited in different tests.
I'm glad the cpu gets part of the score now, I hope they have the weighting right though.
Regards
Andy
How many fps gets a 7800GT? I'm asking because none of the ORB links work for me, anyone can see them? Thanks
//EDIT Quoting gamespot: "3DMark06 is available as a free Basic Edition". But it still forces you to subscribe to gamespot complete. Is there any chance you can download it without paying to gamespot?
Well 3dmark has always got the CPU score wrong. Intel has ALWAYS had a higher cpu score, when the intel scores would always be considerably lower than the AMD despite the higher cpu score. If Futuremark keeps with this tradition, we can all ignore 06 until they fix it.
I wonder how a Dothan does....probably not so well.
Anyway, I find this very intriguing and LIKE the direction it's going in.
EDIT: The one thing I fear is what it'll be 1.5+ years from now, like how 3DMark05 is more dependent on CPU OCing than GPU OCing at the highend (I *don't* like that!), will this become a psuedo-PCMark?
EDIT2: also, let me at that playable game, that looks cool.
EDIT3: seems the installer is over 500MB....heh, released at 9am, installed at 10am, that's not fair :D
Cant w8 to start downloading it :)
I just can`t see the charts of `06. Futuremark says the link is not valid :confused:
This is retarded.
Why are the CPU tests just CPU, they don't even use the GPU. It should be like Aquamark 3, where you get a CPU score, not a CPU benchmark, at 1fps on the best CPU today. Stupid.
It's pretty obvious it's nothing like game performance, and now we can definatly say this for sure, to stop those idiots thinking that it is.
Also seems it's very Nvidia biased. A 7800GTX easily beats a X1800XT. Wonder how much impact overclocked GPUs will have.
They pulled them down....they want an empty ORB when the users get going (yes, they only finished up their ORB testing at 1am EST last night--at around midnight the scores in the *=60s weren't there but at 1am they were ;))
Not really. When it comes down to that type of software rendering it's all small, simple calculations, so sheer clockspeed wins out. I don't think it really says anything about the processing power of your cpu since that type of processing is pretty antiquated.Quote:
Originally Posted by MaxxxRacer
It's out guys :D
http://www.futuremark.com/products/3dmark06/
Well....it seems that futuremark bad habits didn't change.
Intel CPU-s are still faster...
and Dual-core will take over single one (but this is realistic...:)
And SM 2.0 cards scores will be lower than 3.0 capable cards...
And we all know that ATI X1... serie card don't have the same 3.0 shader implementation than Nvidia.....
That's why we see some 256 GTX ahead of X1800 XT....or 7800 GT ahead of X1800 XL....
Then, we should do all the test @ 1280x1024 but with AA and AF maxxed out.
There, ATI would be ahead and the only one to run both HDR AND AA..(or af??:confused: ...don't have the money to buy one of those anyway..)
So, have we all lose our mind?
Futuremark is like a movie...One every year...3D Rendering at its best....period....
And even my X800 GTO² is happy to run the '05 (full-version, of course...;)
Freidns and family like this show.....expensive i should say...25 USD $...
We know how Futuremark cheats with score....why bother finding a reason.....:(
I probably won't be able to run all the new features with a SM 2.0 card but i'll would be happy to see the full-version.....just like that.....
Unless you have 3000 USD $ to throw in your rig.....
Just run '06 for pleasure.....
It's not a test.....neither is it THE test......caus' its too much oriented ;)
It's just a good work......period....but they could do better.....
Anyone wants to compete with futuremark.....
P.S. :my Old-Ma always told me to do things myself when done unproperly....see what i mean.....
LOL, I forgot to mention....this takes up over 1.25GB of HDD space ;)
yea WTF! 576MB download from majorgeeks. The schools network is slow as hell today, or majorgeeks is, cuz i'm only getting 115KB/s, I got 3.4MB/s yesterday downloading premiere from the adobe servers.Quote:
Originally Posted by Vapor
woah really ?Quote:
Originally Posted by Vapor