Printable View
Borrowed these two images from here. Same settings under the single frame 3dmark06 rendering options. Jpeg compression makes them more than a bit limited, but thought I'd put them up anyway. NV drivers 158.16
Both: 3DMark06 1280x1024 8xAA 16xAF
2900XT left. 8800GTX right.
http://img120.imageshack.us/img120/271/xt1ae8.th.jpghttp://img297.imageshack.us/img297/3573/gtx1xv2.th.jpg
http://img153.imageshack.us/img153/2991/xt2xw4.th.jpghttp://img120.imageshack.us/img120/1743/gtx2br1.th.jpg
Well both cards have there positives in that test. I see vary little difference between them just taking a glance at it.
i think the ones on the right have more prominent jaggies
I see no difference, only the light
Yes for OD in CCC, But like I said b4 2 x 6 pin pci-e connections per card will do just fine.Quote:
Originally Posted by mascaras
I been using this since day 1 and have not had any conflicts or issues
to mention. :)
I think the most guys will have a 2 x 6-pin connectors coming from their PSU, so the ideal adapter would be something that takes a single 6-pin and a molex to make an 8-pin... is that the case?Quote:
Originally Posted by fornowagain
Thanks,
Suman
so if 2x6 runs jsut fine even when oced .. why would they include the 8p .,...
yeah and for the adapter i dont want 2x6p .. its just a big waste and stupid .. cuz if ur PSU dont have 8pin u will need a 3x6pin jsut 1to power 1 2900Xt ...
so molex+6p or 2xmolex > 8pin wil certainly be nice
fornowagain, pictures you posted have different jpeg file size
radeon pictures are 3 times smaller
until we don't see bmp source it means nothing
This is completely bogus. If you flip them back and forth, you will see the ATI pictures were slaughtered with heavy JPEG compression. Check the file sizes - the Nvidia pictures are much larger. Someone obviously has an agenda, and I hope no one goes and takes that "comparision" seriously.Quote:
Originally Posted by fornowagain
Yup, see above post. Someone's busted.Quote:
Originally Posted by MAS
:nono:
i don't think anyone tried to fool us, the results were uploaded by two different persons for comparison, unless that becco dude and fornowagain is one and the same.Quote:
Originally Posted by Soulburner
you would've realized that if you visited the thread in the link
Why isn't it a bmp, that's way more usefull when comparing these kind of things.
Arrr.....this isn't looking good, let's hope it's just the game and not the G80 architecture, if R600 wins in this by a considerable margin.....hmm. :(
(Schatten hoch and mittle, is High quality shaders, and medium quality shaders iirc)
DX9 performance is quite bad too, so I guess it's the game that is just another Xbox360 rubbish port.
what does that have to do with the hd 2900??
You have got to be kidding me right? This is the first DX10 game benchmark, DX10 is where the battle is fought, and if R600 does good in this game, and G80 doesn't, then G80 might not be that good in DX10 at all.Quote:
Originally Posted by jimmyz
This is a Xbox360 port, so I wouldn't be suprised at all if this is just a rubbish port. I mean a GTX can run Crysis beautifully in DX10, and this game is just sluggish. Hm, doesn't make any sense.
the graphs you posted don't show the 2900. dx9 is where this battle is being fought like it or not, there are no dx10 games out and won't be for a few months. by the time dx10 has enough games to make it important to switch cards all the cards mentioned will have been passed up by the r650 and 8900. that said i am getting mine monday but will be replacing it when r650 is available.
also elsewhere in this forum is tests with the microsoft dx10 sdk and the radeon beats the g80.
Not true! Call Of Juarez is the first DX10 demo available to the testers of HD2900XT. Hopefully it'll be available to the general public soon!Quote:
Originally Posted by Tim
Do you live in the past or in the present? :p:Quote:
Originally Posted by jimmyz
What I see is the cards were created to handle DX9 but mainly for DX10. So DX10 benchmarks is all that I want to see.
First of all, there are only 64 shader units on the x2900 (not 320) and only 128 shader units on the g80. When they talk about stream processors, thats just trying to mislead. Each nvidia shader can do one MAD only and a MUL. Each ati shader can do 5 mad operations per cycle. Nvidia's MUL cannot be compared to ATI's full MAD, hence why nvidia only has 128 stream processors, unless you speak of MULs only, then you can say it can do 256 simultanously (if they are exclusive).
So if you take the speed differences and the efficiency of ATIs scheduling (which should get better), with a few more tweaks it is clear nvidia will need much more than clock speeds to keep up.
And it doesn't need to be complex. Any vec4 mad will do.
And the other thing of course is, the x2900 has been a dx10 solution from the ground up. Nvidia has only just begun decoupling the shaders.
Quote:
Originally Posted by DilTech
show me the games that are out for dx10.... how is that the past? i also pointed out there are results showing the 2900's dx10 performance already posted, why bother to show a chart with only the g80 line-up ? that was my point. and for what it's worth i am getting a hd 2900 xt crossfire set-up as soon as they are at newegg or microcenter.Quote:
Originally Posted by Metroid
The ATI clearly has the lower texture quality, which I is likely due to JPEG compression. But the ATI has better anti-jaggies support. Now, this is rather subjective overall, because the colors are rendered different with JPEG thus impacting the jaggies. Yet there is one point in the screenshots were there is a difference, which I believe is non-JPEG related. This can be seen on the Canyon Flight screenshot, by looking at the front of the balloon. On the ATI screenshots the spike is larger than on the nVidia screenshots. We need to start taking image quality into account when looking at the benchmarks numbers we are seeing.Quote:
Originally Posted by Soulburner
I was going to pick up a 8800gts screw that :P Im going with this for sure , tomorrow . Hopefully Newegg has some free shipping . :P
No Schatten is Shadow not Shaders!Theres a performance problem with high Quality Softshadows with AA that has nothing to do with DX10.Quote:
Originally Posted by Tim
Well filtering optimizations are enabled on nvidia by default if HQ filtering is not selected. So running default is actually a little advantagous to nvidia. Best thing to do is disable AI on ATI, enable HQ AF on both and be done.
Not really, they are both just as fast as each other in DX9 (the GTS and XT) but they are both MADE for DX10, so we do need DX10 benchmarks, that's why I posted that graph, because it's the first DX10 benchmark that is going to be available to the public. So ofcourse it's important. Both cards will do really well in DX9, that's it. But with UT3, Alan Wake, Crysis, Bioshock and others, we are not going to care about that, two of them will be out before R650 I can guarantee that. I'm not going to upgrade again, not for 1,5-2 years, so ofcourse it's very important to ME what these cards will do. :)Quote:
Originally Posted by jimmyz
It's like pitting an Enzo vs a Zonda on a country road, yes you will get a good idea which one is faster, but put them on a racetrack and you'll really see what the difference is.
Thus it's true, Lost Planet will be available on Tuesday to the public before COJ DX10, hence the first DX10 Demo. ;)Quote:
Originally Posted by Nedjo
Hahaha :lol2: :fact:
Radeon HD 2900XT user guide revealed
http://www.fudzilla.com/index.php?op...d=956&Itemid=1