oh SH*T, Photos of NV35 and prelim benchies!!!!!

http://www.x86-secret.com/

fx5900-1.jpg


http://www.tt-hardware.com/article.php?sid=4316

SHOCK AND AWE!!!!!
 
Impressive score, but a 460/400 9900 Pro should be able to about match it. We'll see if ATi can keep their slim design with such high speeds, or if they'll have to add a two-tier cooler a la nVidia.

Nice to see the market heating up, though. Affordable 9700's all around! :)
 
Only 4 capacitors? Hmm... looking at the pictures of the 5800ultra which has 18 capacitors... things don't seem to add up.

What do you all think?
 
I showed this to Hilbert over at the Guru3D, he said "This should not be out there, the NDA'a are still up and running"...are websites alowed to even acknowledge this piece of info?
 
I'm curious: is there any particular reason as to why NVidia video cards' PCBs are so much huger than ATI cards'?
 
Ostsol said:
I'm curious: is there any particular reason as to why NVidia video cards' PCBs are so much huger than ATI cards'?

I can't find the thread (or remember which board it was on) but if I remember right it's due to nvidia designing in more "tolerance". That's not the word I'm looking for, but it's been a while since I read that thread so my memory's not doing to well, lol.
 
micron said:
I showed this to Hilbert over at the Guru3D, he said "This should not be out there, the NDA'a are still up and running"...are websites alowed to even acknowledge this piece of info?
Well, it's seems that the bench doesn't comes from TT-H but from a friend of the webmaster. Impressive the +20% with the det 50. without the issues of the NV30.
 
http://www.tweaker.ch/3dmark03/index.php

minos8 6420 ATI Radeon 9800Pro 452/742 Intel P4 2.66GHz/533 3043 151
2
Raul 6149 ATI Radeon 9700Pro 410/696 Intel P4 3.06GHz/533 3564 155
3
Hooligan 6133 ATI Radeon 9700Pro 425/742 Intel P4 3.06GHz/533 3564 155
4
crotalb 5964 ATI Radeon 9700Pro 430/724 Intel P4 2.80GHz/533 3370 160
5

I dunno how NV35 will perform on games, how is the new(?) AA, and the UltraShadow marchitecture, but the card seems to be what we expected, a 256bit bus/less noisy/bugfixed NV30 core, nothing more, blah.

Raddy 9700 here i come.
 
I've sent NVIDIA some screenshots of some, errr, odd things that appear to be happening in their later drivers in 3DM03.

As part of the beta team we have access to to beta builds that allow extra functionality for debugging, that non-beta members (i.e NV) don't. When using this mode numerous rendering 'errors' occur with the later drivers. Some are there in the 43.45's, more are there in the 43.51's. These same issues are not present on ATI boards.
 
DaveBaumann said:
I've sent NVIDIA some screenshots of some, errr, odd things that appear to be happening in their later drivers in 3DM03.

As part of the beta team we have access to to beta builds that allow extra functionality for debugging, that non-beta members (i.e NV) don't. When using this mode numerous rendering 'errors' occur with the later drivers. Some are there in the 43.45's, more are there in the 43.51's. These same issues are not present on ATI boards.
Its the applications fault.
nVidia drivers have no errors.
 
So that score is probably still using the FX12 shader hack of the tests (or whatever they've done) that lets the nv30 compete with r300 in 3dmark 03?

Aside from that, i wonder if AA quality has improved. :-?
 
Bambers said:
So that score is probably still using the FX12 shader hack of the tests (or whatever they've done) that lets the nv30 compete with r300 in 3dmark 03?
Which one? It doesn't seems so with det 50.** according to TT-H

"…de jolis chiffres ! Ca faisait longtemps de la part du matériel NVIDIA ! Les images semblent bien, plus de problème de qualité au premier regard. Désolé de ne pas pouvoir te donner des chiffres aussi détaillés que l'autre fois... ...mais les images semblent correctes cette fois. Du vrai floating point à une vitesse décente. Tu sais ce que je pense de ça, et tu partages ma vision des choses je pense (?), c'est pour moi une première carte de NVIDIA qui sera capable d'afficher des graphismes conçus dans le sens de DX9…"
 
DaveBaumann said:
I've sent NVIDIA some screenshots of some, errr, odd things that appear to be happening in their later drivers in 3DM03.

As part of the beta team we have access to to beta builds that allow extra functionality for debugging, that non-beta members (i.e NV) don't. When using this mode numerous rendering 'errors' occur with the later drivers. Some are there in the 43.45's, more are there in the 43.51's. These same issues are not present on ATI boards.
I sent you an email about this Dave. Hope you got it.

I'm not sure if this should be mentioned here due to the delicate situation regarding beta members and the debug build you used. I think as long as it runs "fine" on the public version, it's well, fine. Try the "renaming the exe" experiment I suggested. It'd be interesting to see if my guess about aggressive culling coupled with recognition of exe names (non-aggressive culling in recognized exes but left in for unrecognized exes) has anything to do with the rendering artifacts.

EDIT
Whoops, just got your email Dave. Folks, forget what I said above [edit, again!]although I'm not too far off[/edit] (I think!)
 
Lezmaka said:
Ostsol said:
I'm curious: is there any particular reason as to why NVidia video cards' PCBs are so much huger than ATI cards'?

I can't find the thread (or remember which board it was on) but if I remember right it's due to nvidia designing in more "tolerance". That's not the word I'm looking for, but it's been a while since I read that thread so my memory's not doing to well, lol.
Yep, MuFu said so.

Evildeus said:
Impressive the +20% with the det 50.
Looks more like 10% to me: 6000 -> 6600.
 
Back
Top