eVGA 5900SE or Radeon 9700?

[Pete]

Can we ban the words "nvidiots" and "fanATIcs" completely? That may help avoid a lot of these squabbles.

 

[Evdawg]

Originally posted by: modedepe
I would expect the longevity of the 9700 to be as good as a 5900. It really is in a lot of ways a toss up. As you've said, the 5900 will be faster at doom3 ogl games, whereas the 9700 will be faster at hl2 dx9 games. Really though, rather than looking at a 5900se, you might try getting a regular 5900. They're down to $162 and there are some that come with call of duty.

Yea but i dont think that the ati cards will be as bad at doom3 as the nV were with HL2.

 

[CaiNaM]

Originally posted by: Evdawg

Yea but i dont think that the ati cards will be as bad at doom3 as the nV were with HL2.

i really think it's funny that there's so much focused on how these cards do based upon benchmarks of games which are not only unreleased, they aren't even in late beta form - some run using drivers that are what, 4-5 months old?

while it's safe to assume ati's choice of going the generic dx9 codepath was much smarter than nvidia's proprietary choice (based on gabe newell's comments regarding how difficult and time consuming it was to optimize for nvidia's codepath), and perhaps even suggest the r3xx architecture is more sound than the nv3x's, we really have no clue as to what the final performance will be on these games when these titles actually ship (hell, it's looking like we might see a fall refresh of the next gen cards before we see hl2).

this really isn't an to "defend" nvidia; their performance and their architecture simply could be inferior, however logic would dictate that assumptions made based on the above are harldy derivitive of what the actual performance might actaully be. while image quality remains to be seen, it's been proven that nvidia's efforts in redesigning their drivers have positively affected performance (this article covers how nvidia has started working on shader performance). heck, iq and peformance have BOTH been increased in ut2k3 - by a substantial amount. while skepticism is certainly in order, i don't see how final conclusions can be made at this stage.

what baffles me is why we don't see how these cards run using current drivers and dx9 titles which ARE currently available, such as the 'far cry' demo. surely this would be much more relevant to this topic than making presumtions on performance based on unfinished betas while using outdated drivers...

 

[rbV5]

Originally posted by: Pete
Can we ban the words "nvidiots" and "fanATIcs" completely? That may help avoid a lot of these squabbles.

You got my vote Pete!

 

[Pete]

rbV5: Swish! < /Futurama >

Cainam: I don't know, game engines are generally finished before the actual gameplay is completed, so the D3 and HL2 benchmarks may be fairly close to final numbers on cards that were expecting the tests. The fact that ATi's drivers didn't support its latest card in D3, and nV's drivers promised speed-ups but with missing effects in HL2 makes the early benches seem less than representative on those cards. But I do think nV's D3 numbers and ATi's HL2 numbers are probably fairly accurate.

 

[yhelothar]

is it possible to get COD with an older PNY 5900?

 

[CaiNaM]

Originally posted by: Pete
rbV5: Swish! < /Futurama >

Cainam: I don't know, game engines are generally finished before the actual gameplay is completed, so the D3 and HL2 benchmarks may be fairly close to final numbers on cards that were expecting the tests. The fact that ATi's drivers didn't support its latest card in D3, and nV's drivers promised speed-ups but with missing effects in HL2 makes the early benches seem less than representative on those cards. But I do think nV's D3 numbers and ATi's HL2 numbers are probably fairly accurate.

well, perhaps you might be right, however i've seen too often when ppl base this or that on early alpha's/beta's and even in some cases demo's (the the last few years seems alot of demo's shop after the product does) where test's on these early products change significantly due to either optimizations/bug fixes in the final product or later drivers....

just seems to me people are making a whole lotta hoopala and giving more merit to these earlier results than they have in the past - both from the nv camp (doom3) and ati camp (hl2). would be nice to see results on shipping software when ppl use it as ammo to bash another product

 

[CaiNaM]

Originally posted by: virtualgames0
is it possible to get COD with an older PNY 5900?

don't think so. it was a limted time holiday promo, and pny only offered it with their se product. i've seen a few reports of it being included in the bfg 5900nu, but that's about it...

 

[BenSkywalker]

I don't know, game engines are generally finished before the actual gameplay is completed, so the D3 and HL2 benchmarks may be fairly close to final numbers on cards that were expecting the tests.

Neither D3's nor HL2's engines were complete at the time of the round of benches we saw though. D3 has increased the processor load considerably(so much so that Carmack has voiced his displeasure with it, he was outvoted on some physics effects he didn't think were worth it) which will likely close the boards up a bit in the lower resolutions while the HL2 engine wasn't running HDR yet which will slow all the boards down and it was using the pre compiler Det drivers(not going to perform magic, but based on the numbers we have seen the performance even in pure DX9 mode is up considerably).

Of course, we don't know when we will see either of those games, so it is possible that we could be running R5x0s/NV5xs by then

Right now I'm looking forward to FarCry more then any of the others, mainly because we have some solid proof that that it about to ship and the Crytek engine, to me anyway, looks like it is certainly comparable to HL2's.

 

[Acanthus]

Originally posted by: keysplayr2003
Mods have been petitioned as of 12:55am

how do i show support for the ban?