1800XT vs 7800GTX 512mb.

name='Geomon' said:
Anyone have any result with the overclocking of the 78gtx 512mb?

Watch this space my friend - Jaffa will be testing shortly and will post as soon as.

As I am only interested in SLI I am waiting for the price to drop to a slightly more palletable £425ea.

Mav
 
name='maverik-sg1' said:
Gollum this test IS as close to like for like as its gonna get I feel - it's not too late to cancel your order and re-join the light side of the force (oooops!)

Everyone knows nvidia is the dark side! (smack!)

I'd rather go with ATi because of the software voltage control and better cold scalability (coldbug/90nm etc.) TBH but it's looking harder and harder as master cards have disappeared from the UK. It's actually cheaper to get the CF-DR and both X1800s than an SLi pair of these new GTXs but I may have to go that way in the end.
 
Agreed on the price issue, I'd be running SLi right now, but with Christmas only a month away I reckon it will be more prudent for me to wait for the second card, plus it gives me the chance to wait for the NF4 Experts to arrive in the UK.

I'll add some OC results as soon as I can.
 
As I am only interested in SLI I am waiting for the price to drop to a slightly more palletable £425ea.

Yeah theres no way I would pay that kind of cash for a single vga card! I think its a liberty charging that amount and I wouldn't want to encourage them.

If the gtx 512 cores/ram dont allow for much more room for o'clocking then I may go for the ATI.
 
GoLLuM4444 said:
Good review between these 2 cards:

http://www.driverheaven.net/reviews/7800512dhreview/overclocking.htm

Shame all of it's at stock (except that page). The 7800 wins in nearly all at stock but overclocked it's pretty close and no doubt the ATi will scale further with better cooling. Tough call but I'm still after the X1800s. (smack!) (picture the yellow smiley as green).

Yeah that a pretty good comparison - well it's the only one.

What I enjoyed most about the overclocked comparisons was that the Radeon managed to get 1.8Ghz from the ram at stock cooling - but maxes out at 2Ghz under extreme cooling - impressive.

The GTX was able to produce a 50Mhz overclock, mainly due (I suspect) to the crappy testing that the Nvidia drivers have enabled when overclocking - we all know with rivatuner an extra 30+Mhz is available even with stock cooling on the core and perhaps another 50mhz on the RAM, I dont know if ATI have the same 'testing' on their drivers but it sure is a pain and has to be disabled if you want to do a proper overclocking test.

I liked the min FPS thing - really shows where you should be pitching your settings for a totally playable experience.

Will they scale better? It's hard to say, a real comparsion would have been if they had set memory to same clocks and cores to +50Mhz and tested? Then there's the whole CPU limit again.

Interesting though and the closest direct comparison (apart from Jaffa testing which shows a slightly different story again but one more in line of where we are pitching in terms of benching).

Nice find Gollom mate.

Mav
 
name='maverik-sg1' said:
Will they scale better?

I'd put money on it. GTX cores won't easily scale past 800mhz and we've already seen 1ghz from X1800s. 90nm produces less heat means more voltage usable too. Another key point is that they don't mention crossfire or SLi and I think AFR could be a huge difference between the cards. I've got my heart set on the ATis but it's looking less and less likely that I'll be able to get them.
 
name='GoLLuM4444' said:
I'd put money on it. GTX cores won't easily scale past 800mhz and we've already seen 1ghz from X1800s. 90nm produces less heat means more voltage usable too. Another key point is that they don't mention crossfire or SLi and I think AFR could be a huge difference between the cards. I've got my heart set on the ATis but it's looking less and less likely that I'll be able to get them.

I am not saying they wont clock higher I think thats a given, but I thought that scaling implies a greater FPS percentage per Mhz of overclock, using that definition, I am not yet convinced that they do scale better (could be persuaded either way though).

Sure AFR has massive potential - but crossfire XT's are conspicuous by there absence at the moment with no real sign of when they are expected - although I would hope it will be before xmas.

Mav
 
name='maverik-sg1' said:
I am not saying they wont clock higher I think thats a given, but I thought that scaling implies a greater FPS percentage per Mhz of overclock, using that definition, I am not yet convinced that they do scale better (could be persuaded either way though).

Oh fair point. I was talking about the clocks scaling higher with the cold rather than the fps with the clocks. You can probably get a good idea by the fps change between the 2 cards at stock and overclocked in that review though. In fact I think I'll crunch some numbers now. Watch this space. :cool:
 
Numbers are a little dodgy as they are per mhz core change and ignore mem frequencies. Both mem freqs are at 900ish overclocked but the x1800 is 750 stock as opposed to 850 on the 7800 so it will have gained more in that area. Nevertheless the 100mhz mem change can't explain the big difference can it?

X1800:

Min fps increase/mhz overclock: 0.0411

Avg fps increase/mhz overclock: 0.0822

Max fps increase/mhz overclock: 0.0822

7800:

Min fps increase/mhz overclock: 0

Avg fps increase/mhz overclock: 0.0588

Max fps increase/mhz overclock: 0.1176

Key numbers are NV: 0.0588fps/mhz and ATI: 0.0822fps/mhz. How accurate are these numbers? Depends how much difference mem clocks make. These are probably quite a way off the actual fps increase rate but still are (IMO) too far apart to just say memory explains it away.
 
I'm not sure I was a fan either although it would be a better system to access classifieds than spamming. Neway (offtopic! (offtopic! (offtopic!!
 
Back
Top