5970 benches leaked

[bige420]

Thought you guys would like to know. Looks like one beast of a card.

http://www.guru3d.com/news/radeon---hd-5870-x2-hd-5970-benchmarks-and-images-surcface-/

 

[hallowed_dragon]

Thought you guys would like to know. Looks like one beast of a card.

http://www.guru3d.com/news/radeon---hd-5870-x2-hd-5970-benchmarks-and-images-surcface-/

The only intriguing thing in the whole 'article' is:

He goes on to say, that ATI have something special up their sleeve, but are waiting for Nvidia to release their cards first.

 

[bige420]

Yes, very interesting.

 

[JofaMang]

It looks to be about 3 feet long, though.

 

[bige420]

I know, its kind of ridiculous. Probably wont fit in a lot of cases.

 

[Annisman]

Many of you may be looking at those numbers as good, but I would like to share two benchmarks that I have done for comparison purposes.

Our setups are similar, both cpu's at 4.0Ghz, fast ddr 3 ram, and 2X 5870 cores.

Now these benchmarks I did with the initial driver release for the 5870 release:



Notice how I score higher fps, even with 4xaa (they use 2xaa) all other settings are the same.
24 min, 51 max, 40 avg.



Even if you include a some of the very low min. fps that are probably bug related....look at how much more I score in overall avg and maximum fps. Also on a side note my Stalker Clear Sky score was higher than 99.62% of all entries, check my avatar.

I think they have alot of work to do on the drivers for the X2. And yes, my cards were on stock clocks.

 

[KidHorn]

Interesting, but the clock speeds for the cards seem kind of low to me. Lower than the base rates for the 5870.

 

[JofaMang]

Many of you may be looking at those numbers as good, but I would like to share two benchmarks that I have done for comparison purposes.

Our setups are similar, both cpu's at 4.0Ghz, fast ddr 3 ram, and 2X 5870 cores.

Now these benchmarks I did with the initial driver release for the 5870 release:

[url]http://img3.imageshack.us/img3/8086/crsssssso.th.jpg[/URL]

Notice how I score higher fps, even with 4xaa (they use 2xaa) all other settings are the same.
24 min, 51 max, 40 avg.

[url]http://img687.imageshack.us/img687/7348/stalker.th.jpg[/URL]

Even if you include a some of the very low min. fps that are probably bug related....look at how much more I score in overall avg and maximum fps. Also on a side note my Stalker Clear Sky score was higher than 99.62% of all entries, check my avatar.

I think they have alot of work to do on the drivers for the X2. And yes, my cards were on stock clocks.

Try cutting and pasting those windows out of the original SS, not very readable, even fully zoomed in, it was scaled down so far.

 

[Guest]

not so good

 

[bige420]

The results are on preliminary "leaked" drivers so im sure there will be an improvement once official drivers are released.

 

[EXT64]

I wouldn't be surprised if they have to reduce the clocks a little just to fit in under the power spec. It will be even hard (maybe even worse) for NVidia's dual card as well.

 

[Annisman]

Try cutting and pasting those windows out of the original SS, not very readable, even fully zoomed in, it was scaled down so far.

Fixed, thanks.

 

[Annisman]

not so good

That's what I was thinking, espescially those Stalker numbers, why such a discrepency between the X2 and two 5870's ?

 

[shadow187]

Maybe these aren't the famed x2's? Just a higher-end single-GPU card? (I can't see the benchmarks due to being on a school-compy, so pardon the ignorance!)

 

[cookies]

Seems to be a clock speed thing to me.

HD 5870:
Core Clock: 850mhz
Memory Clock: 1200mhz

HD 5970:
Core Clock: 725mhz
Memory Clock: 1000mhz

 

[jennyh]

Well that's the 5850 clocks so this might be the 5950 not the 5970.

 

[1ce]

The "something special" could be a full clocked version with two 8-pin power connectors, maybe.

 

[chunkymonster]

Please allow me to save some readers the time of reading the article and cut right to the bottom line.

This time AlienBabeltech received some photo's and benchmarks of an 'unknown' user with a RadeonHD 5970...at the moment it should only be taken as unconfirmed rumor so...It is unclear...that ATI have something special up their sleeve

Nothing but speculation on an unreleased product. I will wait for more reliable sources.

 

[jay2tall]

Holy Cow that thing is HUGE! Does it come with it's own defense system?

If I just payed $390 for a 5870, I wonder what they are going to charge for this monster? $700...... $800?

 

[EXT64]

Wow, that is really big. And a backplate too? I see it has a 6 and 8 pin connector, so I guess it is ok on power, but what is the PCI-E card weight spec?!?

 

[jay2tall]

I bet requires room for a full length card including the support at the end. Remember those Full size Workstation cards that had the plastic extension on the end that slid into the support bracket? YEAH, except this is ALL MEAT, and no filler.

 

[redgarl]

Phhhh... lol @ peoples trying to judge the card over some... leaked benches with beta drivers... bravo, simply bravo...

 

[redgarl]

If you want to know the real capabilities of the card, just go check the benches for two 5870 in crossfire, end of story, it will be about the same thing.

 

[daedalus685]

If you want to know the real capabilities of the card, just go check the benches for two 5870 in crossfire, end of story, it will be about the same thing.

While I generally agree with that statement.. there is noone saying that the 5970 has to be at the same clocks as the 5870. There are many factors that may lead to it being slower. We just don't know yet.

 

[Annisman]

If you want to know the real capabilities of the card, just go check the benches for two 5870 in crossfire, end of story, it will be about the same thing.

You do realize that's what I did earlier in the thread comparing my 2X 5870 results, and the X2 results from Guru3D... and guess what ?...... ( hold your breath).......























THE NUMBERS DON'T EQUAL!!!!!!!!!!!!!!!!!!!!

That's what I've been trying to say, either thew new X2 performs worse than 2X 5870 in crossfire, OR they need to do some serious driver work.

Edit: Or these numbers are bogus.