AMD Cypress ''Radeon HD 5870'' Stripped

[sapetto]

If OCed is better than GTX295 when its not OCed its almost the same?

 

[rpsgc]

*sigh* I feared this would happen... the card won't fit inside my case (Antec Solo), it's too long. Maybe the 5850 will maintain a sane size like the current gen.

 

[Hitman.1stGame]

HD 5870 X2 will kill GT300 before it launch

 

[a_ump]

HD 5870 X2 will kill GT300 before it launch

lol that's quite a statement. it won't surprise me if the GT300 is 30-50% faster than the GTX 295. The rumored specs are unreal, more than double the GTX 285's as well as supposedly going with a 512-bit GDDR5 memory bus. 512shaders compared to GTX 285's current 240. GT300 is going to be a monster there is no denying it, however if the HD 5870 is even 10% faster than HD 4890's xfire then the HD 5870x2 should still be able to outperform the GT300 by a decent margin. But my last 2 sentences are all speculation and best guess.

 

[wahdangun]

^
^ but ati is already doubleling it's spec why nvdia can't do the same (with a result of bad yield, power hungry, big die, and hot like hell)

 

[TheMailMan78]

^
^ but ati is already doubleling it's spec why nvdia can't do the same (with a result of bad yield, power hungry, big die, and hot like hell)

My spec has doubled. Damn Burger King. :shadedshu

 

[a_ump]

^
^ but ati is already doubleling it's spec why nvdia can't do the same (with a result of bad yield, power hungry, big die, and hot like hell)

?why can't nvidia do the same, i said they were more than doubling lol. and with a much better memory system than the HD 5XXX series its most definitely going to outperform the HD 5870. They probly are having bad yields which is my suspicion why they can't release till Q1 2010, by then they hope to have enough stock built up and improvements in yields to keep up with supply.

 

[pantherx12]

Just decorative.

http://img.techpowerup.org/090911/bta1q2w.jpg

Looks like they're openings to me not solid bits of plastic

http://www.techpowerup.com/img/09-09-11/132b.jpg

The picture you posted its just a bad angle and the bottom bit of plastic looks as if it is part of the top piece.

They may still be decorative, but they're defitnitely holes.

 

[btarunr]

Looks like they're openings to me not solid bits of plastic

http://www.techpowerup.com/img/09-09-11/132b.jpg

The picture you posted its just a bad angle and the bottom bit of plastic looks as if it is part of the top piece.

They may still be decorative, but they're defitnitely holes.

Yeah, decorative holes that don't participate in the card's internal air-flow.

 

[Nick89]

I'll be getting one when nvidia rolls out there set of cards, that will lower prices.

 

[a_ump]

yea....for decorative purposes is the most retarded reason to put those there. i'd rather they were part of the air flow and i don't see how it'd be that difficult to include them into it.

 

[TheMailMan78]

Looks like they're openings to me not solid bits of plastic

http://www.techpowerup.com/img/09-09-11/132b.jpg

The picture you posted its just a bad angle and the bottom bit of plastic looks as if it is part of the top piece.

They may still be decorative, but they're defitnitely holes.

Mussels mouth is a decorative air hole.

 

[[I.R.A]_FBi]

Mussels mouth is a decorative air hole.

Lool, the only one?

 

[Nick89]

Mussels mouth is a decorative air hole.

Oh man, I laughed my ass off!

Any way looks like I'm going to need to sell my pair of 4870's before nvidia's cards come out.

 

[Benetanegia]

My claim includes 'load'. Again, Max ≠ Load. There's a clear difference between the two which you can look up in our latest reviews. Here's a sample: http://www.techpowerup.com/reviews/Mushkin/GTX_295_Single_PCB/28.html

Trivia includes: GTX 295's max consumption is 320W, and that of GTX 285 is 218W. So Cypress beats them both at max power.

So this is where Cypress will land (red mark), going by AMD's value of 188W max power:

...

Cards never actually consume what the vendor said it would consume. And ALL of them use TDP, or TDP-like numbers and call them Max power consumption.

Anyway, if history should be taken into account... RV790' slide:

Actual consumption:

So there's no way we could know the actual Maximum consumption, but I think that a good estimate would be around RV790's. But if I'm correct and 188w is the TDP, absolute maximum could be anywhere between 225 and 275 W, depending on how much time they thought the card would be 100% loaded and idling during actual use when they calculated the TDP. TDP can be tricky if the variance between idle and load is huge and in RV870 it seems so. The GTX295 is a good example of how much it can change too.

 

[TheMailMan78]

Cards never actually consume what the vendor said it would consume. And ALL of them use TDP, or TDP-like numbers and call them Max power consumption.

Anyway, if history should be taken into account... RV790' slide:

Actual consumption:

So there's no way we could know the actual Maximum consumption, but I think that a good estimate would be around RV790's. But if I'm correct and 188w is the TDP, absolute maximum could be anywhere between 225 and 275 W, depending on how much time they thought the card would be 100% loaded and idling during actual use when they calculated the TDP. TDP can be tricky if the variance between idle and load is huge and in RV870 it seems so. The GTX295 is a good example of how much it can change too.

Hmmmmm Me thinks you're not new here.

 

[btarunr]

Cards never actually consume what the vendor said it would consume. And ALL of them use TDP, or TDP-like numbers and call them Max power consumption.

Anyway, if history should be taken into account... RV790' slide:

Actual consumption:

So there's no way we could know the actual Maximum consumption, but I think that a good estimate would be around RV790's. But if I'm correct and 188w is the TDP, absolute maximum could be anywhere between 225 and 275 W, depending on how much time they thought the card would be 100% loaded and idling during actual use when they calculated the TDP. TDP can be tricky if the variance between idle and load is huge and in RV870 it seems so. The GTX295 is a good example of how much it can change too.

Right, so even with that kind of deviation, it will fall between GTX 275 and GTX 285 at worst, or between HD 4870 and GTX 285 at best.

 

[pantherx12]

Yeah, decorative holes that don't participate in the card's internal air-flow.

That's so rubbish!

I'm so going to find a way to utilise those holes for something!

 

[Benetanegia]

Right, so even with that kind of deviation, it will fall between GTX 275 and GTX 285 at worst, or between HD 4870 and GTX 285 at best.

Yeah, yeah. I wasn't pretending it would consume as much as a GTX295. Just wanted to clarify that it wont be right where you put it either. 250W is already goood!!

And, it's true those things are purely decorative? I know I don't like them...

 

[btarunr]

Yeah, yeah. I wasn't pretending it would consume as much as a GTX295. Just wanted to clarify that it wont be right where you put it either. 250W is already goood!!

I had acknowledged that I'm putting AMD's value in a graph full of TPU's measurements. So the deviation is obvious. What's clear is it's not going to lose to GTX 285 or GTX 295 at max power.

 

[Benetanegia]

I had acknowledged that I'm putting AMD's value in a graph full of TPU's measurements. So the deviation is obvious. What's clear is it's not going to lose to GTX 285 or GTX 295 at max power.

Well, I think this conversation is unfair. I've seen all your posts and it's pretty clear you have "insider" info HAha. I'm guessing it won't lose to those cards. But by the numbers that we have, the rest of us, it could reach even 300W, who knows (always speaking about Wizzard's graph). I mean, look at this:

http://img.donanimhaber.com/resimler/R700_4870-x2-slide.jpg

285 W according to the slide, 381W acording to Wizzards measurements. That's 100 W of difference. In the end it doesn't matter because a card will never reach those levels during actual use, that's just the consumtion it reaches in one odd clock cycle in which all of it's units are working together, and that happens, what, 1 time every complete 3dmark run??

 

[Meizuman]

Bene, Pic No Work

 

[Benetanegia]

Bene, Pic No Work

mm works for me.

Just google HD4870 X2 slide or something... You might even find one with a different number, I don't know. I'm not trying to say "ey guys this is like this!!", just that unless we know something (ahem Btarunr ) everything is possible... Or am I missing something else?

 

[btarunr]

Not much insider info, but I talk to luckier people, in this case, people who made Cypress a substitute for their better-half/teddy for a few days. Anyway, mark your calendar/PIM/Blackberry for 23rd.

 

[imperialreign]

looks like I might ahve to pack some loot away for when the 5870x2s roll out . . .