AMD Justifies Using Liquid Chamber Tech On Upcoming 7900-Series Cards

[qubit]

videocardz.com brings us a photo showing the benefits of AMD's new liquid chamber technology over the standard vapour chamber technology used in current heatpipe cooling solutions. There's quite a few benefits, some of which are no drying out, physical robustness and greater reliability. Indeed, can you imagine the disastrous effect on your graphics card of having the cooling system spring a leak during an intense gaming session? Doesn't bear thinking about… Certainly the era of having to replace your graphics card's shrill and inefficient cooler with a high performance aftermarket one are long over, as stock coolers are now generally of very good quality. The liquid chamber system will be used by all AMD's partners who stick to the reference design. Check out the photo for the full info.



View at TechPowerUp Main Site

 

[qubit]

Thanks again to Damn_Smooth for another great lead!

EDIT: There's now a detailed look at this new technology, here.

 

[satelitko]

So... how exactly does that work, when the GPU is upside-down? Doesn't the wick in the vapor chamber exist for exactly that reason? Someone please enlighten me, as I feel like a moron at the moment

 

[Jstn7477]

Hopefully there aren't chances of defects such as overfilling where your cooler suddenly explodes under a high workload. Anyone that remembers the batches of Nichicon capacitors which weren't actually part of the "cap plague" but were rather overfilled with electrolyte and would pop can probably visualize where I'm going with that.

 

[mediasorcerer]

OOoh cant wait, im going reference next time for sure.

 

[xBruce88x]

Now if AMD could justify the performance of Bulldozer... Seriously.... If I upgrade to another AMD cpu i'll probably just get a high clocked Phenom II x6 and OC it some more. (though i'll probably end up with a 2500k)

yea I'm not so sure how they're gonna get the pool to be on the gpu side when its upside down... unless they're gonna flip the way the PCB is? have the gpu face up instead of down? Doubt it though, would be compatibility issues with several boards/cases... like the ones with the x16 at the top

 

[John Doe]

Hopefully there aren't chances of defects such as overfilling where your cooler suddenly explodes under a high workload. Anyone that remembers the batches of Nichicon capacitors which weren't actually part of the "cap plague" but were rather overfilled with electrolyte and would pop can probably visualize where I'm going with that.

That shouldn't be the case here. The liquid doesn't exist like you think it does, it's inside of the cooler. You have to break the heatpipe for the water to leak. Caps on the other hand are filled with a formula that functions at specific volts/capacitance. This is a solid solution (water inside copper), unlike the liquid inside of a capacitor. I've yet to hear a cooler explode from overheating...

as a side note, that was a specific case with Nichicon's. They're quality caps that're harder to get (due flood) right now.

 

[pantherx12]

Pictures of some of the heatsinks that use this? ( and the liquid chambers them selfs) ?

 

[Wile E]

I don't give a shit is they are made from unicorn jizz and leprechaun farts, so long as they do their job.

 

[pantherx12]

Oh by the by, if you're wondering how this works upside down, the chamber is made of copper so heat will make it round to the other side of the chamber and heat the liquid, once this starts happening the liquid will start to boil ( I'm guessing it won't be water ever in this type of cooler)

causing the liquid to no lower be not quite so still : ] ( I.E it will be bouncing about all over inside the chamber)

 

[satelitko]

Oh by the by, if you're wondering how this works upside down, the chamber is made of copper so heat will make it round to the other side of the chamber and heat the liquid, once this starts happening the liquid will start to boil ( I'm guessing it won't be water ever in this type of cooler)

causing the liquid to no lower be not quite so still : ] ( I.E it will be bouncing about all over inside the chamber)

Yeah, but the point of the system is to boil the liquid, which then goes up as vapor and condense(cool) at the top, then return to the pool. That doesn't work upside-down/the other way around. I'm really curious to find out how they plan on making it work.

 

[John Doe]

Yep. Water is only inside the part of cooler that touch the core of the GPU. The convection system works like a TEC. You know, a plate that sits between the heatsink and the core.

BTW, these heatsinks look somewhat big and effective.

 

[pantherx12]

Yeah, but the point of the system is to boil the liquid, which then goes up as vapor and condense(cool) at the top, then return to the pool. That doesn't work upside-down/the other way around. I'm really curious to find out how they plan on making it work.

TBH I think the diagram is just crappy, if you look up the various bits on bobs written down instead it makes more sense ( at-least to me it does, could be reading into things wrong it's like 6 am here and I've not slept yet)

Pool boiling is the first thing you should look up and read about

 

[John Doe]

It's 28nm anyway, no need for Accelero Xtreme's. This system should be sufficient.

 

[pantherx12]

It's 28nm anyway, no need for Accelero Xtreme's. This system should be sufficient.

I recall people saying the same for 45nm.

Have to remember they'll probably keep within the same TDP as they're current cards.

Otherwise you don't get an upgrade ( well you can get the same card with less heat but they tend to then be sold as midrange cards or sweet spot cards etc)

 

[John Doe]

I recall people saying the same for 45nm.

Have to remember they'll probably keep within the same TDP as they're current cards.

Otherwise you don't get an upgrade ( well you can get the same card with less heat but they tend to then be sold as midrange cards or sweet spot cards etc)

Nope.

http://www.tomshardware.com/news/amd-HD-7900-XDR2-Rambus-Memory,13408.html

45nm was different. nVidia couldn't get it right with two different head managers; that's why the 480 was... what it was.

nVidia claims 3-4 times lower performance per watt (nonsense). But the idea is, they're going to be more efficient. Look at the 580, even it doesn't need a custom cooler. I've seen a few people have it running at 50-55C in games.

 

[de.das.dude]

liquid faster than vapour cuz it has more densitY!!!

 

[pantherx12]

Nope.

http://www.tomshardware.com/news/amd-HD-7900-XDR2-Rambus-Memory,13408.html

45nm was different. nVidia couldn't get it right with two different head managers; that's why the 480 was... what it was.

nVidia claims 3-4 times lower performance per watt (nonsense). But the idea is, they're going to be more efficient. Look at the 580, even it doesn't need a custom cooler. I've seen a few people have it running at 50-55C in games.

But what about 65nm and 90nm?

People have kept saying we'll need less powerful cooling but yet the inverse has happened, nv and nvidia started pushing the clock speeds more and building much more complicated chips, remember when gpus only came in 1 slot?

Just seems tech always get pushed along with the fabrication size.

 

[John Doe]

But what about 65nm and 90nm?

People have kept saying we'll need less powerful cooling but yet the inverse has happened, nv and nvidia started pushing the clock speeds more and building much more complicated chips, remember when gpus only came in 1 slot?

Just seems tech always get pushed along with the fabrication size.

It all comes down on the specific case. Cards that had single slot coolers (8800 GT - 4850 etc) have shown to be ineffective on past occasions. Before those, old GPU's only needed a simple aluminum chipset cooler.

Leakage is the key word. With more effective manufactoring processes, TDP is lowered as well as temps. You're thinking of it upside down.

http://www.overclock.net/t/1126863/real-world-power-usage-user-submitted-values#post_15090607

 

[cdawall]

It all comes down on the specific case. Cards that had single slot coolers (8800 GT - 4850 etc) have shown to be ineffective on past occasions. Before those, old GPU's only needed a simple aluminum chipset cooler.

Leakage is the key word. With more effective manufactoring processes, TDP is lowered as well as temps. You're thinking of it upside down.

http://www.overclock.net/t/1126863/real-world-power-usage-user-submitted-values#post_15090607

With better manufacturing you get lower leakage chips that run at lower voltage that in turn lowers TDP. Yet again manufacturing die size has nothing to do with TDP.

 

[John Doe]

With better manufacturing you get lower leakage chips that run at lower voltage that in turn lowers TDP. Yet again manufacturing die size has nothing to do with TDP.

lol? Both, I'm talking about both. If you shrink down a GPU, you'll lower down it's TDP as well. The 3870, although based on a cheaper PCB, pulled half as much power as the HD 2900 after a big die shrink (with the help of better manufactoring as well).

 

[qubit]

lol? Both, I'm talking about both. If you shrink down a GPU, you'll lower down it's TDP as well. The 3870, although based on a cheaper PCB, pulled half as much power as the HD 2900 after a big die shrink (with the help of better manufactoring as well).

I'll second that. I've still got my 2900 and it runs hot as hell, while giving fairly low framerates. It was the Bulldozer of it's day, really.

It's a beautiful looking card though.

 

[pantherx12]

It all comes down on the specific case. Cards that had single slot coolers (8800 GT - 4850 etc) have shown to be ineffective on past occasions. Before those, old GPU's only needed a simple aluminum chipset cooler.

Leakage is the key word. With more effective manufactoring processes, TDP is lowered as well as temps. You're thinking of it upside down.

http://www.overclock.net/t/1126863/real-world-power-usage-user-submitted-values#post_15090607

I'm not man, yes if they used the exact same design but used a smaller process then tdp will be reduced, but this rarely happens, people take advantage of the smaller size and pack twice as much in as before leading us back to where we were before in terms of TDP and powerdraw.

Might have grabbed the powerdaw rather than TDP but source said it's TDP, the 580 if my source is right has a 244W TDP. The TDP of a 2900xt was 215W.

See what I mean? Heat output stays around the same level for top end cards because like I said they go for double the speed not half the size.


The 3870 was pretty much the same as a 2900xt ( infact it has less transistors, lower rop count) yet was on a smaller process hence it's powerdraw being much lower than that of the 2900xt.


If we were just doing powerdraw ( according to geeks3d)


8800 Ultra 175w 90nm 681M Transistors
9800 GTX+ 141w 55nm 754M Transistors
GTS 285 204w 40nm 1400M Transistors
GTX 480 260w (peak 3D) 40nm 3200M Transistors
GTX 560 Ti 205w (peak 3D) 40nm 1950M Transistors
GTX 580 280w (peak 3D) 40nm 3000M Transistors

2900xt 215w 90nm 700M Transistors
HD 3870 105w 55nm 666M Transistors
HD 4870 157w 55nm 956M Transistors
HD 5870 188w 40nm 2154M Transistors
HD6870 151w 40nm 1700M Transistors
HD 6970 250 (PowerTune +20%) 40nm 2640M Transistors

Hell the 6970 is hotter and more powerhungry then a 2900xt


Fab process just affects how much they can fit in one space, the voltage requirement is reduced as well so yes you can get a lower TDP but not all the time design as it depends on the design of the card it's self.


Excuse me if I've got any odd typos or if this is a bit rambling, no sleep makes panther something something.


*edit* added Nvidia cards, took me longer to collate data.

 

[John Doe]

I'm not man, yes if they used the exact same design but used a smaller process then tdp will be reduced, but this rarely happens, people take advantage of the smaller size and pack twice as much in as before leading us back to where we were before in terms of TDP and powerdraw.

Might have grabbed the powerdaw rather than TDP but source said it's TDP, the 580 if my source is right has a 244W TDP. The TDP of a 2900xt was 215W.

See what I mean? Heat output stays around the same level for top end cards because like I said they go for double the speed not half the size.

The 3870 was pretty much the same as a 2900xt ( infact it has less transistors, lower rop count) yet was on a smaller process hence it's powerdraw being much lower than that of the 2900xt.

If we were just doing powerdraw ( according to geeks3d)

2900xt 215w 90nm 700 M Transistors
HD 3870 105w 55nm 666M Transistors
HD 4870 157w 55nm 956M Transistors
HD 5870 188w 40nm 2154M Transistors
HD6870 151w 40nm 1700M Transistors
HD 6970 250 (PowerTune +20%) 40nm 2640M Transistors

Hell the 6970 is hotter and more powerhungry then a 2900xt

Fab process just affects how much they can fit in one space, the voltage requirement is reduced as well so yes you can get a lower TDP but not all the time design as it depends on the design of the card it's self.

The 3870 doesn't have less ROP's than the HD 2900, it's based on the exact same die. If you noticed, there's not much difference in transistors. So since it performed the same as a HD 2900, most the difference came from a die shrink.

The 6970 isn't more more hungry than a 2900 either. You're just looking at spec sheet and nothing else. Have you read what Pheaderus said? He says it pulls less power than his 4870, although it's rated to have 100W higher TDP.

Look at this. 6970 pulls less power than a 570 although it again is rated higher, has a bigger die and more transistors:

http://www.guru3d.com/article/radeon-6950-6970-review/11

http://www.guru3d.com/article/geforce-gtx-570-review/8

6970 runs hotter than a HD 2900? That's just absurd.

And yes, fab process isn't what ONLY matters. But that's not what I'm talking about.

What I'm saying is they BOTH matter to a degree.

 

[pantherx12]

The 3870 doesn't have less ROP's than the HD 2900, it's based on the exact same die. If you noticed, there's not much difference in transistors. So since it performed the same as a HD 2900, most the difference came from a die shrink.

The 6970 isn't more more hungry than a 2900 either. You're just looking at spec sheet and nothing else. Have you read what Pheaderus said? He says it pulls less power than his 4870, although it's rated to have 100W higher TDP.

Look at this. 6970 pulls less power than a 570 although it again is rated higher, has a bigger die and more transistors:

http://www.guru3d.com/article/radeon-6950-6970-review/11

http://www.guru3d.com/article/geforce-gtx-570-review/8

6970 runs hotter than a HD 2900? That's just absurd.

And yes, fab process isn't what ONLY matters. But that's not what I'm talking about.

What I'm saying is they BOTH matter to a degree.

The 2900 had had 16x2 rops the 3870 just has 16 : ]
Regarding power draw, the newer cards have better power management features, if you switch them off then then a 6970 will draw more power and produce more heat than a 2900xt.

As for a absurd, no it isn't man, slap on the exact same cooler I.E put the 2900xt heatsink on a 6970 and the 6970 will be so much hotter it will switch it's self off. ( The 2900xt cooler may of been copper but it has pretty poor by todays standards, heatpipes were flattened, had a VERY thick base lol)
Infact temperatures with their respective heatsinks
2900xt

6970

Even with superior cooling the 6970 is hotter D:

You'll have to explain you're 570/6970 comparison though because the way you've put it sounds like you're saying the 6970 has more transistors, when it doesn't the 570 has more :S

Oh and by the by I'm not just reading spec sheets I went looking for power draw ratings and went they went onto to look at specifications afterwards just to make the data a bit easier to compare.