[h] 4 Weeks with NVIDIA GeForce GTX 980 SLI

[Keysplayr]

Yeah. Physics is physics.
Next time though, I'd recommend crimping and bending over cutting.

 

[KaRLiToS]

Yeah. Physics is physics.
Next time though, I'd recommend crimping and bending over cutting.

It was impossible to bend

Next time? There won't be a next time.

Never again I will have heatsinks in my rig, just waterblocks.

 

[Keysplayr]

It was impossible to bend

Next time? There won't be a next time.

Never again I will have heatsinks in my rig, just waterblocks.

You can cut away the fins around the pipe first. Then crimp/bend.
But ok. Watercooling is the way.

 

[RussianSensation]

They should really explore the value of after-market 290s in CF and 970 SLI. The 980 SLI and 290X CF setups aren't much faster. Also, 970 Tri-SLI vs. 290s Tri-fire vs. 980 SLI.

$520 R9 290s MSI Gaming
$1160 GTX 980 MSI Gaming

Are you kidding? That's $640 left over for upgrading to 390/390X/GM200 dual-GPUs next year. 980 SLI can run at 1*C, use 1W of power and it still is a horrible purchase all things considered.

HardOCP even said Quad-fire 290X (295X2 in CF) is the fastest setup for 4K they've ever tested. You can now buy 4x 290s for less than it costs to buy dual after-market 980s. Shocking.

 

[Keysplayr]

They should really explore the value of after-market 290s in CF and 970 SLI. The 980 SLI and 290X CF setups aren't much faster. Also, 970 Tri-SLI vs. 290s Tri-fire vs. 980 SLI.

$520 R9 290s MSI Gaming
$1160 GTX 980 MSI Gaming

Are you kidding? That's $640 left over for upgrading to 390/390X/GM200 dual-GPUs next year. 980 SLI can run at 1*C, use 1W of power and it still is a horrible purchase all things considered.

HardOCP even said Quad-fire 290X (295X2 in CF) is the fastest setup for 4K they've ever tested. You can now buy 4x 290s for less than it costs to buy dual after-market 980s. Shocking.

Yeah, which is probably why 290x shouldn't be compared to 980s, but to 970s.
980 is Halo so you pay the Halo sir charge. 970 is more in line with 290x performance.

I'd imagine the GTX 960 will be the direct competitor for AMD 290. There really isn't anything to directly compete with 290 right now.

 

[RussianSensation]

Yeah, which is probably why 290x shouldn't be compared to 980s, but to 970s.
980 is Halo so you pay the Halo sir charge. 970 is more in line with 290x performance.

I'd imagine the GTX 960 will be the direct competitor for AMD 290. There really isn't anything to directly compete with 290 right now.

I think he compared 980s from a standpoint of someone "upgrading" from 290Xs. But I think he should definitely do 970 SLI vs. 290s CF and Tri-SLI 970 and Tri-fire 290s with 980 SLI. Those are more useful reviews.

As far as 960 competing with 290, I think it will depend when AMD launches 300 series. AMD really dropped the ball by being behind NV with 290/290X vs. 780 series. If AMD launches 300 series only by March 2015, it will have to seriously outperform 970/980 at $330/550 price levels or be forced to lower prices on new tech. It would a failure in my books if 390/390X are only 5-10% faster than 970/980 6 months later. At least with 290 they undercut the $650 780 by $250.

 

[96Firebird]

This is even worse though

That heatpipe is sealed off, I have that cooler on my card. It is sealed off internally, unlike the other two which come to a point externally. I'm not sure how the rumor started that it was an open pipe, but its wrong.

 

[3DVagabond]

"Sigh".. The heat pipes are touching "each other" aren't they? You act like there isn't any heat being transferred to the 1st and 5th pipes and only the center pipe and 2/3rds of the 2nd and 4th pipes are collecting any heat from the GPU. Heat is conductive through metal.

You are trying to say that's efficient use? That has to be the worst conceived post you've ever made. The DC cooler is specifically designed to operate most efficiently with the heat pipes in direct contact with the GPU. Have you never read an Asus DC review? Ever? Just once? Again,look at Sapphire Tri-X and Powercolor PCS+ so you can see the results of properly cooling Hawaii.

 

[3DVagabond]

I'm sure he's serious and I agree with him.

Here is the Asus gtx 980 strix :

\\
http://www.techpowerup.com/reviews/ASUS/GTX_980_STRIX_OC/4.html

Hawaii is just a hot chip. We can try to excuse it with saying AMD somehow didn't know what they were doing and it's just 'a lousy cooler'. Then with another stroke of strangeness we have AIB's not 'knowing what they are doing'.
Maybe the obvious common denominator is a very hot chip, and pronounced more when a leaky chip is examined.

We are not talking a 250W chip with the 980.

 

[KaRLiToS]

That heatpipe is sealed off, I have that cooler on my card. It is sealed off internally, unlike the other two which come to a point externally. I'm not sure how the rumor started that it was an open pipe, but its wrong.

Thanks for mentionning my friend.

 

[2is]

"Sigh".. The heat pipes are touching "each other" aren't they? You act like there isn't any heat being transferred to the 1st and 5th pipes and only the center pipe and 2/3rds of the 2nd and 4th pipes are collecting any heat from the GPU. Heat is conductive through metal.

While I'm sure having the heat pipes touch each other is a better solution that not having those heat pipes there at all, I'm also sure that it's a crappier solution than having all the heat pipes make contact with the GPU itself. That is, after all, what they were designed to do. I don't know if you own one of these cards and have a compelling need to defend a hastily implemented cooling solution, but the folks criticizing it have a point. A point backed up by thermal data when compared to a better thought out cooling solution. If you doubt it, there are reviews that will clear it up.

 

[Abwx]

"Sigh".. The heat pipes are touching "each other" aren't they? You act like there isn't any heat being transferred to the 1st and 5th pipes and only the center pipe and 2/3rds of the 2nd and 4th pipes are collecting any heat from the GPU. Heat is conductive through metal.

Most of the thermal contact is under the non contacted pipes, ther is no pressure from the GPU that would held them tightly pressured to the supporting surface, as such thermal conduction between thoses uncontacted pipes and the GPU is just ridiculous.

 

[Keysplayr]

Most of the thermal contact is under the non contacted pipes, ther is no pressure from the GPU that would held them tightly pressured to the supporting surface, as such thermal conduction between thoses uncontacted pipes and the GPU is just ridiculous.

I have no idea what you just said.

I'm now suddenly not sure any of you have ever seen the underside of a heat sink. Aren't they almost always quite larger than the silicon and heatspreaders they are placed upon? Are you trying to tell me that if some of the heatsink is overhanging the GPU/CPU/heatspreader then it isn't making good use of the heatsink? Reach much folks?

The heatpipe design is just fine. One pipe touches the other, actually smashed flat against each other as if a solid block, and that is more than enough to manage thermal conduction.

I'm not an engineer, but I'd like to hear from a neutral one here in the forums.

And I would propose to dare you to hold onto that 1st heatpipe while I hold a blowtorch on the middle one assuming you could only feel the conductive heat and not the radiated heat from the torch itself. I guaranty 3rd degree or greater burns from the 1st pipe. If you've ever sweated copper pipes before or soldered wired before, you'd be hard pressed to disagree successfully with me. You'll try, but won't be very convincing.

 

[KaRLiToS]

This usually don't happen but I agree with Keysplayr (this time).

All the heatpipes are merged together and merged inside the GPU copper base contact. Of course all the pipes will have around the same amount of heat to dissipate.




(Anyway, we are really going far from the original topic)

 

[Silverforce11]

Physics says the efficiency of heat transfer is dependent on a few factors, one of which is distance.

 

[Zardnok]

"Sigh".. The heat pipes are touching "each other" aren't they? You act like there isn't any heat being transferred to the 1st and 5th pipes and only the center pipe and 2/3rds of the 2nd and 4th pipes are collecting any heat from the GPU. Heat is conductive through metal.

Let's try this a different way perhaps since showing pictures is not working. How about showing test reviews of different coolers.

http://www.computerbase.de/2014-05/amd-radeon-r9-290-290x-roundup-test/3/

The page is in German, but you can translate easy enough and the numbers don't change. Note how the cards with custom coolers designed for the 290 that have core-contact with all heat-pipes (Vapor-X, Tri-X, and PCS+) perform MUCH better than the coolers that were re-purposed from other cards and do not have all of their heat-pipes making contact.

Not really sure why this is a point of contention.

 

[Keysplayr]

Let's try this a different way perhaps since showing pictures is not working. How about showing test reviews of different coolers.

http://www.computerbase.de/2014-05/amd-radeon-r9-290-290x-roundup-test/3/

The page is in German, but you can translate easy enough and the numbers don't change. Note how the cards with custom coolers designed for the 290 that have core-contact with all heat-pipes (Vapor-X, Tri-X, and PCS+) perform MUCH better than the coolers that were re-purposed from other cards and do not have all of their heat-pipes making contact.

Not really sure why this is a point of contention.

So it had nothing to do with fan speed profiles, voltages set in the BIOS by the manufacturer? Anything like that over reference models? That's some reasons why it's a point of contention.
But, we are getting hung up on this somehow without purpose. So, I'll not comment on it any further. You may, but I'm done.

 

[Granseth]

(...)

And I would propose to dare you to hold onto that 1st heatpipe while I hold a blowtorch on the middle one assuming you could only feel the conductive heat and not the radiated heat from the torch itself. I guaranty 3rd degree or greater burns from the 1st pipe. If you've ever sweated copper pipes before or soldered wired before, you'd be hard pressed to disagree successfully with me. You'll try, but won't be very convincing.

With you experiment you would see that the 1st heatpipe would get much hotter than the other connected to it.
It helps to have them as they will cool the 1st heatpipe, but not as well as sharing the direct heat from the torch.
I have no idea about how big a difference there would be, but with a heatpipe you would loose some efficiency by having the heatpipe cool another heatpipe instead of cooling the source.

 

[Keysplayr]

n/m

 

[96Firebird]

I always have a little laugh at myself when I hear people talk about the heat pipes not contacting the GPU directly. The thermal paste will have a much larger effect on cooling than the heat pipes. I took heat transfer during my undergrad as a mechanical engineer, but even that isn't needed to understand that the heat pipes are fine not touching the thermal paste. Assuming they are in contact, and the surface finish at the contacting areas is "smooth enough", they are essentially one material. Looking at the image posted in post 10, that looks fine. The EVGA heatsink, however, has a different material between each heat pipe (posted in post 20). That will effect the heat transfer to the third pipe. I'm not even sure what material the buffer is, looks like maybe aluminum holder with copper pipes. Each of these changes in material will have an impact on the heat transfer.

My personal opinion on the matter is that the heat pipe needs to stay in the heatsink as much as possible. All these coolers with the heat pipes traveling outside the heatsink will not cool as well as the coolers with the heat pipes spreading out from a central location. This is why I am stepping-up my ACX 1.0 cooler to the FTW edition, which has the "spider-like" setup. The more time (area) the heat pipe spends in the heat spreader, the better the heat can escape the pipe.

But ultimately, there are too many variables to come to a conclusion on what makes a good heatsink for a GPU. A cooler could have an awesome heatsink, but the fans may be loud and ruin the entire package. Or the heatsink may not be deep enough, or not use an effective material. But to place the blame on heat pipes not contacting the GPU is silly.

 

[KaRLiToS]

I always have a little laugh at myself when I hear people talk about the heat pipes not contacting the GPU directly. The thermal paste will have a much larger effect on cooling than the heat pipes.

Thermal paste doesn't cool anything. It serves as transferring heat.

Thermal grease (also called thermal gel, thermal compound, thermal paste, heat paste, heat sink paste, thermal interface material, or heat sink compound) is a kind of thermally conductive (but usually electrically insulating) adhesive, which is commonly used as an interface between heat sinks and heat sources (e.g., high-power semiconductor devices). The grease gives a mechanical strength to the bond between the heat sink and heat source, but more importantly, it eliminates air (which is a thermal insulator) from the interface area.

 

[96Firebird]

I meant that in terms of heat transfer. If the thermal paste doesn't transfer heat well enough, it doesn't matter how many heat pipes the heat sink has. The heat just won't get there.

 

[KaRLiToS]

I meant that in terms of heat transfer. If the thermal paste doesn't transfer heat well enough, it doesn't matter how many heat pipes the heat sink has. The heat just won't get there.

True. But honestly I don't understand why people are doing a big deal about the heat pipe thingy.

Check the Noctua NH-D15, there is no chance for any chip to directly touch a heat pipe, they are all merged inside the die contact base. Also, an i5 3570k might not have a direct contact with each pipes while a 5960x might.

Same thing goes for the Silver Arrow.

 

[3DVagabond]

True. But honestly I don't understand why people are doing a big deal about the heat pipe thingy.

Check the Noctua NH-D15, there is no chance for any chip to directly touch a heat pipe, they are all merged inside the die contact base. Also, an i5 3570k might not have a direct contact with each pipes while a 5960x might.

Same thing goes for the Silver Arrow.

http://www.frostytech.com/articleimages/201404/NoctuaNHD15_base.jpg
http://archive.benchmarkreviews.com...er-arrow/thermalright_silver_arrow_bottom.jpg

That's fine if that's how it is engineered. Something like the Asus DC cooler though is designed for the heat pipes to make Direct Contact with the GPU.

 

[KaRLiToS]

That's fine if that's how it is engineered. Something like the Asus DC cooler though is designed for the heat pipes to make Direct Contact with the GPU.

Facepalm

You are right, lolll, check this out.