ASUS ROG STRIX , EVGA ACX 3.0 and ZOTAC AMP Extreme 1080 Cards Pictured in All Their Glory

[zMeul]

ZOTAC PGF board:

holy shit power regulators

[Bouzoo]

Not sure if anyone posted but here's HOF. 

 

 

And here's Galax Gamer.

 

 

Personally, that's too much white. I mean it's only white for god's sake.

 

[Curufinwe_wins]

23 hours ago, zMeul said:

ZOTAC PGF board:

holy shit power regulators

Ermg.... wtf....

 

@Lays

[Lays]

21 minutes ago, Curufinwe_wins said:

Ermg.... wtf....

 

@Lays

It's cool and all but from what I've seen it's only going to be a Chinese card. 

 

 

Also, all the phases in the world won't matter if the card is voltage locked like lots of other zotac  cards have been in the past

[zMeul]

some dude got his hands on a STRIX:

 

 

something that bothers me, the "bracket" on the VRAM chips - what the hell, ASUS?! some chips have thermal pads, others not ..

I took a look at the GDDR5X product sheet and I didn't saw normal operating temperature, just the ranges 0-95deg

 

 

ps: 2 useless heat pipes

[zMeul]

GigaByte GTX1080 Xtreme Gaming:

[zMeul]

ASUS ROG STRIX GeForce GTX 1080 offers poor overclocking

 

Quote

ComputerBase confirmed the rumors from PCGamesHardware. Extreme overclocking on STRIX GTX 1080 is poor. All cards are locked to 1.25V maximum, and when the voltages gets close this value, cards become unstable. Neither bypassing VRM controller or removing TDP limit has made any difference.

 

[The Afro Samurai]

9 hours ago, zMeul said:

ASUS ROG STRIX GeForce GTX 1080 offers poor overclocking

 

 

I just read that too... Maybe Gigabyte G1 version doesn't disappoint. I never owned one but people seem to like them.

[Notional]

5 minutes ago, The Afro Samurai said:

I just read that too... Maybe Gigabyte G1 version doesn't disappoint. I never owned one but people seem to like them.

Seems to be a general 1080 issue. Already going past 2ghz seems to be the limit, as the chip sis voltage limited.

[The Afro Samurai]

50 minutes ago, Notional said:

Seems to be a general 1080 issue. Already going past 2ghz seems to be the limit, as the chip sis voltage limited.

So it appears with aftermarket cards pick what brand you prefer and what suits your build scheme is what it's looking like.

[Notional]

1 minute ago, The Afro Samurai said:

So it appears with aftermarket cards pick what brand you prefer and what suits your build scheme is what it's looking like.

Pretty much. Cooling (as in noise) could also be a factor. I mean the good thing about the Asus card is the high boost rate, which is then guaranteed by Asus to be reachable.

[zMeul]

On 5/30/2016 at 10:46 PM, The Afro Samurai said:

I just read that too... Maybe Gigabyte G1 version doesn't disappoint. I never owned one but people seem to like them.

the G1 seems to be just a reference board, it only has a 8pin power

we'll see

 

I searched for a PCB shot, but it appears no one has one

Edited June 8, 2016 by zMeul

[Lays]

11 hours ago, zMeul said:

ASUS ROG STRIX GeForce GTX 1080 offers poor overclocking

 

 

 

2 hours ago, Notional said:

Seems to be a general 1080 issue. Already going past 2ghz seems to be the limit, as the chip sis voltage limited.

 

2 hours ago, The Afro Samurai said:

I just read that too... Maybe Gigabyte G1 version doesn't disappoint. I never owned one but people seem to like them.

 

 

It's present on all 1080 from the looks of it:

 

 

 

13 hours ago, zMeul said:

some dude got his hands on a STRIX:

 

 

something that bothers me, the "bracket" on the VRAM chips - what the hell, ASUS?! some chips have thermal pads, others not ..

I took a look at the GDDR5X product sheet and I didn't saw normal operating temperature, just the ranges 0-95deg

 

 

ps: 2 useless heat pipes

 

That plate seems to have thermal pads contacting every VRAM chip, then the pads ontop of the plate make contact with the heatsink.  So the heat radiating from the VRAM will be transfered to the plate, then the plate will transfer that to the heatsink.  It should be MORE than enough cooling for them.   Some cards don't even have faceplates or VRAM cooling, other than the active airflow they get from the fans on the cooler, and from what I've seen that's perfectly fine.

 

Also the heatpipes aren't completely useless just because they aren't touching the GPU die itself.  The heat from the other heatpipes will transfer into them, aiding in cooling. 

[zMeul]

28 minutes ago, Lays said:

thermal pads contacting every VRAM chip

except this one?!

 

 

now, let's discuss the heatpipes

• big gaps between the pipes

you do need to remember this is direct silicon contact

• two heatpipes don't contact the GPU

the extra heatpipes are basically useless since they do not transfer heat away from the GPU

removing them will matter not to the cooling performance, and I'll bet you the terminals will change very little if not at all if those two weren't present

you'd be better off with more fins that will help with heat dissipation

 

this is not the 1st time ASUS uses direct contact in this manner

GTX980Ti STRIX:

Spoiler

now, let's compare it to MSI's 980Ti:

Spoiler

the difference is that MSI uses a heat spreader to make perfect direct contact with the GPU silicon, and this making way better use of the heatpipes

this is the same method used by Noctua in their CPU coolers:

Spoiler

the difference between MSI's cooling and ASUS is a whopping 6 degrees:

Spoiler

EVGA uses the same design in the ACX coolers:

Spoiler

 

ASUS wants to use direct heatpipe contact? they should've done it like this, with no gaps

Spoiler

 

[RZeroX]

2 hours ago, zMeul said:

 

https://www.reddit.com/r/nvidia/comments/4lfjwc/special_handson_asus_rog_gtx_1080_strix/d3my63f?context=3

Quote

The bottom pipe is not automatically useless because it's not in contact with the gpu. It is still in physical contact with the other heat pipes. The zeroth law of thermodynamics states that if A is in thermal equilibrium with B, and B is in thermal equilibrium with C, then A must be in thermal equilibrium with C.

In this case, A is the "useless" heat pipe, B is the adjacent heat pipe (or all of the other heat pipes considered as a group), and C is the gpu die. So clearly we can see that the "useless" pipe can still be very effective at cooling the gpu even without being in physical contact with it.

The inefficiency of this system is that it's not always at equilibrium. That does not mean that it's super far from equilibrium. Some processes, especially heat transfer processes between high thermally conductive surfaces over a very short distance, can reach equilibrium very quickly.

In the end, I can guarantee that the engineers at Asus are smarter than that, and they wouldn't have designed a useless heat pipe in their product. I am confident that it is responsible for somewhere near 1/5 of the total heat transfer out of the gpu otherwise it simply would not be there.

This is getting a little traction, so I want to clarify. The heat pipe would be better utilized if it was in full contact with the die because it would reach equilibrium faster (it does not have to wait on B to heat up first), but it is still useful how it is now.

 

[zMeul]

1 hour ago, RZeroX said:

https://www.reddit.com/r/nvidia/comments/4lfjwc/special_handson_asus_rog_gtx_1080_strix/d3my63f?context=3

have you read my post?

then how the MSI cooling solution achieves better thermals with the same number of heatpipes? use of a transfer plate that makes direct and perfect contact with the GPU silicon

 

plus: the heatpipes do not cool the GPU, they transfer heat towards the fins; those two extra heat pipes do not provide surface area

 

the worse the gaps are in between those heatpipes, the worse the thermals will be

[zMeul]

EVGA Power Link:

Spoiler

 

 

neat idea 

[Lays]

56 minutes ago, zMeul said:

have you read my post?

then how the MSI cooling solution achieves better thermals with the same number of heatpipes? use of a transfer plate that makes direct and perfect contact with the GPU silicon

 

plus: the heatpipes do not cool the GPU, they transfer heat towards the fins; those two extra heat pipes do not provide surface area

 

the worse the gaps are in between those heatpipes, the worse the thermals will be

 

I'm not sure the integrated heatspreader is the reason for the better thermals, I think it's the lack of fins that are the issue, maybe even the fans?

 

My 980 ti Matrix shares the same "direct copper heatpipe" crap, but it's one of the coolest running 980 ti's on the market. It seems to make slightly better contact with the heatpipes, but it still barely touches the outer ones.

 

 

 

 

 

Here's some pictures of mine for reference:

 

 

Here's a random pic off the net, of where the paste is:

 

 

[Big-G]

28 minutes ago, zMeul said:

EVGA Power Link:

  Reveal hidden contents

 

 

neat idea 

Very nice, and very usable, mmmm... I like

 

9 minutes ago, Lays said:

 

I'm not sure the integrated heatspreader is the reason for the better thermals, I think it's the lack of fins that are the issue, maybe even the fans?

 

My 980 ti Matrix shares the same "direct copper heatpipe" crap, but it's one of the coolest running 980 ti's on the market. It seems to make slightly better contact with the heatpipes, but it still barely touches the outer ones.

...

No contact points with the VRAM or VRMs?

 

Ok, that's it... no Asus GPU Display card in my build ever

[Lays]

1 minute ago, GidonsClaw said:

Very nice, and very usable, mmmm... I like

 

No contact points with the VRAM or VRMs?

 

Ok, that's it... no Asus GPU Display card in my build ever

 

You don't need a contact point on the VRM when there's a heatsink on them...

 

VRAM legitimately doesn't get hot as long as theres air going to it. I'm not sure why everyone has some reasoning behind thinking it's going to be 50000c just because it doesn't get touched by a heatsink.  With the stock air cooler on, they get to about 50-65c in gaming loads. That's perfectly acceptable, VRM only hits about 50-60c.

 

 

 

Look at how I have my card right now, I just got done playing overwatch for multiple hours and I've just been browsing LTT for a bit now. Hottest VRAM temp was 54c, VRM only hit 44c.

 

 

 

Excuse the ridiculous amount of dust, I need to go buy those little air can thingies:

 

All I have is a 120mm fan pointed at the card at an angle, blowing fresh air onto it.

 

[DELTAprime]

I'm new to overclocking so please forgive the stupid question. But why could JayzTwoCents get a 2088mhz oc out of the reference board with custom cooler in his EVGA SC 1080 but Vidcardz couldn't even achieve that using the non-reference design on the Strix? I thought non-reference = better overclocking.

 

Might still go with a Strix OC just because all 1080's seem to be around 2000mhz and I know it's guaranteed to run no slower than 1936mhz in turbo boost.

[zMeul]

36 minutes ago, GidonsClaw said:

No contact points with the VRAM or VRMs?

the VRAM .. so and so, some do some don't 

 

the MOSFETs are assisted with a thermal pad:

[zMeul]

55 minutes ago, Lays said:

but it's one of the coolest running 980 ti's on the market

yes, I agree that MATRIX one has better thermals than STRIX, but not if compared to MSI's Lightning:

[Lays]

4 minutes ago, zMeul said:

yes, I agree that MATRIX one has better thermals than STRIX, but not if compared to MSI's Lightning:

The Matrix is much more user friendly though, in GURU3D's testing I think the Lightning ran like 2c hotter, but 2-3c is no big deal either way.   The Lightning doesn't have unlocked voltage without going to extreme measures, nor the level of support the Matrix does from the OC scene. The Lightning ONLY comes with Hynix memory unless you can somehow coax MSI into sending you a review sample with Samsung on it.

 

 

You can't do this on a Lightning without selling your soul to MSI and signing an NDA for MSI Afterburner Extreme, (If they even respond, they're super flaky about it)

 

On the Matrix, all you have to do is color on 2 copper pads with a pencil, and the card unlocks completely.

 

[Notional]

7 hours ago, RZeroX said:

The bottom pipe is not automatically useless because it's not in contact with the gpu. It is still in physical contact with the other heat pipes. The zeroth law of thermodynamics states that if A is in thermal equilibrium with B, and B is in thermal equilibrium with C, then A must be in thermal equilibrium with C.

In this case, A is the "useless" heat pipe, B is the adjacent heat pipe (or all of the other heat pipes considered as a group), and C is the gpu die. So clearly we can see that the "useless" pipe can still be very effective at cooling the gpu even without being in physical contact with it.

The inefficiency of this system is that it's not always at equilibrium. That does not mean that it's super far from equilibrium. Some processes, especially heat transfer processes between high thermally conductive surfaces over a very short distance, can reach equilibrium very quickly.

In the end, I can guarantee that the engineers at Asus are smarter than that, and they wouldn't have designed a useless heat pipe in their product. I am confident that it is responsible for somewhere near 1/5 of the total heat transfer out of the gpu otherwise it simply would not be there.

This is getting a little traction, so I want to clarify. The heat pipe would be better utilized if it was in full contact with the die because it would reach equilibrium faster (it does not have to wait on B to heat up first), but it is still useful how it is now.

https://www.reddit.com/r/nvidia/comments/4lfjwc/special_handson_asus_rog_gtx_1080_strix/d3my63f?context=3

 

There's a few problems here. I generally agree that adjacent heat pipes will help. However the contact surface between the heatpipes are so small, that there is no way whatsoever that you will ever have thermal equilibrium. So the basics of Zeroth's law of thermodynamics are not fulfilled and this not relevant. What you want to look at are the second law of thermodynamics, meaning a closed system will always move towards maximum entropy (equilibrium of temperature in this case). However since your GPU core will always be hotter than the surrounding area, this will never be achievable.

 

Secondly the gap between the pipes can trap heat, either in a pocket of air, of just in the thermal compound itself. Sure the second law of thermodynamics will kick in to some degree, but it's still not effective.

 

10 hours ago, zMeul said:

except this one?!

now, let's discuss the heatpipes

• big gaps between the pipes

you do need to remember this is direct silicon contact

• two heatpipes don't contact the GPU

the extra heatpipes are basically useless since they do not transfer heat away from the GPU

removing them will matter not to the cooling performance, and I'll bet you the terminals will change very little if not at all if those two weren't present

you'd be better off with more fins that will help with heat dissipation

 

this is not the 1st time ASUS uses direct contact in this manner

GTX980Ti STRIX:

  Reveal hidden contents

now, let's compare it to MSI's 980Ti:

  Reveal hidden contents

the difference is that MSI uses a heat spreader to make perfect direct contact with the GPU silicon, and this making way better use of the heatpipes

this is the same method used by Noctua in their CPU coolers:

  Reveal hidden contents

the difference between MSI's cooling and ASUS is a whopping 6 degrees:

  Reveal hidden contents

EVGA uses the same design in the ACX coolers:

  Reveal hidden contents

 

ASUS wants to use direct heatpipe contact? they should've done it like this, with no gaps

  Reveal hidden contents

 

I mostly agree with you, however you cannot just compare two graphs and claim one is better. We have no idea of the fans capabilities or how fast they are running. If the cooler one is much louder, it's just a question of raising the fan speed on the hotter card to reach the same result.

4 hours ago, GidonsClaw said:

Very nice, and very usable, mmmm... I like

 

No contact points with the VRAM or VRMs?

 

Ok, that's it... no Asus GPU Display card in my build ever

Ikr? My Asus 290 doesn't have any contact to the VRAM what so ever. If you look at it from an angle, you can see the VRAM dies directly. Pathetic. And the VRM runs stupidly hot.

Having the fins going the length of the card is dumb, as heat seems to be trapped between the fans. Having it go the width of the card seems to be a lot more effective (like EVGA does it).

5 hours ago, zMeul said:

EVGA Power Link:

  Reveal hidden contents

 

 

neat idea 

 

Sweet. I never understood why power connectors on GPU's ar eon the side anymore. All cases are made to support long GPU's now, but with them getting wider, it's just inconvenient. Also it makes cables a lot more cluttery. Just like at the R9 Nano, how nice it looks, with the power cable going out the back: