Spin Off: AT's Testing Methods & Uber Mode

[wand3r3r]

Just review in Umber Mode - Period - Then comment about the Noise.

I agree, since the card is loud, why take the quieter "loud" and show the weaker performance? Max it out and complain about the noise.

 

[Z15CAM]

Compare a 70 Challenger 440 RT to a STK 350ci 70 Z28. The Challenger came with everything thing on a TinFoil Frame but the 70 Z28 was Heavier and Engineered to be Modified.

There is no way I will consider a H2O Water Block for a Single GPU card. CPU Yes; but not for Video: Air Flow Case performance is critical if your after Hp-Video at the risk of 3rd Party Coolers which destroys AIB Air Cooled Solutions. Presently I look at a QNIX QX2560x1440 display and LOVE It - Considering a $30 GTX280 powers it.

What I want is an Arctic Accelero R9 290X-Xtreme III card but I have to build it.

------------------------

i7 2700k/ASUS P8Z68-V Pro Gen3/Corsair H110 AIO (running @ .996v/1600Mhz to 1.376v/4600Mhz 24/7 between 36 to 67C), 4 x's 4GB sticks of Samsung MV-3V4G3D-US DDR3 running at 1.34v/1866Mhz 9-9-9-24 1T with 4GB's assigned to a RAMDisk drive to handle Win7 64 Bit Sluff and negate writes to the SSD, Samsung 840 Pro 256 SSD, 2 x's WD5001AALS HDD's in Raid-0, 1 x's WD1002FAEX 1TB, ASUS DRW-24B1ST DVDRW, Hauppauge WinTV HVR-1250 PCIe, 2nd Intel NIC for Dual Networking, XFX 850W Pro Black Black Edition modular PSU, AMD Ref Gigabyte 1000Mhz R9 290X/Accelero Xtreme III, eVGA GTX 280 in reserve for Physic's-X support, Fractal Design ARC Midi R2 case, QX2510 Samsung PLS 2560x1440 res display at 120Hz.

TOTAL Investment on this Platform including 13% HST Tax Plus Shipping is approx $2,534.00 spent over a period of a year.

Aside from Mfgr's 10 to 15% price markup selling to the CDN Market the killer is the 13% HST Tax and cost of shipping in Canada. I pd $3000 for a 386-40 DX in 1994.

 

[Will Robinson]

Where were the howls of protest over the HD5870?

Courtesy of: The Tech Report

 

[Abwx]

This noise debate is moot because the measurements are flawed
as they do not take account of the cards influence on the rest
of the system noise.

In a real PC Gforce cards cooling design choice to blow most of its
hot air inside the PC case will induce higher speed of the case s fans,
wich will completely negate the noise advantage of the card when tested
in isolation , that is , not in real conditions , so much for the reviewers
ability to think as real engineers would do.

 

[96Firebird]

This noise debate is moot because the measurements are flawed
as they do not take account of the cards influence on the rest
of the system noise.

In a real PC Gforce cards cooling design choice to blow most of its
hot air inside the PC case will induce higher speed of the case s fans,
wich will completely negate the noise advantage of the card when tested
in isolation , that is , not in real conditions , so much for the reviewers
ability to think as real engineers would do.

You keep saying this, and you're still wrong. I'm not sure if you think the more times you post it people will start believing you or what, but good luck on trying!

 

[Abwx]

You keep saying this, and you're still wrong. I'm not sure if you think the more times you post it people will start believing you or what, but good luck on trying!

Ok , dont use fans to extract air from the case and then use a 200W card that get 70% of its air blowed inside the case ; that is 140W , the case/air contact surface allow for something like 0.3°C/W thermal resistance ,
after a few time the ambiant inside the case will be 42°C above room ambiant temperature.

Now if you card exhaust 80% of its hot air temp inside
the case will be only 12°C above room temp.

All in all cards that blow air inside the case will yield
a noiser PC case cooling wise so the measurements
of cards in isolation is a worthless metric.

 

[Erenhardt]

Please discuss difference between cooler designs elsewhere. Focus on testing here. Like, why 270 and 270X can't be overclocked?

 

[Vesku]

Overclocked the GTX 760:

http://www.anandtech.com/show/7103/nvidia-geforce-gtx-760-review/17

Anyone able to find the overclocking page in the Anandtech 270(X) review? My browser must be bugging out on me, can't see it.

 

[96Firebird]

Whoa, relax their lynch mob... Sometimes Ryan doesn't have time to get overclocking results included before the NDA lifts. This was the case with the 780 review, where the overclocking review was posted 5 days after the initial 780 review...

 

[bystander36]

Ok , dont use fans to extract air from the case and then use a 200W card that get 70% of its air blowed inside the case ; that is 140W , the case/air contact surface allow for something like 0.3°C/W thermal resistance ,
after a few time the ambiant inside the case will be 42°C above room ambiant temperature.

Now if you card exhaust 80% of its hot air temp inside
the case will be only 12°C above room temp.

All in all cards that blow air inside the case will yield
a noiser PC case cooling wise so the measurements
of cards in isolation is a worthless metric.

The design of the Titan and 780ti blowers does not really allow for any hot air to be blown into the case as the fan is still on the outside of the heat sink. The air hasn't been heated up before it could be blown into the case. I can only assume the open side is for intake, otherwise it doesn't serve a purpose. This isn't like the 690, where the fan is in the middle of the heat sink, and blows air out both directions.

 

[Abwx]

The design of the Titan and 780ti blowers does not really allow for any hot air to be blown into the case as the fan is still on the outside of the heat sink. The air hasn't been heated up before it could be blown into the case. I can only assume the open side is for intake, otherwise it doesn't serve a purpose.

I thought so but some people say it s otherwise , we ll see
once people who have the cards bring some clues , if that s
the case then , as pointed by a member in another thread ,
one is taking advantage to the fact that reviews are done
in open environment noise wise , the very polemical
subject currently.

 

[bystander36]

I thought so but some people say it s otherwise , we ll see
once people who have the cards bring some clues , if that s
the case then , as pointed by a member in another thread ,
one is taking advantage to the fact that reviews are done
in open environment noise wise , the very polemical
subject currently.

My other thought, it may just be so air doesn't build up on that side of the fan from other components which may not build up much heat, but still need to allow for some air circulation. It will be pretty insignificant, but if it can't escape, it builds up a bit.

 

[96Firebird]

My other thought, it may just be so air doesn't build up on that side of the fan from other components which may not build up much heat, but still need to allow for some air circulation. It will be pretty insignificant, but if it can't escape, it builds up a bit.

I was thinking about this too... If air builds up, it can negatively effect the fan. If the 780Ti cooler wasn't open on the end, the pressure difference would be there. I realize it wouldn't be sealed, but "sealed" enough. Could be one reason why the AMD cooler has trouble.

 

[Leadbox]

The design of the Titan and 780ti blowers does not really allow for any hot air to be blown into the case as the fan is still on the outside of the heat sink. The air hasn't been heated up before it could be blown into the case. I can only assume the open side is for intake, otherwise it doesn't serve a purpose. This isn't like the 690, where the fan is in the middle of the heat sink, and blows air out both directions.

How tight a fit is the shroud I wonder, cause i'm guessing that extra venting at the front there probably helps the acoustics of the card an appreciably amount

 

[Lepton87]

No, it really wouldn't. The titan shroud can dissipate over 250W while maintaining quiet sound levels. From my use of the 780 reference briefly, that shroud can be turned up to ridiculously high manual fan levels and still maintain good acoustics. The cut-off for AMD in terms of quietness right now is around 45%.
.

Not really IMHO it's getting unreasonably loud past 55%, while gaming I run the fan at 55% and that's enough to keep it from throttling but at default mode it throttles badly which makes your claims of throttling by no more then 26MHz ridiculous. At 55% after warm-up average clocks are 950MHz in Metro:LL and it goes all the way to 837MHz at default settings at times. Insignificant throttling my ass.

 

[blackened23]

Not really IMHO it's getting unreasonably loud past 55%, while gaming I run the fan at 55% and that's enough to keep it from throttling but at default mode it throttles badly which makes your claims of throttling by no more then 26MHz ridiculous. At 55% after warm-up average clocks are 950MHz in Metro:LL and it goes all the way to 837MHz at default settings at times. Insignificant throttling my ass.

You should get some screenshots up of this supposed throttling in Metro: LL, sounds like you have a defective card. And it definitely isn't loud at 55% fan. I don't ever throttle in Metro: LL.

 

[Vesku]

This is not a thread about reference cooling, pretty sure the mods have already explained this once.

 

[96Firebird]

And it definitely isn't loud at 55% fan.

Judging by his choice in CPU cooler and PSU, his definition of loud is different from most other desktop users.

 

[blackened23]

This is not a thread about reference cooling, pretty sure the mods have already explained this once.

Oh. Okay. So a thread that is about "quiet mode" and "uber mode" testing which takes a reference cooler and spins it at far different speeds is off topic? What does "QUIET MODE" refer to anyway? It sure as heck doesn't refer to the cooler being used does it? Silence versus uber mode testing has everything to do with AMD's reference cooler. If you don't see relevance to reference cooling and uber vs silent mode testing i'd say you're being disingenuous. I'm not really into playing games to silence opinions that I don't like, that isn't my thing. You can say whatever the heck you want in the 780ti thread.

Let's not kid ourselves here, though. When a reviewer has to pick "quiet" or "performance" on a GPU based on the significant compromises AMD made with the reference cooling design, yes, it has absolutely everything to do with the reference cooler, whether you want to admit it or not.

With all that said, i'll admit I pushed the issue too hard earlier. But, it is my opinion, and I say that as someone who has given AMD a ton of money in the past for GPUs. I just want a better product that doesn't ask me to compromise with a BIOS switch - I never had to do that with 7970s. I won't push the issue anymore, but basically, the reference cooler absolutely has relevance to this thread IMHO for reasons i've already mentioned.

 

[GaiaHunter]

It feels to me the only thing that is being silenced is the uber mode performance of the 290X.

 

[Lepton87]

You should get some screenshots up of this supposed throttling in Metro: LL, sounds like you have a defective card. And it definitely isn't loud at 55% fan. I don't ever throttle in Metro: LL.

You have a far superior cooler on your cards. I only increased the fan speed in 5% increments and 55% is the fastest fan speed I can stand, 60% is too much for my ears. It's definitely subjective, what one finds quiet and reasonable will be too loud for somebody else so not much point arguing where it gets loud. For my ears that threshold is past 55%.

BTW. I don't care about stock settings, all I need to do to maintain a quite consistent turbo speed is increase the thermal threshold and set the fan speed at 50% which don't bother me at all since my other fans are already just as loud. I probably don't need to even increase temp target at that fan speed but I did just in case.

 

[Will Robinson]

Oh. Okay. So a thread that is about "quiet mode" and "uber mode" testing which takes a reference cooler and spins it at far different speeds is off topic? What does "QUIET MODE" refer to anyway? It sure as heck doesn't refer to the cooler being used does it? Silence versus uber mode testing has everything to do with AMD's reference cooler. If you don't see relevance to reference cooling and uber vs silent mode testing i'd say you're being disingenuous. I'm not really into playing games to silence opinions that I don't like, that isn't my thing. You can say whatever the heck you want in the 780ti thread.

Let's not kid ourselves here, though. When a reviewer has to pick "quiet" or "performance" on a GPU based on the significant compromises AMD made with the reference cooling design, yes, it has absolutely everything to do with the reference cooler, whether you want to admit it or not.

With all that said, i'll admit I pushed the issue too hard earlier. But, it is my opinion, and I say that as someone who has given AMD a ton of money in the past for GPUs. I just want a better product that doesn't ask me to compromise with a BIOS switch - I never had to do that with 7970s. I won't push the issue anymore, but basically, the reference cooler absolutely has relevance to this thread IMHO for reasons i've already mentioned.

AT VC&G user reports are now coming in about actual noise levels...see the R9 thread...:thumbsup:

 

[Will Robinson]

It feels to me the only thing that is being silenced is the uber mode performance of the 290X.

Its not that its too loud...its that its too fast.

 

[DooKey]

Not really IMHO it's getting unreasonably loud past 55%, while gaming I run the fan at 55% and that's enough to keep it from throttling but at default mode it throttles badly which makes your claims of throttling by no more then 26MHz ridiculous. At 55% after warm-up average clocks are 950MHz in Metro:LL and it goes all the way to 837MHz at default settings at times. Insignificant throttling my ass.

You have a bad card. Neither of my Titans exhibited that kind of behavior. At 55% my Titans were quieter than my current 290s.

 

[Abwx]

It feels to me the only thing that is being silenced is the uber mode performance of the 290X.

And its ability to work on any ATX system , dixit AMD, wich means that its cooling wich blow hot air only out of the case is fully compatible with an ATX form factor that has no fans for case cooling, wich is not true for competing products as well as for open coolers of all kind.