dslreports logo
site
 
    All Forums Hot Topics Gallery
spc

spacer




how-to block ads


Search Topic:
uniqs
2229
share rss forum feed

me1212

join:2008-11-20
Pleasant Hill, MO

(Probably late) Amd 12.11 drivers bring performance increase

AMD's 12.11 "Never Settle" Driver Performance, bring noticeable performance increases. With the new drivers the 7970 competes in performance with the 680, yet is still only priced about the same as the 670 about $100 less than the 680.

Haven't seen any benches for the 7950 vs the 660ti/670 but I assume the same holds true for them as well.

Very interesting, wonder how nvid is gonna respond.

»www.hardwarecanucks.com/forum/ha···nce.html



McBrain
BRB Face Melting

join:2010-05-06
Kalimdor
kudos:2

Anyone using these drivers yet?

Experts say I'll get 15% performance increase from my 7870 GHz Edition...


me1212

join:2008-11-20
Pleasant Hill, MO

I got them the night before I posted this. So far they have been pretty much flawless haven't found any games I cannot play at high-max settings.



Cthen

join:2004-08-01
Detroit, MI
Reviews:
·Verizon Wireless..
reply to me1212

One bug I found (I'll post the amd file name next time it happens) is that upon a system start up or reboot, these drivers will crash explorer.exe in Win7 64bit. ( I run only the 64bit version not sure about 32bit)

Upside is that all I have to do is click "Restart application" when it happens and all is good again.

All my games are set to max settings already but saw no further performance increase on 2 6970's in Crossfire.
--
"I like to refer to myself as an Adult Film Efficienato." - Stuart Bondek



McBrain
BRB Face Melting

join:2010-05-06
Kalimdor
kudos:2
reply to me1212

I'm pretty sure I downloaded the wrong driver last night = /, and forgot to download AMD 12.11 CAP2...

I downloaded and installed the 12.11 Beta driver for Win7 64 bit instead of the one for .NET4...When I logged into WoW, my GPU voltage increased from the stock .82v to 1.22v automatically...this caused my card to run at about 100c with fans blasting at 100%, just sitting idle in Orgrimmar. When I closed the game, the voltage decreased to .82 and temps dropped to my normal idle range of ~38c.

Would uninstalling/deleting the non-.NET4 driver and replacing it with the .NET4 edition as well as adding CAP2 fix this issue?
--
McBrain#1430

Name's Ash...Housewares.


me1212

join:2008-11-20
Pleasant Hill, MO

I don't know, I don't have the .net4 one. I didn't know what it did so I got the other.



McBrain
BRB Face Melting

join:2010-05-06
Kalimdor
kudos:2

Yeah, I don't like my card running at 100c just sitting idle in a game...I'm going to go with the .net4 driver and CAP2 and see where that takes me.

On the bright side, it seems to have fixed my issue of dropping HDMI signal randomly in game.


me1212

join:2008-11-20
Pleasant Hill, MO

Let us know how it goes man, hope it works for you.



McBrain
BRB Face Melting

join:2010-05-06
Kalimdor
kudos:2

Updated to .net4 driver and CAP2 last night...still ran at 1.22v in game and was consistently around 100c...

I decided I would play until I crashed and then use that as a reason to log early and go to bed...

I was stable all night...100c for 4 hours of raiding. So, I guess that's intended.

Next step is liquid cooling, I guess.
--
McBrain#1430

Name's Ash...Housewares.


me1212

join:2008-11-20
Pleasant Hill, MO

Which model do you have, and is it overclocked? It could be the temp sensors are messed up via some glitch in the beta driver.



Ghastlyone
Premium
join:2009-01-07
Las Vegas, NV
kudos:5
reply to me1212

For fuck sakes....100c?



Krisnatharok
Caveat Emptor
Premium
join:2009-02-11
Earth Orbit
kudos:12
reply to McBrain

You could reseat the HSF, or possible RMA the GPU for a faulty sensor.



McBrain
BRB Face Melting

join:2010-05-06
Kalimdor
kudos:2

RMA would be a pain in the ass, and a last resort (obviously). I'm PCSing to Okinawa shortly and I don't think MSI ships to FPO/AP addresses.

I've got the r7870 HAWK GHz Edition.

There must be a glitch in the sensor...there's no way I could run stable for 4 hours at 100c, right?
--
McBrain#1430

Name's Ash...Housewares.



Moos
Tequilablob
Premium
join:2008-12-11
Salt Lake City, UT
kudos:3
Reviews:
·Comcast

You can run stable at 100c for 4 hours. I wouldn't reccomend it longterm but it's not impossible. I'd expect you to start seeing some artifacting at a sustained heat that high though.

First thing i'd try is a complete uninstall and then Fresh install all video card drivers. There are a few methods if your not familiar with it, but it's more than just using the uninstaller. Everytime I have had a weird problem with AMD cards that seems to fix it, and if you did install wrong drivers then updated to new ones some remnants of the old ones may have been left behind. I can post a write-up of the steps I use if you want. You can find a ton of info on it via google too. I would probably also reseat and blow the dust out of the card to just out of habit.



McBrain
BRB Face Melting

join:2010-05-06
Kalimdor
kudos:2

The card was pulled out recently (Tuesday) and the old drivers were completely removed and uninstalled with the assistance of the Driver Fusion utility. There are no remnants of old drivers.

I guess I could delete/uninstall 12.11 beta and go back to 12.08, and see what that does.



Ghastlyone
Premium
join:2009-01-07
Las Vegas, NV
kudos:5

said by McBrain:

I guess I could delete/uninstall 12.11 beta and go back to 12.08, and see what that does.

Sometimes older drivers work better depending on your setup. I ran the Nvidia Beta drivers from June 2012 for awhile. GW2 ran beautiful for me. I remember upgrading to the newer official drivers a month or two later and the game ran like crap.

So I went back to my original ones.


Moos
Tequilablob
Premium
join:2008-12-11
Salt Lake City, UT
kudos:3
Reviews:
·Comcast
reply to McBrain

Have you tried to put your hand on/near your card when it's at 100 to see if it's actually that hot? That way you can rule out the sensor or not. That's boiling water temperature so you should be able to tell with your hand. It sounds like your heatsink needs to be re-pasted and re-seated. Or RMA.

The sensor's are inside the processor so if it's a bad sensor then you have a bad chip. In my opinion it's not acting like a bad sensor though. I'm betting somehow your heatsink shifted/lost contact.



Moos
Tequilablob
Premium
join:2008-12-11
Salt Lake City, UT
kudos:3
reply to McBrain

Try using something like GPU-Z to check your temps. You should have more then 1 sensor inside the chip, I think GPU-z will show you all of them. If all of them are up near 100 then it's not your sensor.



McBrain
BRB Face Melting

join:2010-05-06
Kalimdor
kudos:2

I'm monitoring it right now with CPUID HW monitor. I'm at 33c at idle currently...MSI Afterburner is also reading 33c.

I can dl GPU-Z and see if I get different readings.

When I pulled the card out to change PCIe slots, it wasn't handled rough or anything, and I didn't have any problems with removing it from the old slot and seating it in the new slot.

What makes me think it isn't the card itself is the fact that CPUID HW Monitor and MSI Afterburner read that there's a .4v power increase when I'm in game, which I think is directly related to the high temps.

I'm pretty sure it is the new driver, since I never had a boost in voltage previously.

I'm uninstalling 12.11 beta11 for 12.08 right now.
--
McBrain#1430

Name's Ash...Housewares.



pnjunction
Teksavvy Extreme
Premium
join:2008-01-24
Toronto, ON
kudos:1
reply to McBrain

said by McBrain:

there's no way I could run stable for 4 hours at 100c, right? =

Where I work we check our chips up to 110 or 125C, I have heard that GPUs are rated for 105-120C but I don't have an official number for the 7970.

My 6970 with the reference cooler gets up to 90-92C under load. I've thought about putting a better cooler on it, mostly for noise but also to squeeze some more power out of it.

I see your 7970 already has a big cooler on it, and those typically reduce the load temp to soemthing in the 60s or 70s, so it seems wrong that it should get so hot unless something has gone wrong. Have you checked the temp with different drivers under the same conditions?


McBrain
BRB Face Melting

join:2010-05-06
Kalimdor
kudos:2

Just tried 12.8 drivers...started good with temps in game sitting in the 75-85 range...after a minute or two voltage went up to and stayed at 1.22 and temps were back to 100c...I pulled off the side of the case and touched the card casing. It didn't feel hot enough to the touch for it to be 100c, and it wasn't blowing super heated air from the fans...I'm confused now.

And since I've closed the game I'm back to .82v and 34c idle temp.


me1212

join:2008-11-20
Pleasant Hill, MO

What all programs where you using to monitor the temps of the gpu?



McBrain
BRB Face Melting

join:2010-05-06
Kalimdor
kudos:2

I was just using CPUID Hardware Monitor and MSI Afterburner.

I am downloading GPU-Z now.



McBrain
BRB Face Melting

join:2010-05-06
Kalimdor
kudos:2
reply to me1212

Click for full size
Here's a SS of all three monitoring programs open...This was taken after about two minutes idle in Orgrimmar.

me1212

join:2008-11-20
Pleasant Hill, MO

85C isn't too bad actually, possibly a bit high for that low volts and only 1ghz on a 7000 series card, but not bad based on what I've read.

However, if I'm reading it right gpu-z says your vrm is 255C. That doesn't look right.



Ghastlyone
Premium
join:2009-01-07
Las Vegas, NV
kudos:5
reply to me1212

Damn...that's cooking dude. My shit doesn't even break a sweat playing World of Warcraft.

I can see maybe in a 25man raid during heavy combat. You shouldn't be running that hot just sitting in Org.

Are AMD core voltages supposed to run that high? I thought 1.1v was high, you're showing 1.22v


me1212

join:2008-11-20
Pleasant Hill, MO

I've seen amd cards run at 1.3 and get to the mid 80s. I keep min under 1.1 but still, I've seen a lot higher and about the same temp. the 255C on the vrm is what worries me.



McBrain
BRB Face Melting

join:2010-05-06
Kalimdor
kudos:2

I don't even know what VRM is. That's probably a bad thing. I saw that number and was concerned as well...it didn't budge the entire time I had GPU-Z open, so I'm not sure what to think of it.

If anything was 255c in there I'd be able to physically feel it when I put my hand on the card, wouldn't I?

Oh, I almost forgot. I tried to turn down the voltage in MSI Afterburner, but every time I touched it, it crashed.



Ghastlyone
Premium
join:2009-01-07
Las Vegas, NV
kudos:5

said by McBrain:

Oh, I almost forgot. I tried to turn down the voltage in MSI Afterburner, but every time I touched it, it crashed.

You'll have to adjust your clock speed down before you adjust voltage.

me1212

join:2008-11-20
Pleasant Hill, MO
reply to McBrain

A VRM is a voltage regulator modal. Its what feeds the voltage to your gpu, in this case.