R9 290x black screen

What software do you use to bump up memory/core voltage and clock speed? Also, what is a safe voltage for memory and core?
 
Afterburner.

There is no programable voltage controller for memory on the 290/290x, only for Core.

Voltage is no longer static, it changes, so offset is how you change it.

100mv max.
 
Just a follow up post, I'm pretty much decided that my newly received RMA replacement card is functioning great. I have not tried the older driver but the newer one that addresses black screens has been doing its job (although I did not test with older drivers to induce the black screens, just happy it's not happening anymore). Now if I could just keep this fan noise under control.. anyone have any suggestions? Maybe you could post a screenshot of your custom fan profile in afterburner? I'd have ordered an AC Xtreme3 but they're out of stock everywhere I've checked :-(
 
A custom fan profile isn't going to do much for you. When the fan speed throttles up under load, it is needing as much air flow it can get since it hovers around 94c. You will either have to deal with it or go aftermarket.
 
Ihaveworms said:
A custom fan profile isn't going to do much for you. When the fan speed throttles up under load, it is needing as much air flow it can get since it hovers around 94c. You will either have to deal with it or go aftermarket.
Click to expand...

Since those Accelero's were out of stock everywhere, I picked up a Gelid cooler and some AS thermal adhesive..will post back with impressions. :cool:
 
Update- 2 days so far no more black screens, after installing the new beta drivers, able to run all settings now in ff14, quality has dropped a bit but at this point just happy it's working,
 
The Mac said:
Afterburner.

There is no programable voltage controller for memory on the 290/290x, only for Core.

Voltage is no longer static, it changes, so offset is how you change it.

100mv max.
Click to expand...

The Aux offset will up the memory voltage, as confirmed by GPU-Z's VDDCI voltage reading. Unfortunately +/- 100mV isn't enough to be all that useful.
 
Haven't ran into any black screens since downloading the new beta drivers either. Been playing BF4 and FFXIV
 
SicKlown42012 said:
The Aux offset will up the memory voltage, as confirmed by GPU-Z's VDDCI voltage reading. Unfortunately +/- 100mV isn't enough to be all that useful.
Click to expand...

You are assuming VDDCI is the memory voltage. Unwinder has said there is no memory voltage control on these cards.
 
Forceman said:
You are assuming VDDCI is the memory voltage. Unwinder has said there is no memory voltage control on these cards.
Click to expand...

Then what does the Aux actually control? AMD has implied that the IR controller is functionally similar to the CHiL used on previous designs so it doesn't make a lot of sense that they'd remove memory voltage and use it for another part of the chip without documenting it.
 
Ihaveworms said:
A custom fan profile isn't going to do much for you. When the fan speed throttles up under load, it is needing as much air flow it can get since it hovers around 94c. You will either have to deal with it or go aftermarket.
Click to expand...

Just installed the Gelid cooler, I would suggest following the pictures/instructions over at this thread- on overclock.net.

Instead of using the crappy tape adhesive for the RAM I purchased a vial of AS Thermal Adhesive and used those on the RAM and VRM2 heatsinks.

So far under furmark load I've seen a max of 72C at 74% fan, I pushed the fans to 100% and you can't even hear the difference. The quality of the parts is questionable but it did get the job done, will probably look at purchasing some upgraded fans for the heatsink in the near future.
 
have a few slight flickers nothing really bad, I can just notice it on white, but it seems to dim and pop back to white. otherwise still no black screens, so I'm happy
 
itsbranden said:
Just installed the Gelid cooler, I would suggest following the pictures/instructions over at this thread- on overclock.net.

Instead of using the crappy tape adhesive for the RAM I purchased a vial of AS Thermal Adhesive and used those on the RAM and VRM2 heatsinks.

So far under furmark load I've seen a max of 72C at 74% fan, I pushed the fans to 100% and you can't even hear the difference. The quality of the parts is questionable but it did get the job done, will probably look at purchasing some upgraded fans for the heatsink in the near future.
Click to expand...

What about vrm temps on load?
 
SicKlown42012 said:
Then what does the Aux actually control? AMD has implied that the IR controller is functionally similar to the CHiL used on previous designs so it doesn't make a lot of sense that they'd remove memory voltage and use it for another part of the chip without documenting it.
Click to expand...

I don't know. I heard somewhere it is PLL, but I've never seen it documented. But GDDR5 doesn't run at 1.0V, which is what my VDDCI is, so it's definitely not memory voltage.
 
did anyone here check GPU usage in BF4 with 290/290x?

I am getting GPU usage all over the place, its never steady in BF4 for what ever reason, every second it goes up and down, the graph looks really bad, it drops from top to nearly bottom.

Using MSI Afterburner Beta 17.
 
TroyX said:
did anyone here check GPU usage in BF4 with 290/290x?

I am getting GPU usage all over the place, its never steady in BF4 for what ever reason, every second it goes up and down, the graph looks really bad, it drops from top to nearly bottom.

Using MSI Afterburner Beta 17.
Click to expand...

It's a problem with Afterburner. needs to be updated to work right with the new cards.
 
Its not a problem with Afterburner.
Afterburner is reporting the data the driver returns, it doesnt perform any calculations to get GPU use data.
 
TroyX said:
ok, thanks.
Click to expand...

fwiw, I've been investigating this in other games and some seem to throttle GPU use, but not the clocks so much, and clocking the GPU clocks higher seems to make the GPU use reduce to compensate so performance stays about the same.
Yet other games can max out the GPU and make good use of an overclock.

I was hoping it was the power limit setting or temp related, but I maxed the power slider to +50% and bought an aftermarket cooler, load temps are under 60C always, mostly around 50C.
It makes no difference.

For example, COD Ghosts, Dirt 2 and Unigine Heaven max GPU use.
But Crysis Warhead, X3:TC and Metro LL dont.

This first came to my attention when I noticed things that throttle GPU use arent much faster on my 290 (running as a 290x) than they were on my old 20% clocked GTX580.
CPU cores are not being maxed either.
VRM temps are around 50C, I can make them go to 75C ish with +100mV on the core.

Peculiar.
 
The Mac said:
What about vrm temps on load?
Click to expand...

VRM Temps during real-world gaming (BF4, DayZ, Arma3, etc) will reach around 62-65C MAX. Running furmark VRM1 maxes out at ~100C. VRM2 is usually lower at around 55-60C during benchmarks. I have yet to see the core drop below 940Mhz which is AWESOME, before switching to the aftermarket cooler, I'd see dips down to around 800Mhz or more.

So I'm pretty pleased with this thing, I will say though, the VRM1 temps are kinda high, I put the long VRM attachment thingy that came with the cooler on it, but noticed the mounting holes didn't perfectly line up with the board :mad:. I ended up rigging it up with some copper speakerwire threaded through the holes and clamped down then tied off. It's held on pretty sturdy but if anyone has any ideas for what i could use instead that would be great. I was going to use the narrow heatsink's that I used on VRM2 but I didn't have enough to run the length of VRM1.
 
Nenu said:
Its not a problem with Afterburner.
Afterburner is reporting the data the driver returns, it doesnt perform any calculations to get GPU use data.
Click to expand...

I was keeping it simple. It's not a problem with the cards, it's either a problem with the way afterburner reads the data or the way the drivers send the data. The polling interval is all wrong. If you put the GPU usage sensor to the top of the list in afterburner it minimises the problem.

Some people say it's afterburner, others say its the AMD drivers.
 
Its the actual divers

If the gpu usage is polled too quickly by the driver after another sensor, it returns garbage.

I assume amd will fix it at some point
 
Recently started getting BF4 stopped responding error, i wasnt getting this before, anyone else?
 
I think afterburner is not the problem, It is reporting back 100% usage in Heaven benchmark, and other GPU benchmarks, however in BF4 the GPU usage is all over the place.

It is dropping from 90's to 40's every two seconds, the graph looks like a rollercoaster.

Using 4770k at 4.2ghz, 16gb Ram.

It could be the game or the AMD drivers, or maybe both.
 
Yep, still getting BSODs on my end even with the latest drivers, albeit only periodically--and less frequent than the 9.2 beta drivers.

Went about a week after installing them before I had another BSOD though, and uninstalling/re-installing isn't making any difference.

I'm thinking of doing an RMA because it looks like i've got a faulty card.

The Mac said:
its the drivers, see my prior post.

its been confirmed.
Click to expand...

You sure? I couldn't find which earlier post you were referring to.

If it's purely software related has there been any news from AMD on an ETA for a driver update that'll solve the issue, or is it still "work in progress" at this point?
 
itsbranden said:
VRM Temps during real-world gaming (BF4, DayZ, Arma3, etc) will reach around 62-65C MAX. Running furmark VRM1 maxes out at ~100C. VRM2 is usually lower at around 55-60C during benchmarks. I have yet to see the core drop below 940Mhz which is AWESOME, before switching to the aftermarket cooler, I'd see dips down to around 800Mhz or more.

So I'm pretty pleased with this thing, I will say though, the VRM1 temps are kinda high, I put the long VRM attachment thingy that came with the cooler on it, but noticed the mounting holes didn't perfectly line up with the board :mad:. I ended up rigging it up with some copper speakerwire threaded through the holes and clamped down then tied off. It's held on pretty sturdy but if anyone has any ideas for what i could use instead that would be great. I was going to use the narrow heatsink's that I used on VRM2 but I didn't have enough to run the length of VRM1.
Click to expand...

I am not sure why you didn't just use the same AS Thermal Adhesive that you said you used on the Ram and VRM2 heatsinks..If you mix it properly, and then apply it in the proper timeframe, that stuff will work wonders..I used it on my old 4850 with an Arctic Cooling S1 and it a wonderful job holding the heavy square 'sinks I used on the VRMs.
 
ccityinstaller said:
I am not sure why you didn't just use the same AS Thermal Adhesive that you said you used on the Ram and VRM2 heatsinks..If you mix it properly, and then apply it in the proper timeframe, that stuff will work wonders..I used it on my old 4850 with an Arctic Cooling S1 and it a wonderful job holding the heavy square 'sinks I used on the VRMs.
Click to expand...

The problem is that the long VRM heatsink that came with the Gelid has a sort of indentation and when layed out the heatsink doesn't make contact with the chipset at all.
It's meant to be used with the thin thermal pad that it came with from the looks of it.
I'm going to take the card out later today and inspect it, maybe I can get some thin screws or something to fit through the holes. If that doesn't work I was going to buy those tiny copper heatsinks from Enzotech and attach them with AC Thermal Adhesive.
I am really happy with the price and performance of this mod overall, but a little disappointed that I spent this much on a video card and was forced to look for alternative cooling.
 
You know everytime I start to type a post about my frustrations and aggrevations with this card I just stop and erase it because no one really cares. AMD has my money and I can't get a refund from Newegg. XFX says they can't do anything for me, I have to wait for new drivers.

I've come to the conclusion i'll just take my $475 video card and aftermarket cooler (because they can't get that right either) investment and let it sit on my desk and use my old video card until these people figure out how to fix this POS.

I will NEVER purchase another AMD product again. I learned my lesson the hardrway about AMD processors and now I've learned the hardway about their video cards.
 
Contact AMD for support.
They may want to know that Newegg arent helping.
 
Kc0r8y said:
You know everytime I start to type a post about my frustrations and aggrevations with this card I just stop and erase it because no one really cares. AMD has my money and I can't get a refund from Newegg. XFX says they can't do anything for me, I have to wait for new drivers.

I've come to the conclusion i'll just take my $475 video card and aftermarket cooler (because they can't get that right either) investment and let it sit on my desk and use my old video card until these people figure out how to fix this POS.

I will NEVER purchase another AMD product again. I learned my lesson the hardrway about AMD processors and now I've learned the hardway about their video cards.
Click to expand...

Do a credit card charge-back. If you threaten newegg with a charge-back, chances are they will change their tune - if you approach the CC company with a report of a defective and non working unit, they will take care of you. As far as newegg and AMD? Fuck em.

Trust me on this. Call newegg, threaten a charge-back, and if you do that chances are newegg will change their tune. Because newegg will take a huge hit with any charge back and you'll 99% get your money back from the credit card company; newegg loses more money with a charge-back than they would with a RMA. Even if newegg stands their ground, proceed with the CC charge back and say whatever it takes to get your shit taken care of. Having a non working card is bullshit, and "wait for another driver" isn't an acceptable answer.

You can more or less do the same thing if you paid by paypal, paypal is generally very buyer friendly if you are verified for defective purchases.
 
Kc0r8y said:
I tried, after being on hold for 40 minutes and no one ever answered, I again gave up on them.
Click to expand...

What's the issue here? Why isn't Newegg honoring your RMA replacement? Or were you looking for an RMA refund? I know when I purchased mine it said in big lettering 'Eligible for RMA replacement only'... I RMA'd my card with Newegg with 0 problems (aside from a wait time of around a week or so). My new card performs great and doesn't have the black screen issues.
 
solved my issue, it was just overheating since the fan would never go above 40% even at 93c
so afterburner saved my life here, doesn't get past 65c now, and this has solved my black screens,
 
I get black screens from the following events.

Running Unigine Heaven Full screen, everything else is fine full screen and Heaven runs great windowed.
Memory voltage is too low. (memory voltage increases with VGPU from the results I found, so you may need slightly higher VGPU to fix them).
Memory overclock is too high will definitely give me a black screen.
I crash it when overclocking the GPU, although often it will recover to desktop.

Other than the above, I never get black screens.
 
You must log in or register to reply here.
Back
Top