Topic: GPU usage problem (Read 1375 times)
well never really undervolted, tryed now and nothing happened, so nope back to normal
Have you undervolted the cards? I was having the exact same problem when I undervolted my 5870s to much. The max hashrate was still good, but the gpu load was fluctuating like crazy.
I thought it might be the VRAM temps causing it. mine sits at about 110C. Will try undervolting to see what happens, although i have to wait til tomorrow afternoon to see any results. Will report back. Oh and theres not much point setting your fan above 85%, produces very little extra airflow, just kills the fans faster.
its at 13 , how much lower should it go?
Change the intensity to a lower setting and see what it does to the temperatures, this will lower your hashrate but help you figure out if the VRM temperatures are throttling your GPUs
theres a 200MM fan ontop of the GPUS and a HUGE box fan blowing on to them too
I think 112 and 114 is too hot and the GPU will throttle, can you increase the airflow? Or try setting the fans to 100% manually if they're not already at 100%. Can you add a fan to blow the air through the cards out?
Tab "Sensors", scroll down, look for "VRM temperature 1 and 2", all the way at the bottom.
Mine are between 90-105 degrees. Very hot, but they're designed to withstand somewhere between 125 and 140 degrees.
Mine are between 90-105 degrees. Very hot, but they're designed to withstand somewhere between 125 and 140 degrees.
I have a Sapphire R9 280x Vapor X with the problem described in the OP and this was the cause. When VRM1 reached 110 deg C the card would throttle the GPU usage down to 64%. I undervolted the card with Sapphire TRIXX (1087 VDD on that program), used the '--gpu-powertune -20' flag in cgminer, and also used '--gpu-fan 63' instead of '--auto-fan' to fix the overheating. Also make sure you have good airflow over the GPU.
I also set the fan to 100 and still no difference and the GPU has good airflow
Sapphire TRIXX utility, not sure if its also possible in MSI Afterburner. Some of the R9 280xs are voltage locked at the BIOS, so you'll have to check that on your card also. I adjust the voltage in TRIXX, then make sure the VDDC voltage drops when looking at it in GPU-Z.
https://www.sapphireselectclub.com/ssc/TriXX/TriXX.aspx
https://www.sapphireselectclub.com/ssc/TriXX/TriXX.aspx
how did you undervolt?
Tab "Sensors", scroll down, look for "VRM temperature 1 and 2", all the way at the bottom.
Mine are between 90-105 degrees. Very hot, but they're designed to withstand somewhere between 125 and 140 degrees.
Mine are between 90-105 degrees. Very hot, but they're designed to withstand somewhere between 125 and 140 degrees.
I have a Sapphire R9 280x Vapor X with the problem described in the OP and this was the cause. When VRM1 reached 110 deg C the card would throttle the GPU usage down to 64%. I undervolted the card with Sapphire TRIXX (1087 VDD on that program), used the '--gpu-powertune -20' flag in cgminer, and also used '--gpu-fan 63' instead of '--auto-fan' to fix the overheating. Also make sure you have good airflow over the GPU.
yup im pretty sure they're overheated :/ it's at 112 and 114
Tab "Sensors", scroll down, look for "VRM temperature 1 and 2", all the way at the bottom.
Mine are between 90-105 degrees. Very hot, but they're designed to withstand somewhere between 125 and 140 degrees.
Mine are between 90-105 degrees. Very hot, but they're designed to withstand somewhere between 125 and 140 degrees.
Is it possible (you can see this with the GPU-Z program) that your Voltage regulators are overheating. Normal GPU sensors don't show these values. You can view them with: http://www.techpowerup.com/gpuz/
I have gpuz , but where do I check that my voltage regulators are overheating?
Is it possible (you can see this with the GPU-Z program) that your Voltage regulators are overheating. Normal GPU sensors don't show these values. You can view them with: http://www.techpowerup.com/gpuz/
nevermind figured it out, well I've officially tried EVERYTHING. nothing's working. I guess that's it :/
If you're using the MSI Afterburner program, there's a setting that synchronizes, clock speed and voltage settings accross all of your cards. Perhaps you can try that?
http://event.msi.com/vga/afterburner/download.htm
http://event.msi.com/vga/afterburner/download.htm
ya I'm doing that as we speak though there's something happening, everytime I save and apply my settings (highering my power usage) it saves but then after I run cgminer the settings reset. Why?
If you're using the MSI Afterburner program, there's a setting that synchronizes, clock speed and voltage settings accross all of your cards. Perhaps you can try that?
http://event.msi.com/vga/afterburner/download.htm
http://event.msi.com/vga/afterburner/download.htm
ya ik nothing worked, did it work for you?
--gpu-engine gpu1,gpu2
e.g
--gpu-engine 1080,1080
the -d 1 is telling cgminer to only use gpu 1, so get rd of that or change it to -d 1,2 (depending on your setup of course)
e.g
--gpu-engine 1080,1080
the -d 1 is telling cgminer to only use gpu 1, so get rd of that or change it to -d 1,2 (depending on your setup of course)
Jump to: