770 throttling at 73C?

[mrmonmon2]

well back to 1384 core mem still same full round of valley and nothing happening

[Keudn]

well back to 1384 core mem still same full round of valley and nothing happening

I really don't know whats holding me back, I'm about ready to message one of the guys who made the custom BIOS I'm using and see if they can help...

[mrmonmon2]

so at 1.2 volts since it wont let me go higher, mem 8110 , core 1384, valley reported it at 1488, 61 max for about a second rest lower.

 

sop I have no problems, making one last vally run with fans at 100% 90% and mem at 8160

[mrmonmon2]

so my core clock has been maxing out at 1437 even though its set at 1384, 1.2 volts 8160 mem clock.

 

no problems here, or need to put ice in my water here now, try putting it back to the normal bios and ocing then.

 

anyways time to try and oc my processor more now since im in the cold,.

 

Good Luck Bud!!

[Keudn]

so my core clock has been maxing out at 1437 even though its set at 1384, 1.2 volts 8160 mem clock.

 

no problems here, or need to put ice in my water here now, try putting it back to the normal bios and ocing then.

 

anyways time to try and oc my processor more now since im in the cold,.

 

Good Luck Bud!!

Alright thanks for the help!

[mrmonmon2]

It was no problem, I've been meaning to see what my rig can do.

[Striker579]

Some common causes of gfx card performance capping are thermals and power. From what i read in this thread thermals are not an issue so it may revolve around the power. That said more is not always better, i suspect voltage reliability might be the culprit, if not other power issues include total power limits and operating voltage limits. For you i would check for clean power from the wall to your PSU, then to the card. Good luck and sorry if I waste your time due to my misdiagnosis.

[Keudn]

Some common causes of gfx card performance capping are thermals and power. From what i read in this thread thermals are not an issue so it may revolve around the power. That said more is not always better, i suspect voltage reliability might be the culprit, if not other power issues include total power limits and operating voltage limits. For you i would check for clean power from the wall to your PSU, then to the card. Good luck and sorry if I waste your time due to my misdiagnosis.

Any advice on how to test for clean power? That could very well be why the clock speed drops if the card can't keep a constant voltage when it is that high, that also would explain why the higher I set it the more it drops

[Striker579]

Any advice on how to test for clean power? That could very well be why the clock speed drops if the card can't keep a constant voltage when it is that high, that also would explain why the higher I set it the more it drops

 

My method involves the old trusty mulimeter, i dont know if there is a faster, better way but start from the wall outlet and look for steady and strong voltage over 110 VAC, unless you live in 220 land. Then move to your card and check the 6 and/or 8 pin connectors that you are using for the same clean power and then you can use the voltage hard points on your card if you have any or just use a software monitor (less accurate than putting a multimeter to the hard points) to see if your voltage is staying in an expected range at idle then again under load. 

 

If everything checks out until you throw a load on the card it may be a fault within the power supply, try switching to a different set of PCI-E cables or changing out at least one cable if possible as they might be on the same rail causing a problem if the psu is already starting to fail. 

 

2 more troubleshooting options, plug in to a different outlet in the room and change out the power cable from wall to your psu, that one is a hard one to catch and cheap/fast to rule out.

[Keudn]

My method involves the old trusty mulimeter, i dont know if there is a faster, better way but start from the wall outlet and look for steady and strong voltage over 110 VAC, unless you live in 220 land. Then move to your card and check the 6 and/or 8 pin connectors that you are using for the same clean power and then you can use the voltage hard points on your card if you have any or just use a software monitor (less accurate than putting a multimeter to the hard points) to see if your voltage is staying in an expected range at idle then again under load. 

 

If everything checks out until you throw a load on the card it may be a fault within the power supply, try switching to a different set of PCI-E cables or changing out at least one cable if possible as they might be on the same rail causing a problem if the psu is already starting to fail. 

 

2 more troubleshooting options, plug in to a different outlet in the room and change out the power cable from wall to your psu, that one is a hard one to catch and cheap/fast to rule out.

Will do when I get home, if im plugged into a surge strip could it affect anything? My two monitors, computer, desk lamp, and speakers are all on the same strip

 

Edit : Plugging into a separate wall outlet instead of my surge strip actually worked, there were 2 small brief dips but it could easily be from the GPU getting to 77C since my room isn't cold right now. I'm going to push the card as far as it will go tonight and see if I can keep any drops from occuring, but this may have actually fixed it!

[Striker579]

Will do when I get home, if im plugged into a surge strip could it affect anything? My two monitors, computer, desk lamp, and speakers are all on the same strip

 

Edit : Plugging into a separate wall outlet instead of my surge strip actually worked, there were 2 small brief dips but it could easily be from the GPU getting to 77C since my room isn't cold right now. I'm going to push the card as far as it will go tonight and see if I can keep any drops from occuring, but this may have actually fixed it!

 

I am glad to hear you are making headway, keep trying to eliminate variables like you said the power strip, they get old or fail eventually so that's a good one to eliminate. If your home has old wiring you might be drawing too many amps from the circuit you are on or might have clean power problems no matter what you do. 

[Keudn]

I am glad to hear you are making headway, keep trying to eliminate variables like you said the power strip, they get old or fail eventually so that's a good one to eliminate. If your home has old wiring you might be drawing too many amps from the circuit you are on or might have clean power problems no matter what you do. 

I figured out a solution (kind of). Using the separate wall outlet and keeping the GPU under 70C (which meant it was too cold in my room to be in it while benchmarking) kept it from underclocking, but I noticed something strange. It would work fine and keep a constant clock speed UNTIL it crashed, at that point it would underclock constantly. Restarting fixed this, until it crashed again, which while overclocking is pretty common. But after many restarts and freezing my fingers I got the benchmark speed I was looking for, and if I want to go farther I probably have to mod my BIOS as I think what is happening is the BIOS is hitting the TDP limit of the card and stopping it from going over. Anyway thank you for the help!