Originally Posted by PrettyDancer
So did you guys noticed a decrease in the max over clock you can do with the latest drivers ? I used to be able to run 1520 / 8000 before for an hour of FFXV in 4K and now I'm down to 1505 / 8000. Max temps 72 (dunno if this is good ?)
I'm a bit disappointed with my card, given it has 79,8% ASIC quality...
Drivers lately haven't gone through the greatest QC and sometimes they can absolutely "make or break" your experience. There are people rolling back all the time and stopped even wanting to try new drivers. Having said that, the latest drivers v365.19 have been pretty stable for me I was able to make the 1640Mhz SLI O/C video with this latest driver which was the same maximum speed I was able to get with the older driver I had used when I first overclocked that high. I don't think I am experiencing lower overclocking with this driver anyway at least @ 4K (I mostly test @ 4K).
Your ASIC% is excellent but that's indirectly, not directly related to your maximum stable overclock. If you were stable at a higher speed before, you have to also ensure you are comparing the same benchmark/game using identical settings. The maximum overclock you can reach is going to always be specific to that software when you "dial it in". For example I can run Catzilla 4K @ 1640Mhz but the maximum I could run FFXIV is closer to 1610Mhz.. GTA V / Witcher 3 is pretty demanding and I might have to drop to 1580Mhz.
Some people prefer to NOT dial things in.. they just want a single speed that will work with everything.. for that you have to find your maximum stable overclock under the worst case scenario. A hot room, the most demanding GPU workload you can throw at it (please don't use furmark/kombuster they are really only good for heat and power testing IMO) over an extended period.
The way I test stability is to strap into the gaming chair, open my most graphically intensive game, max out ALL the settings to make it hard for my GPUs and just play the game and have fun. Keep OSD temp monitoring on. Expect some crashes and possibly artifacts. If they happen, just reboot, lower the overclock -10Mhz (I think you really need to test in 10Mhz increments, 5Mhz is just a waste of time) and strap yourself back in for some more gaming. When you've been able to have a good 1-2 hour intense gaming session without issue you've found the stable O/C.
It's always possible some new benchmark or game will push our GPUs harder than ever and we may need to back down the overclock even more.. it's part of what we have to do when overclocking
Originally Posted by harney
Re this 980ti g1
At stock bios core B 1491 8000 mem but with with V rel
With bios mod Core B 1555+ 8000mem no caps prefs still in testing phase all this as this is a new 2nd hand card but thus far seems to be handling very well.
will be pushing this more today.
Its performing much better than the other G1 i had which had a way much higher ASIC ... had the same thing with ASIC before with a few 970's i was system building with ....turned out 1 of the lower ASIC one's performed better ...
The other g1 i had was 76 ASIC but did not do well as this one with only an ASIC of 65.7
So i take all this ASIC thing with a grain of salt...
Thanks for the feedback.
ASIC% is just the voltage that is required to reach the maximum overclock. The value of that maximum overclock is silicon lottery.
Originally Posted by PrettyDancer
Laithan, I have a questio regarding my G1 980Ti. My 1500/8000 O/C crashed after 2H of FFXV. When I'm putting 1525/8000, it crashes in 10 mins. I'm trying 1495 now. If it crashes 2H into a bench, can it be that the issue is a PSU issue ?
No I think it is heat. Especially FFXIV in a loop (OMG) I hope you are monitoring temps closely.
The stress testing that you are doing is pretty brutal but it does bring up a good point. A 980Ti on AIR is a challenge to keep happy. As good as the cooler is over a long period of time the ambient temps in the room are going to increase (the heat output from these cards is significant) and if you are looping tests you're going to need to compensate for all the additional heat. The case would need to be open (mandatory), supplemental fans on the GPU backplate (Ghetto tech) and/or other fans dedicated to keep the GPU(s) cool. Fan blowing in the room and airflow (open doors in the room). Otherwise to be honest, I would expect it
to crash.Also JIC important tip: ALWAYS reboot after a crash
. As tempting as it is to "keep on going" it's not a good idea and you can expect weird problems. A simple reboot always clears everything up.