Overclock.net banner
2001 - 2020 of 2525 Posts

·
PC Evangelist
Joined
·
47,600 Posts
Has the TimeSpy issue been resolved with overclock? I am able to get 2850MHz with the older driver.
 

·
Registered
Joined
·
64 Posts
Has the TimeSpy issue been resolved with overclock? I am able to get 2850MHz with the older driver.
I think it still persists with the latest driver. I get best results with the latest driver though, around 21K @2620 core / 2110mhz vram. With the previous one I had 2656mhz as the max, but graphics score was lower. Which old driver are you using?
 

·
Registered
5800X | 6900XT | X570 Crosshair Impact | Teamgroup 3600Cl14 32GB | Cooling: 2x360 | 2x560 | Mora 420
Joined
·
26 Posts
I’m thinking of replaceming the pads of the asrock phantom gaming 6900xt bykski block and backplate with thermalright odyssey 12.8W/m-k pads when I receive it. Do you think 1.5mm pads should suffice?
 

·
PC Evangelist
Joined
·
47,600 Posts
I think it still persists with the latest driver. I get best results with the latest driver though, around 21K @2620 core / 2110mhz vram. With the previous one I had 2656mhz as the max, but graphics score was lower. Which old driver are you using?
I am using 21.3.1. I am getting 22.1K with 2850MHz Core. I will give the new driver a try and see if there is more performance even if I have to lower core clock.
 

·
Registered
Joined
·
969 Posts
Hi guys looking for some insight i have a powercolor RX 6800 XT Red Devil the card looked like it has a great cooler on it so i bought it hoping to get extra OC out of it but i can tell you that the heatsink is crap, stock voltage's and clocks it will hit about 106c on hot spot on time spy, so i tried to OC it using using MPT as taking the voltage down on the slider does nothing with an overclock of 2650mhz, it will still use all the full voltage range of 1150v without MPT, so set maximum GFX voltage in MPT @1120v and then i set the slider to 975v, it will pass time timespy extreme stress test all day long but on the timespy benchmark if i bench it maybe 6 times in a row it will pass but it might have an error occurred maybe once out of the 6 runs on the second graphic test my score is around 17000 with powerlimit set to 0 temps are around 97c, setting power limit at 2650mhz pushed temps up to around 106c so i don't bother using any powerlimit

But here is the thing that is confusing me, i just for fun set a min frequency of 2350 and max of 2450 in wattman then using the same MPT @1120V and leaving the slider at 975v and wow hot spot temp dropped into the low 80's and my timespy benchmark score actually went even higher than when i set it to 2650mhz and its pulling about 30 -40 watts less power, on the tomb raider benchmark i only lost one FPS @ 4k on my OLED tv compared to my 2650mhz, yet i lost 6 fps @ 1440p on my monitor which i don't care about anyhow as i only game on my OLED tv so i thought wow i will just use it at 2450mhz as it runs way cooler and and time spy bench has got better and i have only lost one FPS on Tomb raider @4k on my OLED tv

So i wondered if i even need to use MPT at all now, so i took a note of max vcore voltage in HWINFO64 after a run of a tomb raider benchmark using MPT set at maximum GFX voltage @1120v which was 0.954v then i deleted the SPPT in MPT rebooted and ran a tomb raider bench again and max vcore was indeed the same in HWINFO64 (0.954v) so i thought great i dont need MPT and i am happy to just leave here for 24/7, but here is the bit that REALLY is confusing me!

Now i have deleted the SPPT in MPT so the card is set back to max "standard" Vcore voltage of 1150v with the slider still set in wattman @ 975v i CAN NOT run a successful bench of timespy it will crash in seconds on the first graphic test which i have found in the past is the easier one of the two to pass!,

Yet if i use MPT again and set back max vcore voltage to 1120v with the slider STILL set at 975v it passes timespy bench every time and also the timespy extreme stress test every run with a 99.8% pass
yet with stock voltage set by NOT using MPT it will fail every time with the slider still set @ 975v, BTW there are no crashes in any game it seems only time spy likes MPT set at 1120v in MPT for it to pass any of its runs

So can anyone give me any insight as to why this could be happening? as in my mind what difference in having max voltage set back to standard when its not pulling anymore voltage in tomb raider benchmark or under load regardless of using MPT or not?, any insight would be helpful cheers

I am using drivers 21.4.1 and my system is a brand new build using a brand new Cosiar RM 850 X and a fresh install of windows
 

·
Registered
Joined
·
37 Posts
Hi guys looking for some insight i have a powercolor RX 6800 XT Red Devil the card looked like it has a great cooler on it so i bought it hoping to get extra OC out of it but i can tell you that the heatsink is crap, stock voltage's and clocks it will hit about 106c on hot spot on time spy, so i tried to OC it using using MPT as taking the voltage down on the slider does nothing with an overclock of 2650mhz, it will still use all the full voltage range of 1150v without MPT, so set maximum GFX voltage in MPT @1120v and then i set the slider to 975v, it will pass time timespy extreme stress test all day long but on the timespy benchmark if i bench it maybe 6 times in a row it will pass but it might have an error occurred maybe once out of the 6 runs on the second graphic test my score is around 17000 with powerlimit set to 0 temps are around 97c, setting power limit at 2650mhz pushed temps up to around 106c so i don't bother using any powerlimit
Lots to unpack here. First sharing my experience with 3DMark TimeSpy tests. I can pass TimeSpy stress test with certain clocks but fail in TimeSpy GT2 test run in a loop. It might pass for a setting for 30-40 loops and then fail at the 6th loop for the same setting the next run. In summary, various 3DMark tests are good for checking clock stability but they are not precise and consistent. A good OC will not fail TimeSpy stress test but a bad OC might or might not. Run TimeSpyGT2 in loop to get the most accurate result regarding clock stability as far as 3dMark/TimeSpy is concerned.

About your junction temps, the default fan curve is very conservative for all GPUs these days and biased towards being quiet rather than keeping the card cool. Adjust fan curves and you should see a lot of improvement
But here is the thing that is confusing me, i just for fun set a min frequency of 2350 and max of 2450 in wattman then using the same MPT @1120V and leaving the slider at 975v and wow hot spot temp dropped into the low 80's and my timespy benchmark score actually went even higher than when i set it to 2650mhz and its pulling about 30 -40 watts less power, on the tomb raider benchmark i only lost one FPS @ 4k on my OLED tv compared to my 2650mhz, yet i lost 6 fps @ 1440p on my monitor which i don't care about anyhow as i only game on my OLED tv so i thought wow i will just use it at 2450mhz as it runs way cooler and and time spy bench has got better and i have only lost one FPS on Tomb raider @4k on my OLED tv

So i wondered if i even need to use MPT at all now, so i took a note of max vcore voltage in HWINFO64 after a run of a tomb raider benchmark using MPT set at maximum GFX voltage @1120v which was 0.954v then i deleted the SPPT in MPT rebooted and ran a tomb raider bench again and max vcore was indeed the same in HWINFO64 (0.954v) so i thought great i dont need MPT and i am happy to just leave here for 24/7, but here is the bit that REALLY is confusing me!

Now i have deleted the SPPT in MPT so the card is set back to max "standard" Vcore voltage of 1150v with the slider still set in wattman @ 975v i CAN NOT run a successful bench of timespy it will crash in seconds on the first graphic test which i have found in the past is the easier one of the two to pass!,

Yet if i use MPT again and set back max vcore voltage to 1120v with the slider STILL set at 975v it passes timespy bench every time and also the timespy extreme stress test every run with a 99.8% pass
yet with stock voltage set by NOT using MPT it will fail every time with the slider still set @ 975v, BTW there are no crashes in any game it seems only time spy likes MPT set at 1120v in MPT for it to pass any of its runs

So can anyone give me any insight as to why this could be happening? as in my mind what difference in having max voltage set back to standard when its not pulling anymore voltage in tomb raider benchmark or under load regardless of using MPT or not?, any insight would be helpful cheers

I am using drivers 21.4.1 and my system is a brand new build using a brand new Cosiar RM 850 X and a fresh install of windows
Higher clocks require higher watts to work properly. I have seen temprature and power requirements increase drastically for more than 2600MHz in-game clocks. The closer you are to default frequencies the cooler the card. Instability in high core clocks due to lack of power (you have not mentioned if you increased power limits) can lead to lower minimum FPS and lower scores.
Not sure about what's happing with MPT and volategs but I don't see a harm in keeping MPT settings if it helps you get good temperatures.
 

·
Registered
Joined
·
969 Posts
Lots to unpack here. First sharing my experience with 3DMark TimeSpy tests. I can pass TimeSpy stress test with certain clocks but fail in TimeSpy GT2 test run in a loop. It might pass for a setting for 30-40 loops and then fail at the 6th loop for the same setting the next run. In summary, various 3DMark tests are good for checking clock stability but they are not precise and consistent. A good OC will not fail TimeSpy stress test but a bad OC might or might not. Run TimeSpyGT2 in loop to get the most accurate result regarding clock stability as far as 3dMark/TimeSpy is concerned.

About your junction temps, the default fan curve is very conservative for all GPUs these days and biased towards being quiet rather than keeping the card cool. Adjust fan curves and you should see a lot of improvement


Higher clocks require higher watts to work properly. I have seen temprature and power requirements increase drastically for more than 2600MHz in-game clocks. The closer you are to default frequencies the cooler the card. Instability in high core clocks due to lack of power (you have not mentioned if you increased power limits) can lead to lower minimum FPS and lower scores.
Not sure about what's happing with MPT and volategs but I don't see a harm in keeping MPT settings if it helps you get good temperatures.
Thanks the reply i don't touch the power limit as is only sends temps up for me and i get the best ever time spy extreme results at 2450mhz with 99.8% results however with my 2650mhz it was never that good and once it failed due to the fact it was under 97% but never a crash and that only happened once and the next run which was done straight after was fine again, as for fans i have them set at about 70% which is not to loud for me at all, i am indeed very happy to leave them at that due to much lower temps and not much of a change in loss of frames in games @4k compared to my 2650mhz OC, setting MPT back to stock voltage and the resulting instability is really confusing though and it was that that made me post to see if anyone could shed some insight

Furthermore i only found out about the SAM feature yesterday and turned that on in the bios and got 3 extra frames with tomb raider @ 4k and 9 extra FPS @1440p so that was a nice surprise

As for 3Dmark how do you loop a run of the second Time spy graphic test as that's what i think you are referring too when you say loop TimeSpyGT2 ?, i dont see an option for me to do that in my Advance 3dmark or am i looking in the wrong place?
 

·
Registered
Joined
·
92 Posts
Hi guys looking for some insight i have a powercolor RX 6800 XT Red Devil the card looked like it has a great cooler on it so i bought it hoping to get extra OC out of it but i can tell you that the heatsink is crap, stock voltage's and clocks it will hit about 106c on hot spot on time spy, so i tried to OC it using using MPT as taking the voltage down on the slider does nothing with an overclock of 2650mhz, it will still use all the full voltage range of 1150v without MPT, so set maximum GFX voltage in MPT @1120v and then i set the slider to 975v, it will pass time timespy extreme stress test all day long but on the timespy benchmark if i bench it maybe 6 times in a row it will pass but it might have an error occurred maybe once out of the 6 runs on the second graphic test my score is around 17000 with powerlimit set to 0 temps are around 97c, setting power limit at 2650mhz pushed temps up to around 106c so i don't bother using any powerlimit

But here is the thing that is confusing me, i just for fun set a min frequency of 2350 and max of 2450 in wattman then using the same MPT @1120V and leaving the slider at 975v and wow hot spot temp dropped into the low 80's and my timespy benchmark score actually went even higher than when i set it to 2650mhz and its pulling about 30 -40 watts less power, on the tomb raider benchmark i only lost one FPS @ 4k on my OLED tv compared to my 2650mhz, yet i lost 6 fps @ 1440p on my monitor which i don't care about anyhow as i only game on my OLED tv so i thought wow i will just use it at 2450mhz as it runs way cooler and and time spy bench has got better and i have only lost one FPS on Tomb raider @4k on my OLED tv

So i wondered if i even need to use MPT at all now, so i took a note of max vcore voltage in HWINFO64 after a run of a tomb raider benchmark using MPT set at maximum GFX voltage @1120v which was 0.954v then i deleted the SPPT in MPT rebooted and ran a tomb raider bench again and max vcore was indeed the same in HWINFO64 (0.954v) so i thought great i dont need MPT and i am happy to just leave here for 24/7, but here is the bit that REALLY is confusing me!

Now i have deleted the SPPT in MPT so the card is set back to max "standard" Vcore voltage of 1150v with the slider still set in wattman @ 975v i CAN NOT run a successful bench of timespy it will crash in seconds on the first graphic test which i have found in the past is the easier one of the two to pass!,

Yet if i use MPT again and set back max vcore voltage to 1120v with the slider STILL set at 975v it passes timespy bench every time and also the timespy extreme stress test every run with a 99.8% pass
yet with stock voltage set by NOT using MPT it will fail every time with the slider still set @ 975v, BTW there are no crashes in any game it seems only time spy likes MPT set at 1120v in MPT for it to pass any of its runs

So can anyone give me any insight as to why this could be happening? as in my mind what difference in having max voltage set back to standard when its not pulling anymore voltage in tomb raider benchmark or under load regardless of using MPT or not?, any insight would be helpful cheers

I am using drivers 21.4.1 and my system is a brand new build using a brand new Cosiar RM 850 X and a fresh install of windows
The stock Red Devil heatsinks are definitely crap. I bought a waterblock for my 6900 XT Red Devil Ultimate. It looks like they should be good in theory, but maybe there is some sort of mounting problem with them.
 

·
Registered
Joined
·
37 Posts
Thanks the reply i don't touch the power limit as is only sends temps up for me and i get the best ever time spy extreme results at 2450mhz with 99.8% results however with my 2650mhz it was never that good and once it failed due to the fact it was under 97% but never a crash and that only happened once and the next run which was done straight after was fine again, as for fans i have them set at about 70% which is not to loud for me at all, i am indeed very happy to leave them at that due to much lower temps and not much of a change in loss of frames in games @4k compared to my 2650mhz OC, setting MPT back to stock voltage and the resulting instability is really confusing though and it was that that made me post to see if anyone could shed some insight
The reason you failed TimeSpy stress test with 2650mhz is : required power was not being supplied which leads to --> clock instability which in turn leads to unstable fps --> which leads to TimeSpy marking your OC as fail.
Furthermore i only found out about the SAM feature yesterday and turned that on in the bios and got 3 extra frames with tomb raider @ 4k and 9 extra FPS @1440p so that was a nice surprise

As for 3Dmark how do you loop a run of the second Time spy graphic test as that's what i think you are referring too when you say loop TimeSpyGT2 ?, i dont see an option for me to do that in my Advance 3dmark or am i looking in the wrong place?
There you go.
2514549
 

·
Registered
Joined
·
43 Posts
I came into some cash the other day and decided to order an Asus ROG Strix 6900XT LC. Should be here in a couple days. Looking forward to see how it compares to the evga 2080ti hybrid it'll be replacing.

sam
 

·
Never Finished
Joined
·
2,346 Posts
So, I was having some issues with random shutdowns, probably OCP tripping because of the card. Setting the minimum clockspeed to 100mhz lower than what the max is set to has so far fixed it for me. Clockspeeds and voltages aren't jumping all over the place and I haven't run into a shutdown since. Might be worth a try for anyone else having power spike issues.
 
2001 - 2020 of 2525 Posts
Top