Thats is drivers related, not power management itself. Some monitoring softwares or third party drivers can interfere too. I'd start the diagnosys of the problem by changing drivers and see if happens with also an other version.
You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
Guys, just throwing my 2 cents here. I got the game last week-end and was concerned about the overclock on my GTX 970 overheating in this game after reading this thread and the warning of reports from Nvidia tweaking guide. My GPU is overclocked at the BIOS level by about 15%, so a significant overlock, but the fan profile has been set up accordingly with the BIOS tweaker.
I can report that after a couple days of intense playing, the temperature never reached higher than 80C (and it's fairly warm weather here) and I did not experience any crash or driver issue. I am not suggesting that the issue is not there, but just that it really doesn't seem to affect everyone. The overclock in itself is unlikely to cause the issue, but the overall temperature rise might be. However, it should be mentioned that if the game really is pushing your GPU to unconventional level of temperature, I am sorry to say this to you but your overclock is the problem. No software can push your hardware beyond the max limit of what it should be able to tolerate, it's just not how electronic works. If you overheat, you have an unsuitable / unstable overclock or lack appropriate cooling. Saying no other games does that to your GPU means only that no other games pushes your GPU to the max. Third party tools should be able to help you test your overclock. I can personally say that EVGA OC Scanner X brings my GPU to level I have never seen any game push it to, and that's exactly how you should test to.
If you have any question, you can ask me.
Regards,
I can report that after a couple days of intense playing, the temperature never reached higher than 80C (and it's fairly warm weather here) and I did not experience any crash or driver issue. I am not suggesting that the issue is not there, but just that it really doesn't seem to affect everyone. The overclock in itself is unlikely to cause the issue, but the overall temperature rise might be. However, it should be mentioned that if the game really is pushing your GPU to unconventional level of temperature, I am sorry to say this to you but your overclock is the problem. No software can push your hardware beyond the max limit of what it should be able to tolerate, it's just not how electronic works. If you overheat, you have an unsuitable / unstable overclock or lack appropriate cooling. Saying no other games does that to your GPU means only that no other games pushes your GPU to the max. Third party tools should be able to help you test your overclock. I can personally say that EVGA OC Scanner X brings my GPU to level I have never seen any game push it to, and that's exactly how you should test to.
If you have any question, you can ask me.
Regards,
Sorry for reviving this thread but i thought i need to post here aswell.
Got the exact same issue with the game with a Sapphire 7970 Dual-X.. if i overclock it anything above 1100 20% power and 1550 on the memory it CTD´s after a few minutes while the audio keeps on going.
It is NOT my cards or the Overclock settings fault.. I can benchmark the gpu with 1200/1575 voltage 1.225 for 2 hours on end in furmark @4k and in DA:I without any problems.. Heaven benchmark is also no issue (all applications heavily use the GPU at 99-100% all of the time).
Either cdpr or amd has to look at this issue, its not normal.
EDIT: Gpu never reaches above 68 degrees
Got the exact same issue with the game with a Sapphire 7970 Dual-X.. if i overclock it anything above 1100 20% power and 1550 on the memory it CTD´s after a few minutes while the audio keeps on going.
It is NOT my cards or the Overclock settings fault.. I can benchmark the gpu with 1200/1575 voltage 1.225 for 2 hours on end in furmark @4k and in DA:I without any problems.. Heaven benchmark is also no issue (all applications heavily use the GPU at 99-100% all of the time).
Either cdpr or amd has to look at this issue, its not normal.
EDIT: Gpu never reaches above 68 degrees
Not sure how much value my input will have on this conversation, but here it goes:
Specs:
i7 5820k (12 CPU's @ 3.3GHz per)
MSI GTX 980
Intel X99 mobo
16GB RAM DDR4
120GB SSD (TW3 lives in this SSD)
27" BenQ 1440p monitor @ 144 refresh rate
Overclock settings (via MSI Afterburner):
Voltage +30Mv
Power limiter 115%
Temp target 80degs Celsius
Core Clock +115 MHz
Mem Clock +250 MHz
Run the game with everything on Ultra except for foliage/shadows at high and no hairworks.
My GPU stays cool at 71degs C and my GPU fans never go up past 86%. I use a pretty steep fan curve as well (100% at 80degs C). No cashes. No artifacting.
With all of these settings, I'm able to play with 57-65 FPS (even in Novigrad) on my 1440p monitor. The only time FPS dips to 52-55 FPS is when I'm in a swamp with fog while its raining.
I believe I'm running on the latest NVIDIA drivers as well. Not sure why I'm not having similar issues with the OC set up.
Specs:
i7 5820k (12 CPU's @ 3.3GHz per)
MSI GTX 980
Intel X99 mobo
16GB RAM DDR4
120GB SSD (TW3 lives in this SSD)
27" BenQ 1440p monitor @ 144 refresh rate
Overclock settings (via MSI Afterburner):
Voltage +30Mv
Power limiter 115%
Temp target 80degs Celsius
Core Clock +115 MHz
Mem Clock +250 MHz
Run the game with everything on Ultra except for foliage/shadows at high and no hairworks.
My GPU stays cool at 71degs C and my GPU fans never go up past 86%. I use a pretty steep fan curve as well (100% at 80degs C). No cashes. No artifacting.
With all of these settings, I'm able to play with 57-65 FPS (even in Novigrad) on my 1440p monitor. The only time FPS dips to 52-55 FPS is when I'm in a swamp with fog while its raining.
I believe I'm running on the latest NVIDIA drivers as well. Not sure why I'm not having similar issues with the OC set up.
The card doesn't have to overheat to become unstable. Some minor instabilities that might have gone unnoticed in other games or benchmarks do actually crash this game. Sure, it probably could've been remedied at the expense of efficiency, but why? Your video card becomes unstable and that's the issue. The game is not at fault.
BTW. I can overclock my video card to the point where all textures are screwed and geometry becomes rather whimsical, but the game still doesn't crash.
BTW. I can overclock my video card to the point where all textures are screwed and geometry becomes rather whimsical, but the game still doesn't crash.
Overclocking is just not a given on any hardware and any game. Manufacturers set clock frequency and voltage specifications such that every one of their devices will meet their performance specifications all the time. If you then overclock or overvoltage, sometimes your hardware will exceed the maximum performance of which that device is capable, and it will fail in some cases. Nobody but yourself is responsible for those failures. No developer can prevent them.
I would like to give something in consideration if you go for an overclock.
Not every batch of GPU chips are exactly the same. The tolerance can vary in the production process. That means that some GPU's can be overclocked to an rather high level, while others crash at even an slight overclock level. Most times manufacturers of video cards have an "special" series, that is factory-overclocked. They achieve that by selecting those batches that have an high overclock tolerance. That means often that those GPU's are "cherry-picked" from several batches, to be put in higher clocked (and thus higher priced) cards.
The result can be that the slightly cheaper, not factory overclocked cards, have an smaller tolerance against overclocking, because they get the GPU's that did not meet the required tolerance that was needed for those overclocked cards. Keep in mind that those crashes are often an result of signal timing mismatches, that have little of nothing to do with the temperature of the card. Keep also in mind that manufacturers that supply only one "standard" version of an card, can have mixed batches.
This can finally result in cards that are crashing when slightly overclocked, and cards that have an high tolerance when overclocking - even when both use the exact same GPU chip version number. When coming to overclocking some will have luck with their cards, while others can have crash after crash.
Not every batch of GPU chips are exactly the same. The tolerance can vary in the production process. That means that some GPU's can be overclocked to an rather high level, while others crash at even an slight overclock level. Most times manufacturers of video cards have an "special" series, that is factory-overclocked. They achieve that by selecting those batches that have an high overclock tolerance. That means often that those GPU's are "cherry-picked" from several batches, to be put in higher clocked (and thus higher priced) cards.
The result can be that the slightly cheaper, not factory overclocked cards, have an smaller tolerance against overclocking, because they get the GPU's that did not meet the required tolerance that was needed for those overclocked cards. Keep in mind that those crashes are often an result of signal timing mismatches, that have little of nothing to do with the temperature of the card. Keep also in mind that manufacturers that supply only one "standard" version of an card, can have mixed batches.
This can finally result in cards that are crashing when slightly overclocked, and cards that have an high tolerance when overclocking - even when both use the exact same GPU chip version number. When coming to overclocking some will have luck with their cards, while others can have crash after crash.
Last edited:
Just to give my input, I've been running a small overclock (~80MHz, both core and memory) with my Radeon R9 380 for several weeks now with no issues, and as far as I know the 380 (aka 285) isn't exactly known for its overclocking abilities. I'd echo what others have already stated, just because you're stable in Furmark doesn't mean you're going to be stable in everything, overclocking is a fickle thing!
Went at my 960 now that its been a couple moths running stable at stock....
EVGA FTW edition runs at 1418/3505 by default.
Furmark/3DMark/GTA5/MGS5/etc all run perfect and 100% stable at up to 1550/3780Mhz
Witcher 3 runs fine but I get atifacts and failures to load textures (red boxes everywhere) at anything above 1525/3700.
SO yeah, it seems it is a little more sensitive than everything else, even furmark.
O_O
EVGA FTW edition runs at 1418/3505 by default.
Furmark/3DMark/GTA5/MGS5/etc all run perfect and 100% stable at up to 1550/3780Mhz
Witcher 3 runs fine but I get atifacts and failures to load textures (red boxes everywhere) at anything above 1525/3700.
SO yeah, it seems it is a little more sensitive than everything else, even furmark.
O_O
Hello my friends, I'm bringing back this old thread to offer my input and ask a few questions...]
I too am getting a crash to desktop from playing with two Titan x's in SLI. I overclockd them 20%, I put the max voltage up to 110% in MSI Afterburner and all my games have been running smoothly, no artefacts or crashes to speak of at all.
I'm just getting this crash to desktop with the sound continuing for a little while. Very strange. I thought with a bad overclock one would experience artefacts and glitching beforehand on screen?
I also have a question in general about overclocking if you would be so kind as I am slightly unsure of the 'Power Limit' setting. I have moved it to 110% as I was told to in a guide. I was under the impression that one would increase the core and memory as far as one can take it while maintaining a decent temperature. My temperature maxes out at about 80C playing Witcher overclocked, this is with two Titan x cards in SLI, the top one getting hotter obviously but never over 80. The power limit I see maxes at about 112% sometimes. Is this too much? Other than the crash I have mentioned I get zero problems with the overclock in Witcher or any other game. I get sick frames too and just worry I am damaging the cards with that Power Limit setting.
Thank you.
I too am getting a crash to desktop from playing with two Titan x's in SLI. I overclockd them 20%, I put the max voltage up to 110% in MSI Afterburner and all my games have been running smoothly, no artefacts or crashes to speak of at all.
I'm just getting this crash to desktop with the sound continuing for a little while. Very strange. I thought with a bad overclock one would experience artefacts and glitching beforehand on screen?
I also have a question in general about overclocking if you would be so kind as I am slightly unsure of the 'Power Limit' setting. I have moved it to 110% as I was told to in a guide. I was under the impression that one would increase the core and memory as far as one can take it while maintaining a decent temperature. My temperature maxes out at about 80C playing Witcher overclocked, this is with two Titan x cards in SLI, the top one getting hotter obviously but never over 80. The power limit I see maxes at about 112% sometimes. Is this too much? Other than the crash I have mentioned I get zero problems with the overclock in Witcher or any other game. I get sick frames too and just worry I am damaging the cards with that Power Limit setting.
Thank you.
You are probably better off asking these questions on another website like toms hardware or OCU, CDPR staff are unlikely to help you void your warranty on 2000$ of GPU.
As for upping the voltage on your gpu or cpu, yes you are shortening the lifespan of the cards, when you heat them to those temperatures for extended periods it degrades the connections on the pcb.
EDIT:
I can't help but feel worried for someone who bought 2000$ of GPU and then followed some internet guide on overclocking them..Best be very careful with increasing voltage pal only takes .1 of a volt to fry it. I suggest you just tweak the frequency and not the voltage until you understand better what you are doing with them.
Not to mention you should not be touching the voltage unless you have a solid cooling solution in place. Stock fans are no good for serious overclocking which is what you are attempting.
As for upping the voltage on your gpu or cpu, yes you are shortening the lifespan of the cards, when you heat them to those temperatures for extended periods it degrades the connections on the pcb.
EDIT:
I can't help but feel worried for someone who bought 2000$ of GPU and then followed some internet guide on overclocking them..Best be very careful with increasing voltage pal only takes .1 of a volt to fry it. I suggest you just tweak the frequency and not the voltage until you understand better what you are doing with them.
Not to mention you should not be touching the voltage unless you have a solid cooling solution in place. Stock fans are no good for serious overclocking which is what you are attempting.
Last edited:
Last night I had a playing session that lasted 2 and half hours. I was still in Velen and Oxenfurt area preparing to leave for Touisant. No crashes or anything similar whatsoever.
This morning I started playing with an overclocked card as well and have experienced two crashes with variously colored artifacts on the screen and a windows warning specifying that the display driver stopped working.
I will play the game with stock factory settings next and see how it performs.
This morning I started playing with an overclocked card as well and have experienced two crashes with variously colored artifacts on the screen and a windows warning specifying that the display driver stopped working.
I will play the game with stock factory settings next and see how it performs.