xixou Posted May 20, 2018 Share Posted May 20, 2018 Hi, Nicehash2 freezes when aida64 sensor panel is running. It occurs within 5 min when using google chrome together with nicehashv2 using equihash for the GPU and cryptonightv7 for the CPU , using Nvidia drivers 391.35. When using nvidia drivers 397.64, this is even worst, there are BSOD generated. Using win10 home 64 bits, latest aida64 5.97.4633 beta 5 GTX1070 and 1950X CPU. My sensor panel config is attached. Sensor panel Refresh speed set to 2 seconds. 2018-05-20.sensorpanel Quote Link to comment Share on other sites More sharing options...
xixou Posted May 20, 2018 Author Share Posted May 20, 2018 Someone on nvidia forum stated that nicehash uses low level access to the GPU. https://forums.geforce.com/default/topic/1052670/geforce-drivers/official-397-64-game-ready-whql-display-driver-feedback-thread-released-5-9-18-/27/ Can the low level access of aida64 be changed into an indirect level access? Quote Link to comment Share on other sites More sharing options...
sblantipodi Posted May 20, 2018 Share Posted May 20, 2018 same problems here. I have freeze and than BSOD with latest nvidia driver + aida. If I use an older driver or if I close AIDA the problem does not happen. Quote Link to comment Share on other sites More sharing options...
Forceflow Posted May 21, 2018 Share Posted May 21, 2018 Just chiming in, I've been experiencing similar crashes with a similar monitoring tool (GPU-Z). See my reports (and minidumps) over at the Techpowerup forums or the Geforce board. I've ruled out pretty much everything, it's definitely something wrong in the latest series of Nvidia drivers. Quote Link to comment Share on other sites More sharing options...
xixou Posted May 21, 2018 Author Share Posted May 21, 2018 I see 4 low level settings under aida64 stability settings preferences: MSR PCI SMBUS SENSOR They are all ticked active. For my sensor panel, the PCI and SMBus are not relevant, so I disabled them(unticked). So far I was not experiencing the freeze anymore using combination of nicehash v2 and google chrome using Nvidia drivers 391.35. Could you try at your side sblantipodi and Forceflow ? Quote Link to comment Share on other sites More sharing options...
Squall Leonhart Posted May 21, 2018 Share Posted May 21, 2018 i can't say i ever had issues with 391.35, it might be just your unique mining configuration in that case Xixou, Quote Link to comment Share on other sites More sharing options...
Squall Leonhart Posted May 21, 2018 Share Posted May 21, 2018 20 hours ago, xixou said: Someone on nvidia forum stated that nicehash uses low level access to the GPU. https://forums.geforce.com/default/topic/1052670/geforce-drivers/official-397-64-game-ready-whql-display-driver-feedback-thread-released-5-9-18-/27/ Can the low level access of aida64 be changed into an indirect level access? not nicehash itself, the claymore binary and a few others have nvapi interaction to monitor card thermals and disable p2 cuda. Quote Link to comment Share on other sites More sharing options...
xixou Posted May 21, 2018 Author Share Posted May 21, 2018 Do you know how to disable nicehash v2 low level accesses? I twitted nicehash but got no answer. BTW, on aida64 restart, my gpu sensors were gone, so I had to also select PCI low level access to make it work, so only the low level SMB is unticked at the moment. I will try and see if it is still stable Quote Link to comment Share on other sites More sharing options...
Squall Leonhart Posted May 21, 2018 Share Posted May 21, 2018 I don't, not without recompiling the claymore binary. Quote Link to comment Share on other sites More sharing options...
xixou Posted May 21, 2018 Author Share Posted May 21, 2018 so far so good at the moment with the SMB unticked. Quote Link to comment Share on other sites More sharing options...
xixou Posted May 22, 2018 Author Share Posted May 22, 2018 Are aida64 devs not responding anymore? Quote Link to comment Share on other sites More sharing options...
Squall Leonhart Posted May 23, 2018 Share Posted May 23, 2018 Fiery is a busy man. Quote Link to comment Share on other sites More sharing options...
xixou Posted May 24, 2018 Author Share Posted May 24, 2018 Same issue with 397.93 Quote Link to comment Share on other sites More sharing options...
Squall Leonhart Posted May 26, 2018 Share Posted May 26, 2018 Quote HeavyHemi wrote: On Wed 5/23/2018 10:54:52 PM your computer crashed or a problem was reported crash dump file: C:\WINDOWS\MEMORY.DMP This was probably caused by the following module: aida64.sys (0x0000000000000000) Bugcheck code: 0x1E (0xFFFFFFFFC0000005, 0x0, 0x8, 0x0) Error: KMODE_EXCEPTION_NOT_HANDLED Bug check description: This indicates that a kernel-mode program generated an exception which the error handler did not catch. This might be a case of memory corruption. This may be because of a hardware issue such as faulty RAM, overheating (thermal issue) or because of a buggy driver. A third party driver was identified as the probable root cause of this system error. It is suggested you look for an update for the following driver: aida64.sys . Google query: aida64.sys KMODE_EXCEPTION_NOT_HANDLED A couple of hours of leaving AIDA64 open which was entirely accidental. https://forums.evga.com/FindPost/2823774 Quote Link to comment Share on other sites More sharing options...
xixou Posted May 28, 2018 Author Share Posted May 28, 2018 8 days later no reply from the dev, I won't renew my license when expired ... This is sad. Quote Link to comment Share on other sites More sharing options...
Fiery Posted May 30, 2018 Share Posted May 30, 2018 I'm sorry for the late reply. We've already got a few topics about the same bug of the recent ForceWare drivers. The issue is not related to AIDA64 at all -- which can be considered a problem since we cannot fix it in our end. Downgrading ForceWare to 388.31 solved the issue: Quote Link to comment Share on other sites More sharing options...
xixou Posted June 2, 2018 Author Share Posted June 2, 2018 Can you help Nvidia to reproduce it such that they fix the issue at their end? Downgrading to such old driver is not an option as programs like nicehash2 and miners don't support it anymore. Quote Link to comment Share on other sites More sharing options...
Fiery Posted June 3, 2018 Share Posted June 3, 2018 On 6/2/2018 at 8:33 AM, xixou said: Can you help Nvidia to reproduce it such that they fix the issue at their end? Downgrading to such old driver is not an option as programs like nicehash2 and miners don't support it anymore. Since the issue affects any software that uses NVAPI (nVIDIA's own API), it should be pretty easy to reproduce it in nVIDIA's own labs. We can however be of assistance if asked. We're always open to work with hardware manufacturers, but nVIDIA is not famous for being open to such cooperations I'm afraid. Quote Link to comment Share on other sites More sharing options...
xixou Posted June 7, 2018 Author Share Posted June 7, 2018 391.35 are fine too and I can use nicehash2 with it. Quote Link to comment Share on other sites More sharing options...
Squall Leonhart Posted June 8, 2018 Share Posted June 8, 2018 nvidia are still claiming no repro, so i think they are missing a particular configuration requirement to bring it out. They might need one of our cfg files and sensor panels or whatever Quote Link to comment Share on other sites More sharing options...
xixou Posted June 9, 2018 Author Share Posted June 9, 2018 Or Could they bring us a debug drivers version that can show them more data when crash occurs? Quote Link to comment Share on other sites More sharing options...
p1r473 Posted June 11, 2018 Share Posted June 11, 2018 I have been having an issue for the last 2 driver releases 397.93+397.64 where Windows is hanging/freezing constantly. 30 seconds after I log into Windows, roughly around the time Aida64 loads, it will hang for like 5 minutes. Then it will be fine, but it will freeze again and again later for a few minutes each time. There does not seem to be anything related to the drivers in the event viewer at the time of the hang. 391.35 works perfectly and I keep having to roll back to it. I have done the driver upgrades with DDU, and also without my SLI bridge. Nothing is helping. I have 2x 1080 TI in SLI. Quote Link to comment Share on other sites More sharing options...
Squall Leonhart Posted June 12, 2018 Share Posted June 12, 2018 On 6/11/2018 at 1:24 PM, p1r473 said: I have been having an issue for the last 2 driver releases 397.93+397.64 where Windows is hanging/freezing constantly. 30 seconds after I log into Windows, roughly around the time Aida64 loads, it will hang for like 5 minutes. Then it will be fine, but it will freeze again and again later for a few minutes each time. Totally different problem. Quote Link to comment Share on other sites More sharing options...
xixou Posted June 13, 2018 Author Share Posted June 13, 2018 On 6/11/2018 at 5:24 AM, p1r473 said: I have been having an issue for the last 2 driver releases 397.93+397.64 where Windows is hanging/freezing constantly. 30 seconds after I log into Windows, roughly around the time Aida64 loads, it will hang for like 5 minutes. Then it will be fine, but it will freeze again and again later for a few minutes each time. There does not seem to be anything related to the drivers in the event viewer at the time of the hang. 391.35 works perfectly and I keep having to roll back to it. I have done the driver upgrades with DDU, and also without my SLI bridge. Nothing is helping. I have 2x 1080 TI in SLI. Yes, I get the same. When BSOD comes, this is when it freezes longer. No issue with driver 391.35 Quote Link to comment Share on other sites More sharing options...
Fiery Posted June 21, 2018 Share Posted June 21, 2018 We believe we've found the root cause of the issues concerning Windows 10 April 2018 Update (Version 1803 / Build 17134) based systems. The timer that AIDA64 uses for BCLK measurement is also used by a new Win10 component. Most likely that component is used for GPU profiling. And when AIDA64 and Win10 both use the timer for their own purposes in the same time, the system could become unstable. We've had to disable polling for BCLK, so from now on AIDA64 will only measure BCLK once, and then use the last value until AIDA64 gets restarted. Here's the new AIDA64 beta build that includes the workaround: https://www.aida64.com/downloads/latesta64xebeta Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.