Cuda error 702

[mbutler2]

Had the same problem and the roll back worked for me.

Interestingly, I was only getting the 702 error in the live viewer window. I could still team render with Octane fine without the error.

Details: Asus EWS/USB3.1 Motherboard, with one 1070 onboard and four 1080TIs outboard. Was getting 702 errors in Octane live viewer and also while crypto mining. (But was fine team rendering Octane.) Latest windows 10 updates as of right now, which is January 2018.

Rolled back NVidia driver from 388.71 to 382.33 and seems to be working so far. This driver was so old that it didn't even come up in the NVidia search, I had to google the driver version specifically to find it. But seems to have worked.

Sounds like that registry edit mentioned here could be a better way to fix, I'll have to try that when things slow down.

On a side note, I think my future motherboards will be the cheaper slower Z270A, not these fancier EWS models. Z270A's might not be able to hold the faster CPUs but they've proven to be very solid reliable multi-GPU machines for me.

Thanks for the help mates!

[dmartin73]

Hi Orb101,

I have the same computer than you with 4x1070 TI. I have Error cuda 702 in slave also. I need to decrease the amount of polygon for a successfull render.

Have you find another solution about it please ?

Best regards,

[mbutler2]

An update to my driver rollback a few days ago... Using an older driver is now creating instability elsewhere, like in After Effects, so not as successful as I had hoped. I'll need to change to a newer driver and keep trying things.

[orb101]

Hi guys

I found that rolling back to some 2017 drivers on the slave machine caused it to become stable for network rendering.

However, if i use the GPU's locally in cinema4d on that machine, I get a crash.

If I update to latest drivers, local use of the GPU's seems stable, but 702 occurs on net render.

Frustrating.

I actually had all the hardware tested in the machine, and everything passed.

I'm even using a different set of 1080TI cards and I still have the same issue.

[orb101]

Wonder if everyone posting here has the same mobo?

[Fake Pilot]

I also have (had) the exact same problem with my x4 1080 Ti. Installed the 382.33 driver, custom install, checked "Clean Install".
Now all my Nvidia cards are rendering at full! God, I've had this problem for months... Was driving me crazy.

Finally.

[belly80naz]

Maybe I solved moving the Titan Xp card in the PCI-E slot number 1 and the 2x GTX 980Ti in the 3 and 4 slots number.
And putting the HDMI port for my display into the Titan Xp.

But with 4 videocards all the same I do not know how to solve.

[Fake Pilot]

Just wanted to post an update. Whenever I render and/if I have After Effects opened, my computer or GPU crashes.
If After Effects is not, then everything works fine.

[paride4331]

Hi guys

I found that rolling back to some 2017 drivers on the slave machine caused it to become stable for network rendering.

However, if i use the GPU's locally in cinema4d on that machine, I get a crash.

If I update to latest drivers, local use of the GPU's seems stable, but 702 occurs on net render.

Frustrating.

I actually had all the hardware tested in the machine, and everything passed.

I'm even using a different set of 1080TI cards and I still have the same issue.

Hi,
The specific error message is related to a timeout issue. What happens is that the Windows system instantly disable the GPU drivers, if it does not receive a signal after some seconds.
By default the TdrDelay is set to 2s, i.e. if one kernel call takes more than 2s, Windows kills the driver.
If you install the Standalone via installer, we change some Windows registry to better work with GPU Rendering.
To change the timeout, either run the Standalone installer or manually set the TdrDelay key to something like 10s or so. See here:
https://msdn.microsoft.com/en-us/librar ... s.85).aspx
In the Standalone installer we set these values:
WriteRegDWORD HKLM ""SYSTEM\CurrentControlSet\Control\GraphicsDrivers"" TdrLevel 3
WriteRegDWORD HKLM ""SYSTEM\CurrentControlSet\Control\GraphicsDrivers"" TdrDelay 10jh
Some users have set the TdrDelay at 60, to solve their stability issues with CUDA.
viewtopic.php?f=85&t=63700&p=327361&hilit=+tdr#p327361
Regards
Paride

[eyelovemotion]

That worked for me!

https://www.pugetsystems.com/labs/hpc/W ... ience-777/

1. Test after editing the registry i have no more crashes in cinema 4d Plugin 3.08.
Seems to have worked - i did not downgrade the drivers. i use the newest one frome Nvidia.

Problem seems to have disappeared after the registry change.

Perfect Tipp!

Thank You