2019-06-17 09:59:54 |
spear |
bug |
|
|
added bug |
2019-06-17 10:03:38 |
spear |
description |
No matter if I tried it with 410.x series, 415.x series, 418.x or 430.x series after an S3 (suspend to RAM) ... everything that needs CUDA and GPU acceleration stop working. will update with appropriate dmesg output |
No matter if I tried it with 410.x series, 415.x series, 418.x or 430.x series after an S3 (suspend to RAM) ... everything that needs CUDA and GPU acceleration stop working. will update with appropriate dmesg output.
Fun fact: I just tried adding info putting this system into S3 and after waking up all the CUDA and driver related stuff also worked again ...
this is weird !! |
|
2019-06-17 16:19:47 |
spear |
description |
No matter if I tried it with 410.x series, 415.x series, 418.x or 430.x series after an S3 (suspend to RAM) ... everything that needs CUDA and GPU acceleration stop working. will update with appropriate dmesg output.
Fun fact: I just tried adding info putting this system into S3 and after waking up all the CUDA and driver related stuff also worked again ...
this is weird !! |
I you watch a movie that uses CUDA/VDPAU (nvdec for mpv, vdpau for VLC) decoding with {410,415,418,430} and pause the movie and do a S3 (suspend to RAM) ... everything that needs CUDA and GPU acceleration stop working. will update with appropriate dmesg output.
Maybe VDPAU can recover but CUDA can't
Got the extra outut from dmesg - it prints graphics cards serial no and bus id
```
NVRM: Xid (PCI:0000:1f:00): 31, Ch 00000003, intr 10000000. MMU Fault: ENGINE HOST6 HUBCLIENT_HOST faulted @ 0x1_001a1000. Fault is of type FAULT_PDE ACCESS_TYPE_READ
NVRM: Xid (PCI:0000:1f:00): 31, Ch 0000007c, intr 10000000. MMU Fault: ENGINE HOST0 HUBCLIENT_HOST faulted @ 0x2_00800000. Fault is of type FAULT_PDE ACCESS_TYPE_READ
``` |
|
2019-06-17 16:20:07 |
spear |
description |
I you watch a movie that uses CUDA/VDPAU (nvdec for mpv, vdpau for VLC) decoding with {410,415,418,430} and pause the movie and do a S3 (suspend to RAM) ... everything that needs CUDA and GPU acceleration stop working. will update with appropriate dmesg output.
Maybe VDPAU can recover but CUDA can't
Got the extra outut from dmesg - it prints graphics cards serial no and bus id
```
NVRM: Xid (PCI:0000:1f:00): 31, Ch 00000003, intr 10000000. MMU Fault: ENGINE HOST6 HUBCLIENT_HOST faulted @ 0x1_001a1000. Fault is of type FAULT_PDE ACCESS_TYPE_READ
NVRM: Xid (PCI:0000:1f:00): 31, Ch 0000007c, intr 10000000. MMU Fault: ENGINE HOST0 HUBCLIENT_HOST faulted @ 0x2_00800000. Fault is of type FAULT_PDE ACCESS_TYPE_READ
``` |
I you watch a movie that uses CUDA/VDPAU (nvdec for mpv, vdpau for VLC) decoding with {410,415,418,430} and pause the movie and do a S3 (suspend to RAM) ... everything that needs CUDA and GPU acceleration stop working. will update with appropriate dmesg output.
Maybe VDPAU can recover but CUDA can't
Got the extra outut from dmesg - it prints graphics cards serial no and bus id
<code>
NVRM: Xid (PCI:0000:1f:00): 31, Ch 00000003, intr 10000000. MMU Fault: ENGINE HOST6 HUBCLIENT_HOST faulted @ 0x1_001a1000. Fault is of type FAULT_PDE ACCESS_TYPE_READ
NVRM: Xid (PCI:0000:1f:00): 31, Ch 0000007c, intr 10000000. MMU Fault: ENGINE HOST0 HUBCLIENT_HOST faulted @ 0x2_00800000. Fault is of type FAULT_PDE ACCESS_TYPE_READ
</code> |
|
2019-06-17 16:42:08 |
spear |
tags |
|
nvidia s3 |
|
2019-06-17 16:44:47 |
spear |
description |
I you watch a movie that uses CUDA/VDPAU (nvdec for mpv, vdpau for VLC) decoding with {410,415,418,430} and pause the movie and do a S3 (suspend to RAM) ... everything that needs CUDA and GPU acceleration stop working. will update with appropriate dmesg output.
Maybe VDPAU can recover but CUDA can't
Got the extra outut from dmesg - it prints graphics cards serial no and bus id
<code>
NVRM: Xid (PCI:0000:1f:00): 31, Ch 00000003, intr 10000000. MMU Fault: ENGINE HOST6 HUBCLIENT_HOST faulted @ 0x1_001a1000. Fault is of type FAULT_PDE ACCESS_TYPE_READ
NVRM: Xid (PCI:0000:1f:00): 31, Ch 0000007c, intr 10000000. MMU Fault: ENGINE HOST0 HUBCLIENT_HOST faulted @ 0x2_00800000. Fault is of type FAULT_PDE ACCESS_TYPE_READ
</code> |
I you watch a movie that uses CUDA/VDPAU (nvdec for mpv, vdpau for VLC) decoding with {410,415,418,430} and pause the movie and do a S3 (suspend to RAM) ... everything that needs CUDA and GPU acceleration stop working. will update with appropriate dmesg output.
Maybe VDPAU can recover but CUDA can't
Got the extra outut from dmesg - it prints graphics cards serial no and bus id
dmesg output after S3 with a NVDEC movie playing
"Could not consume packet." Restart mpv or any other CUDA based application:
<code>
NVRM: Xid (PCI:0000:1f:00): 31, Ch 00000003, intr 10000000. MMU Fault: ENGINE HOST6 HUBCLIENT_HOST faulted @ 0x1_001a1000. Fault is of type FAULT_PDE ACCESS_TYPE_READ
NVRM: Xid (PCI:0000:1f:00): 31, Ch 0000007c, intr 10000000. MMU Fault: ENGINE HOST0 HUBCLIENT_HOST faulted @ 0x2_00800000. Fault is of type FAULT_PDE ACCESS_TYPE_READ
</code> |
|
2019-06-17 16:47:00 |
spear |
tags |
nvidia s3 |
cuda cuvid nvdec nvidia s3 |
|
2019-06-17 16:47:29 |
spear |
summary |
proprietary nvidia drivers become unusable after an S3 |
proprietary nvidia drivers become unusable after STR/S3 |
|