InternalError:cudnn poolforward launch failed

Question

InternalError:cudnn poolforward launch failed

weihua04 opened this issue 4 years ago · 12 comments

i have installed tf-gpu 12.0, cuda9, tf-compression 10.0 , and when used the code below for test:
python OpenDVC_test_video.py --path BasketballPass --model PSNR --metric PSNR --l 1024

the information is :

"cudnn poolforward launch failed"

does anyone has good idea to solve it?

Answer 1 · 2020-09-18T06:25:48.000Z

Hi, What kind of GPU are you using?-------- 原始邮件 --------发件人： weihua04 <notifications@github.com>日期： 2020年9月18日周五 08:12收件人： RenYang-home/OpenDVC <OpenDVC@noreply.github.com>抄送： Subscribed <subscribed@noreply.github.com>主题： [RenYang-home/OpenDVC] InternalError:cudnn poolforward launch failed (#5) i have installed tf-gpu 12.0, cuda9, tf-compression 10.0 , and when used the code below for test: python OpenDVC_test_video.py --path BasketballPass --model PSNR --metric PSNR --l 1024 the information is : ![Uploading 5074244104368395475.jpg…] does anyone has good idea to solve it? —You are receiving this because you are subscribed to this thread.Reply to this email directly, view it on GitHub, or unsubscribe.

Answer 2 · 2020-09-18T07:07:28.000Z

Hi, What kind of GPU are you using?-------- 原始邮件 --------发件人： weihua04 notifications@github.com日期： 2020年9月18日周五 08:12收件人： RenYang-home/OpenDVC OpenDVC@noreply.github.com抄送： Subscribed subscribed@noreply.github.com主题： [RenYang-home/OpenDVC] InternalError:cudnn poolforward launch failed (#5)
i have installed tf-gpu 12.0, cuda9, tf-compression 10.0 , and when used the code below for test:
python OpenDVC_test_video.py --path BasketballPass --model PSNR --metric PSNR --l 1024
the information is :
![Uploading 5074244104368395475.jpg…]
does anyone has good idea to solve it?

—You are receiving this because you are subscribed to this thread.Reply to this email directly, view it on GitHub, or unsubscribe.

GeForce RTX 2080

Answer 3 · 2020-09-18T07:50:14.000Z

I am not sure whether it is because of insufficient memory. Is there an error if running on CPU?

…

On Fri, 18 Sep 2020 at 09:07, weihua04 ***@***.***> wrote: Hi, What kind of GPU are you using?-------- 原始邮件 --------发件人： weihua04 ***@***.***日期： 2020年9月18日周五 08:12收件人： RenYang-home/OpenDVC ***@***.***抄送： Subscribed ***@***.***主题： [RenYang-home/OpenDVC] InternalError:cudnn poolforward launch failed ( #5 <#5>) i have installed tf-gpu 12.0, cuda9, tf-compression 10.0 , and when used the code below for test: python OpenDVC_test_video.py --path BasketballPass --model PSNR --metric PSNR --l 1024 the information is : ![Uploading 5074244104368395475.jpg…] does anyone has good idea to solve it? —You are receiving this because you are subscribed to this thread.Reply to this email directly, view it on GitHub, or unsubscribe. GeForce RTX 2080 — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#5 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AES6XUFLODER3SDE5LHPUBLSGMBL5ANCNFSM4RRPXXTA> .

-- Ren Yang Doctoral student Computer Vision Laboratory | ETH Zürich Sternwartstrasse 7 | ETF D 114.2 | 8092 Zürich, Switzerland

Answer 4 · 2020-09-18T08:41:18.000Z

I am not sure whether it is because of insufficient memory. Is there an
error if running on CPU?
…

CPU is ok, I have tried tf 12.0, but tf-gpu12.0 have problem

Answer 5 · 2020-09-18T09:00:48.000Z

Maybe try on a GPU with larger memory (e.g., 12GB or more).
I haven't met that problem, so I am not sure what's wrong it is.

Answer 6 · 2020-09-18T09:59:34.000Z

Maybe try on a GPU with larger memory (e.g., 12GB or more).
I haven't met that problem, so I am not sure what's wrong it is.

you run the test.py in docker or the host machine?

Answer 7 · 2020-09-18T10:10:57.000Z

I run on the host machine, not on docker-------- 原始邮件 --------发件人： weihua04 <notifications@github.com>日期： 2020年9月18日周五 11:59收件人： RenYang-home/OpenDVC <OpenDVC@noreply.github.com>抄送： Ren Yang <r.yangchn@gmail.com>, Comment <comment@noreply.github.com>主题： Re: [RenYang-home/OpenDVC] InternalError:cudnn poolforward launch failed (#5) Maybe try on a GPU with larger memory (e.g., 12GB or more). I haven't met that problem, so I am not sure what's wrong it is. you run the test.py in docker or the host machine? —You are receiving this because you commented.Reply to this email directly, view it on GitHub, or unsubscribe.

Answer 8 · 2020-10-14T10:04:08.000Z

thanks for reply. I think maybe it's the version problem, Could u tell me what's your minor version of CUDNN CUDA, and tensorflow? thank u very much.

mine:
CUDA: 9.0.176_384.81
CUDNN: V7.1.4
tensorflow: 1.12.0

Answer 9 · 2020-10-15T04:42:32.000Z

I meet the same problem.

Answer 10 · 2020-10-15T08:51:11.000Z

thanks for reply. I think maybe it's the version problem, Could u tell me what's your minor version of CUDNN CUDA, and tensorflow? thank u very much.

mine:
CUDA: 9.0.176_384.81
CUDNN: V7.1.4
tensorflow: 1.12.0

Mine is:
CUDA: 9.0, V9.0.176
CUDNN:
#define CUDNN_MAJOR 7
#define CUDNN_MINOR 6
#define CUDNN_PATCHLEVEL 0

#define CUDNN_VERSION (CUDNN_MAJOR * 1000 + CUDNN_MINOR * 100 + CUDNN_PATCHLEVEL)

tensorflow: 1.12.0

Answer 11 · 2020-10-28T07:51:39.000Z

I meet the same problem.

I solved the problem by change version of cudnn，my environment is in docker ,
the environment of docker is :
CUDA: 9
CUDNN: 10.0 V7.4.2

Answer 12 · 2020-10-28T08:00:47.000Z

I meet the same problem.

I solved the problem by change version of cudnn，my environment is in docker ,
the environment of docker is :
CUDA: 9
CUDNN: 10.0 V7.4.2

Happy to hear that :-)

Mine is: CUDA: 9.0, V9.0.176 CUDNN: #define CUDNN_MAJOR 7 #define CUDNN_MINOR 6 #define CUDNN_PATCHLEVEL 0

Mine is:
CUDA: 9.0, V9.0.176
CUDNN:
#define CUDNN_MAJOR 7
#define CUDNN_MINOR 6
#define CUDNN_PATCHLEVEL 0