Comments (22)
Could you share the full command line, and any log output before the crash? Or does it crash immediately?
From: iamkkk [mailto:[email protected]]
Sent: Thursday, January 28, 2016 16:14
To: Microsoft/CNTK [email protected]
Subject: [CNTK] can not run cntk binary (#52)
hi, I am trying the cntk binary file and when I try to run it on gpu, I encounter the the following error message:" CNTK: Win32 exception caught (such an access violation or a stack overflow"
I am using nvidia k5000 and cuda70 on Windows 10
Does anyone know what is the problem?
thanks a lot
—
Reply to this email directly or view it on GitHubhttps://github.com//issues/52.
from cntk.
I run"cntk configFile=../Config/FeedForward.config" and the program crashes immediatly. no log output either. The command prompt only print build infom,then "MPIWrapper: initializing MPI" then"CNTK: Win32 exception caught (such an access violation or a stack overflow)"
I tried to run simple2d on cpu and It worked fine.
from cntk.
Could you try with a non-existent config file? (change the path to a file that does not exist) and let me know whether it still crashes? In that case, it is a setup error, I would bet a missing delay-loaded DLL.
From: iamkkk [mailto:[email protected]]
Sent: Thursday, January 28, 2016 16:45
To: Microsoft/CNTK [email protected]
Cc: Frank Seide [email protected]
Subject: Re: [CNTK] can not run cntk binary (#52)
I run"cntk configFile=../Config/FeedForward.config" and the program crashes immediatly. no log output either.
—
Reply to this email directly or view it on GitHubhttps://github.com//issues/52#issuecomment-176498646.
from cntk.
the following is erroe message when i try a non-existent config file
Build info:
Built time: Jan 26 2016 04:34:07
Last modified date: Mon Jan 25 09:01:43 2016
CUDA_PATH: C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v7.0
CUB_PATH: C:\src\cub-1.4.1
CUDNN_PATH: C:\NVIDIA\cudnn-4.0\cuda
Build Branch: HEAD
Build SHA1: 56a2a15f64676ea4c0e0a0a681a57b19a46f64c6
Built by svcphil on DPHAIM-22
Build Path: c:\jenkins\workspace\CNTK-Build-Windows\Source\CNTK\
[CALL STACK]
>Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::GPUMatrix::AssignNCEUnnormalizedEval
-Microsoft::MSR::CNTK::GPUMatrix::AssignNCEUnnormalizedEval
-Microsoft::MSR::CNTK::GPUMatrix::AssignNCEUnnormalizedEval
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::BaseMatrix::GetSizeAllocated
-BaseThreadInitThunk
-RtlUserThreadStart
attempt: error opening file '../Config/FeedForwar1d.config': No such file or directory, retrying 2-th time out of 5...
[CALL STACK]
>Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::GPUMatrix::AssignNCEUnnormalizedEval
-Microsoft::MSR::CNTK::GPUMatrix::AssignNCEUnnormalizedEval
-Microsoft::MSR::CNTK::GPUMatrix::AssignNCEUnnormalizedEval
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::BaseMatrix::GetSizeAllocated
-BaseThreadInitThunk
-RtlUserThreadStart
attempt: error opening file '../Config/FeedForwar1d.config': No such file or directory, retrying 3-th time out of 5...
[CALL STACK]
>Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::GPUMatrix::AssignNCEUnnormalizedEval
-Microsoft::MSR::CNTK::GPUMatrix::AssignNCEUnnormalizedEval
-Microsoft::MSR::CNTK::GPUMatrix::AssignNCEUnnormalizedEval
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::BaseMatrix::GetSizeAllocated
-BaseThreadInitThunk
-RtlUserThreadStart
attempt: error opening file '../Config/FeedForwar1d.config': No such file or directory, retrying 4-th time out of 5...
[CALL STACK]
>Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::GPUMatrix::AssignNCEUnnormalizedEval
-Microsoft::MSR::CNTK::GPUMatrix::AssignNCEUnnormalizedEval
-Microsoft::MSR::CNTK::GPUMatrix::AssignNCEUnnormalizedEval
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::BaseMatrix::GetSizeAllocated
-BaseThreadInitThunk
-RtlUserThreadStart
attempt: error opening file '../Config/FeedForwar1d.config': No such file or directory, retrying 5-th time out of 5...
[CALL STACK]
>Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::GPUMatrix::AssignNCEUnnormalizedEval
-Microsoft::MSR::CNTK::GPUMatrix::AssignNCEUnnormalizedEval
-Microsoft::MSR::CNTK::GPUMatrix::AssignNCEUnnormalizedEval
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::MatrixBase::operator=
-Microsoft::MSR::CNTK::BaseMatrix::GetSizeAllocated
-BaseThreadInitThunk
-RtlUserThreadStart
EXCEPTION occurred: error opening file '../Config/FeedForwar1d.config': No such file or directory
Usage: cntk configFile=yourConfigFile
For detailed information please consult the CNTK book
"An Introduction to Computational Networks and the Computational Network Toolkit"
from cntk.
I have the same problem, but some differences.
I run the samples for speech as the readme and get the info,"MPIWrapper: initializing MPI, CNTK: Win32 exception caught (such an access violation or a stack overflow)", the enviroment is Win 7 64bit, only CPU, and had installed the cuda7.0 toolkit, and also vs2013.
Are there some other setup should be do?
Thanks.
from cntk.
I will see if I can add a more clear error message. I suspect this is due to a missing DLL that is being delay-loaded, likely a missing MPI runtime.
From: haixinguan [mailto:[email protected]]
Sent: Wednesday, February 3, 2016 19:07
To: Microsoft/CNTK [email protected]
Cc: Frank Seide [email protected]
Subject: Re: [CNTK] can not run cntk binary (#52)
I have the same problem, but some differences.
I run the samples for speech as the readme and get the info,"MPIWrapper: initializing MPI, CNTK: Win32 exception caught (such an access violation or a stack overflow)", the enviroment is Win 7 64bit, only GPU, and had installed the cuda7.0 toolkit, and also vs2013.
Are there some other setup should be do?
Thanks.
—
Reply to this email directly or view it on GitHubhttps://github.com//issues/52#issuecomment-179592364.
from cntk.
A fix for MPIWrapper has been put into master. Haixin, iamkkk, would you mind trying again?
Haixin should get a proper error message saying that MPI is not installed. For iamkkk, I would like to know whether this is the same issue.
from cntk.
I am using the binary file(https://github.com/Microsoft/CNTK/releases). it seems that that file has not been updated yet.
p.s.I am pretty sure that I installed MPI
from cntk.
I just added more detailed error reporting for Win32 exceptions. Will land soon, but only in source for now.
From: iamkkk [mailto:[email protected]]
Sent: Thursday, February 4, 2016 15:45
To: Microsoft/CNTK [email protected]
Cc: Frank Seide [email protected]
Subject: Re: [CNTK] can not run cntk binary (#52)
I am using the binary file(https://github.com/Microsoft/CNTK/releases). it seems that that file has not been updated yet.
p.s.I am pretty sure that I installed MPI
—
Reply to this email directly or view it on GitHubhttps://github.com//issues/52#issuecomment-180106222.
from cntk.
I get the same error and I noticed that MSMPI.DLL is not accessible by the binary even though MS MPI installation is finished successfully.
from cntk.
I fixed something a few days ago—the actual initialization of MPI was guarded by a catch of a failed delay-load, but there was a check outside whether MPI was already initialized, which was not caught.
If your version is older than that commit, would you mind trying whether latest master still have this problem?
commit 1e65a41
Author: Frank Seide [email protected]
Date: Thu Feb 4 08:51:08 2016 -0800
fixed bug in catching when MPI runtime is missing
From: raminia [mailto:[email protected]]
Sent: Monday, February 8, 2016 15:13
To: Microsoft/CNTK [email protected]
Cc: Frank Seide [email protected]
Subject: Re: [CNTK] can not run cntk binary (#52)
I get the same error and I noticed that MSMPI.DLL is not accessible by the binary even though MS MPI installation is finished successfully.
—
Reply to this email directly or view it on GitHubhttps://github.com//issues/52#issuecomment-181614628.
from cntk.
Will do. However, in the process of getting dependencies ready to build CNTK I realized that I've missed an important step which is installing both _SDK _*and * _runtime _ versions of the MS-MPI from the link. That means installing both .exe and .msi files. After doing so CNTK started normally.
However, when testing AN4 dataset example, I landed on another exception (below) which I believe is a separate issue but it emphasizes that I need to build CNTK for better customization.
EXCEPTION occurred: Gradient quantization is unsupported in CNTK binaries built without quantized gradient aggregation support!
from cntk.
This is related to the 1-bit SGD method. You will need to get that through a separate submodule.
https://github.com/Microsoft/CNTK/wiki/Enabling-1bit-SGD
From: raminia [mailto:[email protected]]
Sent: Monday, February 8, 2016 16:23
To: Microsoft/CNTK [email protected]
Cc: Frank Seide [email protected]
Subject: Re: [CNTK] can not run cntk binary (#52)
Will do. However, in the process of getting dependencies ready to build CNTK I realized that I've missed an important step which is installing both _SDK _*and * _runtime _ versions of the MS-MPI from the linkhttps://na01.safelinks.protection.outlook.com/?url=https%3a%2f%2fmsdn.microsoft.com%2fen-us%2flibrary%2fbb524831(v%3dvs.85).aspx&data=01%7c01%7cfseide%40microsoft.com%7c209132d7e71840146adc08d330e724bf%7c72f988bf86f141af91ab2d7cd011db47%7c1&sdata=xryYx%2bziUnS%2fkoquK0j8x4tUMvxcVuAea8iNKn%2ftZcs%3d. That means installing both .exe and .msi files. After doing so CNTK started normally.
However, when testing AN4 dataset example, I landed on another exception (below) which I believe is a separate issue but it emphasizes that I need to build CNTK for better customization.
EXCEPTION occurred: Gradient quantization is unsupported in CNTK binaries built without quantized gradient aggregation support!
—
Reply to this email directly or view it on GitHubhttps://github.com//issues/52#issuecomment-181637731.
from cntk.
@iamkkk - are you experiencing issues to run the CNTK binary? Please let us know, otherwise we'll close this isuse...
from cntk.
I downloaded CNTK-2016-02-08-Windows-64bit-GPU.zip and got the error "mpihelper: msmpi.dll missing
mpiaggregator: MPI_Init, MPI error 16
from cntk.
Hello!
You need to have MPI installed: https://msdn.microsoft.com/en-us/library/bb524831(v=vs.85).aspx
Tell us, whether it fixes it.
Tank you,
Alexey
from cntk.
hi, the following is the command I typed:D:\cntk\Examples\Other\Simple2d\Data>cntk configFile=../Config/Multigpu.cntk
Build info:
Built time: Feb 8 2016 00:54:07
Last modified date: Sun Feb 7 16:51:01 2016
Build type: Unknown
Build target: Unknown
With 1bit-SGD: no
CUDA_PATH: C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v7.0
CUB_PATH: C:\src\cub-1.4.1
CUDNN_PATH: C:\NVIDIA\cudnn-4.0\cuda
Build Branch: HEAD
Build SHA1: 2f9a48c71dc0a6097498cb7e90ac3b151ab536dd
Built by svcphil on LIANA-09-w
Build Path: c:\jenkins\workspace\CNTK-Build-Windows\Source\CNTK\
MPIWrapper: initializing MPI
mpihelper: msmpi.dll missing
mpiaggregator: MPI_Init, MPI error 16
[CALL STACK]
>Microsoft::MSR::CNTK::TracingGPUMemoryAllocator::operator=
-Microsoft::MSR::CNTK::TracingGPUMemoryAllocator::operator=
-Microsoft::MSR::CNTK::TracingGPUMemoryAllocator::operator=
-Microsoft::MSR::CNTK::TracingGPUMemoryAllocator::operator=
-Microsoft::MSR::CNTK::TracingGPUMemoryAllocator::operator=
-Microsoft::MSR::CNTK::TracingGPUMemoryAllocator::operator=
-Microsoft::MSR::CNTK::TracingGPUMemoryAllocator::operator=
-BaseThreadInitThunk
-RtlUserThreadStart
EXCEPTION occurred: mpiaggregator: MPI_Init
Usage: cntk configFile=yourConfigFile
For detailed information please consult the CNTK book
"An Introduction to Computational Networks and the Computational Network Toolkit"
from cntk.
Sorry for asking again - do you have MS-MPI installed? (see my previous post)
from cntk.
I installed Microsoft MPI SDK(7.0.12437.6). there are 2 files in the download website. I downloaded and installed the .msi file, is it ok? Or should I install the MSMPISetup.exe
from cntk.
Ah, you need run time. (Same download page, file MSMpiSetup.exe)
from cntk.
it's working! thanks a lot!
from cntk.
Great! I will make binary setup instructions clearer. Thank you. I will close the request now.
from cntk.
Related Issues (20)
- Value goes invalid when using TestMinibatch
- Request for a no-opencv dotnet release
- CNTK C# Crash when layer is deeper
- Convolution 1D CNTK C++ HOT 1
- Add support to release linux aarch64 wheels
- Error while deploying MS Teams Bot with SSO
- how to install in python3.8
- A model causes CNTK crash with cudnnSetPoolingNDDescriptor when invoking the pooling operator
- This repo is missing important files HOT 1
- API document issue due to syntax in source code HOT 1
- Error: could not find all specified 'to_nodes' in clone. Looking for ['relu5'], found [None]
- SequenceClassification.py does not work out of box HOT 1
- Microsoft has absolutely no potential
- development of a neural network for object search
- join vcpkg
- cntk crash and pycharm process finished with exit code -1066598274 (0xC06D007E)
- ConvolutionTranspose2D outputs normally when num_filters is tuple and dilation is 0
- program crash when get gradient of `ConvolutionTranspose2D`
- when shape contains negative integer, input_variable should throw an error HOT 1
- MAX_POOLING crash, when pooling_window_shape contains 0 HOT 1
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from cntk.