Next-Gen debugger fails to start
I installed the latest driver for Nsight 5.5 in order to use the next-gen debugger, but unfortunately it's not working. The legacy debugger is working fine. I get the following message: "Could not initialize driver for debugging. Debugging has been automatically stopped." Windows: 7 Pro SP1 GPU: Quadro P6000 Driver: 388.59 SDK & Toolkit: 8.0 Nsight: 5.5.0.18014 Visual Studio: 2015 14.0.25431.01 Update 3 Extra info: - GPU for driving graphics is an AMD. - TDR is disabled. Thanks, Andreas
I installed the latest driver for Nsight 5.5 in order to use the next-gen debugger, but unfortunately it's not working. The legacy debugger is working fine.
I get the following message:
"Could not initialize driver for debugging.
Debugging has been automatically stopped."


Windows: 7 Pro SP1
GPU: Quadro P6000
Driver: 388.59
SDK & Toolkit: 8.0
Nsight: 5.5.0.18014
Visual Studio: 2015 14.0.25431.01 Update 3

Extra info:
- GPU for driving graphics is an AMD.
- TDR is disabled.


Thanks,
Andreas

#1
Posted 01/22/2018 10:06 AM   
Hello, Next-gen debugger only works under TCC mode, you must install 2 GPUs, one for display and use your P6000 as second card, then set it to TCC mode. If your legacy debugger works well, I suggest you still to use it, nex-gen is for Tesla card, nex-gen will support wddm in next version of nsight. Best Regards Harry
Hello,

Next-gen debugger only works under TCC mode, you must install 2 GPUs, one for display and use your P6000 as second card, then set it to TCC mode.

If your legacy debugger works well, I suggest you still to use it, nex-gen is for Tesla card, nex-gen will support wddm in next version of nsight.

Best Regards
Harry

#2
Posted 01/23/2018 09:26 AM   
Hi Harry, I switched to TCC mode and it worked. I had already switched to using another card for the graphics. Thank you very much. Best Regards, Andreas
Hi Harry,

I switched to TCC mode and it worked. I had already switched to using another card for the graphics. Thank you very much.

Best Regards,
Andreas

#3
Posted 01/23/2018 10:04 AM   
Hello, I have the same issue - same error message, but this time with a Quadro in WDDM and a Tesla K20c in TCC. Legacy debugger is working fine as well. Windows 7 x64 Nsight version: 5.5.0.1814 Driver version: 390.65 Any idea what could be the problem ? Thanks, Antoine
Hello,

I have the same issue - same error message, but this time with a Quadro in WDDM and a Tesla K20c in TCC. Legacy debugger is working fine as well.

Windows 7 x64
Nsight version: 5.5.0.1814
Driver version: 390.65

Any idea what could be the problem ?

Thanks,
Antoine

#4
Posted 02/01/2018 11:16 AM   
Are you sure the debugger is using the Tesla and not the Quadro for debugging?
Are you sure the debugger is using the Tesla and not the Quadro for debugging?

#5
Posted 02/01/2018 12:01 PM   
I can only say that I'm never using the Quadro when selecting the device. Moreover, when running my application with the profiler, only the Tesla shows usage. Quadro context is never mentioned in any call.
I can only say that I'm never using the Quadro when selecting the device. Moreover, when running my application with the profiler, only the Tesla shows usage. Quadro context is never mentioned in any call.

#6
Posted 02/01/2018 04:37 PM   
[quote=""]I can only say that I'm never using the Quadro when selecting the device. Moreover, when running my application with the profiler, only the Tesla shows usage. Quadro context is never mentioned in any call.[/quote] Hello, Next-gen debugger only work with pascal above card, so you cannot debug it on k20c, profiler should work on any device, could you please check if your app only runs on one GPU? I guess your app only use one card, you can set CUDA_VISIBLE_DEVICES to let it run on another card. Best Regards Harry
said:I can only say that I'm never using the Quadro when selecting the device. Moreover, when running my application with the profiler, only the Tesla shows usage. Quadro context is never mentioned in any call.


Hello,

Next-gen debugger only work with pascal above card, so you cannot debug it on k20c, profiler should work on any device, could you please check if your app only runs on one GPU? I guess your app only use one card, you can set CUDA_VISIBLE_DEVICES to let it run on another card.

Best Regards
Harry

#7
Posted 02/02/2018 02:23 AM   
Hello Harry, Thanks for the quick feedback. I supposed the requirement was stated on the page but I suppose I missed it. I'm ok with the profiler behavior. I just don't want to use the Quadro at all for CUDA. Best regards, Antoine
Hello Harry,

Thanks for the quick feedback. I supposed the requirement was stated on the page but I suppose I missed it.

I'm ok with the profiler behavior. I just don't want to use the Quadro at all for CUDA.

Best regards,
Antoine

#8
Posted 02/02/2018 09:03 AM   
Hello there, I'm facing the same issue here. I want to debug a Windows CLI project. From [url]http://docs.nvidia.com/nsight-visual-studio-edition/Content/Debug_External_Applications.htm[/url] I learned that that's only possible with Nsight 5.5. However, I "only" have one GPU in my machine a no option to upgrade. Does that mean, that I can't debug a CLR project?
Hello there,

I'm facing the same issue here.
I want to debug a Windows CLI project. From http://docs.nvidia.com/nsight-visual-studio-edition/Content/Debug_External_Applications.htm I learned that that's only possible with Nsight 5.5. However, I "only" have one GPU in my machine a no option to upgrade.

Does that mean, that I can't debug a CLR project?

NVidia Flex for .Net now available FREE & OPEN SOURCE at: https://github.com/HeinzBenjamin/FlexCLI

#9
Posted 02/13/2018 06:33 PM   
[quote=""]Hello there, I'm facing the same issue here. I want to debug a Windows CLI project. From [url]http://docs.nvidia.com/nsight-visual-studio-edition/Content/Debug_External_Applications.htm[/url] I learned that that's only possible with Nsight 5.5. However, I "only" have one GPU in my machine a no option to upgrade. Does that mean, that I can't debug a CLR project?[/quote] Yes, if you want to use next-gen debugger, you can use the legacy debugger to debug you cuda app, it can works on wddm mode
said:Hello there,

I'm facing the same issue here.
I want to debug a Windows CLI project. From http://docs.nvidia.com/nsight-visual-studio-edition/Content/Debug_External_Applications.htm I learned that that's only possible with Nsight 5.5. However, I "only" have one GPU in my machine a no option to upgrade.

Does that mean, that I can't debug a CLR project?


Yes, if you want to use next-gen debugger, you can use the legacy debugger to debug you cuda app, it can works on wddm mode

#10
Posted 02/14/2018 02:33 AM   
[quote=""][quote=""]Hello there, I'm facing the same issue here. I want to debug a Windows CLI project. From [url]http://docs.nvidia.com/nsight-visual-studio-edition/Content/Debug_External_Applications.htm[/url] I learned that that's only possible with Nsight 5.5. However, I "only" have one GPU in my machine a no option to upgrade. Does that mean, that I can't debug a CLR project?[/quote] Yes, if you want to use next-gen debugger, you can use the legacy debugger to debug you cuda app, it can works on wddm mode[/quote] Thanks a lot for the answer. If I use the legacy debugger it doesn't hit any break point. I have three layers in my project: .Net (CS) for testing .cu for actual functions and a C++/CLI layer to connect the two While the standard VS debugger goes as deep as into the CLI layer before it loses track, Nsight (legacy) doesn't hit any break point in any of these layers. I wonder if it's because I didn't tune the settings properly, or I'm using a late Windows 10 Pro Insider Preview Built (16299.248)? (referring to this thread [url]https://devtalk.nvidia.com/default/topic/971720/nsight-visual-studio-edition/nsight-visual-studio-ignoring-break-points/[/url]) The Next-Gen debugger throws "Could not initialize driver for debugging". Any thoughts? My card is a Quadro K2100M btw.... driver version 390.77 with Nsight 5.5.0.18014
said:
said:Hello there,

I'm facing the same issue here.
I want to debug a Windows CLI project. From http://docs.nvidia.com/nsight-visual-studio-edition/Content/Debug_External_Applications.htm I learned that that's only possible with Nsight 5.5. However, I "only" have one GPU in my machine a no option to upgrade.

Does that mean, that I can't debug a CLR project?


Yes, if you want to use next-gen debugger, you can use the legacy debugger to debug you cuda app, it can works on wddm mode


Thanks a lot for the answer. If I use the legacy debugger it doesn't hit any break point.
I have three layers in my project:
.Net (CS) for testing
.cu for actual functions
and a C++/CLI layer to connect the two

While the standard VS debugger goes as deep as into the CLI layer before it loses track, Nsight (legacy) doesn't hit any break point in any of these layers. I wonder if it's because I didn't tune the settings properly, or I'm using a late Windows 10 Pro Insider Preview Built (16299.248)? (referring to this thread https://devtalk.nvidia.com/default/topic/971720/nsight-visual-studio-edition/nsight-visual-studio-ignoring-break-points/)

The Next-Gen debugger throws "Could not initialize driver for debugging".
Any thoughts?

My card is a Quadro K2100M btw.... driver version 390.77 with Nsight 5.5.0.18014

NVidia Flex for .Net now available FREE & OPEN SOURCE at: https://github.com/HeinzBenjamin/FlexCLI

#11
Posted 02/24/2018 10:05 AM   
[quote=""][quote=""][quote=""]Hello there, I'm facing the same issue here. I want to debug a Windows CLI project. From [url]http://docs.nvidia.com/nsight-visual-studio-edition/Content/Debug_External_Applications.htm[/url] I learned that that's only possible with Nsight 5.5. However, I "only" have one GPU in my machine a no option to upgrade. Does that mean, that I can't debug a CLR project?[/quote] Yes, if you want to use next-gen debugger, you can use the legacy debugger to debug you cuda app, it can works on wddm mode[/quote] Thanks a lot for the answer. If I use the legacy debugger it doesn't hit any break point. I have three layers in my project: .Net (CS) for testing .cu for actual functions and a C++/CLI layer to connect the two While the standard VS debugger goes as deep as into the CLI layer before it loses track, Nsight (legacy) doesn't hit any break point in any of these layers. I wonder if it's because I didn't tune the settings properly, or I'm using a late Windows 10 Pro Insider Preview Built (16299.248)? (referring to this thread [url]https://devtalk.nvidia.com/default/topic/971720/nsight-visual-studio-edition/nsight-visual-studio-ignoring-break-points/[/url]) The Next-Gen debugger throws "Could not initialize driver for debugging". Any thoughts? My card is a Quadro K2100M btw.... driver version 390.77 with Nsight 5.5.0.18014[/quote] Quaro K2100M is a kepler card actually GK106, I'm not sure if the latest nsight can work with it as we are only testing gk104 and gk107 now. AFAIK, the legacy debugger can debug the .net exe + cuda dll, you must make sure that your cuda codes are contained in a local c++ not cli dll file, then use your .net exe to load it via DllImport. You can set the bp in cuda dll then use legacy debugger to debug the .net exe, the bp should be hit.
said:
said:
said:Hello there,

I'm facing the same issue here.
I want to debug a Windows CLI project. From http://docs.nvidia.com/nsight-visual-studio-edition/Content/Debug_External_Applications.htm I learned that that's only possible with Nsight 5.5. However, I "only" have one GPU in my machine a no option to upgrade.

Does that mean, that I can't debug a CLR project?


Yes, if you want to use next-gen debugger, you can use the legacy debugger to debug you cuda app, it can works on wddm mode


Thanks a lot for the answer. If I use the legacy debugger it doesn't hit any break point.
I have three layers in my project:
.Net (CS) for testing
.cu for actual functions
and a C++/CLI layer to connect the two

While the standard VS debugger goes as deep as into the CLI layer before it loses track, Nsight (legacy) doesn't hit any break point in any of these layers. I wonder if it's because I didn't tune the settings properly, or I'm using a late Windows 10 Pro Insider Preview Built (16299.248)? (referring to this thread https://devtalk.nvidia.com/default/topic/971720/nsight-visual-studio-edition/nsight-visual-studio-ignoring-break-points/)

The Next-Gen debugger throws "Could not initialize driver for debugging".
Any thoughts?

My card is a Quadro K2100M btw.... driver version 390.77 with Nsight 5.5.0.18014


Quaro K2100M is a kepler card actually GK106, I'm not sure if the latest nsight can work with it as we are only testing gk104 and gk107 now.

AFAIK, the legacy debugger can debug the .net exe + cuda dll, you must make sure that your cuda codes are contained in a local c++ not cli dll file, then use your .net exe to load it via DllImport.

You can set the bp in cuda dll then use legacy debugger to debug the .net exe, the bp should be hit.

#12
Posted 02/24/2018 11:24 AM   
Thanks a lot, I'll try that, would you generally say, that's best practice for .Net and Cuda? I'm starting a new project so wouldn't mind exploring a totally different work flow. I was looking into things like Alea GPU, managedCuda or altimesh, but can't quite tell, which is the best bet in terms of performance and ease of development / debugging. I want to write something like a physics engine for .Net with minimal data transfer between host and device, so would like to have access to device memory to manage these things myself..
Thanks a lot, I'll try that,

would you generally say, that's best practice for .Net and Cuda? I'm starting a new project so wouldn't mind exploring a totally different work flow. I was looking into things like Alea GPU, managedCuda or altimesh, but can't quite tell, which is the best bet in terms of performance and ease of development / debugging. I want to write something like a physics engine for .Net with minimal data transfer between host and device, so would like to have access to device memory to manage these things myself..

NVidia Flex for .Net now available FREE & OPEN SOURCE at: https://github.com/HeinzBenjamin/FlexCLI

#13
Posted 02/24/2018 12:21 PM   
[quote=""]Thanks a lot, I'll try that, would you generally say, that's best practice for .Net and Cuda? I'm starting a new project so wouldn't mind exploring a totally different work flow. I was looking into things like Alea GPU, managedCuda or altimesh, but can't quite tell, which is the best bet in terms of performance and ease of development / debugging. I want to write something like a physics engine for .Net with minimal data transfer between host and device, so would like to have access to device memory to manage these things myself..[/quote] Sorry, my mistake, I just saw you only have on GPU, nsight cannot debug the CLR project, I'm afraid you cannot debug the code like managedCuda, nsight can only debug the code in local c++ project, the .net exe + cuda dll I told you before is actually the same thing, cu codes are contained in a local dll and the .net exe project is only a wrapper.
said:Thanks a lot, I'll try that,

would you generally say, that's best practice for .Net and Cuda? I'm starting a new project so wouldn't mind exploring a totally different work flow. I was looking into things like Alea GPU, managedCuda or altimesh, but can't quite tell, which is the best bet in terms of performance and ease of development / debugging. I want to write something like a physics engine for .Net with minimal data transfer between host and device, so would like to have access to device memory to manage these things myself..


Sorry, my mistake, I just saw you only have on GPU, nsight cannot debug the CLR project, I'm afraid you cannot debug the code like managedCuda, nsight can only debug the code in local c++ project, the .net exe + cuda dll I told you before is actually the same thing, cu codes are contained in a local dll and the .net exe project is only a wrapper.

#14
Posted 02/26/2018 02:02 AM   
[quote=""][quote=""]Thanks a lot, I'll try that, would you generally say, that's best practice for .Net and Cuda? I'm starting a new project so wouldn't mind exploring a totally different work flow. I was looking into things like Alea GPU, managedCuda or altimesh, but can't quite tell, which is the best bet in terms of performance and ease of development / debugging. I want to write something like a physics engine for .Net with minimal data transfer between host and device, so would like to have access to device memory to manage these things myself..[/quote] Sorry, my mistake, I just saw you only have on GPU, nsight cannot debug the CLR project, I'm afraid you cannot debug the code like managedCuda, nsight can only debug the code in local c++ project, the .net exe + cuda dll I told you before is actually the same thing, cu codes are contained in a local dll and the .net exe project is only a wrapper.[/quote] Okay thanks a lot anyways. I can't avoid some messy work around then, but okay... Best
said:
said:Thanks a lot, I'll try that,

would you generally say, that's best practice for .Net and Cuda? I'm starting a new project so wouldn't mind exploring a totally different work flow. I was looking into things like Alea GPU, managedCuda or altimesh, but can't quite tell, which is the best bet in terms of performance and ease of development / debugging. I want to write something like a physics engine for .Net with minimal data transfer between host and device, so would like to have access to device memory to manage these things myself..


Sorry, my mistake, I just saw you only have on GPU, nsight cannot debug the CLR project, I'm afraid you cannot debug the code like managedCuda, nsight can only debug the code in local c++ project, the .net exe + cuda dll I told you before is actually the same thing, cu codes are contained in a local dll and the .net exe project is only a wrapper.


Okay thanks a lot anyways. I can't avoid some messy work around then, but okay...
Best

NVidia Flex for .Net now available FREE & OPEN SOURCE at: https://github.com/HeinzBenjamin/FlexCLI

#15
Posted 02/26/2018 09:18 AM   
Scroll To Top

Add Reply