# cudnnGetConvolutionForwardWorkspaceSize return zero

**URL:** <https://forums.developer.nvidia.com/t/cudnngetconvolutionforwardworkspacesize-return-zero/204761>\
**Category:** cuDNN\
**Created:** [March 1, 2022, 12:19pm UTC](https://forums.developer.nvidia.com/t/cudnngetconvolutionforwardworkspacesize-return-zero/204761 "2022-03-01T12:19:16Z")\
**Posts on this page:** 4\
**Page:** 1

<div class="post-metadata">

**Author:** ![2391553280](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@2391553280](https://forums.developer.nvidia.com/u/2391553280)\
**Post date:** [March 1, 2022, 12:19pm UTC](https://forums.developer.nvidia.com/t/cudnngetconvolutionforwardworkspacesize-return-zero/204761/1 "2022-03-01T12:19:16Z")

</div>

//intput set  
const int in\_n = 1;  
const int in\_c = 3;  
const int in\_h = 56;  
const int in\_w = 56;  
cudnnTensorDescriptor\_t in\_desc;  
bool input\_c = false;

CUDNN\_CALL(cudnnCreateTensorDescriptor(&in\_desc));  
CUDNN\_CALL(cudnnSetTensor4dDescriptor(  
in\_desc, CUDNN\_TENSOR\_NHWC, CUDNN\_DATA\_FLOAT,in\_n, in\_c, in\_h, in\_w));

//kernel set  
const int filt\_k = 4;  
const int filt\_c = 3;  
const int filt\_h = 3;  
const int filt\_w = 3;  
cudnnFilterDescriptor\_t filt\_desc;  
CUDNN\_CALL(cudnnCreateFilterDescriptor(&filt\_desc));  
CUDNN\_CALL(cudnnSetFilter4dDescriptor(  
filt\_desc, CUDNN\_DATA\_FLOAT, CUDNN\_TENSOR\_NCHW,  
filt\_k, filt\_c, filt\_h, filt\_w));

//get workspacesize  
size\_t ws\_size;  
CUDNN\_CALL(cudnnGetConvolutionForwardWorkspaceSize(  
cudnn, in\_desc, filt\_desc, conv\_desc, out\_desc, algo, &ws\_size));

when is input h and w set 56 ws\_size is 16832，and input h and w set 96 ws\_size is 0 ，and input h and w set 150 ws\_size is 0，input h and w set 151 ws\_size is 5.01MB，so i don’t know why this happen.

---

<div class="post-metadata">

**Author:** ![spolisetty](https://sea2.discourse-cdn.com/nvidia/user_avatar/forums.developer.nvidia.com/spolisetty/32/14047_2.png) [@spolisetty](https://forums.developer.nvidia.com/u/spolisetty)\
**Post date:** [March 8, 2022, 6:15am UTC](https://forums.developer.nvidia.com/t/cudnngetconvolutionforwardworkspacesize-return-zero/204761/2 "2022-03-08T06:15:36Z")

</div>

> [@2391553280](#):
>
> when is input h and w set 56 ws\_size is 16832，and input h and w set 96 ws\_size is 0 ，and input h and w set 150 ws\_size is 0，input h and w set 151 ws\_size is 5.01MB

Hi,

Could you please confirm were you able to compute results successfully even it returns zero, there could be the possibility of less than MB returns zero. And also could you please share the output log for the above h and w inputs.

Thank you.

---

<div class="post-metadata">

**Author:** ![2391553280](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@2391553280](https://forums.developer.nvidia.com/u/2391553280)\
**Post date:** [March 9, 2022, 8:17am UTC](https://forums.developer.nvidia.com/t/cudnngetconvolutionforwardworkspacesize-return-zero/204761/3 "2022-03-09T08:17:35Z")

</div>

when is input h and w set 56 ws\_size is 16832kb，and input h and w set 96 ws\_size is 0 ，and input h and w set 150 ws\_size is 0，input h and w set 151 ws\_size is 5.01MB(this is turn to MB)  
i confirm result is successfully, i use cudnn 7.6.0, There is a range h size and w size generate to ws\_size zero

---

<div class="post-metadata">

**Author:** ![spolisetty](https://sea2.discourse-cdn.com/nvidia/user_avatar/forums.developer.nvidia.com/spolisetty/32/14047_2.png) [@spolisetty](https://forums.developer.nvidia.com/u/spolisetty)\
**Post date:** [March 9, 2022, 8:50am UTC](https://forums.developer.nvidia.com/t/cudnngetconvolutionforwardworkspacesize-return-zero/204761/4 "2022-03-09T08:50:05Z")

</div>

> [@2391553280](#):
>
> 7.6.0

Could you please try on the latest cuDNN version 8.3.2, and let us know if you still face this issue.

Thank you.