Trying to configure my de1-soc board with oneAPI

[Ratan]

I am trying to run the sample program vector-add which was obtained from the oneapi-cli menu and I am just not sure if the de1-soc board is an officially supported board for the oneapi base toolkit. Well, I assumed that I had to follow the development kit first and then move onto the deployment kit which includes quartus prime and the BSP(along with the development flow) that needs to be used to deploy a custom logic onto an FPGA. 

Can anyone tell me if im moving in the right direction? and whether if I have to do the jupyterlab intro along with the FPGA emulation to actually understand how to get the de1-soc and oneapi to work together ? 

[BoonBengT_Intel]

Hi @Ratan,

Thank you for posting in Intel community forum and hope all is well.

Unfortunately DE1-SoC is not an officially supported board for oneAPI toolkit for FPGA as there is no BSP available.

Readily available support board with BSP can be found in the link below:

- https://www.intel.com/content/www/us/en/developer/tools/oneapi/fpga.html (For non PAC product, please scroll to the 'Choose an FPGA Platform' section)

- https://www.intel.com/content/www/us/en/docs/oneapi/programming-guide/2023-1/fpga-bsps-and-boards.html (For PAC product)

Yes, you are in the right path, as you will need to decide on the board/platform required. Followed by installing the oneAPI toolkit with FPGA add-on for development.

Jupyter lab are recommended as it would be a more hands-on way to get familiar with oneAPI build flow.

Other than that would recommend going through this 'Using FPGAs with the Intel® oneAPI Toolkits' video., which would be beneficial as start. 

- https://www.youtube.com/watch?v=SU9S-PG_W9A

Hope that clarify

Best Wishes

BB

[BoonBengT_Intel]

Hi @Ratan,

Greetings, just checking in to see if there is any further doubts in regards to this matter.

Hope your doubts have been clarified.

Best Wishes

BB

[Ratan]

Thank you for the information on the de1-soc board. I will probably have to switch boards.

As an alternative, my advisor told me to continue with jupyterLabs but I find myself to be in a confusing position. 

I currently completed the following steps on JupyterLabs:

oneAPI_Essentials->Introduction_to_Jupyter->INtroduction_to_Jupyter.ipynb

Browsed the section intro_to_jupyter.

After this I clicked on get started with module1 but that did not give me much information. It just informed me about SYCL and the XPU structure and functioning with oneAPI.

I remember reading somewhere that the next step was :

FPGA emulation using oneAPI base toolkit.

I do have my oneAPI base toolkit installed and have tested the sample project vector-add on my linux system but honestly have a little trouble following advice/instructions from the webpage.

how do I proceed in jupyterLabs and do you believe that this is a good way to fully understand the development/deployment flow without my board?

I am asking this question since this was advertised on : https://www.intel.com/content/www/us/en/developer/tools/oneapi/training/fpga-development-flow.html ->

2. Open the FPGA_Development_Flow folder, and then select FPGA_Emulation_Using_Intel_oneAPI_Base_Toolkit.ipynb.

Thank you for all your help!

[BoonBengT_Intel]

Hi @Ratan,

Good to know that you are exploring with the oneAPI with the jupyterLabs, that is definitely a good place to start.

The subsequent place which will help you to proceed further understanding on the development flow would be running the oneAPI sample design below:

- https://github.com/oneapi-src/oneAPI-samples/tree/master/DirectProgramming/C%2B%2BSYCL_FPGA

On the link above to go the GettingStarted --> fpga_compile which is a good example to help you understand the development flow. Once you are familiar with the compile flow, you can moved on to the features or tools.

Hope that clarify.

Best Wishes

BB

[Ratan]

So I was exploring the FPGA_emulation_Using_Intel_oneAPI_Base_Toolkit.ipynb under HOUgh transform by replicating the steps from the video :

I do not understand the path he takes in 17:23 of the video. Where does he source it? Is this within JupiterLabs ? That part is not made very clear unfortunately. Can I please get some help ?

https://www.intel.com/content/www/us/en/developer/videos/accelerate-fpga-programming.html#gs.2h9o8d

[Ratan]

I am talking in reference to when he creates the tester directory and then proceeds to source it as something a tar file downloads else where? I am not sure what this does or if its relevant to the steps after... Could you please clarify this as well?

Also do I follow the steps shown in this video and then emulate my code on the software end or do I refer to the link you have attached above regarding 'fpga_compile'? 

[BoonBengT_Intel]

Hi @Ratan,

For the mention source command to get the hough example, its more of a command which already pre-setup in the jupyter environment to download the hough example files. You can look at it as a easier way of manually go to the github link for hough example from the web browser, download the hough example source code and copying/uploading it to the jupyter environment path.

As for the emulation build, since you have started with the steps in the video, would recommend to complete them as they will interactive and smaller breakdown of the learning by section. The link send above are more towards a shortcut which you can jump into compilation with the assumtion you are familiar with the oneAPI toolkits compiler, code structure and etc.

Hope that clarify.

Best Wishes

BB

[Ratan]

As I was following the steps in this particular video, he had a smooth run with the simulation.

For me, unfortunately I failed at a point.

In the section all the way towards the end in the path testdir/FPGA_Emulation_Using_Intel_Base_ToolKit.ipynb ,

i got the following error:

## u196294 is performing Hough Transform compilation emulation notebook.
icpx: warning: use of 'dpcpp' is deprecated and will be removed in a future release. Use 'icpx -fsycl' [-Wdeprecated]
src/original/hough_transform.cpp:8:10: fatal error: 'CL/sycl/INTEL/fpga_extensions.hpp' file not found
#include <CL/sycl/INTEL/fpga_extensions.hpp>
         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~1 error generated.
/bin/bash: bin/hough_transform.emu: No such file or directory

whereas I should've seen the output:

'VERIFICATION PASSED'

I followed all the steps correctly. Can I please know how top solve this issue?

how I should find the path to/include the header files so that I can see the correct output?

[Ratan]

[BoonBengT_Intel]

Hi @Ratan,

Noted, unfortunately the mention hough example might be a bit dated, hence there are component/libraries which has been updated.

To fixed that there is a few lines of codes which needed the path updates.

More details of what to changed can be found in the link below:

- https://community.intel.com/t5/Intel-High-Level-Design/Hough-transform-on-Jupyter-oneAPI-DevCloud/m-p/1376210

Please do go through and let me know.

Best Wishes

BB

[BoonBengT_Intel]

Hi @Ratan,

Good day, just following up on the previous clarification.

Please do let us know if there is anything else we can help with.

Best Wishes

BB

[Ratan]

Yeah I asked all the following questions on the link you provided me. It was a resolved case but I decided to ask my questions on there because you can have better references.

Please do check as I am eagerly waiting for your response.

Thank you so much.

[BoonBengT_Intel]

Hi @Ratan,

I see, after going through the post mention in the other thread it seems that you are still having a compilation issues on the code, where compiler are complaining 'fatal error that says: CL/sycl/INTEL/fpga_extensions.hpp . file not found'.

It seems that the changes are not taking effect as compiler are still getting the previous path and throwing error. (i.e., instead of '#include <sycl/ext/intel/fpga_extensions.hpp>)

Please do ensure that correct .cpp or check all other .cpp files to ensure changes are made.

Hope that clarify.

Note: please do reply to this thread for me to better observe the issues you have.

Best Wishes

BB

[Ratan]

I just reverted back to the old configuration of including #include <CL/sycl/INTEL/fpga_extensions.hpp>  , i also have it saved in the testdir directory but I have no errors like before. no fatal error display. It just runs through all the cells but has no 'VERIFICATION PASSED' output... I dont know if i should skip the hough transform emulation for now? it seems to behave weirdly and a trial and error process didnt exactly work and I am not quite sure how to check all other .cpp files?

any advice on this?

Followed the video again to check if i missed any steps but there is nothing that I missed. The process has alternating behavior.

[BoonBengT_Intel · Accepted solution]

Hi @Ratan,

You should changed all the includes statement based on what is being mention previously to fixed the fatal error mention.

Apologies as I am confused, are you saying you have reverted back to the old include statement and issues is fixed?

Unsure of what stage you are in now, please do share the project that you are facing issues with, so that we can better understand the current stages you are at.

Hope to hear from you soon.

Best Wishes

BB

[Ratan]

 I am at the same stage, which is;

I was performing emulation of hough transform for FPGA's on Jupyterlab but then all i had to do was run the cells:

I get the following error ->

## u196294 is performing Hough Transform compilation emulation notebook.
icpx: warning: use of 'dpcpp' is deprecated and will be removed in a future release. Use 'icpx -fsycl' [-Wdeprecated]
src/original/hough_transform.cpp:8:10: fatal error: 'CL/sycl/INTEL/fpga_extensions.hpp' file not found
#include <CL/sycl/INTEL/fpga_extensions.hpp> //try this and prev
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1 error generated.
/bin/bash: bin/hough_transform.emu: No such file or directory

I am not sure how to solve this. I am unable to find the header files that they are supposed to include. I cant even find the parent directory which is supposed to be 'CL'.

Has anyone done fpga emulation using jupyterLabs and gone through the same problem?
i am eagerly looking for help, thanks.

I have been stuck on this for more than a week now but I had to inform you that the conversation that you linked me to, I tried that even included the ! mkdir bin statement which was used to solve the other users problem. unfortunately that did not solve mine, since i already had a directory named bin within the testdir directory.

can you do a new tutorial of the fpga emulation and cell running process to get the expected output - 'VERIFICATION PASSED'

I dont know which path to check and how to overcome this issue.

Thank you for all your help till now as well, I really appreciate it.

[BoonBengT_Intel]

Hi @Ratan,

Noted on the stage that you are still stucked.

I have pulled the similar hough example mention, and have made the necessary changes line #9, #72 and #76.

Attached are the zip files that contain hough_transform.cpp that I have changed from my end which is successfully compiled and run emulation.

Would suggest perhaps to compare with your .cpp files to see if any differences.

Once .cpp files are made, please go ahead to create a bin directory as below:

With that you would be able to run the emulation and get the following result as below:

Note: you can ignore the warning for now as the example code might have some outdated code.

Best Wishes

BB

[BoonBengT_Intel]

Hi @Ratan,

Greetings, as we do not receive any further clarification/updates on the matter, hence would assume challenge are overcome. Please login to ‘https://supporttickets.intel.com’, view details of the desire request, and post a feed/response within the next 15 days to allow me to continue to support you. After 15 days, this thread will be transitioned to community support. For new queries, please feel free to open a new thread and we will be right with you. Pleasure having you here.

Best Wishes

BB

p/s: If any answer from the community or Intel Support are helpful, please feel free to give best answer or rate 4/5 survey.