LabVIEW FPGA Idea Exchange

cancel
Showing results for 
Search instead for 
Did you mean: 
kdnelson

NI Cloud Compile Farm: take advantage of cached cores

Status: New

When you compile on your own machine, the output of the "coregen" step is cached and used on subsequent compiles.  This saves considerable time.  The fact that this does not happen on the NI cloud service erases any speed gains (and more).

 

You could:

 - Cache cores at NI.  Save for xx days since last use (or forever if space is not an issue)

 - Transfer cores when they exist (cached locally) to compile server along with other intermediate files

 

3 Comments
RSteinbrück
Member

Hi kdnelson,

 

This thread is somewhat old, but maybe I'm lucky and you're still around.

 

I set up a Compile Farm Server and an according Linux Compile Worker in our intranet. This Worker has of course a core cache folder. Now I'd like to fill this core cache in advance because otherwise the first compile of a new project takes forever due to the "coregen". Where do I find the cores? Are they installed with LabVIEW FPGA on the Develpment system? Can they simply be transfered to the Linux Worker?

 

Regards

Reinhard

kdnelson
Member

I don't think cached cores can help you on the first compile.  In my experience the cores are created by the compiler and cached, I suppose, because they are identified as self-contained units that probably won't need re-compiling next time around.  So caching them helps only on subsequent compiles.  I don't know if there's a way to provide a bunch of pre-compiled cores for common functions--my feeling is that the cores are still pretty project-specific.  But if there were a way, that would be a great addition to this idea: prepopulate the cache as well as retain any other cores generated by a compile.

 

Nan.Yang
Member

Hi kdnelson,

 

There is performance improvement in the FPGA Compile Cloud after NIWeek 2016. Check out the latest compile cloud!