Force the Compile Worker to Work Faster

[muahang1234]

The process of creating the FPGA bit files takes a long time even with a simple FPGA design.  I monitor the system resources when the  Compile Worker creates the bit file. Only a few percentage of the CPU is being used. Is there a way to force the Compile Woker works faster because there is system resources available?

My system is dual Xeon 6 Physical Core with 16Gb RAM, running LabVIEW 2011 version 90 days come with the Evaluation Rio single board.

[Thoric · Accepted solution]

The FPGA code compilation is performed not by the LabVIEW compiler but a compiler from Xilinx, specific to their FPGA chipsets.

The problem you see is the Xilinx compiler is single thread, and therefore will only use one core, and also can be seen at times to 'dawdle'. There's nothing you can do to push the compiler to work harder by encouraging the use of other resources. If you cores are 3.2GHz each, then compiling on your dual Xeon 6 will be no faster than on a single-core 3.2GHz CPU.

What you need for minimizing your compile times is the fastest core you can get! 

[muahang1234]

Thank you very much for the response. The problem is it does not even use all power of a single core - most of the time, I only see it using 3-5%.  

[Hooovahh]

The rumor I heard is that FPGA compilation is more memory instensive then processor intensive.  Your 16GB of ram will go far but if it won't fit, the compiler will continue to try (for probably 4 hours or so) trying to optimize to get it to fit.

[Thoric]

I'm not sure what the reasons are for the compile worker to reduce CPU usage at times, but as Hoovah points out the process is varying in it's approach and resource usage, and consequently it can't make use of the CPU all the time. Until Xilinx create something revolutionarily better, you just have to live with it I'm afraid!