LabVIEW FPGA Compilation: PC requierements

[BeñatA]

Hi,

I am going to upgrade some of my PC characteristics due to very slow and sometimes unsuccessful FPGA compilation. 

I am adding more RAM because 4 Gb are not enough. This should speed up the process.

But what about the processor or the hard drive?

Which are the most important characteristics that a processor needs to compile: cores, clock, bus speed...?

Would a solid state hard drive speed up all the process?

Thanks a lot,

Beñat 

[James_McN · Accepted solution]

Hi Beñat,

I would always suggest a solid state now! But this doesn't tend to be one of the bottlenecks for FPGA compilation.

Most of the compilation process doesn't scale across multiple cores well so fewer, higher-frequency cores tend to be the better option.

RAM is also a resource constraint, so as you say increasing this will also help.

Another option is NIs compile cloud but that may depend on what software licenses you already have.

[GerdW · Accepted solution]

Hi Benat,

---

@BeñatA wrote:

Which are the most important characteristics that a processor needs to compile: cores, clock, bus speed...?

Would a solid state hard drive speed up all the process?

---

• Bitfile creation mostly benefits from high clock frequencies, but mainly uses just one core. So any modern CPU would be fine, with some advantages with "not so many cores, but higher clock frequencies"…
• A SSD always improves computer performance when compared to HDDs.
• I don't think you can fiddle with "bus speed" as you could 20 years ago, so it's not important…

[Terry_ALE]

RAM is the big one.

Which FPGA are you using?

For example:

KU035, KU040, or KU060 which uses 7 to 11 GB (typical, peak), 5785

Kintex-7 XC7K410T which uses 3 to 5 GB (typical, peak), 7976, 6592

Virtex-7 X690T which 5 to 7 GB (typical, peak) - VST, ATCA

ZU28DR which uses 14 and 17 GB (typical and peak) - x410

If you have less than the above, the compile will error out.

Xilinx has a table with this info but the link has changed.

[ZYOng]

Run the compilation on a Linux OS.

NI LabVIEW FPGA Compilation Options

NI invests in supporting the FPGA compile worker on the Linux OS. Depending on your FPGA VI complexity and logic, you can expect the Linux compile worker to reduce compile times by approximately 20 to 50 percent, compared to a Windows compile worker. Xilinx originally designed and optimized the compilation tools for the Linux OS, which explains this substantial reduction in compile time.

Setting up a Linux-Based FPGA Compile Worker