faster fpga compile with use of ramdisk

[JFTICH]

I have read the other posts in the forums regarding faster FPGA compiles...

It sounds like the Xilinx tool only uses 1 core processor. (I guess it will be great when they can use 2, 4, or 8+++).

I pumped my ram up to 3GB as well as a RAMDISK from gigabyte. Link to amazon's site:

http://www.amazon.com/Gigabyte-GC-RAMDISK-i-RAM-Hard-Drive/dp/B000EPM9NC/ref=pd_bbs_sr_1?ie=UTF8&s=e... 

So in a machine with 1GB of ram and a 2.4GHz processor, my compile time is about 110 minutes.

Pump that up to 3GB of ram and it goes down to 70 minutes.

Throw in the RAMDISK and move the temp file locations for the xilinx server and client to the RAMDISK drive and that compile time goes down to 50-55 minutes.

[JFTICH]

Oops! Accidently posted it before finishing...

I also tried moving my page file to the RAMDISK without any noticeable performance improvements.

  My particular compile starts the page file at 300MB and goes up to 1.2GB during the compile.

There appears to be a lot of times in either the LabVIEW client or the actual xilinx server where the processor usage on my machine is relatively low (less than 5%).

There also appears to be a lot of idle time too...

I wonder if there is a setting in the ini file to push the compiler to work the system harder.

Does anyone have some repeatable results with overclocked 6GHz + systems???

[Graziano]

Hi JFTICH,
  maybe you posted to the wrong board.... FPGA has little to do with FieldPoint!

graziano

[JFTICH]

No kidding! That was not intentional on my part. Thanks for the heads up, I will see about reposting it... and deleting this one...

[Graziano]

I think you can't delete posts