I am working on a design that takes 18 minute to compile.
As my current PC a Dell Precision 3620 (E3-1270 V5 3.6 GHz) is more than 2 years old, I purchased a HP Z4 i9-7940X. Both have 16G of ram, SSD etc. Based on the specs the i9 would compile about 2.5 times (PassMark spec) faster than the Xeon. The 7940X has 14 cores, the Xeon has only 4.
To my complete and utter surprise there was practically no difference in compile time between the 2 computers; The differences in compile time were within 1 minute on an 18 minute scale. (The price was the biggest difference.)
Therefore my question to the developers of Quartus & the forum is: does it make sense to invest in a more recent Xeon based computer? Will Quartus be able to use a 6 Core Xeon faster? I read in similar threads that Quartus is only partly multithreading but that Intel is working on this. Are there any changes to come in the near future? If I look at the load distribution of the Xeon during compile, I have the impression that load is much better distributed on the different cores, is this an impression, or is this really the case?
I posted a similar thread in the processor's forum, but I think, since the question is more Quartus related, in this forum I will have more success.
Maximum processor that Quartus can support is 16 cores https://www.intel.com/content/dam/www/programmable/us/en/pdfs/literature/hb/qts/archives/qts-qpp-5v1..., 6.3.3 and the recommended RAM size can be found in here http://fpgasoftware.intel.com/requirements/18.1/. I could not tell explicitly how much compilation time difference between PCs that have different cores.