Code running faster on older computer

Newbie82

Prominent
Apr 20, 2017
2
0
510
I am using a 32 bit optimization program and a 32 bit simulation program for my project. The two program interact with each in the following manner. The optimization program generates random numbers, add them to an array, and send them to the simulation program. The simulation program opens a text file, read it, replace the random numbers, run the simulation, obtain performance values and send them back to the optimization software. The two program are connected through a dll file (that's how the optimization program works) that is coded in c# .Net Framework 3.5. I have been using these two programs for the past few years. On my older machine i7 3770 12 GB RAM, it is used to take 7 hours to do this cycle 100000 times. The CPU usage was on average 70% during the run.
Recently, I bought two computers i7 6770 16 GB RAM and AMD Ryzen 7 1700x 16 GB RAM. I assumed that these two computers would outperform the old one but I was wrong. Running exactly the same two programs for the same 100000 times took 11 hours and 14 hours; respectively. The CPU usage is less than 20% during the run.

I have used 4GB Patch by NTCore on both programs over the three machines. I am sure that the two new computers have more computational power than the old one but I cannot seem to find out why it is taking longer to run the programs. Any ideas?

Please keep in mind that I am not a programmer nor a computer/software engineer, so I would appreciate it if you can be thorough and patient.

 

Newbie82

Prominent
Apr 20, 2017
2
0
510


The old one had Windows 7 Enterprise while the new ones have Windows 7 Pro and Windows 10 Pro.
The new ones have a higher HDD speed. The CPU usage is less than 10%
 
It's quite possible that your application(s) are not behaving well in multi-thread environment. I haven't read any reviews on Ryzen, but it's quite possible that it's single-thread performance is lower than that on i7. Unless the data you're exchanging over files is in megabytes, I don't think difference in I/O performance will matter.