Optimizing multiple systems in multiple analysis windows in parallel

Hello,

I have a few very simple AFLs to optimize. Those all have vast number of potential combinations of the parameters to optimize.

My PC has one CPU that can run 8 threads at a time.

When I start optimization of one AFL then all 8 threads combined use only up to 20% of the CPU.

In order to utilize the CPU more - I opened another APX+AFL in separate tab and started optimization in the second tab as well.

Now the CPU is still used only up to 25%.
First APX time estimate for optimization, increased from 3:20 to 4:10.
Second APX of similar complexity as compared to the first AFL - shows 4 hours estimate to optimize.

So optimizing multiple systems in parallel does not seem to help in better utilizing the CPU.

What else can I do to ensure more efficient use of the CPU?

Note: RAM is only 59% in use.

Please DO READ Knowledge Base article on that subject. It contains ALL answers:

Your systems are apparently too simplistic to saturate CPU, you are constrained by memory (RAM) bandwidth. In today's computers single core is able to fully saturate RAM bandwidth.

1 Like