When I started with optimization everything worked fine until I increased the permutations or the time frame. Then at a certain point the CPU usage went down to 1 %, the HDD usage to 100% and the PC was blocked.
I could solve it by changing the optimization method from "Exhaustive" to "Exhaustive (non-Parallel)". This reduces the speed of the optimization for smaller data as the CPU is only used partially, but makes it possible to run manually several optimizations in parallel.
The automatic parallel exhaustive optimization is a tremendous feature for PCs with a large RAM (or maybe even with an SSD), but as soon as the parallel optimization required swapping the PC is blocked completely. For that reason the default optimization method should be set to "Exhaustive (non-Parallel".
Additionally the display of "Elapsed" and "Remaining" should be enhanced by showing day additionally to hours, minutes and seconds.
I could solve it by changing the optimization method from "Exhaustive" to "Exhaustive (non-Parallel)". This reduces the speed of the optimization for smaller data as the CPU is only used partially, but makes it possible to run manually several optimizations in parallel.
The automatic parallel exhaustive optimization is a tremendous feature for PCs with a large RAM (or maybe even with an SSD), but as soon as the parallel optimization required swapping the PC is blocked completely. For that reason the default optimization method should be set to "Exhaustive (non-Parallel".
Additionally the display of "Elapsed" and "Remaining" should be enhanced by showing day additionally to hours, minutes and seconds.
Rename
If increasing the size of the simulation problem significantly slows down the execution, then you've most likely hit the limit of your on-chip cache memory. I would reduce the memory footprint of your problem until you've reach maximum speed.
Also, you should be using a workstation with at least a i7 Core processor so you maximize your on-chip cache size. Now the i9 Core processor does have a larger L3 cache, but it's not that much larger. The i9 is really featuring more parallel processor cores, which you do not need if your bottleneck is the amount of L3 cache memory. You might even find execution faster by disabling some of those processor cores so they aren't all competing for the same, limited, L3 cache memory. You'll need to experiment.
Also, you should be using a workstation with at least a i7 Core processor so you maximize your on-chip cache size. Now the i9 Core processor does have a larger L3 cache, but it's not that much larger. The i9 is really featuring more parallel processor cores, which you do not need if your bottleneck is the amount of L3 cache memory. You might even find execution faster by disabling some of those processor cores so they aren't all competing for the same, limited, L3 cache memory. You'll need to experiment.
Your Response
Post
Edit Post
Login is required