In NXPowerLite for File Servers v7.1.4, it is now possible to optimize more than one file at a time. By default, NXPowerLite for File Servers will run two optimizers in parallel. However, you may wish to change this setting. This article shows you how you can do this, both for evaluation and optimization runs.
This requires version 7.1.4 or later. If you have an earlier version please click here to download the latest version.
Changing the number of concurrent optimizers
- Close the dashboard, and from the services console (services.msc) stop the NXPowerLite File Server service.
- Open the registry editor, and navigate to the appropriate key:
- HKEY_LOCAL_MACHINE\Software\Neuxpower\NXPowerLite Fileserver\7.0 (on 32-bit systems)
- HKEY_LOCAL_MACHINE\Software\Wow6432Node\Neuxpower\NXPowerLite Fileserver\7.0 (on 64-bit systems)
- Create or edit a DWORD key called "Max optimizer threads", with a value set to the number of optimizers that you wish to run concurrently.
Now launch the NXPowerLite for File Servers dashboard and resume or start the run as appropriate.
Viewing the concurrently-running optimizers
Each optimizer thread launches a separate instance of the process 'Optimizer.exe', which you can view from your preferred process monitoring tool (e.g. Task Manager).
Optimizer.exe is only run while it is actually optimizing an individual file or zip file, and exits once that file is complete. At times, therefore, you will see your configured number of Optimizer.exe instances running, at other times there might be fewer running.
Choosing how many optimizers to run concurrently
As stated, by default NXPowerLite for File Servers now runs with a default of 2 concurrent optimizers and on many systems that will be sufficient to see a substantial increase in performance, compared with earlier versions.
NXPowerLite can generate heavy CPU and disk I/O traffic. If you do increase the number of concurrent optimizer processes, we recommend monitoring your CPU and disk I/O to ensure that this is not becoming a bottleneck affecting performance of the system as a whole. In our testing, we have found that you can increase the number of concurrent optimizer processes to 4, before other constraints impact on the performance.