There are a lot of questions on Claro’s performance and how to optimize it.
Here are a few insights into the matter.
Most people realize that image enhancement is a resource intensive process. A lot of calculations done with many millions of pixels. Claro offers image enhancement based on analysis. The image analysis adds to the processing time of images, but is an absolute necessity in automated photo retouching.
For your reference: an 8 mega-pixel image takes about 15 seconds to process on my MacBook, a dual core machine with 8 GB of RAM and an SSD drive.
The machine should be equipped with a 64bit OS and a 64bit Java Runtime Environment before installing Claro.
Claro can process up to 4 images simultaneously. This is a major advantage over our competitors. Each concurrent image process as we call it, needs about 2GB of RAM, and more if you process very large images. To actually use this performance capability of Claro, you need a server with a minimum of 4 processor-cores and 12 GB of RAM. A fast hard-drive is also important; SSD drives do wonders for the performance.
During the installation process, the number of concurrent image processes and the maximum amount of memory per process can be set. There are also two pre-sets available (advice from the installer based on the specs of the machine you are installing on): “workstation” and “dedicated server”.
“Workstation”: Many trial installations are done on a workstation. If that is the case, choose this option. Only one image will be processed at a time. This makes sure the workstation can still be used for other tasks as well; Claro does not consume all of the systems resources. However, if you are looking at the performance, take into account that on a dedicated server it can be 4 times as fast.
“Dedicated server”: based on the system specs, the installer will propose up to 4 concurrent processes and RAM per process. This will get the best performance possible on that machine.
It is an important item with any trial installation of Claro. Need any specific advice? Contact [email protected]