CapMonster2 system requirements

When using the program on virtual systems, we strongly recommend enabling hardware virtualization.

Software virtualization uses low-level instructions, which will slow down recognition by tens of times.

Also, if your virtual system supports VT-d virtualization technology, then it is recommended to use this particular standard.

The video card can be used to recognize local captchas, saves up to 40% of the processor load (depends on the CPU and GPU), and also gives an increase in the recognition speed depending on the video card model.

CapMonster2 supports CUDA compatible Nvidia graphics cards, list on the official website. This video card must have the latest drivers installed. When choosing a video card, pay attention to the number of CUDA cores, memory size and bit depth.

For example, using an Nvidia GeForce GTX 750 video card, recognition will be 5-15% faster than using only a CPU (4-core Intel Core i5 3.1 GHz processor) with a stable load of different types of captchas. An eight-core Xeon will already outperform this video card in performance, but you will still save CPU resources, both computational load and memory.

But we do not recommend buying a super-powerful video card separately only for using CapMonster2, since the program does not use all the resources of the video card to the fullest. You will not be able to transfer all software processes to the GPU. some of the processing will still be done on the processor. In addition, you simply cannot use all the capabilities of the video card on small local captchas.

It is recommended to buy in portions, gradually increasing their quantity and testing the program. Even powerful hardware does not always allow recognition of complex captchas in several hundreds of threads. For example, when sending captchas to a server, the speed of processing a large number of requests will depend on the network equipment and network load, as well as the ability of the software from which the captchas are sent to support multiple connections. A situation may arise that the network will hang, overloaded with requests, and the powerful server processor will be idle.

In the CapMonster2 recognition settings, there is an option Use adaptive system for setting the degree of parallelism. If you recognize a large number of captchas of various types or the load on CapMonster2 is unstable, then using this option will help you save RAM usage by optimizing the amount of allocated memory for modules with both high and low load.

We do not recommend using this setting when you recognize one or two types of captchas with a constant load, since in this case the use of the adaptive system may decrease the recognition performance.

It is important to take into account that the captcha database is constantly updated, that is, new modules are automatically downloaded to your computer, so you need to have a constant free space on your hard disk.