Testing setup specifications - ABBYY Documentation

The tests were carried out in ABBYY’s own enterprise environment using the testing methodology described above. We used a standard batch of 69 grayscale pages from the SingleEntryPoint project.

Diagram of the FlexiCapture 12 versus 11 test environment, where the combined Application, Processing, and Licensing Servers connect to File Storage over SAS3 and to the Database, and reach Load Generators, the Backend System, and ABBYY FlexiCapture Processing Stations through 1 Gb/s network switches.

Machine role	Requirements
ABBYY FlexiCapture Servers: Application Server Processing Server Licensing Server	Intel Xeon® E5-2680 v4 2.4 GHz 8 virtual CPU cores 32 GB RAM 2 NICs, 1 Gb/s each Windows 2016
Database Server	Microsoft SQL Server Developer 2016, or Oracle 12c Enterprise Edition Intel Xeon® E5-2680 v4 2.4 GHz 8 virtual CPU cores 128 GB RAM 2 NICs, 1 Gb/s each Windows 2016
FileStorage	SSD disks running on SAS3 with: Continuous reading at 3000 MB/s, writing at 2660 MB/s Random access to a single stream with a 4 KB block size: reading at 19.4 MB/s, writing at 26.1 MB/s. The data acquired using CrystalDiskMark 6.0.0
Typical Processing Station and workload generator	Intel Xeon® E5-2680 v4 2.4 GHz 8 virtual CPU cores 32 GB RAM 1 Gb/s NIC
Backend	Shared SMB network folder where processing results are exported Intel Core i5-2400, 3.1 GHz 4 physical CPU cores No Hyper-Threading 4 GB RAM 1 Gb/s NIC Windows 2008 R2 SP1

Testing FlexiCapture 12 against FlexiCapture 11 at the ABBYY bench (Small)

Testing results at the ABBYY bench

⌘I