Increasing the frequency will help in increasing the frequency. As per your host application you are performing memory copy and comparison in the same for loop where you are performing SPI write operation. Moving the data processing to a different thread will give some improvement in the throughput. Once SPI write is done you may trigger the second thread to copy the data and start processing.
I understand what you said. you are right moving the data processing would be better.
But there is something wrong may be with my code or it is a limitation of the CY7C65211 + SPI.
I am expecting to download the 32 K byte of data from the EEPROM at least at 100 K byte / S and I am not.
Most USB to SPI chip that are USB 2.0 full speed generally perform poorly, because them have internally a very small buffer and
USB latency constraints. I was hoping that the CY7C65211 could do better.
I am writing a blog post "USB to SPI for .NET" - http://madeintheusb.blogspot.com/2016/02/usb-to-spi-for-net.html
For example the Microchip MCP2210, is very disappointing. But the FTDI FT232H give great performance.
I would like to evaluate fully the performance of the CY7C65211 + SPI.
Could you put me in contact with somebody that is specialized with the CY7C65211 + SPI.