I’m trying to tightly bound memory usage for a new application which shares rx buffers with another important one. We have multiple NT cards/firmware. In testing, NTAPI mmapped exactly `Num RSS Queues * Size` from the .ini file. I then set these limits for production and saw NTAPI fail to mmap – it was actually requesting exactly 2x that amount.

How can I determine how much shared memory NTAPI will use? Is this for a ring buffer optimization, or something else?