Version 2.2
Performance
Native C++ Version
The following tables show the estimated PrivMX Endpoint performance on chosen platforms. The assumptions of the tests are as follows:
- each operation is repeated 100 times
- the file used for the test (if applicable) has a size of 1 MB
- network latency is negligible, because the tests are performed using a local server instance
getMessage()/sendMessagegets/sends a 1KB message- entries used by the
readEntry()/sendEntry()have a 1MB file attached
Platform: Core i7 12gen
| Container | Operation (x100) | min (ms) | max (ms) | avg (ms) |
|---|---|---|---|---|
| Thread | getMessage | 19 | 44 | 25 |
| Thread | sendMessage | 7 | 18 | 8 |
| Store | readFromFile | 53 | 75 | 58 |
| Store | writeToFile | 51 | 78 | 63 |
| Inbox | readEntry | 71 | 137 | 81 |
| Inbox | sendEntry | 59 | 116 | 73 |
Platform: ARMv7 32bit (RPi2)
| Container | Operation (x100) | min (ms) | max (ms) | avg (ms) |
|---|---|---|---|---|
| Thread | getMessage | 122 | 163 | 142 |
| Thread | sendMessage | 119 | 161 | 136 |
| Store | readFromFile | 1251 | 1476 | 1292 |
| Store | writeToFile | 763 | 851 | 801 |
| Inbox | readEntry | 1594 | 1853 | 1682 |
| Inbox | sendEntry | 817 | 900 | 854 |
More examples on different platforms are in progress, this page will be updated.