The typical read benchmark currently is easiest to perform when we do a write , drop cache and read test. the dropping of the cache requires elevated privileges. a technique other tools use is to write an amount of data larger than the systems' memory to flush the disk cache and then read data not in the cache to measure the read disk performance from that.
This mode which could be the new default would free our need to have the user run in administrator mode to drop the disk cache for the read benchmark.
the pollution purge operation:
- get the size of system memory
- create a single file that is x1.1 of system memory
- delete the created file
this would replace the existing disk cache drop which requires admin and would be used if the program is not run in admin mode
Objectives:
have @jslcom or @jamesmarkchan review implementation
The typical read benchmark currently is easiest to perform when we do a write , drop cache and read test. the dropping of the cache requires elevated privileges. a technique other tools use is to write an amount of data larger than the systems' memory to flush the disk cache and then read data not in the cache to measure the read disk performance from that.
This mode which could be the new default would free our need to have the user run in administrator mode to drop the disk cache for the read benchmark.
the pollution purge operation:
this would replace the existing disk cache drop which requires admin and would be used if the program is not run in admin mode
Objectives:
have @jslcom or @jamesmarkchan review implementation