old-code memory hog?
Posted: 2009-08-12 03:27:08
Yesterday I converted a one-page WordPerfect document. I did what needed to be done, and closed the file. Nisus Writer Pro was still running, but no other documents were open.
This morning I was using the Activity Monitor to check memory usage by a program that seemed to be malfunctioning, and I noticed a process called "soffice.bin". It was using 0.7% of the CPU time and 122.89 MB of real memory (and 895 MB of virtual memory!). It was also a Power PC process rather than Intel. It was called by Nisus Writer, so I assume it is part of the conversion system.
Coming from the days when 128K memory was huge, and the best programs were written in assembler, 122 MB is a hair-raising amount of memory to be hanging onto when not in use. And while 0.7% CPU load may seem small, if all of the programs and processes I used were this demanding "just-in-case", the total load would make a real-world difference.
So: Why is this still PPC instead of "universal" code?
Why doesn't it unload itself when it is not being used?
and, why does it need to sip away at the CPU instead of just waiting until it is called?
Then whole thing seems wasteful and sloppy... unless it normally does not operate this way.
I suspect that the conversion coding is farmed out to someone else. Perhaps it's time to crack the whip and demand better coding.
This morning I was using the Activity Monitor to check memory usage by a program that seemed to be malfunctioning, and I noticed a process called "soffice.bin". It was using 0.7% of the CPU time and 122.89 MB of real memory (and 895 MB of virtual memory!). It was also a Power PC process rather than Intel. It was called by Nisus Writer, so I assume it is part of the conversion system.
Coming from the days when 128K memory was huge, and the best programs were written in assembler, 122 MB is a hair-raising amount of memory to be hanging onto when not in use. And while 0.7% CPU load may seem small, if all of the programs and processes I used were this demanding "just-in-case", the total load would make a real-world difference.
So: Why is this still PPC instead of "universal" code?
Why doesn't it unload itself when it is not being used?
and, why does it need to sip away at the CPU instead of just waiting until it is called?
Then whole thing seems wasteful and sloppy... unless it normally does not operate this way.
I suspect that the conversion coding is farmed out to someone else. Perhaps it's time to crack the whip and demand better coding.