shared mem (again)
08-Aug-93 23:41:54
Sb: #50890-shared mem (again)
Fm: Lee Hart [Microsoft] 76150,2536
To: James Zaun 71652,1762
Something I want to make clear here: You are measuring the number of task
switched in Win32s (which depends on the rate at which each application
empties its message queue in turn) and when you are running memory copies it
slows this down – this makes sense to me. Am I missing something here? It
seems like almost any API usage would radically change the number of times
the current task is changed each second.
Lee