Hi,
thank you. The 2 answers I received so far already highlight the astonishing differences among native w7 systems:
One machine manages to invoke callbacks at the exact rate I requested (10, 22, 117): sync.c:558: thread 13316 hpctime 664 delta 10 sync.c:558: thread 13316 hpctime 296 delta 22 sync.c:558: thread 13316 hpctime 642 delta 117 and shows the useage of several threads when the callbacks is too busy: sync.c:558: thread 17404 hpctime 374 delta 10 sync.c:570: thread 11912 busy 12ms sync.c:558: thread 11912 hpctime 384 delta 10 sync.c:570: thread 17404 busy 12ms sync.c:558: thread 17404 hpctime 394 delta 10 sync.c:570: thread 11912 busy 12ms
Another machine only knows 15.625ms intervals: sync.c:583: TimerQueue period 10ms with 0 stress, flags 0 sync.c:558: thread 5512 hpctime 15 delta 15 sync.c:558: thread 5252 hpctime 15 delta 0 sync.c:558: thread 5512 hpctime 30 delta 15 sync.c:558: thread 5252 hpctime 46 delta 16 sync.c:558: thread 5512 hpctime 46 delta 0 sync.c:558: thread 5512 hpctime 61 delta 15 sync.c:558: thread 5512 hpctime 77 delta 16 sync.c:558: thread 5252 hpctime 77 delta 0 or when asked for a 117ms period: sync.c:558: thread 5252 hpctime 529 delta 109 sync.c:558: thread 5252 hpctime 654 delta 125 and struggles with more threads due to the irregularities: sync.c:558: thread 5252 hpctime 0 delta 0 sync.c:558: thread 6088 hpctime 0 delta 0 sync.c:570: thread 6088 busy 7ms sync.c:570: thread 5252 busy 7ms sync.c:558: thread 5252 hpctime 16 delta 16 sync.c:570: thread 5252 busy 7ms sync.c:558: thread 5252 hpctime 31 delta 15 sync.c:558: thread 6088 hpctime 31 delta 0 sync.c:570: thread 5252 busy 7ms sync.c:570: thread 6088 busy 7ms yet the average result is ok: sync.c:601: 61 callbacks within 593/600ms, avg 9.721/10
Thank you guys, I don't need more tests. I think I've made my point. CreateTimerQueue ought to be changed.
Regards, Jörg Höhle