2009-01-06, 22:30
Platform: Ubuntu 8.10 on Athlon X2 6000+, SVN 16927
Hi,
I've just noticed a lot of those in the debug log, stemming from an .avi playback session:
I've looked into the implementation of QueryPerformanceCounter(), and it seems to call through to clock_gettime with CLOCK_MONOTONIC, which is documented to not go backward, ever, so this looks slightly worrying.
Best Regards,
Olli
Hi,
I've just noticed a lot of those in the debug log, stemming from an .avi playback session:
Code:
20:23:14 T:3051738944 M:793210880 WARNING: QueryPerformanceCounter moving backwords by 27080 ticks with freq of 1000000000
20:24:33 T:3051738944 M:796422144 WARNING: QueryPerformanceCounter moving backwords by 62160 ticks with freq of 1000000000
20:24:59 T:3051738944 M:797814784 WARNING: QueryPerformanceCounter moving backwords by 41160 ticks with freq of 1000000000
20:25:12 T:3051738944 M:797814784 WARNING: QueryPerformanceCounter moving backwords by 34280 ticks with freq of 1000000000
20:25:48 T:3051738944 M:797052928 WARNING: QueryPerformanceCounter moving backwords by 10600 ticks with freq of 1000000000
I've looked into the implementation of QueryPerformanceCounter(), and it seems to call through to clock_gettime with CLOCK_MONOTONIC, which is documented to not go backward, ever, so this looks slightly worrying.
Best Regards,
Olli