Hello Scott, Scott Dwyer wrote:
I was not too long ago doing some benchmarking with different versions and options, and can say that I have seen some very nasty plots of the head bouncing all over the place.
Splitting the blocks of a drive with more than (logfile_size / 2) errors tends to be "bouncy", but I am working to make the accesses more linear.
I know one thing that helps is to increase the logfile-size option.
This is why I have multiplied by ten the default value of logfile-size.
But I do agree that some of us would like more control, and are sometimes unsure of how to properly use the options to get it.
I guess we need to discover the way of getting such control while still providing a default algorithm good enough for all devices.
Antonio, I did just finally do a quick benchmark of 1.18-rc1 and do not have any anomalies to report.
Thanks for the feedback. :-) Best regards, Antonio. _______________________________________________ Bug-ddrescue mailing list [email protected] https://lists.gnu.org/mailman/listinfo/bug-ddrescue
