https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=244356
--- Comment #90 from Olivier Certner <olivier.free...@free.fr> --- More tests with SD_128G. Stick is dd-zeroed entirely before each test. 1. Did again the previous test (UFS+SU, 64k blocks and fragments), to confirm and get a feeling of possible variations. Results are worse in the sense that the bandwidth dip happened much earlier, after less than 240MiB had been transferred (less than 11min after start). Bandwidth is approximately the same (4.4MiB/s overall, 27.8MiB/s without stalls). Got a lot more errors in dmesg. 2. Did again the exFAT test. Bandwidth is 18.7MiB/s (50.3MiB/s not counting stalls). This is a bit worse than the previously reported results (comment #86), but still compatible (and still much better than UFS, even 64k). More importantly, I see absolutely no dmesg error at all. So the stick probably is not "failing", it just doesn't handle 64k writes well, whereas 128k is OK. 3. Did a `yes | dd` on the whole stick on a very recent Linux laptop: Got 129MiB/s! It seems I did not do a whole `yes | dd` on FreeBSD, only one for about ~9GiB, and for this one I had seen 62.6MiB/s. So in fact the difference is huge, contrary to what I thought, and much bigger than that reported by Gary (comment #87). I will do a full `yes | dd` on FreeBSD to see. Did not have time to make any progress on GEOM_CACHE, but I should be able to this week. -- You are receiving this mail because: You are the assignee for the bug. _______________________________________________ freebsd-usb@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-usb To unsubscribe, send any mail to "freebsd-usb-unsubscr...@freebsd.org"
