The math all looks good. Unfortunately, Robin's test is not indicative of the correct use case. I suspect most people would be OK with the idea that you can only handle about 50 tracks on an external USB2 drive. The harder results to explain come from internal drives. I will definitely get my own mac mini with a spinner inside running tomorrow and re-run the test there with your additional line, and then redo the math you just did above. We know by existence proof (aka Logic) that the system can handle 128 or more tracks.
On Wed, Mar 11, 2015 at 10:08 PM, Hamilton Feltman <[email protected]> wrote: > > > On Mar 11, 2015, at 6:53 PM, Paul Davis <[email protected]> > wrote: > > > > robin reports back that your dd-tester gives: > > > > 41.1806 MB/sec write > > 42.6064 MB/sec read > > > > (Western Digital 6400AAV - we can't find specs online) > > > > He also tried it formatted as FAT32, with essentially no change in the > results. > > > > Ok, this is enough information to calculate if your drive is performing > within spec. Looking at the first example (block size 65536) > > From your test: > $ ./run-readtest.sh -f 10 -d /Volumes/OSXext/disktest/ 65536 131072 4096000 > # Building files for test... > # Blocksize 65536 > # Min: 8.9910 MB/sec Avg: 9.4645 MB/sec || Max: 0.890 sec > # Max Track count: 51 @ 48000SPS > # Sus Track count: 49 @ 48000SPS > 65536 8.9910 9.4645 0.8898 0.00912 > > So: > file size: 10485760 > num files: 128 > total bytes: 1342177280 > seeks = 1342177280 / 65536 > seeks: 20480 > total_time: 270.484 > > Drive reports: > 42.6064 MB/sec read (44676048 bytes/sec) > Time required to stream 1342177280 bytes -> 30.04 seconds > > Blocksize 65536 > Number of seeks 20480 > Time taken: 270.484 seconds > Time available for seeks: 270.484 - 30.04 = 240.444 seconds > 240.444 / 20480 -> 11.74 ms seek time > > Out of 270 seconds it took to read the 1.25GB, the drive had to spend 240 > seconds performing 20480 seeks, with an average seek time of 11.74 > milliseconds. > This drive seems to be performing well. > > Regards, > Hamilton > >
_______________________________________________ Do not post admin requests to the list. They will be ignored. Coreaudio-api mailing list ([email protected]) Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/coreaudio-api/archive%40mail-archive.com This email sent to [email protected]
