lkpio.blogg.se

That's not really very fast for Thunderbolt.

Internal laptop hd (7200 rpm, sata): Write=42.99 Mb/sec, Read=38.09 Mb/secĮxternal G-Raid (esata): Write=134.76 Mb/sec, Read=192.32 Mb/secĮxternal Seagate hd (laptop drive, USB-2): Write=33.59 Mb/sec, Read=36.38 Mb/secĮxternal G-Raid (Firewire 800): Write=60.79 Mb/sec, Read=66.17 Mb/secĮncrypted sparsebundle image on external G-Raid above (esata): Write=68.66 Mb/sec, Read=81.33 Mb/sec Here's what I get using this method (and dividing by 1048576 to get Mb/sec): This one-liner will test the write speed, clear the cache, properly test the read speed, and then remove tstfile to reclaim disk space:ĭd if=/dev/zero bs=1024k of=tstfile count=1024 & purge & dd if=tstfile bs=1024k of=/dev/null count=1024 & rm tstfile The proper way to do the read test is to be to dd the tstfile created by the write benchmark into /dev/null (but only after clearing the RAM cache by using the 'purge' command). Using /dev/zero as dd's input and output file doesn't hit the disk at all and will return ridiculous speeds like 15-20 GB/sec. The read speed test is flawed as written. Mac OS X Hints editor - Macworld senior contributor