| |||
| [Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] |
| I think all other things being equal it might make a reasonable basis for comparison. But you would need to be very careful that all other things were in fact equal. For one thing it seems to degrade reasonably quickly if anything else is going on. This is a Mac Mini, 1.42GHZ G4 512MB 80G. First run I was playing X11 gnuchess at the same time, unfortunately lost on a sort of nifty back rank mate. You wouldn't think real I/O bound anyways. FileOutputStream write 8M (avg) 161 FileInputStream read 8M (avg) 196. FileOutputStream fill and write 8K avg 215 FileInputStream read and check 8K avg 475 Time: 56.044 Tried some subsequent tests without anything I knew of going on... FileOutputStream write 8M (avg) 113 - 82 - 80 FileInputStream read 8M (avg) 135 - 160 - 164 FileOutputStream fill and write 8K avg 130 - 115 - 189 FileInputStream read and check 8K avg 286 - 291 - 308 Time: 34.484 - 37.426 - 38.552 Hmm, better all the way around. Last one pretty poor for some reason unknown to me. Let's try one where one where we fire up Safari at the same time and do a little browsing... FileOutputStream write 8M (avg) 309 FileInputStream read 8M (avg) 156 FileOutputStream fill and write 8K avg 160 FileInputStream read and check 8K avg 267 Time: 47.213 The initial launch seemed to hit heavy on the write time and have only so-so, if any, impact on subsequent results. But this still slowed total time a fair amount. So you would I think need to stress that testing should take place on an idle machine, or carefully compare it against the same other stuff going on. Or, you are going to get results where to a large extent it depends on what else might be going on? (usually attributed to Mark Twain or Benjamin Disraeli): "There are three kinds of lies: lies, damned lies, and statistics." I always thought that was Disraeli but googling seems to show that even that is subject to speculation. |
_______________________________________________ Do not post admin requests to the list. They will be ignored. Java-dev mailing list (email@hidden) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/java-dev/email@hidden This email sent to email@hidden
| References: | |
| >McDougall's Disk Masher (was: Mac OS X File IO Performance Issue?) (From: Craig Mattocks <email@hidden>) |
| Home | Archives | FAQ | Terms/Conditions | Contact | RSS | Lists | About |
Visit the Apple Store online or at retail locations.
1-800-MY-APPLE
Contact Apple | Terms of Use | Privacy Policy
Copyright © 2007 Apple Inc. All rights reserved.