Tests for Computer Equipment
This is a place to post hardware tests. I'll start with a test with a brand-new scoring system for system IO speed. Hopefully with the new FSB and all we'll see a great change in scores between current hardware and new G5s. However, to make the comparison valid, a good cross-section of current hardware needs to be tested first.
I wanted to make this test more relevant, so it covers the spectrum of available hardware where low is 10 and high is 100, without everything jumping off the scale.
This first test takes 1 GB of disk space and about two minutes of time. Don't worry. None of the space is used up permanently.
http://www.myersdaily.org/joseph/ref...iletest.tar.gz
5953 bytes
MD5 (filetest.tar.gz) = 282672f491a3a5f44b0d27b116e519dc
In the folder named "filetest" just open the file which is named "mkfil.command." Make sure you quit anything you don't want to get in the way of the test. (A README file is also included.)
The summary is saved into a file called "mkfil-tests.out." Post this result plus all of your system info if you run the test. Thanks in advance.
I hope that I can contact nVidia and get some graphics tests, also. I hate the graphics tests that take huge, specialized software to run.
I wanted to make this test more relevant, so it covers the spectrum of available hardware where low is 10 and high is 100, without everything jumping off the scale.
This first test takes 1 GB of disk space and about two minutes of time. Don't worry. None of the space is used up permanently.
http://www.myersdaily.org/joseph/ref...iletest.tar.gz
5953 bytes
MD5 (filetest.tar.gz) = 282672f491a3a5f44b0d27b116e519dc
In the folder named "filetest" just open the file which is named "mkfil.command." Make sure you quit anything you don't want to get in the way of the test. (A README file is also included.)
The summary is saved into a file called "mkfil-tests.out." Post this result plus all of your system info if you run the test. Thanks in advance.
I hope that I can contact nVidia and get some graphics tests, also. I hate the graphics tests that take huge, specialized software to run.
Comments
System Specs
Originally posted by cookies
This is a place to post hardware tests. I'll start with a test with a brand-new scoring system for system IO speed. Hopefully with the new FSB and all we'll see a great change in scores between current hardware and new G5s. However, to make the comparison valid, a good cross-section of current hardware needs to be tested first.
If you want to make a disk test meaningful you need to take into account how full the disk is. A disk that is almost full can be signficantly slower than an empty one. To level the field the test should be run on a newly initialized empty disk.
in regards to disk fullness: a complete performance evaluation needs a test of performance under many parameters. It is up to a tester to enjoy the speed of their machine and submit their own machine's information. To show the disk space and disk information, type "df ." or "df -i ." on the same disk where the test is run.
I would like to have all system information automatically included in the log file, but this information should not be "spoofed" or stolen by my script without an informed decision being made by the user. The only way to guarantee that this decision was made with full knowledge and consent is to ask the user to perform the step of reporting system information on their own.
This is an objective performance test, not a disk manufacture comparison. However, it is true that this is not a fair test to evaluate the FSB. Sorry about that.
However, it is a very useful test with a scoring system extremely carefully designed as a ranking system. In fact, it was borrowed from what will be used in the future for graphics tests.
You can look at the forumla and derivation of scores inside of setconst.pl or mkfil.command, which are within the distribution.
QS867 gig ram
WD Caviar 8mb cashe HD
Maxtor 60gb slave
MDD DP1250/1GB RAM, IBM 120GB drive.
G4 Cube, 450Mhz, 384MB RAM, 20GB HDD.
Wait, how did I get about the same score as 1.25Ghz DP G4?
QS dual 800, IBM 40GB 60GXP boot drive with 4.9 GB free.
Interesting...
Total time 155. Score is 35.2022402446234
I have a Beige G3 MT 266 Rev. A Radeon ME, 768MB RAM, 80GB 7200ROM and 120GB 7200RPM HD'
Total time 175. Score is 28.672480406192
And here's a link to the detailed log from the program:
http://www.myersdaily.org/joseph/ref...-tests.out.log
BTW, here's notes about the hard drive upgrade/installation process for an iMac.
http://www.myersdaily.org/joseph/notes/macdrives.html
Originally posted by CubeDude
Wait, how did I get about the same score as 1.25Ghz DP G4?
This test depends almost none on what kind of processor the computer uses, but it is very interesting to see how the actual results come in.
Processor time compared to disk/system access time is about 0.03 seconds out of eight seconds total.
Total time 130. Score is 45.4937356699112
Total time 114. Score is 53.6087437531454
Total time 125. Score is 47.8881278986832
Total time 116. Score is 52.5200494665448
Total time 45. Score is 108.804788989531
QS 933Mhz
Maxtor 120GB HD
MDD DP1250, 1GB RAM, 120GB IBM, 15GB free
Journaling off:
Total time 91. Score is 67.8744656930345
Total time 86. Score is 71.4467837450546
Total time 85. Score is 72.1834998536362
Total time 87. Score is 70.7175866764994
Journaling on (ouch!):
Total time 160. Score is 33.4421386458426
Total time 170. Score is 30.1815489241596
This seems to be a purely disk intensive test, CPU is barely spiking...
Total time 24. Score is 134.961004141906
With ~100% CPU usage
Also I'm curious as to how it would run on a WebDAV, I know DAVs inflate disk performance too, so if anyone has one enabled, try it!
Originally posted by chych
Hmm just for kicks I tried running it on a RAM disk:
Hey, cool. I didn't think about that. Maybe this would make a good RAM and general system test, too. I wonder whether it would need adjustment, however. Obviously, the score system would have to be changed in order to make the values relevant instead of comparing to a hard disk.
I bet the G5s will have an even bigger advantage in a RAM test than in a disk test.
BTW, I just ran it on another computer, with an old, old disk (G3 PowerMac, 400 MHz, 6 GB).
Score 21.9598. Total time 201.
Originally posted by Fluffy
All of my runs.
Total time 130. Score is 45.4937356699112
Total time 114. Score is 53.6087437531454
Total time 125. Score is 47.8881278986832
Total time 116. Score is 52.5200494665448
Hmm, seems inconsistent. This is inconvenient of course, but I wonder if launching the computer in single user mode would result in more similar results.
(type ">console" as the login name, and then type user name and password)
Total time 112. Score is 54.720005711744
Total time 114. Score is 53.6087437531454
Total time 118. Score is 51.4534645443258
Mine seems pretty consistant.
G4/533DP
Seagate Barracuda IV 80Gig (ST380021A)