|
|
Favor: Powerbook compile test
|
|
|
|
Registered User
Join Date: Apr 2003
Location: The Internets
Status:
Offline
|
|
I am thinking of switching from a desktop to a alu powerbook 1.67 and want to see what sort of speed hit i will take for compiling apps etc.. There is one benchmark i want to know before i commit to purchasing.
Is there anyone out there with a alu powerbook 1.5 or 1.67 that would be to compile a copy of php for me as a benchmark test? The instructions below will not overwrite your php if you have it installed as we stay in the same folder when we compile.
I tried doing this at an apple store last night but there are no dev tools installed and no one i know owns a new powerbook
I would consider it a favor and would pass along all my karma points to you if you could perform this test.
You need to have dev tools to perform this test.
1) Download the PHP source:
http://www.php.net/get/php-4.3.10.ta...php.net/mirror
2) Wait for it to unzip etc... and find the folder that just downloaded. Mine is on the desktop.
3) open terminal (Applications ->Utilities ->terminal)
4) Type in "cd " <---notice the space after "d"
5) drag the php folder to the terminal as a shortcut so its the current dir in terminal and type return.
6) Copy and paste and hit return the following statement:
./configure --prefix=/www --enable-module=so
It will take a minute to run thru all the settings.
Get your timer ready:
7) type "make" return
It should take less then 4 minutes.
How long does it take on your machine.
On a dual g4 867 (10.3.8) php i get 3:42. I just want to be doubly sure of what I am getting into so once again, thanks for whomever can perform this test.
|
|
|
|
|
|
|
|
|
Dedicated MacNNer
Join Date: Nov 2003
Location: Oslo, Norway
Status:
Offline
|
|
Here are the results from my php-4.3.10 compilation on a PowerBook G4 1.5GHz, 1GB Ram and 80GB 5400rpm hdd.
./configure started 20:38:56
./configure ended 20:40:09
make started 20:40:09
make ended 20:42:10
The 1.67GHz model should be a little bit faster.
|
|
|
|
|
|
|
|
|
Registered User
Join Date: Apr 2003
Location: The Internets
Status:
Offline
|
|
so that would be
2 minutes 1 sec.
+ 8 or 9 % faster for a 1.67 pb
THANK YOU SO MUCH albook!!
Dang. they are pretty darn fast.
I think gcc is by default single proc so that means i was hoping that a 1.67 would be "about' twice as fast as a 867 and it bascially looks that way....
sweet.
thanks again.
|
|
|
|
|
|
|
|
|
Dedicated MacNNer
Join Date: Nov 2003
Location: Oslo, Norway
Status:
Offline
|
|
You�re welcome!
Yes, it sure seems like you only get to use the power from one of your cpu�s.
I say, go go go... To your nearest Apple Store!
|
|
|
|
|
|
|
|
|
Registered User
Join Date: Apr 2003
Location: The Internets
Status:
Offline
|
|
Amazon is running a rebate on the 1.67 pb.
$150 rebate
http://www.macprices.com/g4tracker.shtml
so a PB 1.67 comes in at $2149.... but it ends 4/5/2005 so of course apple will ship quad proc at NAB! But by then i may have my powerbook already
But the powerbook is much lighter so.....
|
|
|
|
|
|
|
|
|
Junior Member
Join Date: Jan 2005
Location: RD Land
Status:
Offline
|
|
5 minutes and 35 seconds on a Powerbook G4 550 with half a gig of ram. LOL.
-RD
|
|
|
|
|
|
|
|
|
Registered User
Join Date: Apr 2003
Location: The Internets
Status:
Offline
|
|
Originally posted by Rampant Desire:
5 minutes and 35 seconds on a Powerbook G4 550 with half a gig of ram. LOL.
-RD
whoah.
anyone want to try this on a dual 2.0 g5 just for kicks?
|
|
|
|
|
|
|
|
|
Junior Member
Join Date: Jan 2005
Location: RD Land
Status:
Offline
|
|
Originally posted by osxisfun:
whoah.
anyone want to try this on a dual 2.0 g5 just for kicks?
Yeah, I guess those times aren't half bad for a 550 G4. If your dual 867 takes over three minutes, it's likely that the compile is only using one processor. If it were utilizing both, logic would dictate that your duallie would be almost as fast as the latest Powerbook. Of course there are many other factors involved such as bus speed, cache sizes, other running processes in the background, inefficiency of the second process on your system.
That said, it would be interesting to see the times on a G5, or more specifically, the difference between a dual and single. I can run a test on a Rev. A Single 1.8 G5 PowerMac this weekend and post the results.
-RD
|
|
|
|
|
|
|
|
|
Registered User
Join Date: Apr 2003
Location: The Internets
Status:
Offline
|
|
coool. yes i heard bus speed is only 15% of real world speed. I have no idea if this is true or under what context (ie more for video compression, less with text editing)
i look forward to your results.
|
|
|
|
|
|
|
|
|
Registered User
Join Date: Apr 2003
Location: The Internets
Status:
Offline
|
|
any update?
I'd love to see your numbers on the G5. If you are too busy I understand...
|
|
|
|
|
|
|
|
|
Senior User
Join Date: Jan 2002
Location: Laurentia
Status:
Offline
|
|
On a 1.8 Ghz *single* processor G5
1.5 Gig RAM
Running safari and web server (taking minor traffic)
Compile time for php 4.3.10 was 1 min 40 sec
|
|
|
|
|
|
|
|
|
Registered User
Join Date: Apr 2003
Location: The Internets
Status:
Offline
|
|
Originally posted by cambro:
On a 1.8 Ghz *single* processor G5
1.5 Gig RAM
Running safari and web server (taking minor traffic)
Compile time for php 4.3.10 was 1 min 40 sec
wow, thank you!
so its really linnear. (unless one can trick out gcc to do dist processing then a dual would matter.)
I could live with this (PB 1.67 compile speeds) since you get mobility.
Thank you again.
|
|
|
|
|
|
|
|
|
Junior Member
Join Date: May 2002
Location: Germany/Heidelberg or Germany/Dresden
Status:
Offline
|
|
Originally posted by osxisfun:
On a dual g4 867 (10.3.8) php i get 3:42. I just want to be doubly sure of what I am getting into so once again, thanks for whomever can perform this test.
Try "make -j <number of jobs>" in order to create more jobs running simultaneously. If you try at least 2 jobs, both of your processor could compile the source code. That should speed up things a little bit.
In fact, "make" then just starts <number of jobs> gcc processes. More info: "man make".
Kind regards,
Lars
|
|
|
|
|
|
|
|
|
Registered User
Join Date: Apr 2003
Location: The Internets
Status:
Offline
|
|
Originally posted by lars-man:
Try "make -j <number of jobs>" in order to create more jobs running simultaneously. If you try at least 2 jobs, both of your processor could compile the source code. That should speed up things a little bit.
In fact, "make" then just starts <number of jobs> gcc processes. More info: "man make".
Kind regards,
Lars
thank you for that tip lars:
G4 dual 867 = 118 secs
Big Big change.
But....
I just tried it on my app and i am running into directory problems when i use -J 2. I did a quick search and others have said multiple dirs can give you some problems. (all the php stuff was in the same dir)
I am going to look further into a workaround for my app.
thanks again.
|
|
|
|
|
|
|
|
|
Junior Member
Join Date: Jan 2005
Location: RD Land
Status:
Offline
|
|
Originally posted by osxisfun:
any update?
I'd love to see your numbers on the G5. If you are too busy I understand...
I'm sorry, I wasn't able to run the test because the G5 didn't have the Dev tools installed, and I didn't have a cd handy.
Again, I apologize.
|
|
|
|
|
|
|
|
|
Forum Regular
Join Date: Apr 2003
Status:
Offline
|
|
Dual 2ghz G5 ....
time make -j 2
68.090u 13.380s 0:43.52 187.2% 0+0k 2+107io 0pf+0w
|
|
|
|
|
|
|
|
|
Registered User
Join Date: Apr 2003
Location: The Internets
Status:
Offline
|
|
Originally posted by johnt519:
Dual 2ghz G5 ....
time make -j 2
68.090u 13.380s 0:43.52 187.2% 0+0k 2+107io 0pf+0w
hi john,
was the total time 43:54 or 13.3?
ps. no problem rampant!
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Forum Rules
|
|
|
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts
|
HTML code is Off
|
|
|
|
|
|
|
|
|
|
|
|