Looks like no one’s replied in a while. To start the conversation again, simply ask a new question.

Level 9

50,900 points

Quickie Performance Results for 64-Bit vs 32-Bit

I have both Geekbench 32-bit and 64-bit. So, I made four quick benchmark passes. Here they are with their Geekbench scores.

64-bit SL/64-bit Geekbench--4,263
64-bit SL/32-bit Geekbench--3,910
32-bit SL/64-bit Geekbench--4,159
32-bit SL/32-bit Geekbench--3,809

Pending additional testing, these results suggest what I would hope for. 64-bit programs are faster than 32-bit programs whether run in 64-bit or 32-bit mode. There is only a small performance penalty for running 64-bit programs in 32-bit mode. Lastly, 32-bit programs run faster in 64-bit mode than in 32-bit mode. Apple couldn't have planned these outcomes any better IMO:) 🙂 🙂 😉

Message was edited by: donv (The Ghost)

15" 4 GB 2.8 GHz 500 GB MacBook Pro with LaCie 1TB external 7200 RPM RAID 0, Mac OS X (10.6), PowerBook 540c; iBook G4; 2 Dell Laps; 1 HP Lap; Mega Gateway FX Desk

Posted on Aug 28, 2009 5:50 PM

34 replies

Samsara

Level 4

1,321 points

Aug 31, 2009 8:11 AM in response to donv_the_ghost

And you're a good guy. 🙂
Isn't there a Latin term for you? Deux Ex Machina doesn't quite fit.

donv_the_ghost Author

Level 9

50,900 points

Aug 31, 2009 4:54 PM in response to Samsara

HaHaHa And, thank you. The Ghost was my alter ego in my dark period. 🙂 I didn't study Latin. But, how about "He who lives in the land of whatever."

Samsara

Level 4

1,321 points

Aug 31, 2009 5:19 PM in response to donv_the_ghost

I remembered a term called "Ghost in the Machine", but the term I found meant "God from the Machine".

Dark period huh, I hope that's over with. I've had too many to know that dark really means hellish. At least that's what mine were like and they still happen from time to time, frequently followed by the opposite.

Take care, Donv (The Ghost)
Steve

Aug 31, 2009 5:21 PM in response to donv_the_ghost

I have an EFI-32 only 2006 edition Mac Pro, and I am noticing 10% or so differences with Geekbench between 32 and 64 bit:

Rosetta (Just for the heck of it): 2612
32-bit: 5631
64-bit: 6196

Kevin Horn

Level 4

1,836 points

Sep 1, 2009 2:06 PM in response to donv_the_ghost

Thanks for taking the time to do the tests and post them here, all very interesting.

And, in case someone doesn't know this, Activity Monitor tells you which apps are 64 bit, and which processes as well. Right now none of my third party apps are 64 bit.

Kevin

donv_the_ghost Author

Level 9

50,900 points

Sep 1, 2009 2:44 PM in response to Kevin Horn

You are very welcome. I won't be doing any more. I just posted the results of a test that was unfavorable to SL. And, the thread was deleted after comments from some heavies and my replies to them. So much for free speech, open mindedness, objectivity, honesty, etc., etc.

Kevin Horn

Level 4

1,836 points

Sep 1, 2009 3:45 PM in response to donv_the_ghost

I saw a thread of yours that I couldn't get in to and was wondering what that was about.

Apple does stuff like that here from time to time, but not all that often.

When they came out with the new Aluminum iMacs myself and several others ranted for quite a long time about Apple putting a cheap TN screen in the 20" that was darker at the top and washed out at the bottom and completely got rid of the matte screen completely. We filled up 3 or 4 big folders before they stopped it.

Kevin

donv_the_ghost Author

Level 9

50,900 points

Sep 1, 2009 4:13 PM in response to Kevin Horn

Yes, they do. I was told that the currently deleted thread was not on topic. It seemed pretty pertinent to using SL. But, then it did not answer a question. As though many are not interested in the relative performance of SL and L.

Samsara

Level 4

1,321 points

Sep 1, 2009 7:32 PM in response to donv_the_ghost

One also has to realize that the Moderators will sometimes let things go until a discussion member complains about it. That's the really sad thing, Tattletales.
Well, at least you know that you're appreciated by the majority here, that's the important thing. 🙂

Sep 2, 2009 8:20 AM in response to donv_the_ghost

If you're willing to repost the deleted results, I'd be interested to see them.

Kort

Level 3

828 points

Sep 2, 2009 9:01 AM in response to donv_the_ghost

Me too Don. I also appreciate you taking the time to post and discuss them and the ones at the beginning of this thread.

I noticed another thread about the topic of people that had little or no problems with their SL installs. It was closed by the moderators when it reached page 10. I would have thought they'd appreciate a little positive PR at a time like this.
😉

donv_the_ghost Author

Level 9

50,900 points

Sep 2, 2009 12:31 PM in response to Samsara

I didn't really have a dark period. I had some dark days here and changed my name to The Ghost. Maybe, being a God from the machine would be pretty good. 🙂

Message was edited by: donv (The Ghost)

donv_the_ghost Author

Level 9

50,900 points

Sep 2, 2009 12:49 PM in response to Mike Connelly

Mike and Kort. I would love to post them here, but I'd probably just get in trouble again. I'll just give a brief summary. Please don't make a big deal of my deleted thread, my results, or whatever.

I did some incredibly carefully controlled xBench tests of SL vs L. L handily produced the higher average overall score. I wondered why. So I paid attention to the sub-results. SL was slightly better for the first four, as I recall, tests, but then L took over "winning" handily on tests that surprised me a great deal--Quartz Graphics, OpenCL, and User Interface.

I basically ask the question of why xBench would produce these results in my post. I still wonder because they are hard to believe. Several other more elaborate tests show SL to be the winner on everything. And, you may have seen my other results (linked above) that imply that SL is the better performer.

Do you have any ideas about why xBench might have produced the results it did on the three tests specifically mentioned above. I'd really like to figure this out. I suspect that, for some reason, xBench is biased against SL on these tests, but I have no idea what might produce the bias. These are the issues I was raising in my post, although I don't think I was clear enough and may have expended too many words on the write up.

Message was edited by: donv (The Ghost)

donv_the_ghost Author

Level 9

50,900 points

Sep 2, 2009 12:52 PM in response to lurkingbf

Thank you for posting your results. They are interesting.

sinX

Level 1

20 points

Sep 4, 2009 10:21 AM in response to donv_the_ghost

I do know that on 10.5.7 they made some changes to quartz extreme that made graphics in the UI and otherwise unstable. It may be worth it to test 10.5.6 and see the diiference.

Quickie Performance Results for 64-Bit vs 32-Bit