Latest or all posts or last 15, 30, 90 or 180 days.
2024-03-28 08:50:55
Designed for the most demanding needs of photographers and videographers.
877-865-7002
Today’s Deal Zone Items... Handpicked deals...
$398 $328
SAVE $70

$3399 $2999
SAVE $400

$2997 $2997
SAVE $click

$348 $248
SAVE $100

$999 $699
SAVE $300

$5999 $4399
SAVE $1600

$1049 $879
SAVE $170

$4499 $3499
SAVE $1000

$999 $849
SAVE $150

$999 $799
SAVE $200

$5999 $4399
SAVE $1600

$799 $699
SAVE $100

$1199 $899
SAVE $300

$1099 $899
SAVE $200

$348 $248
SAVE $100

$1601 $998
SAVE $603

$3399 $2999
SAVE $400

$3997 $3697
SAVE $300

$5999 $4399
SAVE $1600

$1397 $997
SAVE $400

Mac OS X 10.5.7: hidden changes

I had put together a short piece on Photoshop filter speed (not yet published), and the scalability across multiple CPU cores. Yesterday, I completed various benchmarks, using Mac OS X 10.5.6, the same day Apple released the 10.5.7 update.

Today, I updated to Mac OS X 10.5.7. All the test numbers changed! And not by small amounts, and not always for the better.

I had previously written When More is Less, showing that disabling half the CPUs could dramatically speed up certain Photoshop operations on a dual-CPU Mac Pro Nehalem. Well, that gap has narrowed considerably; the disabling trick still improves matters, but only by a small amount.

Clearly the Apple elves have been at work, because various Photoshop filter tests on my dual-CPU (16 virtual core) Mac Pro Nehalem now run a smidgen faster with all 16 virtual cores, or a lot slower with some cores disabled, as compared to 10.5.6. As they say on the internet, WTF?!!

I don’t have a single-CPU Mac Pro Nehalem, so I don’t know if those users are, as they say, screwed by 10.5.7. Probably not, disabling installed CPUs is not the same as not having them to start with, but it’s certainly possible.

I rebooted a couple of times, and double-checked my findings—same deal. I’m seeing up to twice as slow execution time with some Photoshop filters when some CPUs are disabled, as compared to Mac OS X 10.5.6.

Does this matter?

Probably not for normal users, who won’t ever disable CPU cores. But it shows that loose screws are rattling around inside Mac OS X, not exactly confidence inspiring. The numbers should not have changed unless Apple changed the way OS X schedules CPUs/threads.

We can hope that Snow Leopard (Mac OS X 10.6) will be a lot more intelligent about “scheduling” CPU cores and/or in the efficiency of related matters (synchronization efficiency, memory allocation, pinning threads to a core, etc).

Bottom line: Mac Pro Nehalem dual-CPU systems are probably better off with 10.5.7, but test your own system under 10.5.6 before upgrading to 10.5.7, just to be sure you’re not getting a downgrade. I should have cloned my system onto a backup drive using the Voyager Q first; at this point it’s way too much work to create a 10.5.6 system again.


View all handpicked deals...

Sony WH-1000XM5 Noise-Canceling Wireless Over-Ear Headphones (Black)
$398 $328
SAVE $70

diglloyd Inc. | FTC Disclosure | PRIVACY POLICY | Trademarks | Terms of Use
Contact | About Lloyd Chambers | Consulting | Photo Tours
RSS Feeds | X.com/diglloyd
Copyright © 2022 diglloyd Inc, all rights reserved.