New Toys
As I am sure you’ve both heard, Apple released a few new toys today.
iPod Hi-Fi is an iPod speaker system featuring two 80mm mid-range speakers and one 130mm woofer. According to Mr. Jobs it’s “audiophile quality” sound. Any iPod with a dock connector can plug into it, as can almost any device via an audio-in mini-jack (so you iPod shuffle users are not SOL, nor are you AirPort Express users). The down side? It’s three hundred and fifty frickin’ dollars (USD, plus any tax), which seems excessive. Right up until you compare it to the next item released today, that is.
Leather cases for your iPod from Apple. They’re good looking and made from real Italian cow hide, but they don’t close at the top and they cost ninety nine US dollars. Now, I won’t deny that they are fairly attractive cases, but this just seems like fanboy milking to me.
Lastly, Apple refreshed the Mac mini line with Intel CPUs. The new specs are:
- 1.5 GHz Intel Core Solo or 1.66 GHz Intel Core Duo
- 667 MHz system bus
- 512 MB DDR2 667 RAM
- 60 GB or 80 GB hard drive
- Slot-loading combo drive or slot-loading super drive
- Integrated Intel GMA950 graphics
Now, I wasn’t surprised that it’s the Core Solo CPU in the low end machine, but the Core Duo in the higher end one caught me off guard. Not only is the Core Solo more than adequate for everything Apple advertises the Mac mini as being a good machine for, but having a Core Duo in it conceivably takes away from the mind share of the iMac (as in, the iMac has the same CPU as Apple’s lowest end machine).
There is also the integrated graphics card to think about, which is good enough for the video play back that Apple is now hyping for the Mac mini, but it seems funny given that it was just a year ago Apple was mocking low end PCs for having integrated graphics on their Mac mini page.
The price jumped too, to $599 USD and $799 USD respectively for the low and high end models.
Rosetta Performance
We released Geekbench Preview 2 for Rosetta yesterday, and we’ve started receiving performance results for Rosetta. Here are the results we received from Geekbench Preview 2 and Geekbench Preview 2 for Rosetta running on an iMac Core Duo with the following configuration:
- iMac Core Duo
- Intel Core Duo 2GHz
- 2GB RAM
- Mac OS X 10.4.5 (Build 8G1454)
We’re using the baseline scores (not the raw scores) from each benchmark, so in all cases a higher score is better. We’ve computed each benchmark’s score running under Rosetta as a percentage of its score running natively, and again, higher is better.
| Benchmark | Threads | Native | Rosetta |
|---|---|---|---|
| Emulate 6502 | 1 | 183 | 85 (46%) |
| Emulate 6502 | 4 | 346 | 162 (46%) |
| Blowfish | 1 | 350 | 3865 (1104%) |
| Blowfish | 4 | 604 | 4589 (759%) |
| bzip2 Compress | 1 | 139 | 59 (42%) |
| bzip2 Compress | 4 | 263 | 120 (45%) |
| bzip2 Decompress | 1 | 139 | 69 (49%) |
| bzip2 Decompress | 4 | 269 | 135 (50%) |
| Mandelbrot | 1 | 148 | 74 (50%) |
| Mandelbrot | 4 | 281 | 139 (49%) |
| Latency | 1 | 586 | 251 (42%) |
| Read Sequential | 1 | 353 | 179 (50%) |
| Write Sequential | 1 | 161 | 134 (83%) |
| Stdlib Allocate | 1 | 143 | 10 (6%) |
| Stdlib Allocate | 4 | 156 | 9 (5%) |
| Stdlib Write | 1 | 153 | 151 (98%) |
| Stdlib Copy | 1 | 176 | 173 (98%) |
| Stream Copy | 1 | 119 | 88 (73%) |
| Stream Scale | 1 | 119 | 71 (59%) |
| Stream Add | 1 | 174 | 91 (52%) |
| Stream Triad | 1 | 177 | 83 (46%) |
I’m impressed with Rosetta; Geekbench performance running under Rosetta is 40% to 80% of what it is running natively. Plus, running Geekbench under Rosetta is comparable to running Geekbench natively on a Power Mac G5 1.6GHz (our baseline system), at least in the single-threaded tests.
Still, two benchmarks stand out. Stdlib Allocate’s performance is incredibly poor under Rosetta, while Blowfish’s performance is incredibly (and unbelievably) good. While we don’t have an Intel-based Mac at our disposal for development[1], my guess is a poor standard library implementation under Rosetta is holding back the Stdlib Allocate benchmark, while a bug in our Blowfish benchmark (or even Rosetta) is inflating the Blowfish benchmark.
Overall, I don’t think Rosetta performance will be a concern for almost anyone with an Intel-based Mac.
[1] Donations towards a Mac Mini Core Solo gratefully accepted, of course!
Geekbench Preview 2 for Rosetta
Rosetta performance seems to be an area of concern for a number of Mac users; several people have asked us about measuring Rosetta performance with Geekbench. Currently, Geekbench doesn’t have any benchmarks that test Rosetta performance so, as a solution, we’ve released Geekbench Preview 2 for Rosetta, which is a PowerPC-only version of Geekbench Preview 2.
Geekbench Preview 2 for Rosetta will run natively on PowerPC-based Macs, but will run under Rosetta on Intel-based Macs, and should give an idea how efficient Rosetta is at translating PowerPC instructions.
So, if you’ve got an iMac Core Duo or a MacBook Pro, download it and give it a try.
MS re-design of iPod packaging
Ever wonder what goes through the minds of Microsoft’s marketing people when they are designing packaging? Well wonder no more! Watch this video: Microsoft re-designs the iPod packaging. It’s only funny because the process is so bloody believable.
MacBook Pro Geekbenched
Since we released Geekbench Preview 2 we’ve started receiving Geekbench results for MacBook Pros, so we thought we’d compare a MacBook Pro against a recent PowerBook G4.
Setup
Here is the configuration of our two test machines:
PowerBook G4
- PowerPC G4 1.5GHz
- 1.25GB RAM
- Mac OS X 10.4.5 (Build 8H14)
MacBook Pro
- Intel Core Duo 2.0GHz
- 1GB RAM
- Mac OS X 10.4.5 (Build 8G1453)
We’re using the baseline scores (not the raw scores) from each benchmark, so in all cases a higher score is better. We’ve computed the MacBook Pro’s score as a percentage of the PowerBook G4’s score, and again, higher is better.
It’s worth mentioning that Geekbench is a universal binary and does not measure Rosetta performance at all.
Results
| Benchmark | Threads | PowerBook G4 | MacBook Pro |
|---|---|---|---|
| Emulate 6502 | 1 | 76 | 161 (211%) |
| Emulate 6502 | 4 | 75 | 344 (458%) |
| Blowfish | 1 | 232 | 292 (125%) |
| Blowfish | 4 | 186 | 556 (298%) |
| bzip2 Compress | 1 | 82 | 129 (157%) |
| bzip2 Compress | 4 | 81 | 244 (301%) |
| bzip2 Decompress | 1 | 69 | 120 (173%) |
| bzip2 Decompress | 4 | 67 | 255 (380%) |
| Mandelbrot | 1 | 67 | 124 (185%) |
| Mandelbrot | 4 | 67 | 277 (413%) |
| Latency | 1 | 79 | 520 (658%) |
| Read Sequential | 1 | 43 | 277 (644%) |
| Write Sequential | 1 | 71 | 140 (197%) |
| Stdlib Allocate | 1 | 171 | 141 (82%) |
| Stdlib Allocate | 4 | 171 | 143 (83%) |
| Stdlib Write | 1 | 31 | 134 (432%) |
| Stdlib Copy | 1 | 38 | 166 (436%) |
| Stream Copy | 1 | 29 | 113 (389%) |
| Stream Scale | 1 | 29 | 115 (396%) |
| Stream Add | 1 | 19 | 183 (963%) |
| Stream Triad | 1 | 19 | 169 (889%) |
Conclusion
The MacBook Pro outperformed the PowerBook G4 in almost every benchmark (especially the multi-threaded benchmarks). Since all of the MacBook Pro’s baseline scores are over 100, it even outperformed our baseline system, a Power Mac G5 1.6GHz! The only benchmark where the PowerBook G4 outperformed the MacBook Pro, Stdlib Allocate, depends more on library performance than raw hardware performance.
If you’re upgrading from a PowerBook G4 (or even an early Power Mac G5), you’ll certainly notice how much faster the MacBook Pro is, especially with multi-threaded applications.
Google Page Creator
Google, in furtherance of their long term world domination goals, has released a WYSIWYG web page editor called “Google Page Creator”.
It is of course a beta, and is unsupported in Safari (as most Google products seem to be when they first arrive).
Here’s mine! It took about 60 odd seconds to create, including logging in.
Poésie de Apple
Just catching up on some reading this afternoon I stumbled across something interesting, apparently in OS X for x86 there is a message to would be hackers in the form of a poem which is decrypted via the TPM:
Your karma check for today:
There once was a user that whined
his existing OS was so blind,
he’d do better to pirate
an OS that ran great
but found his hardware declined.
Please don’t steal Mac OS!
Really, that’s way uncool.
(C) Apple Computer, Inc.
Which is about two pounds of awesome in a one pound sack.
WordPress
We’ve switched from TextPattern to WordPress (and from Textile to Markdown), so if something seems broken, feel free to blame it on WordPress rather than our technical competence (or lack thereof).
Web Inspector
Web Inspector is a cool feature in Safari that lets you browse the DOM hierarchy of a web page in a compact window. It’s only enabled in the nightly builds of WebKit, though, so you’ll need to grab a build to give it a try.
Entries