Friday, December 23, 2005

Aperture 1.0.1 update removes "panties"

I was poking around in the Aperture 1.0 application binary (/Applications/Aperture/Contents/MacOS/Aperture) and noticed some very strange "code" in there, about panties and Cap'n Funktographer and his Peanut Butter Elephant.

Unfortunately, this very interesting code has been removed in the new Aperture 1.0.1 Update (11.4 MB) .

Fortunately, the new update does have some nice improvements:
Aperture 1.0.1 Update addresses a number of issues related to reliability and performance. It also delivers improved image export quality and metadata handling.

Among the key areas addressed are:

- White balance adjustment accuracy and performance
- Image export quality
- Book and print ordering reliability
- Auto-stacking performance
- Custom paper size handling

A quick test confirms that JPEG exports are vastly improved:

Aperture 1.0:

Aperture 1.0.1:

(Click on pictures to enlarge)

Speed in some areas is also somewhat improved, and Auto-stacking is now fast. However, RAW conversions were not updated, as these will require an OS update (since RAW support is a function of the OS, not the application). Also, EXIF data is still not retained with exported TIFF files.

It was nice to see that Apple has not (yet) prevented the latest iBook (with Radeon 9550) from running Aperture, despite what the new Aperture Compatibility Checker 1.0.1 says when run on that iBook. Not surprisingly however, this update disables the hack that allowed Aperture 1.0 to run on other unsupported hardware (like G4 Power Macs).

[Update 2005-12-23]

P.S. I almost forgot... The Aperture binary also has references to some current and (very) old PowerPC and x86 CPUs.

I do not know the significance of these references, but they're in both Aperture 1.0 and 1.0.1.

Monday, December 19, 2005

Yonah is fast. Very fast.

AnandTech has published updated Yonah benchmarks, which show that Yonah Core Duo 2.0 GHz performs very well, even against its desktop counterparts like the dual-core 2.0 GHz Athlon 64 X2 CPU (with 1 MB L2) in professional applications.

Some are also reporting that Yonah will appear in an iBook in January. As I don't feel that any iBook would get a dual-core chip, I surmised that either these predictions are wrong, the iBook would get the more expensive 2 MB single-core Yonah instead of the cheaper 1 MB L2 version, or else the roadmap was wrong. Now I'm thinking that unless the product is in fact a PowerBook, perhaps the last option was the correct one. Researching this further, I see that some are reporting that in fact the cheaper 1 MB single-core Yonah will be released in January too. If that is true, then it would be perfect for a Macworld iBook Yonah release.

If this Yonah iBook prediction does come true, one wonders what will happen with the rest of the laptop lines. It would be nice to see dual-core Yonah PowerBooks released at the same time too, preferably including a 13" model. Even if there are issues with compatibility with legacy code, Apple could always continue to sell one or two PowerPC models for those who need perfect PowerPC compatibility.

Saturday, December 17, 2005

iPod shuffle 1 GB completely sold out

It seems the 1 GB was quite a popular item for Christmas shoppers. It is now completely sold out at Apple. This suggests we may see an updated version of the shuffle in 3 weeks, at Macworld.

Apple to kill Aperture's iBook support, and clues to a widescreen 13" PowerBook?

According to the first iteration of Apple's Aperture Compatibility Checker, Aperture supports the latest iBook. However, according to Aperture Checker 1.0.1, that is no longer the case:

Aperture 1.0 does in fact currently support screen sizes under 1024x768, and installs fine on current iBooks with the Radeon 9550. However, if updated versions of Aperture enforce the new 1280x780 requirement, that means that all current iBooks will be excluded. That isn't terribly surprising I suppose, since Aperture's technical specifications have never included the iBook as a supported machine despite what the Aperture Checker software said.

It is curious that the resolution of 1280x780 was chosen however. There are no Macs in existence with a screen resolution of 1280x780. Perhaps this is a clue to a new 13.3" PowerBook? I wish, although I will note that most 13.3" screens are 1280x800, the same 16:10 aspect ratio as most of Apple's current widescreen displays.

Even if the 1280x780 restriction is enforced at install, one wonders if it still may be possible to install the software an a current iBook without hacks to Aperture, just by installing the Screen Spanning Doctor and then plugging in an appropriately sized external display. That won't work though if Aperture 1.0.1 includes checks for the resolution within the app itself and not just in the installer.

GPU-assisted H.264 acceleration has arrived

H.264 acceleration on GPUs is now a reality, courtesy of ATI... if you have an X1x00 series GPU. This is just in time too for Blu-ray and/or HD-DVD, as these technologies have mandatory support the very compute-intensive H.264 codec.

To give you an example of how well this technology works, AnandTech has benchmarked a real world system decoding Apple's H.264 HD QuickTime videos. The H.264 speedup is impressive, and shows a significant decrease in CPU utilization during playback of these videos.

Unfortunately, the lower end X1300 GPU does not support H.264 acceleration above 576p. The X1600 or higher is required to decode 720p, which is disappointing since one would think that a 13.3" PowerBook with a 720p-compatible 1280x800 screen could get a Mobile X1300 within the year, but would not get an X1600 any time soon.

Friday, December 09, 2005

Aperture runs on the G4 Cube

To my surprise, Aperture runs on my Cube, and runs tolerably. I had to hack it to get it to run, given that the Cube does not have official Aperture support, but everything works just fine (albeit fairly slowly). This Cube does have a 1.7 GHz G4 which is adequate as far as G4 speeds go, but the real surprise is that Aperture has no problem with this Cube's GeForce 2 MX, a GPU that is several generations behind even the GeForce FX 5200 (which is also unsupported). The GeForce 2 MX does not even fulfill the hardware requirements for Core Image's GPU-based acceleration. That likely means that Aperture supports Core Image's Altivec fallback on the CPU, despite the fact the application is built never to run on a machine like this. At least this GPU has Quartz Extreme support though. One wonders if Aperture could run on even an older GPU like the ATI Rage 128, which does not even support Quartz Extreme.

I won't detail the Aperture hack here, but suffice it to say that there are explanations available elsewhere out there online, and a valid serial number is still required of course. For the sake of all the people out there with adequately spec'd G4 Power Macs, I just hope that Apple doesn't choose to disable the hack in the 1.0.1 update to Aperture.

Now I said it's tolerable on the Cube, but in truth maybe I think it's tolerable just because I didn't test it for too long. Fortunately, I already have a faster and fully supported iMac (G5 2.0 & Radeon 9600 128 MB). Aperture is not particularly fast on my iMac either, but at least it's a noticeable improvement, although nowhere near as fast as the dual G5 2.3 GHz Power Mac with GeForce 6600 256 MB I had the chance to test out with this application.

Wednesday, December 07, 2005

No PowerBook G5 ever

In case you've been hiding under a rock and somehow still believe a PowerBook G5 is coming, this c|net article should finally convince you otherwise. Michael Mayer, CEO of Freescale, states that IBM chose not to finish the work on a laptop version of the G5, partially so that they could focus more of their resources on chips for game consoles. He should a pretty good idea of the goings-on at IBM too, as he was general manager of IBM Microelectronics until November 2003.

He also states that Jobs wanted to go with Intel chips many years ago, but it was the promise of the G5 that kept Apple on PowerPC. That is not a complete surprise, since Rhapsody was even released to developers on Intel, and we are all well aware of Apple's Marklar project. Furthermore Jobs has had a long and good relationship with Intel, partially because of his days at NeXT, and because of personal relationships with Intel brass. However, back then Apple was in a precarious position, and in my opinion it would have been a lot more risky then for Apple to go Intel than now. I'm sure Jobs didn't need a huge amount of convincing to stick with PowerPC, especially when it was IBM doing the convincing with the potential of their G5.

One other interesting point made in the article is that Apple has a contract with Freescale that doesn't end until 2008. I'm not sure what that means, but I would be somewhat surprised to see G4 Macs sold after 2006. That said, there will be a lot of G4 Mac repairs needed for years to come. It is possible that this contract is simply to ensure adequate repair parts are available until institutional G4 Mac AppleCare contracts expire.

Intel Macs: 666?

The DigiTimes reports that while new Intel Macs may be demoed at Macworld 2006 in January, they will not be released until June 6, 2006. That long of a lag time seems strange, but it's interesting that 06/06/06 date was mentioned (and not just for religious reasons). First of all, the date is exactly one year from the announcement of the switch to Intel. More important however is the fact that "666" has a special place in Apple's history, since the original Apple 1 was released at $666.66 (in 1976, making 2006 the 30th year Apple has been selling computers).

If that date is true, it could make sense for iBooks and Mac minis, since the cheaper single-core 1 MB L2 version of Yonah is scheduled for release in Q2 2006. However, I still hold out hope for a 13" PowerBook announcement in January. Such a machine could use a dual-core 1.66 GHz Yonah, or perhaps more likely, a single-core 1.66 GHz Yonah with 2 MB L2. Both chips are on target for release in volume in January.

One curious tidbit in the article is the claim that the new Macs will sport an "Intel Inside" logo. That seems particularly out of character for Apple, given Apple's extreme emphasis on aesthetics for its computers.

The article should be taken with a grain of salt though, since when it comes to Apple prognostication, The Digitimes has no better a record than the usual rumour sites, and it could just be that the sources quoted are simply repeating the June 6 date suggested last year.

Sunday, December 04, 2005

Hints to an Aperture hack

A lot of people (including myself) have been complaining that Aperture has unnecessarily strict hardware requirements. Well, it seems the Aperture Compatibility Checker may provide some hints as to how to ease some of these restrictions:

The above is the result I get with my modified Compatibility Checker on my fully supported iMac (20" G5 2.0 GHz, Radeon 9600 128 MB, OS X Tiger 10.4.3, 1.5 GB RAM).

[Update 2005-12-04]

More fun with the Aperture Compatibility Checker. :)

Saturday, November 26, 2005

Aperture compatibility, and it's shipping?

According to the Apple Store front page, Aperture is now shipping. However, when you attempt to purchase it, it says it's shipping in one week. Furthermore, Amazon has a November 30 release date, and there have not yet been any reports of end users receiving their copies.

In addition, reports have been coming in about hardware compatibility with Aperture as determined by Apple's Aperture Compatibility Checker. Some confirm my observation that 2005 iBooks are listed as supported, but it seems the iBooks are the only big surprise. As expected from the minimum requirements, prelimary reports indicate that these Macs are NOT supported:

1) Any 12" PowerBook
2) Any G4 iMac
3) Any G4 Power Mac
4) G5 iMacs with GeForce FX 5200 Ultra (August 2004)
5) G5 Power Macs with GeForce FX 5200 Ultra (2003 & 2004)

I can understand Apple's decision not to support 12" PowerBooks or other Macs with GeForce 5200 series GPUs. Also, the G5 Power Macs are not a issue, since one can easily upgrade the GPU. However, given Apple's decision to support G4 PowerBooks, I do not understand their decision to exclude all G4 Power Macs, including duals. Perhaps Apple might have a change of heart later on and allow support of Power Macs with fast G4s and appropriate GPUs like the Radeon 9600 Pro or better, but I wouldn't count on it.

[Update 2005-11-26]

Aperture is now shipping. Emails went out today to some end users indicating their orders shipped today, with expected delivery beginning early next week.

[Update 2005-11-27]

I almost forgot... While I don't have end user confirmation of this yet, Aperture will not likely run on any 1.6 GHz Power Mac either, regardless of the video card. The 1.6 GHz G5 does not meet Apple's minimum CPU speed specification.

Intel iBook for January?

Several websites are reporting that Apple is preparing for a release of an Intel iBook in January. If it is only the iBook that is updated in January with an x86 chip, then this would be an interesting situation for Apple. According to Intel's mobile CPU roadmap January will see the release of a single-core and a dual-core Yonah CPU, both with 2 MB L2 cache. The iBook will undoubtedly get a single-core CPU, but I would have expected it to get the cheaper single-core Yonah with 1 MB L2. However that chip isn't scheduled for release until Q2 2006. One of several things could be true: 1) There is no January iBook, or 2) The iBook will get a souped up 2 MB L2 single-core Yonah, or 3) The roadmap is wrong. A souped up Yonah single-core would not be extremely cheap. One would hope that Intel in the iBook doesn't mean increased prices over what current iBooks cost. Perhaps there is a fourth option: Maybe the Intel machine for January isn't an iBook at all. Besides the possible Intel Mac mini, it could be that Apple will introduce a replacement to the 12" PowerBook. This would make sense, since at the last PowerBook update the 12" saw no significant upgrades.

The other interesting part about this iBook prediction is the fact that there is no mention of any upgrade to the PowerBook line. If the iBook (or smallest PowerBook) did get the single-core Yonah with 2 MB L2 in January, it'd be at 1.66 GHz. Such a chip would be faster than the G4 in any current PowerBook in existence (at least with software that runs natively on x86). For this reason one would expect that after such an update, it wouldn't be too long before the (larger) PowerBooks got upgraded too. However, I will admit that one problem with this reasoning is the spectre of software compatibility. PowerBook users often need pro Mac software, and an early PowerBook Intel migration would mean that some pro software may not work on these machines. However, if the Altivec'd Rosetta rumours are true, then this pro software issue could be somewhat less significant.

Friday, November 25, 2005

Rosetta now supports Altivec?

The OS X 86 Project is reporting that Rosetta now supports Altivec in Mac OS X 10.4.3 for x86. Previously, Rosetta could only emulate the G3, and thus any PowerPC-only application that required the G4's Altivec instructions to run would not run at all through Rosetta on OS X for x86. If Rosetta has now gained the ability to dynamically translate Altivec, that would be a huge benefit during the transition period. While Altivec translation is bound to be slow, slow functionality is better than no functionality. This claim of Altivec support is unconfirmed however, and Apple's developer site still states that Altivec is not supported. (Note that such support is not required at all for the iApps. All the iApps, including iTunes, now apparently have universal binaries with native support for x86.)

The same article also states that ATI video cards for PCs are now also supported, in the latest seeds of OS X x86. If true, this would reduce the need for dedicated Mac video cards. People could simply go out and buy the latest and greatest PC video cards at the local video store, and plug them right into their Power Macs. Presumably any such support would be best on video cards utilizing reference designs, but the selection of PC video cards based off reference designs is still bigger than the selection of Mac video cards. However, this claim of PC video card support is also unsubstantiated.

We may find out the truth in a few short months though, as x86 Macs are (hopefully) coming relatively soon.

Sunday, November 20, 2005

Install Front Row on any Mac with Front Row Enabler

Apple's newest iMac G5 with built-in iSight and remote control comes with a new media centre application called Front Row. This application has not been released for any other Mac. However, Apple has recently released a free Front Row 1.0.1 update which can be installed on any Mac using Pacifist and Front Row Enabler.

Above is a picture of Front Row running on an older G5 iMac. While Front Row can be controlled with the keyboard, it's not all that useful without the remote control. However, it at least gives us an idea what Front Row is like without the need to find a new iSight iMac to test.

Aperture supports the iBook

Apple has now posted a Compatibility Checker on its website, which checks your system's specifications and determines if Apple's new flagship photography application Aperture will run on your system.

Much to my surprise, my (July 2005) iBook is supported:

Aperture supports the iBook.

Although this iBook's 1.33 GHz G4 meets the minimum CPU speed requirements and its Radeon 9550 does include the hardware (Core Image) video features necessary for Aperture, this GPU is not listed as supported on Apple's Aperture tech specs page. Furthermore, the page states that only 15" and 17" PowerBooks are supported. That G4 PowerBook (or G5 desktop) requirement is echoed by the Compatibility Checker when run on unsupported machines (like my 1.7 GHz G4 Cube with GeForce2 MX):

Aperture doesn't like G4 Power Macs.

Assuming the Compatibility Checker is correct and Aperture will run on the iBook, the small 1024x768 screen resolution, the relatively slow 1.33 GHz CPU, the slow Radeon 9550 GPU, and the anemic 32 MB GPU RAM will surely make Aperture on the iBook a less than ideal experience. Still, I am pleased. Being able to run it slowly is better than not being able to run it at all.

One wonders if 12" PowerBooks with the GeForce FX Go5200 or G5 iMacs (and G5 Power Macs) with the GeForce FX 5200 Ultra are also supported. The GPUs in these machines are also compatible with the hardware video features required by Aperture, although they are not listed on the Aperture tech specs page either, and they are even slower than the Radeon 9550. I also wonder if some G4 Power Macs are supported. The more recent G4 Power Macs have fast enough CPUs and some may be running reasonably fast Core Image compatible GPUs as well, but Apple makes no mention whatsoever of supporting G4 Power Macs with Aperture.

Sunday, November 13, 2005

Quad G5 scores 1101 in Cinebench

Bare Feats has posted an impressive Power Mac G5 Quad 2.5 GHz Cinebench score of 1101. The previous highest Mac score was held by a dual G5 2.7, and was less than 700. The 1101 score by the Quad G5 is about the same as one would get from a quad 2.4 GHz Opteron 280 machine with the 32-bit version of Cinebench. (The Opteron scores higher with the 64-bit version of Cinebench under 64-bit Windows, but there is no 64-bit Mac OS X version with which to compare.)

Although the single processor score was not posted in his review, Rob at Bare Feats was kind enough to tell me via email that it is 359, and thus the multiprocessor speedup is 3.07. This disproves the theory held by some that Cinebench was being held back by poor support of quad CPUs in Mac OS X. This idea was put forward when another quad G5 tested got scores of 359 and 1016, with a low multiprocessor speedup of 2.83. While the speedup of 2.83 is low, Bare Feats' speedup of 3.07 for the Quad 2.5 is quite good, and in line with what would be expected with quad machines based off dual-core Opterons and Xeons. The large variation for the multiprocessor score is curious however, although not unexpected since previous dual G5 Macs also had large variations in the multiprocessor scores.

In addition, the Quad G5 2.5 970MP has been tested by Luxology and it turns out that it is more than twice as fast as the Dual G5 2.5 970FX in modo. It seems the 970MP's increased L2 cache (for each core) really helps performance in this application. It is also interesting to watch their load balancing video which demonstrates that all four cores are at 100% utilization during the render.

Wednesday, October 26, 2005

nVidia GeForce 7800 GT now available for order

The Mac version of the PCIe nVidia GeForce 7800 GT is now available for order with new Power Macs at the Apple Store. The 7800 GT is a US$350 upgrade from the 6600, and a US$400 upgrade from the 6600 LE. Unfortunately, this upgrade changes the ship times to 6-8 weeks, which means that 7800 GT endowed Power Macs won't ship until December.

However, this GPU might be well worth the wait for those of you interested in running software like 3D games, Motion, or Aperture, since the 7800 GT offers a very significant 3D performance boost over the 6600 series GPUs.

IBM releases details of G5 970MP

Yesterday, IBM's Norman Rohrer presented the details of their new 2.5 GHz dual-core G5 970MP at the Fall Processor Forum in San Jose.

It looks like they've gotten the power utilization down quite nicely in these chips, as the new 2.5 GHz puts out only 100 Watts maximum with both cores active. Estimates are that when the 2.5 GHz 970FX first came out it was somewhere in the range of 80-100 Watts, and that's for only a single core chip, with less cache, and also on a 90 nm process. The 2.5 GHz 50 Watt per core power spec for the 970MP likely means something like a 2 GHz low power single-core laptop G5 or a lower clocked dual-core laptop chip could be produced. In fact, Rohrer states that "low power, high performance" dual-core G5s will be available at 1.2 GHz and above. However, it is very possible that Apple won't release such a machine and will wait until dual-core Yonah or dual-core 64-bit Merom chips are available from Intel before moving away from the G4 in their laptops. As for the Power Macs, one suspects that Apple will wait until the Woodcrest Xeon chips are available before switching away from PowerPC. These dual-core chips are the server version of Merom, and are planned to have a maximum power utilization of about 80 Watts.

Rohrer also released SPEC CPU2000 scores yesterday for the 970MP. While the integer score of 1438 in mediocre, the floating point score of 2076 is quite impressive. Given the excellent fp scores, it's likely that the improvement isn't due only to the L2 cache boost. Other factors are probably also contibuting to the better fp performance (including for example possible improvements in IBM's XL compilers). In comparison, the current 2.26 GHz Pentium M 780 scores 1812 int, a significant improvement over the G5. However, Pentium M's fp performance suffers in comparison. A 2 GHz Pentium M Dothan has an fp score closer to 1100. Fortunately for Apple, Yonah and Merom are to get significant improvements in fp performance, according to Intel.


(Edited 2005-11-07 to correct Dothan fp score.)

Monday, October 24, 2005

New dual-core 64-bit PowerPC chips for 2006

The new startup fabless semiconductor company P.A. Semi, today announced a new dual-core 64-bit PowerPC chip, to be released late next year. The 65 nm PA6T-1682M chip features dual 2 GHz 64-bit PowerPC cores with a shared 2 MB L2 cache, dual VMX units, dual DDR2 memory controllers, and several different I/O configurations. The chip apparently has not yet taped out, but is estimated to have per core SPECint performance of 1100 and per core SPECfp performance of 2255, all with a "typical" power usage of 13 Watts. While the SPECint performance is OK, it isn't great. However, if the chip were to live up to the hype and achieve the estimated 2255 score for floating point performance, that would be truly impressive, whether the score is peak or base, especially when the power utilization is considered.

This is in stark contrast to Freescale's chips. The G4 currently also has reasonable integer and Altivec performance, but its floating point performance is poor. Freescale's next generation e600 line, due soon, will not significantly alter this situation since it essentially a tweaked dual-core version of the G4 in many ways. IBM's G5 provides floating point performance in spades, but IBM has not been able to provide a fast low power G5 option.

What does this all mean for Apple? Probably nothing, as Apple has already announced the switch to Intel. It should also be noted that P.A. Semi's chip will not ship for at least a year, and by that time Intel's dual-core 64-bit low power Merom should be available. Merom will provide vastly superior SPECint performance and at least reasonable SPECfp performance, albeit at higher power usage. It doesn't look like P.A. Semi's chips were ever in the cards for Apple. That would be fine for P.A. Semi however, since it seems that one of the main targets would be Freescale's embedded business, and they likely would not compete with the likes of Intel for the personal computer market. Another main target would be supercomputing clusters, where floating point performance and power efficiency are amongst the primary concerns.

Note again though that what we have here is still just prelimary information on a chip not due to sample for a year. However, P.A. Semi tomorrow will present the details of the chip at the Fall Processor Forum in San Jose. Hopefully we'll have a clearer idea of this chip's capabilities and its intended market tomorrow.

PCIe peripherals already available for Macs

Well, that was quick. The new PCIe Power Macs (including the Quad) have only been out a few days, and Black Magic Design already has PCIe pro video capture cards for it.

For those who still need PCI-X compatibility however, Apple still sells the dual 2.7 GHz G5 970FX Power Macs, at $2799 with a Radeon 9650, 512 MB RAM, 16X SuperDrive, and 250 GB hard drive.

Sunday, October 23, 2005

QuickTime Pro - Top 10 at the Apple Store

Apple's QuickTime is doing very well these days, as evidenced by the brisk sales of QuickTime Pro licences at the Apple Store. Right now, QT 7 Pro for Mac OS X is third on the list, and QT 7 Pro for Windows is sixth. Certainly, a significant part of its recent success can be attributed to Apple's inclusion of both MPEG4 and H.264/AVC video support in its newest iPod, which also happens to be the number one seller at the Apple Store. QT 7 Pro offers the user the ability to export video from various sources to various different formats, and includes an "Export Movie to iPod" option for iPod-compatible 320x240 H.264 video.

Apple Store Top 6

However, if you are looking to save some money, a number of free H.264 encoder options do exist. These include x264, which is used in ffmpegX and Handbrake, as wll as Michael Ash's excellent and free app, QTAmateur, which allows access to Mac OS X's built-in QuickTime export features (including iPod H.264 support) without the need for a QuickTime Pro licence.

By the way, Apple's description for the iPod in the top ten list is wrong. None of Apple's iPod models support Firewire now. While both the iPod nano and the iPod can charge over Firewire, there is no longer any support for Firewire file transfers or music synching.

Friday, October 21, 2005

Mac nVidia GeForce 7800 GT to ship in November?

Rumour has it that the Mac nVidia GeForce 7800 GT that is mentioned all over Apple's website, but which cannot be ordered at this time, may be shipping as early as November.

Considering how much faster this GPU is compared to the stock GeForce 6600 and 6600 LE, for some people the wait might be worth it. For those who don't need the extra GPU speedhowever, the good news is that initial reports state the 6600 series cards are fanless.

[Update 2005-10-21]

Here is Apple's GeForce 7800 GT manual:

7800 GT

The Quad Power Mac G5 is a bargain

Apple's Power Macs aren't usually known for being the cheapest on the block, but the new Power Mac G5 Quad brings quad machines into the mainstream. Not only can it be a good value as an ultra high end personal desktop computer (if for instance you need the extra speed for applications such as video encoding, etc.), it is quite the bargain as a professional 3D workstation. Let's compare the Power Mac configured as a 3D workstation to similarly configured AMD Opteron workstations:

Apple Power Mac G5 Quad

2.5GHz dual dual-core IBM PowerPC G5 970MP
Quadro FX 4500 512 MB GDDR3
250 GB Serial ATA hard drive
16x SuperDrive

Boxx 7400

2.2 GHz dual dual-core AMD Opteron 275
nVidia Quadro FX 4500 512 MB
250 GB Serial ATA hard drive
16x SuperDrive

2.4 GHz Quad-core AMD Opteron 280
nVidia Quadro FX 4500 512 MB GDDR3
250 GB Serial ATA hard drive
16x SuperDrive

The quad 2.2 GHz Opteron is over 20% more expensive than the quad 2.5 Mac.

In terms of non-workstation desktops however, one big drawback for some is that the available PCIe video cards for the Mac are few. Aside from the the very expensive Quadro FX 4500, the only option is the nVidia GeForce 6600 (or the 6600 LE on the low end Power Mac). The 6600 a fast card, but it is not state of the art. Apple advertises that a Power Mac with the nVidia GeForce 7800 GT is up to 84% faster than with the GeForce 6600, but unfortunately the 7800 GT is not currently available. That's too bad, because for applications such as Motion and Apple's new flagship photo application, Aperture, a 7800 GT would be most desirable. In fact, both these applications list the 7800 GT as recommended. Therefore it's likely it will become available within a couple of months. (Even if doesn't then perhaps ATI will come in and fill the void with the Radeon X1800. An added bonus with this ATI GPU is that it supports H.264 decode and transcode acceleration, which would be very useful if Apple chooses to support it.)

On the bright side, if you buy a Power Mac with the 6600 now, and then get the 7800 GT later, you could always run both the cards simultaneously, with a total of four Cinema Displays. :)

Thursday, October 20, 2005

New iPod uses Broadcom BCM2722 chip

Ars Technica has taken a look inside the new iPod. It uses the new Broadcom BCM2722 chip, which was just announced yesterday. This chip can decode MPEG4 video up to 640x480, but H.264 video only up to 352x288. Apple rates the iPod at 480x480 MPEG4 and 320x240 H.264.

So, as expected, the new iPod is incapable of decoding significantly higher resolution H.264 video than its 320x240 spec. Perhaps next year...

Wednesday, October 19, 2005

New Quad Power Mac G5 and PowerBooks

I've been talking about the Quad Power Mac and new PowerBooks for just about forever now, but it has finally happened and fortunately I won't need to eat my hat. ;) Some pertinent details:

Quad 2.5GHz ($3299)
Dual dual-core 2.5 GHz 970MP with 1 MB L2 cache per processor
1.25 GHz bus for each processor
512 MB DDR2-533, with support up to 16 GB, ECC memory optional
PCIe nVidia GeForce 6600 256 MB
Support for professional 3D workstation graphics cards (Quadro option)
Three free PCIe slots
250 GB hard drive
2 hard drive bays only (albeit with up to 1 TB storage)
16X SuperDrive
Dual Gigabit Ethernet

Dual 2.3GHz ($2499)
Single dual-core 2.3 GHz
1.15 GHz bus

Dual 2.0GHz ($1999)
Single dual-core 2.0 GHz
1.0 GHz bus
160 GB hard drive
PCIe nVidia GeForce 6600 LE 128 MB

It looks like Apple finally has a "real" pro machine, with PCIe workstation graphic cards available and ECC support. Three hard drive bays would have been nice however. The 2.5 GHz speed was as expected.

12" PowerBook ($1499)
Slot-load 8X SuperDrive is now standard
80 GB hard drive (up from 60 GB)
Otherwise no significant change in specs from before.

15" PowerBook
1440x960 screen
Mobility Radeon 9700 128 MB (up from 64 MB)
7200 rpm 100 GB upgrade option

17" 1680x1050
1680x1050 screen
120 GB hard drive
7200 rpm 100 GB up/downgrade option

The clock speeds of the PowerBooks are the same at 1.5, 1.67, and 1.67 respectively. No specifics are listed but it looks like Freescale wasn't able to get the 90 nm G4 7448 out in time.

Many will welcome the higher resolution screens in the 15" and 17" PowerBooks, but I wonder how some people will react to the higher pixel density, which will make stuff on screen appear smaller. My guess is that they will be OK with it too, since Apple isn't using the insanely high pixel density screens that some PC laptops use.

The 12" PowerBook is looking pretty long in the tooth though. Basically, it is the 12" iBook with a SuperDrive and a couple of other things, but a worse GPU. It's interesting to see that it didn't even get DDR2 support, or even the 7200 rpm hard drive upgrade option. (The DDR2 support is probably irrelevant however. It's running at 333 MHz anyway.)

One wonders if a 13" widescreen PowerBook is coming early next year.

Sunday, October 16, 2005

Yonah power specs confirmed

Intel Inside®AnandTech has confirmed the previous CNET article about Yonah's power utilization, and adds that the numbers are not absolute maximum values, but are in fact TDP:
E: TDP >50W
T: TDP 25W to 49W
L: TDP 15W to 24W
U: TDP <14W
49 Watts for the T series is disappointing, since it's much higher than Apple's preferred maximum usage of 30-ish Watts. (The L series are much lower power, but its clockspeeds are also low.) The good news is that some sites are reporting that most of the T chips will be closer to 31 Watts, as originally expected. The question remains however, if the 31 Watt spec as listed on the previous roadmap for the 2.16 GHz Yonah chip is accurate, or if that number has since creeped up too.

The other question is whether or not Apple will actually use these chips in PowerBooks. These chips are due out in the first quarter of 2006. By then I suspect that many pro-oriented applications will not have yet been ported to x86 Intel. Pros who are dependent upon such applications could thus have issues running their applications on Intel PowerBooks initially, and those issues would range from just poor performance all the way to programs not running at all. Will Apple risk this with a PowerBook release in the first half of 2006? My guess is probably yes, but I am by no means sure of this.

I suppose Apple could release a 1.9 GHz G5 PowerBook on Wednesday, and then another one in early 2006, before releasing an Intel PowerBook later on that supported 64-bit. (The 2006 Q1 Yonah chip is 32-bit.) However, ever since Jobs announced the Intel switch, I've thought the next PowerBook will likely have a G4, a 32-bit chip, and there will be no G5 PowerBook at all.

Saturday, October 15, 2005

Thoughts on the video iPod and future directions

In my previous iTunes 6 article I lamented the fact that Apple has chosen to limit video downloads to files with a resolution of 320x240 (or smaller), and expressed hope that Apple would release 720x480 (480p) standard definition video in the future through iTunes.

Unfortunately, Apple's video format of choice, H.264, is extremely hard to decode. The new iPod is likely hardware limited to speeds sufficient only to decode H.264 files up to 320x240 at 30 frames per second. Thus, Apple has a hard cap on the video resolution and bitrate. I also mentioned in the article that while 320x240 was watchable on a computer, it is of fairly low quality. Similarly, that output would look mediocre to poor, played from an iPod on a medium-sized or larger TV. Now, before you bring this up, I am well aware of the similar arguments made when it was announced that Apple had chosen 128 Kbps AAC for audio for iTunes. I have always felt that 128 Kbps AAC, although not perfect, was adequate for this purpose. Video is a very different situation however, and I feel that 320x240 is subjectively much lower quality for video playback than 128 Kbps AAC is for audio playback.

To rectify this issue, Apple would need to release a new iPod that supported decoding of higher resolution video. That video could played at native resolution from the iPod's video output to a TV for optimium quality, but could be resized on-the-fly for the iPod's small screen. Apple could choose one of several options:

1) Orphan the current iPod in the future, and only release higher resolution video for newer iPods and computer playback.
2) Start releasing multiple versions of the same video at the iTunes video store, with smaller ones for the 5G iPods, and larger ones for 6G iPods and beyond.
3) Provide software for the consumer to transcode higher bitrate video files to lower bitrates video files for the 5G iPod.
4) Stick with 320x240.

Option #4 could happen, but I just hope not. Option #3 seems unnecessarily complex, and would be very slow too. It's easier just to have the video pre-recorded and ready for download. This brings us to Option #2. This could be doable especially for a short time, during a transition period, but it also seems unnecessarily complex for the average consumer. A variation on Option #2 would be to have iPods remain at 320x240, but have larger videos for computer set top boxes or future Mac mini-like home theatre machines, but that also seems like a fairly cumbersome solution. We're left with Option #1, which would be dependent solely on having new iPod hardware, capable of fast enough decoding to make it work.

It appears that implementing Option #1 would be a ways off yet, since there may be no available chips that are currently both fast enough for higher datarate H.264 decoding, while remaining low power enough to be useful in an iPod. Also, the higher the resolution, the harder the decode, and the longer it would take before such an iPod could be released.

With this and other issues in mind, it seems I may have been too quick to push for 480p material. After looking at the potential internet bandwidth requirements, I think 480p may be probitive, due to cost, storage requirements, and download times. Bandwidth is increasing but the reality for the near and mid-term is that bandwidth will continue to be an issue. Reasonable quality 480p H.264 might take as much as 20 MB per minute of video, and at today's home broadband speeds in North America, that could mean 30-60 minutes of download time for a single TV episode and that's clearly not ideal.

Fortunately, 480x360 video and 480x270 widescreen video actually looks reasonably good on medium-sized screens, if encoded at high enough bitrates. Such video is in the ballpark range of about 8 MB per minute, and full length 1 hour TV episodes would be somewhere in the range of about 350 MB each. (A TV hour is only about 44 minutes. Yes there are that many commercials.) This essentially halves download time compared to 480p, yet still maintains reasonably quality, at a level that is noticeably better than 320x240 (or 320x180) even to the eyes of average customers. Interestingly, the PSP, which Sony is marketing partially as a portable movie device, has a widescreen resolution of 480x272. Obviously Sony feels that the 480xXXX screen size is a good one. Apple wouldn't have to use the same size screen, but could create an iPod that can downsample 480x360 H.264 to the current 320x240 screen size, to maintain the iPod's nice form factor yet give it support for better quality video for TV display. Such files would also look good played back from iTunes directly on the computer's screen. Ironically, before iTunes 6 launched, Apple itself encoded and distributed music videos at 480x360, so it seems that Apple too felt the 480xXXX resolution was the sweet spot. Unfortunately, hardware limitations have Apple's hands tied... for now. (Ironically, the iPod is already able to play back up to 480x480 MPEG4, but Apple seems to have made up its mind not to use MPEG4 for this. The present and the future for Apple is H.264.) Still, I would be overjoyed if Apple could somehow implement a system with 640x480 "fullscreen" and 640x360 (or even 720x480) widescreen standard definition video. 480x360 is a good compromise appropriate for next year, but a maximum 640x480 resolution would be almost ideal from an image quality point of view, other issues notwithstanding. As for high definition, let's not even go there.

How all this pans out in the next year or two will be very interesting. This week's iTunes 6 launch isn't yet the revolution for video, but several of the pieces are now in place, and the revolution may just start soon, with better quality video, better video selection, and a simple download concept to provide reasonable quality video for our computers, for our iPods, and for our TVs. And indeed there is hope. Steve Jobs didn't push this iPod as a video iPod. He pushed it as an iPod which also happens to play video. Perhaps 2006 will see the appearance of a "real" iPod video.

[Update 2005-10-15]

The first comment in response to my article is an interesting one:
Jobs doesn't feel the portable video device has much usefullness or market demand. The new iPod is more of a trojan horse to establish momentum and move the content providers to get into the action. Apple's working with Intel on future technologies so we may see better video quality in future hand helds; but the bigger market is home consumption. "Front Row" is only the first small step. The delivery and processing of HD video is the ultimate target. As Intel moves into the TV chip business we may well see Apple move with them.
I will respond by saying that there are couple of fundamental issues which make video downloads problematic:

1) 99.9% of TVs do not have any sort of computer connected to them
2) Apple's video download service offers no way of burning the content to video DVD.

Most people, like me, hate watching movies and TV shows on their computers. and require the use of a TV for this purpose. Apple has a few options to solve this problem.

1) Create a new set top box that can access content on the iTunes Store. Basically it would be a set top box with recording capabilities for video on demand. However, this is remarkably similar to the boxes already provided by the cable companies.
2) Sell a Mac mini-type computer for use with the TV. The computer could function as sort of a set top box as above, or stream video from a 2nd computer on which the videos are hosted. This would work but would require some sort of network connection. Wireless would be doable, but it would depend upon the size the house, the strength of the wireless signal, and having multiple computers.
3) Sell an iPod which can output higher resolution video.

Options #1 and #2 are in many ways very similar in concept, and just differ in implementation. Option #3 is quite a different beast, but has a lot of potential to move a lot of iPods. If I knew I could buy missed TV show episodes for $2 to play on my TV (which has no computer hooked up to it), I'd be buying quite a few of them, and I would buy a new iPod too, to act as the very convenient and portable source.

However, this requires good video output from the iPod, and 320x240 does not qualify as good, at any bitrate.

There is also another option, a revamped Airport Express that also does video. This seems like the logical progression of Airport, but Apple may wait until 802.11n is available before moving on and selling this. However, when they do release an Airport Video Express, a single file size for video downloads (that worked on the Airport Video Express, computer screens, and the iPod) would still be preferred, since it would mean less confusion for consumers.

[Update 2005-10-16]

I have been told that most digital broadcast video is 352x480 (60 fps interlaced). That's just 2% less pixels displayed per second compared to the 480x360 (30 fps progressive scan) I was advocating, so it sounds like the digital broadcasters think that is the sweet spot too. (FYI: I'm not a video guru, nor do I play one on iTunes TV, but apparently 352 is used because it is a multiple of 16, the size of a single 4x4 macroblock.)

Oh and just one more thing... Ironically, Apple's free H.264 stream of the "One more thing..." keynote is 640x360, three times the resolution of Apple's $1.99 videos on iTunes. :)

Friday, October 14, 2005

Safari is no longer a memory hog

The latest official release version of Safari suffers from memory leak problems, which means that after several hours of surfing, the memory usage of Safari can go up to several hundred megabytes.

However, the Safari team is now releasing binaries of its Webkit nightly builds, and the latest builds seem to have largely fixed the issue. Even after a day of use of the Friday, October 14 Webkit build, the memory usage has stayed within reasonable amounts, mostly in sub-100 MB territory.

The gold-rimmed Webkit icon is rather fetching too. :)

Oh and One more thing... yet again

Apple is getting people together for yet another invitation-only press event next week to announce "Apple’s latest pro innovations". This is only one week after the "One more thing..." event which took place two days ago, where Apple announced new consumer products and services: iMacs, video-capable iPods, and a video download store.

If we don't see new Power Macs (quad G5 that is, with PCIe and DDR2) and PowerBooks, I'll eat my hat. We shall see in a few days, but in the meantime I'm off to look up some hat recipes.

[Update 2005-10-14]

The Apple Store now states the 15" & 17" PowerBooks and all the Power Macs will take 3-5 business days after purchase before they are shipped. Such increases in shipping times for stock configurations of Macs are often a sign of imminent updates.

About this (Quad Power) Mac

Quad G5

I don't know if this is a fake or not. However, even though it would be easy to fake this (especially since it says DDR and not DDR2), I wouldn't be completely surprised if it's a real screen capture either.

We have all been predicting the quad Power Mac for just about forever now. It's about time it finally shows up.

Wednesday, October 12, 2005

New Apple trailers site - HD galore

Apple is pushing HD H.264 in a big way, and has put up a new QuickTime trailers site with lots of 720p and 1080p HD trailers. It's too bad Apple's new iTunes TV show store however does not even use 480p standard definition video.

iTunes 6 is out, now with TV show downloads

iTunes 6 was released along with new iPods that support video playback. Included in iTunes 6 is support for video purchases, including TV shows. Unfortunately, the selection of TV shows is small, and they are not available in places like Canada. Unfortunately, the videos are only 320x240, which although (barely) acceptable for now, are only about one quarter (or less) of the resolution of 480p DVD quality video. One can only hope that once this technology matures, internet bandwidth increases, and H.264 decoding speeds of the iPod improve, Apple will offer 480p downloads and the selection of available titles will increase.

Also, as expected, iTunes refuses to manage DivX files, but that shouldn't impact most customers. Unlike the past with MP3 music files, these formats of video files are not that important overall to the general paying public.

[Update 2005-10-12]

The biggest size H.264 videos the iPod can decode is 320x240 at 768 Kbps and 30 fps, hence the 320x240 size limit at the iTunes Store. I hate to say it, but I hope Apple decides to orphan these iPods in a year or two, by releasing a new widescreen iPod line that can decode 720x480 H.264 at 2 Mbps and 30 fps (and resize it in realtime to fit-to-screen), along with video content to match.

I just downloaded the Pixar Luxo Jr. short, and although it was watchable, it definitely wasn't pretty at that low resolution on my 20" iMac.

Is this the new iPod video?

I don't know where this originally came from, and I don't know if it's real, but we'll know very soon.

[Update 2005-10-12]

Indeed it is absolutely real. :)

Apple reports stellar quarterly performance, and the stock drops like a rock

Apple yesterday released its 2005 fiscal Q4 and fiscal numbers. Despite the stellar numbers, AAPL dropped like a rock in after hours trading, by almost 11% to US$45.99.

Sales of iPods were up 220% to 6.451 million units since the previous quarter last year, and sales of Macs were up 48% to 1.236 million units. However, the numbers, particularly for the iPod, were evidently simply not good enough for the investors. It seems the street believed the analysts' unrealistic rosy predictions for iPod sales and when those numbers were not met, decided to lock in profits. That is reasonable I suppose though, as gains on AAPL have been astronomical since last year. As they say, buy on rumours and sell on news.

Before the Apple financial conference call, I had made a (somewhat) flippant prediction at Ars Technica of 1234567 Macs and 6543210 iPods sold during the quarter. Ironically, that prediction beat all the analysts. ;) Sometimes it pays not to think too hard. :)

The big news for today however is what's on deck for Apple's One More Thing event, scheduled for 10 am in San Jose. I had predicted new Power Macs and PowerBooks, but just about all the rumour sites now state I am wrong, and this is about iPods. Perhaps, but even if so, we should expect new Power Macs and PowerBooks sometime soon.

Tuesday, October 11, 2005

Yonah is a power hog?

CNET News has published an article about the power output of Intel's next generation dual-core Yonah chips. It suggests that early predictions of Yonah power utilization as implied by Intel's Thermal Design Power (TDP) values (see diagram) may be somewhat optimistic:
The "T" class of Yonah chips, which are expected to be fitted in most business notebooks, will come with a maximum power consumption of between 25 and 49 watts. Right now, single-core Pentium Ms top out at 27 watts.
If true, that's not entirely unexpected given that these chips have two complete and independent cores, but it is still disappointing. Historically, Apple has preferred CPUs with a max power utilization under 30 Watts. However, the specifications published by CNET indicate that dual-core 2+ GHz Yonah chips are far above that 30 Watt barrier.

It does make one wonder how Apple will deal with this issue. Will Apple will choose to use Intel's fastest and hottest Yonah chips in some PowerBook models, or will Apple choose to use only slower and/or low voltage (LV) chips to save power?

[Update 2005-10-13]

The DigiTimes echoes CNET's article, but goes on to suggest that these numbers are actual TDP values, and not absolute maximum power utilization values. (TDP is a better representation of what would be expected as an effective maximum with real-world software, but it is lower than the absolute maximum power utilization possible, with say a power virus.)

Tuesday, October 04, 2005

Apple to release just "One more thing" next week

Apple will be hosting an invite-only event next week on October 12. No details are given, but this "One more thing..." event will be held at the California Theater in San Jose, California. The invitation shows theater curtains too, and because of this some people surmise that this may be the launch of something video related, such as the iTunes Video Store and the iPod video.

Before anyone jumps for joy at the idea of a new iPod that can play BitTorrented DivX TV shows, please note that even if Apple were to release an iPod video, it would not likely support playback of these files. Unlike the situation with MP3 many years ago, there is no single universal video format which geeks and casual users alike use. If Apple were to do this, Apple would likely stick with its own QuickTime codecs for commercial video content, especially since QuickTime is already one of the top two formats on the web for it. Indeed, Apple's coup would be to gain the rights to distribute that commercial video content through an iTunes Video Store, not to play existing ripped video. In short, the fight of the future is between Apple's QuickTime and Microsoft's Windows Media. No other video format matters.

If this is purely a non-Mac announcement however, one wonders what happened with the dual dual-core (quad) G5 970MP Power Macs and the G4 7448 PowerBooks. I think we can say that it's likely these products will be released, but it's just a matter of when.

[Update 2005-10-05]

Think Secret suggests the event is to announce the new Power Macs and PowerBooks.

Friday, September 30, 2005

New PowerBooks are a comin'

Many of us had predicted that Steve Jobs would unveil new PowerBooks at Apple Expo, but he then went and cancelled his keynote. One can only guess that there was some unforeseen problem which caused a delay in the release of these (and other) new Macs. However, new updated PowerBooks may be coming very soon.

First of all, now if one orders the upper end 15" PowerBook with any upgrade option, shipping times increase to 2-3 weeks. The shipping times are much lower for other build-to-order Macs. Furthermore, AI reports that the as yet unreleased OS X 10.4.3 contains references to new PowerBook models. The last PowerBook update was in January, 9 months ago.

What these PowerBooks will get as upgrades is unknown, but I would expect:

Up to 1.8 GHz G4 7448, with 1 MB L2 cache and 200 MHz bus.
512 MB RAM built-in on all PowerBooks. (The 12" has 256 MB built-in.)
New GPUs, especially on the 12" (which has a worse GPU than the iBook).
Possibly a higher screen resolution option for the PowerBook 17".

These would likely represent the last of the PowerPC based PowerBooks. It's likely that Apple will release new PowerBooks in the first half of 2006 which utilize Intel's new dual-core Yonah laptop CPUs. Since the new Yonah machines will be much faster than any of today's single-core PowerPC PowerBooks, many may want to wait until the Intel PowerBooks are released before upgrading. However, those with the least bit of concern for software compatibility or who require Classic support would be best advised to buy the last of the PowerPC PowerBooks. In the near term, the current G4 7447A and upcoming G4 7448 PowerBooks will encounter fewer compatibility problems than next year's Intel PowerBooks.

New machines may come as early as October, although it would not be surprising to see the upgraded PowerBooks hit the streets a short while later.

Mac mini updated... Kinda sorta

Apple has begun shipping upgraded Mac minis recently. Upgrades include a slightly faster CPU (up from 1.42 GHz to 1.5 in the top model), Bluetooth 2.0, an 8X SuperDrive, and a version of the Radeon 9200 updated with 64 MB memory (up from 32 MB). What's odd about this upgrade is that the Apple Mac mini page still lists the old Mac mini specifications. That may suggest that there is still a significant amount of inventory of the old model in the channel, and that Mac mini sales are perhaps not as good as Apple had hoped.

All of the upgrades are welcome, but the most significant upgrade is the video memory. 32 MB is not ideal for larger screen sizes, since things such as Exposé can slow down dramatically when there are several large windows open. 64 MB provides much more breathing room, especially when a large LCD screen such as Apple's 20" Cinema Display is used. However, it is a shame that the upgraded Mac minis still do not include a GPU that fully supports all of the features of Core Image. The Mac mini is Apple's last Mac that doesn't. Even the iBook now sports a fully Core Image compliant Radeon 9550.

Wednesday, September 14, 2005

IBM publishes 2.2 GHz G5 970FX power specs

IBM has published the IBM PowerPC 970FX Microprocessor Design Guide, in which it lists a 60 Watt maximum power specification for the 2.2 GHz 970FX.

This is in additon to the previously published 39 Watt specification for the 2.0 GHz 970FX.

One wonders what all of this means for the as yet unreleased 970MP, in terms of top speed and power utilization. Given that Apple already has 2.7 GHz G5 Power Macs for sale, one would hope that a dual-core version of the G5 would debut at 2.5 GHz, with acceptable power requirements.

Saturday, September 10, 2005

iPod nano: The beginning of the end of Firewire?

Apple's release of the iPod shuffle marked Apple's first iPod product without the support of Firewire. That was not totally unexpected however, considering the iPod shuffle is essentially a USB 2.0 pen drive that plays music as well. Implementing Firewire on such a product would have been quite problematic, given its overall design and form factor.

However, Apple has now released the iPod nano, the replacement to the very popular iPod mini, which also does not support Firewire synchronization. This is especially surprising considering it sports the normal iPod dock connector used in other units. One presumes that part of the reason for this is the iPod nano's very small size. The included chipset probably doesn't support Firewire, and adding full Firewire support would have meant a larger unit and higher cost.

Similarly, Apple's and Motorola's new ROKR iTunes phone doesn't support Firewire synchronization either.

It is somewhat unfortunate that Apple has gone this route. For fastest synchronization, USB 2.0 is required, but most of Apple's installed base of Macs do not support USB 2.0. (All currently shipping Macs do support USB 2.0 however.) Apple came late to the USB 2.0 party, but now has embraced it head-on at the expense of Firewire. Ironically, Firewire is a standard that Apple helped create.

The silver lining however is that synchronization of the iPod nano will still work on older Macs using USB 1.1 (albeit slowly), and the iPod nano is absolutely gorgeous. I want one... or three.

Thursday, August 18, 2005

Apple begins iMac G5 Repair Extension Program for Video and Power Issues

Apple today released a FAQ about their new iMac G5 Repair Extension Program for Video and Power Issues:
The iMac G5 Repair Extension Program for Video and Power Issues applies to first generation iMac G5 computers that have video or power-related issues as a result of a specific component failure. If your iMac G5 is exhibiting any of the symptoms listed below and your computer's serial number is within the noted ranges, your computer may be eligible for repair, free of charge. If Apple or an Authorized Apple Service Provider (AASP) determines that your iMac G5 computer is eligible as part of the program, the repair will be covered by Apple even if your iMac G5 is out of warranty. This is a worldwide Apple program.

Affected systems will exhibit one of the following video- or power-related symptoms :

Scrambled or distorted video
No video
No power

Note: If your iMac G5 is not experiencing any of these symptoms, you do not have to contact Apple or any Apple Authorized Service Provider.

Which iMac G5 computers are affected by the iMac G5 Repair Extension Program for Video and Power Issues?
The program is available for certain iMac G5 models that were sold between approximately September 2004 and June 2005 featuring 17- and 20-inch displays with 1.6GHz and 1.8GHz G5 processors.

The affected iMac G5 computers have serial numbers where the first 5 digits fall into the ranges noted below.

Serial Number ranges:

W8435 - W8522
QP435 - QP522
CK435 - CK522
YD435 - YD522

Somehow this all seems vaguely familiar.

[Update 2005-08-20]

The repair program includes the first G5 iMac models released in August 2004. It is not clear if the more recent G5 iMac models released in May 2005 are included. Newer 2.0 GHz iMacs do indeed fall into the serial number ranges, but Apple specifically leaves out 2.0 GHz model iMacs in the program description.

Wednesday, August 17, 2005

Apple's Security Update 2005-007 kills 64-bit support

Reports online state that apps that are compiled specifically for 64-bit G5 systems will no longer run after the latest Security Update 2005-007 in installed. Affected programs include many command line applications as well as the 64-bit version of Mathematica 5.2.

That something as obvious as this could be missed suggests that Apple really needs to revamp their software quality assurance program.

[Update 2005-08-18]

Apple has released Security Update 2005-007 Version 1.1 to correct the problem:

Security Update 2005-007 delivers a number of security enhancements and is recommended for all Macintosh users.

Security Update 2005-007 v1.1 replaces Security Update 2005-007 v1.0 for Tiger systems Mac OS X v10.4.2. Users who have already installed v1.0 on Tiger systems should install v1.1.

Security Update 2005-007 v1.1 provides a combined 32- and 64-bit version of LibSystem to replace the 32-bit version that was delivered in v1.0. No other changes have been made in version 1.1.

Monday, August 08, 2005

Core Image, Monitor Spanning, and the new iBook's Radeon 9550

Some of you asked about Core Image after we posted our previous article on the new iBooks released in July which sport the Radeon 9550. As we mentioned in our article updates, yes indeed it does fully support Core Image in hardware, and this is best demonstrated by the ability of the iBook to display the ripple effect in Dashboard:

(Click on picture for larger version.)

Update [2005-08-09]

The desktop spanning hack also works fine on the new iBooks. With Screen Spanning Doctor 0.3.3 I can extend my iBook's desktop onto an external 17" 1280x1024 LCD without any problems. I have not tried clamshell mode however, since I do not use this feature.

Two finger scrolling rocks my world

Although much of the recent scrolling hoopla is with Apple's new scrolling Mighty Mouse, I'm here to extoll the virtues two finger scrolling available with Apple's trackpads on their laptops.

Originally introduced with the last PowerBooks, Apple's two finger scrolling is now standard across all its laptops, so the last iBooks got it too. I have just purchased a 12" iBook and have come to love Apple's implementation of trackpad scrolling.

I have never liked the limited up/down scrolling on scroll mice, and didn't like the trackpad scrolling hacks with 3rd party apps either. Thus, I thought I wouldn't like trackpad scrolling, but it's a world of difference. Not only is it very smooth, but it also works in all directions. This is especially useful for large photographs, say for example in Preview or Safari. With two finger scrolling, I can scroll up, down, right, left, sideways, or even in circles. Diagonal and circular scrolling do not work quite as smoothly in Photoshop CS, but I'm sure it will improve with subsequent updates to Photoshop and/or the scrolling driver. (I have not yet tried Photoshop CS2 however.)

It's these little features that continue to make Apple's laptops so desirable.

Friday, August 05, 2005

G5 970FX 2.0 is 39 Watts max

IBM has published an article titled Thermal Considerations: PPC970FX which states the 2.0 GHz G5 has a maximum power output of 39 Watts (105 C part). This number is significantly lower than the 50 Watts listed for previous non-low power 2.0 GHz 970FX chips. Further details are not provided, but it could represent a specific low power version of the chip, which would be sold in parallel with higher power parts.

The question now is what wattage a low power 1.8 GHz 970FX and a 2.5 GHz 970MP would put out.

Tuesday, July 26, 2005

New iBooks and Mac minis released today

After 40 weeks since the last iBook release, Apple finally updated its iBook line. The release is pretty much in line with previous predictions:

12" screen
1.33 GHz G4 7447A
40 GB HD
512 MB RAM
Core Image GPU (Radeon 9550, better than the predicted GeForce FX 5200U)

We also see built-in Airport Extreme (as before), included Bluetooth 2.0, Sudden Motion Sensor technology to protect the hard drive, and the scrolling trackpad. What we don't see is 64 MB RAM for the GPU. Given the relatively low resolution of the screen, 32 MB is usable, even though 64 MB would have been preferred. We didn't see a widescreen form factor, but that seemed unlikely in the first place. The price is unchanged, at US$999.

The 14" iBook now has a built-in SuperDrive as well, whereas it used to cost $1499 with the SuperDrive. If you don't need the SuperDrive, you can select a combo drive to bring down the current 14" model's price to $1199. Interestingly, the 14" iBook has a 142 MHz bus, which doesn't exist on any other Mac as far as I know.

Overall, the new iBooks look like a very nice buy. I bought the new iBook 12" in fact.

However, the Mac mini updates, although reasonable, are more of a mixed bag. They didn't get a clockspeed upgrade, and they did not get a fully Core Image compliant GPU. In the era of Tiger, I would have hoped that all Macs would support the full Core Image feature set in hardware. More important however is the fact that the Mac mini did not receive an increase to the amount of GPU RAM included. While 32 MB is passable on the iBook (with its built-in 1024x768 screen), it is too little for a 2005 desktop in my opinion. Running Exposé with 32 MB on a 17" screen is not ideal (although acceptable for many), and can be outright painful on a 20" screen.

The 1.42 GHz Mac mini seems like a more reasonable upgrade. Apple has stopped including the modem on those units, but instead have included internal Airport Extreme and Bluetooth, a nice upgrade. Thus, it seems the new bang-for-the-buck Mac mini unit is the $599 1.42 GHz Mac mini. It's still stuck with the 32 MB GPU RAM though.

One wonders if part of the reason the Mac mini did not get a speedbump relates to the power output and cost of the CPU. Freescale's datasheet claims the 1.42 GHz G4 7447A has a maximum power output of 30 Watts. Keeping the CPU at 1.42 GHz keeps costs down, and makes it easier to keep Mac minis from overheating. Mind you, the 12" PowerBook handles the 1.5 GHz G4 just fine, and the 15" PowerBook goes up to a 1.67 GHz G4.

Because the Mac mini's upgrades are not as impressive as they could have been, those looking at the Mac mini really should look hard at the iMac G5 too. It is a very good value, and in my opinion represents the better buy for most customers.

[Update 2005-05-27]

Sergey asked if I'm sure the Radeon 9550 supports Core Image. The answer is yes, I'm sure it has full hardware Core Image support. It seems that Apple's Core Image page just hasn't been updated yet.

By the way, my iBook shipped today. :)

[Update 2005-05-27 Part 2]

Apple's Core Image page has now been updated to include the Radeon 9550.

Sunday, July 24, 2005

Confidential IBM roadmap: 970GX coming, 970MX proposed

IBM has published a previously confidential PowerPC roadmap from last year which covers things from 970FX power specs to future chips such as the long rumoured 970GX. There is also mention of a proposed 970MX, which appears to be a dual-core chip with speeds over 3 GHz and 2 MB shared L2. (This should not be confused with the 970MP, which has 1 MB L2 dedicated to each core.)

Despite the claim that the 970GX will have general availability in 2005, we have not yet had previous official confirmation of this chip. However, there has been an unofficial mention of the 970GX. Also, it remains to be seen whether or not the 970MX will ever see the light of day. In the meantime, we are still waiting the appearance of 970MP Macs, which I believe will become available within six months.

In terms of power specs, IBM claims that the G5 970FX is superior to Freescale's current G4 7447A at most clock speeds, measured by performance per Watt. What is interesting though is that even IBM concedes that Intel's Pentium M Dothan is superior with this measure at moderate speeds and possibly also at higher speeds. However, these power specs do not seem to include numbers for IBM's new low power G5 970FX. For example, the new 1.6 GHz low power G5 is listed at 21 Watts max, but the roadmap lists it at 31 Watts max, nearly 50% higher. Still, it is likely that Dothan still wins in terms of performance per Watt overall.

Thursday, July 21, 2005

IBM publishes low power 970FX power specifications

IBM has released several new documents about the PowerPC 970FX and its low power variant. One of the new documents is a new 970FX datasheet which includes typical and maximum power specifications of the new low power G5 970FX.

The low power 970FX can operate with a voltage as low as 1.0 V. At 1.6 GHz, the typical power usage is 17 Watts, with a maximum power usage of a mere 21 Watts. If this maximum power usage value is accurate, then that is probably about a one-quarter power savings over previous parts, and somewhat better than I had previously thought. Although no specifications are given for a low power 1.8 GHz G5 970FX, one might expect that say a 1.1 V part would come in at close to or perhaps less than my previously predicted 30ish Watts. Such a chip would make for a reasonable laptop CPU. However, as previously stated, Apple may choose to use the G4 7448 instead (also at 1.8 GHz) for the next iteration of the PowerBook, before the switch to Intel.

IBM has not yet published power specifications for their dual-core 970MP. If the single-core 970FX 2.2 GHz puts out 60 Watts max (according to the datasheet), then it seems a dual-core 2.5 GHz 970MP might be quite a hot chip. On the other hand, Apple claims the 2.3 GHz 970FX is only 55 Watts, and has done so for the past six months, so it is possible the dual-core 970MP 2.5 doesn't have to be overly hot. We won't know for sure, however, until IBM publishes its specs.

New iBooks and Mac minis soon

The iBook recently went to 5-7 day shipping for stock units at the Canada Apple Store and 7-10 days shipping for CTO units at both the Canada and US Apple Stores. The Mac mini also just went to 5-7 days shipping.

I'd expect new updates within the next couple of weeks, and perhaps as early as Tuesday.

For both the new Mac mini and new iBook, I hope to see a Core Image-compliant GPU (eg. GeForce FX 5200, possibly with 64 MB RAM) and an increase in the stock memory to 512 MB. I would also expect to see new G4 updates, utilizing faster clocked MPC7447A chips. Otherwise I don't expect major changes, besides maybe an increase in the hard drive size, especially for the low end iBook (which is currently at 30 GB). The widescreen iBook that some have predicted seems very unlikely for this update.

[Update 2005-07-22]

Think Secret states that new iBooks and Mac minis are due next week, and will include a US$699 model for the Mac mini.

Thursday, July 07, 2005

IBM announces dual-core G5 970MP and low power G5 970FX

Dual-core G5 970MP
IBM today finally officially announced the long-rumoured dual-core G5 970MP at the 2005 Power Everywhere Forum. (Japanese press release) The chip is expected to hit 2.5 GHz and will sport a total of 2 MB L2 cache, as we had predicted. I would expect to see a quad G5 Power Mac based on this chip relatively soon (within the next 6 months).

IBM also announced a low power version of the G5 970FX, with a rated power utilization of 16 Watts at 1.6 GHz. This specification is IBM's "typical" power, which is significantly lower than its maximum power. A reasonable estimate might be that a 1.8 GHz G5 970FX could be produced with a maximum power utilization not too much higher than 30 Watts (which is appropriate for a laptop). A laptop G5 at 1.8 GHz would be in agreement with earlier predictions.

While a Power Mac based on the dual-core 970MP seems obvious, it's not so clear for a new PowerBook based on the low-power 970FX. A G5-based PowerBook would require an entirely new design, and such a machine is unlikely to be released if the switch to Intel in PowerBooks came early next year. It is possible that Apple may choose to release a speed-bumped PowerBook using the current design and the new G4 7448, which is a direct replacement for the 7447A used in current PowerBooks. In 2006 Apple could release an x86 PowerBook based on the new dual-core Pentium M Yonah to replace the G4-based PowerBooks, skipping the G5 entirely.

Thanks to M. Isobe and, who first broke the 970MP and low power 970FX news.

[Update 2005-07-07]

The English press release is now available:

Momentum for Power Architecture Technology Accelerates in Japan

IBM today announced the newest member of the Power Architecture family of microprocessors -- the PowerPC 970MP. The new processor is a dual-core version of IBM's award winning PowerPC 970FX, targeted for clients who desire a low-cost, high performance, 64-bit, symmetric multiprocessing (SMP)-capable system in a small package with ranges from 1.4 to 2.5 GHz. The microprocessor also provides power-saving features that system architects can use to dynamically control the system power.

The IBM PowerPC 970MP microprocessor builds on the proven 64-bit IBM Power Architecture family and is designed for entry level servers as well as to provide new levels of performance and power management for the embedded marketplace. The increased computing density of the PowerPC 970MP brings a new level of performance to a variety of applications, from HPC clusters to demanding embedded system applications such as high performance storage, single board computer and high performance networking applications.

Each of the two 64-bit PowerPC 970MP cores has its own dedicated 1MB L2 cache, resulting in performance more than double that of the PowerPC 970FX. This design provides clients with a wide range of performance and power operating points that can be selected dynamically to match system processing needs. The frequency and voltage of both cores can be scaled downward to reduce the power during periods of reduced workload. For further power savings, each core can be independently placed in a power-saving state called doze, while the other core continues operation. Finally, one of the cores can be completely de-powered during periods of less stringent performance requirements.

IBM also announced today new low-power extensions to its award-winning PowerPC 970FX offering. This newest offering is targeted for clients who desire a low-cost 64-bit processor featuring high performance, a sub-20 Watt power envelope and SMP. The new offering is targeted to provide an operating power of 13W at 1.4 GHz and 16W at 1.6GHz under typical workloads. The microprocessor also provides power-saving features that system architects can use to dynamically control the system power.

The 64-bitPowerPC 970FX microprocessor builds on the proven 64-bit IBM Power Architecture family and is suited to embedded applications including imaging and networking, and provides new levels of performance and power management for the embedded marketplace. Designed to run at frequencies up to 2.7 GHz, the PowerPC 970FX includes a 512KB L2 cache, provides native 64-bit and 32-bit application compatibility and uses a high bandwidth processor bus capable of delivering up to 7.1 GB/s to keep the processor core and the SIMD/Vector engine fed with data. The processor core can dispatch five instructions per cycle, and issue one instruction per cycle to each of its ten execution units, including two fixed point, two floating point, two load store, two vector and two system units. The L1 instruction cache holds 64 KB, the L1 data cache holds 32 KB, and each processor has its own dedicated 1MB L2 cache.

Sunday, July 03, 2005

The real iTunes phone unveiled

Sane minds have prevailed. Pictures at Engadget show that (at least one version of) the Motorola iTunes phone is indeed a white E790 and not the aesthetically-challenged Xbox phone.

More pictures of the phone are posted at the Hebrew language forum,

[Update 2005-07-04]

Apparently, claims the phone will be called the "E1", but details are otherwise scarce.

Thursday, June 30, 2005

IBM talks about the mobile G5, and of future Macs

In response to Apple's announced plans to switch to Intel, IBM has issued some statements to eWeek about its PowerPC lines and Apple.

Not surprisingly, IBM counters Apple's claims of Intel's superiority in terms of the power efficiency of future CPUs:
"Apple positioned it that way in the public," Adkins said in an interview with But "Obviously I have a different point of view, because there's nothing about power architecture that limits you in any way in terms of power management or power efficiency."
IBM also confirms that they were working on a mobile G5:
"They had Freescale primarily for the low-end and mobile solutions, and they really had IBM focus more on PowerBook, xServe and iMac. That's where we collaborated deeply with Apple," Adkins said.


Intel, as Adkins sees it, won out mainly based on its mobile chip technology and the way Intel chips fit into Apple's software planes, even though he maintains that IBM has the capability to deliver a product such as a mobile PowerPC 970 chip.
At this point I would be surprised to see a mobile G5 in a Mac. However, if Apple does release a G5 PowerBook, I'd consider buying it. If IBM can get its power utilization under control, it could be a reasonable chip for a laptop (at least in 2005).

In addition, IBM goes on to state:
"We still have a number of products we're doing with them," Adkins, who declined to elaborate on the exact ones, said. Thus "this is not a transition that's going to occur over night."
We already knew that Apple would continue to ship G5-based machines for quite some time, but the wording of this statement could suggest that the dual-core G5 970MP is still coming. The release of the 970MP by IBM and its utilization by Apple in Macs would make practical sense for both sides, especially in light of Apple's current transitions plans. Apple will not finish the transition until 2007, and still needs to release additional PowerPC desktops in the interim. IBM and Apple have a contract to work together until then, and the 970MP is the perfect fit.

Wednesday, June 29, 2005

Developers unimpressed by PS3 and Xbox 360 CPUs

AnandTech today posted (and then promptly removed) an article detailing how developers dislike the CPUs in the PS3 and Xbox 360.

The developers appear less than impressed with the raw performance of the PPE unit(s) in these CPUs. They also say the SPE units in the PS3's Cell CPU are difficult to use in game code.
Right now, from what we’ve heard, the real-world performance of the Xenon CPU is about twice that of the 733MHz processor in the first Xbox. Considering that this CPU is supposed to power the Xbox 360 for the next 4 - 5 years, it’s nothing short of disappointing. To put it in perspective, floating point multiplies are apparently 1/3 as fast on Xenon as on a Pentium 4.

The reason for the poor performance? The very narrow 2-issue in-order core also happens to be very deeply pipelined, apparently with a branch predictor that’s not the best in the business. In the end, you get what you pay for, and with such a small core, it’s no surprise that performance isn’t anywhere near the Athlon 64 or Pentium 4 class.

The Cell processor doesn’t get off the hook just because it only uses a single one of these horribly slow cores; the SPE array ends up being fairly useless in the majority of situations, making it little more than a waste of die space.
Even though much of the article should be considered speculation since it does not contain any hard data (partially because the hardware hasn't even been officially released yet), we should not be surprised that some developers feel this way, given the differences in the designs of these chips from current general purpose CPUs. Nonetheless, it's not surprising that Apple's engineers chose not to use Cell in the Mac.

[Update 2005-06-30]

We now have an explanation as to why the article was pulled: "PS3 article is pulled for now because Anand is worried about MS tracing his anonymous insider."