8 GB Partition Issue Impacts OS 9, RAM for G3 iMacs, Success with Open Firmware Hack, and More
Dan Knight - 2007.12.10
- 8 GB Partition Problem with OS 9
- Sources for 256 MB Modules for Tray-loading iMacs
- RAM for G3 iMacs
- Success with the Open Firmware Hack
- Leopard Running Well on 800 MHz iBook, 533 MHz Power Mac
- No iPhone for Me Either
- Millions vs. Thousands of Colors
From Ken Watanabe:
Excellent article on making the most of older Mac OS X-capable hardware. This type of knowledge is needed, because old Macs keep getting sold and resold to new Mac users, instead of being put into the landfill like PCs.
In the section, "A Glitch with OS X," you provide information that is not accurate. First, the 8GB issue is not a problem related to Mac OS X at all. It is a limitation of the old IDE controller. During the startup sequence (before the OS takes over), the IDE controller in these affected Macs only allows access to the first 8 GBs (actually about 7.7 GBs when counting like humans) of disk space. Therefore, if a file needed during startup resides beyond the 8GB limit, the Mac would hang or crash. Windows PCs with the same IDE controller had similar issues. After startup is done and the OS takes over control, the entire disk (up to other IDE hardware limit of 128GB total size) was accessible.
The second often-repeated misconception is that this problem only affects Mac OS X and is not an issue for Mac OS 9. Actually, it is, but it is not as obvious because the Mac OS 9 installer does not prevent such installations (the Mac OS X installer does). As long as all files in the Mac OS 9 System Folder needed during startup remain within the 8 GB limit, the user will not experience any problems. However, as the disk gets filled and a system extension (or some other needed file) later gets installed or moved above the 8GB limit, the user may suddenly experience the "gray screen" or other crash during startup. I participated in a discussion about this problem on Apple Discussions. Here is a link to the thread with more details.
One of the other participants confirmed what I had tested several years ago, to prove to myself that the problem affected both Mac OS 9 and X.
Thanks for Low End Mac. It's one of my favorite and most visited websites.
Thanks for sharing this information. I've just updated the article and will add this to the profiles of the tray-loading iMacs, beige G3s, WallStreet PowerBooks, and clamshell iBooks.
One of the joys of the Internet is that there's so much good information out there and so many people who have become expert in some area or another, so when we publish an article that misses something most of us have never encountered, someone like you steps up to the plate, provides the info, and lets us spread the word to others.
The new warnings should make Mac-ing safer for those older G3 Macs with larger than 8 GB hard drives.
I am writing in response to Dan Knight's article Bringing G3 iMacs and Other G3 Macs into the Tiger Age from Friday, 12/7/2007.
Toward the end of the article, you wrote that 256 MB memory modules for the tray loading iMacs can be picked up "for under $32 nowadays". I was just wondering where you and/or Bill can obtain the 256 MB PC66 SO-DIMM modules that these iMacs are supposed to require.
I semi-recently upgraded my Tangerine 333 Mhz Rev. D tray loading iMac to 256 MB, up from the 96 MB that the unit came with when I purchased it second-hand. Each slot on the processor card took what I assume is a standard 128 MB PC66 SO-DIMM purchased through Star Micro via www.pricewatch.com. However, the only 256 MB PC66 modules that I can find through this site (and the many others I have looked at online) are priced at $50 or more each.
The pricing that you state for the other module sizes, 128 MB for $15 and 64 MB for $9, matches what I have been able to find exactly.
However, I have noticed that 256 MB PC100 SO-DIMMs do fit into the $32 price range. Could that be what Bill is referring to? If so, will a slot-loading iMac take PC100 modules instead of the "required" PC66? I'd be very interested in knowing if this is a possibility, as it would give me the opportunity to speed things up a bit more on the OS X 10.3.9 setup that that machine is currently running.
Thank you in advance for any light you can shed on this situation!
We check memory pricing using ramseeker.com, a website that was developed especially to support the Macintosh platform. As I write this, prices for 256 MB modules compatible with the tray-loading iMac range from $31.49 to $77.20.
I use the term compatible, because they may not be PC66 memory. Omni Technologies, which has the lowest price today, notes that they are either PC100 or PC133, both of which should work just fine in a tray-loading iMac. (You can almost always use faster memory of the same type - pin configuration, etc. Slower memory will cause nothing but problems.)
Also, these prices may not be for the low profile modules that fit the second RAM socket on these iMacs. Again, you'll have to check with the vendors. For instance, Other World Computing, which has the second-lowest price on full-sized RAM modules, charges $37.99 for the low profile modules.
From Paul Gorski:
I was surprised to hear that some people have had problems putting 512 MB (2 x 256) of RAM in tray-load iMacs, and even more surprised to hear folks had problems with 1 GB (2 x 512) in slot-load G3 iMacs. I've only had one such memory related problem in a G3 iMac, having dealt with and serviced dozens of G3 iMacs. I'm setting up two G3 slot-loads right now with 1 GB of RAM.
The only problem I've had was that one tray-load iMac wouldn't boot into Mac OS 9.1 from a Diskwarrior startup CD (but would 9.0 and 9.2 and Mac OS X 10.2 and 10.3) with two "unmatched" (different vendor) 256 MB modules installed.
Thanks for writing. I'm one of the people who has had problems with 256 MB modules in tray-loading iMacs, but as I no longer have them, I can't go back to determine why. I haven't heard of any widespread problems getting 1 GB up and running in the slot-loaders, and your advice to always match RAM is one of the best ways of avoiding the problems that can crop up with unmatched modules.
From Matthew Bionda:
I've just used the firmware method to install Leopard on an 800 MHz eMac with the minimal 512 MB of RAM.
I previously tried burning a modified DVD from my MacBook, but the eMac would spit it out whenever I tried to start up from it. In any case, the firmware way is by far the easiest way!
One problem, however, is sleep. When I try waking it from sleep, it powers on, but there is only a blank screen and it needs a hard restart. I'm not sure if anyone else has had that problem.
Otherwise, the performance is very acceptable, even Cover Flow in the Finder runs very well. It's not as snappy as Panther, but it has breathed new life into it and the new version of Safari is very speedy.
Thanks for sharing you success. Your problem with Sleep is baffling, as the GeForce graphics used in the first generation eMac and the ATI graphics used in later eMacs are fully supported in Leopard.
Season"s Greetings from South Jersey! Just thought I would throw into the mix with regard to the Big Cat. I was originally a little miffed with Apple when they came out with their stringent hardware specs for Leopard. My beloved and has been oft-lamented, venerable Pismo finally gave up the good fight this past September after seven honorable years of service (although I believe she is salvageable, I think I just need a new power/sound card), because of which I found a good deal on a nice little 12" 800 MHz G4 iBook that had been upgraded to 640 MB of RAM and an AirPort Extreme card.
In my naivety, I thought "Oh well, this is a plenty powerful machine . . . I"ll definitely be safe when Leopard comes to town." So when the official specs for Leopard came out, I was livid. Luckily, we receive a bulk OS X license through our AMP at work, so I am fortunate to be able to test things out before plunking down my own money on upgrades. I had been hesitant to force an upgrade on my machines just because it was all a little too complicated for my tastes, with all of the cloning and firewire drives and such. I'm a simple man and like to keep things that way.
I knew there had to be a simpler method to bypass the installer on these "old" machines. Needless to say, I was elated when you had posted Mr. McDermond's Open Firmware workaround - it worked (almost) brilliantly. I did experience one problem however, and I wonder if it was a typo on his part as well as if other"s experienced the same. When I typed the last line of code to boot from the CD, it would shoot back an error stating that it couldn't perform such an action. However, I found that if I substituted that last boot line with a simple "mac-boot" and summarily held down the "c" key right after - it loaded into the installer perfectly. So perhaps you might want to include an editor's note about this little tidbit, in case others are running into the same problem.
I did a fresh install on my second partition, which I always keep for file storage as well as for tinkering around with new OS releases (I remember I didn't finally remove Panther off of my machines until 10.4.7). I was immediately shocked at just how massive this OS is . . . even with all of the extra printer drivers, languages, fonts, and X11 deselected, I think it still required some 4.X GBs of disk space!
All in all, it is running just fine on not only the iBook, but equally as well on my 533MHz "Digital Audio" Power Mac. The only thing I"ve found to be "broken" by Leopard thus far is the free scrolling trackpad utility on my iBook - FFScroll, which is a major bummer, as I have come to rely heavily on the convenience - luckily the shareware utility SideTrack has been patched to work with Leopard, although I'll have to pony up $15 or live with the registration notifications if the authors of FFScroll don't get things working. I would say that any changes in speed, for better or worse, are largely imperceptible. The only noticeable change in speed is at startup and shutdown: Leopard is much slower than Tiger in this regard - startup takes a full minute on the iBook and slightly longer on the PM. Shutdowns also take longer, as OS X now updates the boot cache on the boot drive or something like that, whereas before the system would be down in under ten seconds. I'm not really upset by the lack of speed gains, and I think it's downright silly that some people are . . . we were all spoiled with 10.1 through 10.4. While those previous optimizations were nice, it was almost unnatural - rarely, if not never, do major upgrades consistently run faster on the same hardware. Frankly, I'm of the mindset that we should just be happy it doesn't run horribly slower than Tiger, and to a certain extent that it even runs at all - especially on the G4.
Despite those sentiments, in the end, I really can't understand Apple's seemingly arbitrary cutoff point simply because they did choose to continue support for the G4. Strike that - I can understand, because Apple is a business, they want their customers to buy new Intel machines, and they want to eventually stop spending time and resources on developing for PPC. But from a strict hardware point of view, it makes no sense. I would say in my own personal assessment - any AGP-based G4 with at least 32 MB of VRAM will run Leopard as comfortably as it did Tiger. Apple should have only set that 867 MHz as a "recommendation", just because many Mac users are apt to stick with what they have rather than run out and by the newest hardware, because of that they're ultimately losing out on a lot of upgrade revenue.
As for me, I'm just glad someone figured out a simple way to get Leopard installed on these unsupported machines.
If we don't speak before then, I hope you and your family have a joyous Christmas and an ever-so wonderful New Year.
Thanks for sharing you discoveries and observations. None of us will ever fully understand why Apple drew the line at 867 MHz, especially when the dual 800 MHz Quicksilver Power Mac is much more powerful than a single-processor 867 MHz system. Whatever, at least we now have several ways to get around the limitations of the installer, although McDermond's is definitely the easiest.
I've added a note to McDermond's article about the boot-mac command. Thanks!
From Scott Kitts:
I live in Northern California (a mere 3 hours from One Infinite Loop), and their is no AT&T coverage. AT&T's map is a little deceiving. Their map shows everywhere you're AT&T phone will get a signal, not everywhere AT&T offers contracts. Case in point: Their map shows coverage in Idaho, south-western Oregon, and northwest California; but, in fact, this is a "partnered area" where you can only get a GSM phone contract from Edge Wireless. Since Edge Wireless is not AT&T, iPhones simply are not available.
I'm not sure I'd bother with a GSM phone in rural areas anyway. I had a friend visiting from northern Oregon, and he has an iPhone. He never got more than 1 bar of signal strength and couldn't make or receive calls anywhere in my area, while my CDMA based US Cellular RAZR V3m never went below 3 bars with crystal clear call quality. Since the most important part of any phone is the ability to make and receive phone calls, I'll stick a CDMA phone for now, even if I could get an iPhone.
I wouldn't hold your breath for better support for our V3m's from Apple any time soon. In Apple's view, all non-iPhones are unworthy of any true Mac user's time. Therefore, they will get only rudimentary support in iSync (just like support for non-iPods got put on the back burner in iTunes once the iPod came out). It's truly pathetic that Apple won't spend the time on iSync's device profiles to support more (and offer better support) for more phone models. Especially since all we're talking about is adding text files that describe each phone's capabilities in order to get them to work or to even be recognized by iSync.
Your experience with syncing is similar to mine, only reversed. I can synch both calendars and contacts "without any problems" with a microSD card in place, but get nothing but problems without one in place. And I say "without any problems" in quotes for a reason. If it does synch, I get everything; but it usually takes three or four tries before it will actually synch without failing to start. I should mention I use BlueTooth to synch, not the USB cable. Curiously, BlueTooth file sharing (including the auto push) has never failed. Even weirder, my father and I have identical RAZRs. His will only synch via USB (but again, BlueTooth file sharing functions perfectly). Calendar always gets synced, but only a random subset of his contacts will ever synch. Go figure.
I guess we'll have to see what goodies await us five years from now. Keep up the good work, and hopefully you won't get flamed too badly for not bowing to the iPhone god in your post!
Thanks for sharing your frustration with Apple and AT&T. It's silly that Apple has drivers for hundreds (perhaps thousands) of printers and digicams but has been so stingy in supporting mobile phones. I've had to hack stock scripts to get iSync to work with my Motorola V262 (see Getting iSync to Work with an Unsupported Cell Phone).
I have no problem with Apple monopolizing the MP3 player and digital online music markets, as iTunes works well for me, most of my music is ripped from CD, iPods are great music players, and I plan on owning Macs forever. I bristle at Apple's exclusive contracts with AT&T, Orange, Rogers, and other vendors in other countries. It was bad enough when AT&T had a two-year exclusive deal, as I was content to wait and have an Alltel contract a long ways from expiring, but when it jumped to five years, it told me that Apple was more interested in control than unit sales - over 2/3 of the mobile market doesn't use AT&T.
Let's hope that our various national governments will make it illegal for mobile phones to be locked to a specific carrier (and usually crippled in the deal) and for wireless services to refuse to support phones not locked to their service.
In response to your comment asking for benchmarks to see if switching to thousands of colors instead of millions made any performance difference, I ran Xbench and Geekbench tests at both settings.
Here are the results:
- Geekbench Score @ Millions of colors: 430
- Geekbench Score @ Thousands of colors: 435
- Xbench Score @ Millions of colors: 27.23
- Xbench Score @ Thousands of colors: 25.97
I don't know why Xbench performed worse at thousands of colors, but to me, the difference seems negligible for the casual web browsing that I use this machine for. Also, the difference between thousands and millions of colors is very noticeable, so I prefer millions of colors. I hope this helps!
Thanks for sharing your findings. I did a little testing on my own Power Mac last Friday, but using Tiger. I ran Xbench and Let 1000 Windows Bloom (average of three tests). With the latter program, more colors was always faster, but with Xbench I saw a small improvement going from millions of colors to thousands or 256:
- 1000 Windows: millions, 18.70; thousands, 19.58; 256, 23.03
- Xbench Quartz: millions, 129.92; thousands, 113.47; 256, 89.81
- Xbench OpenGL: millions, 108.42; thousands, 108.30; 256, 40.14
- Xbench User Interface: millions, 248.08; thousands, 229.11; 256, 225.04
Visually, there's not a lot of difference between 16-bits (thousands of colors) and 24-bits (millions) for regular work - nor is there much of a speed difference, so there's little benefit (and sometimes a loss) from switching to thousands of colors. As for 256 colors, it looks pretty bad.
Dan Knight has been publishing Low End Mac since April 1997. Mailbag columns come from email responses to his Mac Musings, Mac Daniel, Online Tech Journal, and other columns on the site.
Recent Low End Mac Mailbag columns
- Tying up Loose Ends on Bitty Boxen, Buttercup, and the State of Tech, 2002.12.02. Mandrake Linux rocks, Buttercup gets a new motherboard, and the sorry state of the tech sector today.
- Joomla and WordPress Highly Recommended Content Management Systems, 2008.09.09. Low End Mac share their suggestions for a powerful, friendly, WYSIWYG content management system.
- Apple Will Not Abandon Optical Drives, the Mac Drought, Purposeful Mac Acquisition, and More, 2008.09.04. Also Mac OS X 10.5 on a G4-upgraded Blue & White G3 and problems using a flat panel display with a Quadra 700.
- More in the Low End Mac Mailbag index.
Links for the Day
- Mac of the Day: PowerBook 165c, introduced 1993.02.10. The first color PowerBook had an attractive screen, slow graphics.
- May 25 in LEM history: 99: OS X and the econoMac - iMac iMpact - 00: Is a Mac better than a PC? - 01: 1993: PowerBook vs. ThinkPad - Old files on a new Mac - 05: Mac minis in classroom a big hit - Of mice and keyboards - 06: The best 'Book for my needs - 07: More RAM vs. matched RAM - Running OS X from flash memory
- Support Low End Mac
Recent Content on Low End Mac
- World Book Encyclopedia 2012 DVD, Tommy Thomas, Reviews, 2013.03.05. "You may be asking yourself, in an age of Wikipedia and instant information, is World Book still relevant?"
- Vintage Computer Festival SouthEast, April 20-21, 2013, Simon Royal, Mac Spectrum, 2013.02.25. Old Apple gear and old PCs.
- iMessage: The Ultimate Messaging Service?, Simon Royal, Mac Spectrum, 2013.02.21. In most ways, Apple's iMessage is far superior to BlackBerry Messenger.
- More links in our archive.
- Best Mac mini Deals
- Best 13" MacBook Pro Deals
- Best Intel iMac Deals
- Best iPod touch Deals
- Best iPhone Deals
- Best iPod nano Deals
- Best iPod classic Deals
- Best Apple TV Prices
- More deals in our archive.
Low End Mac Reader Specials
Cult of Mac
Shrine of Apple
The Mac Observer
Accelerate Your Mac
The Vintage Mac Museum
Mac Driver Museum
System 6 Heaven
System 7 Today
the pickle's Low-End Mac FAQ