Talk:Power Macintosh G3 beige

From Wikipedia, the free encyclopedia
  (Redirected from Talk:Power Macintosh G3)
Jump to: navigation, search
WikiProject Apple Inc. (Rated Start-class, Mid-importance)
WikiProject icon This article is within the scope of WikiProject Apple Inc., a collaborative effort to improve the coverage of Apple, Macintosh, iOS and related topics on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
Start-Class article Start  This article has been rated as Start-Class on the project's quality scale.
Checklist icon
 Mid  This article has been rated as Mid-importance on the project's importance scale.
 


Untitled[edit]

It is stated that the Beige G3 Mac takes 66MHz RAM. I'm rather dubious of this, because some sites (e.g. this one) states it uses 100MHz RAM. Also, I have a pair of PC133s in my own Beige G3 Mimitower along with a PC100 bar, and they work fine. -- RAMChYLD 15:30, 15 Apr 2005

  • Isn't faster RAM backwards-compatible with slower clock speeds? (e.g. I'm pretty sure I've got some 133 Mhz RAM in a x86 100 MHz motherboard, and it seems to work fine). I've always assumed that if the RAM is the same speed as the motherboard, or faster, that it'll work OK. -- All the best, Nickj (t) 05:42, 16 Apr 2005 (UTC)
    • Yes, I understand that most modern RAM is backwards compatible with older specifications (which is why I was daring enough to put in 133MHz RAM into mine). Was just a little confused because some websites (like the one in the link) claims that the Beige G3 runs on 100MHz RAM and not 66MHz. -- RAMChYLD 14:57, 24 Apr 2005

Here are the specs from Apple. [1] Says that they use 100 MHz RAM, although they have a 66 MHz bus. AlistairMcMillan 13:17, August 9, 2005 (UTC)

Hint: you should mention about OldWorld and NewWorld differencies.

Of course they are backwards compatable ;-)

Yes, the computer can use up to PC133, but it will always run at 66MHz (actually, in the G3's case, 66.83Mhz), unless you overclock it. The original Beige G3s (even the last ones produced) shipped with 10ns PC66 CL2 SDRAM. Angelic Wraith

Fixing this. jhf 18:59, 15 November 2006 (UTC)

Adding reference to copper technology?[edit]

I noticed that there is a link to the copper metal wiki but could someone please add a reference to the how copper is used in the G3 processors manufactured by IBM? mattabat

Explain to me what you mean. You need help explaining it, or you don't know how to link a page you wish to? Dan 00:35, 11 February 2006 (UTC)

I don't know *of* a page to link to. mattabat

You're looking for Copper-based chips (I already linked there). -- grm_wnr Esc 11:35, 16 February 2006 (UTC)

Images and splitting[edit]

Okay, Angelic Wraith, any particular reason to throw out two images ([2])? -- grm_wnr Esc 23:34, 13 March 2006 (UTC)

Yep. As it was, they took up too much space, and we didn't need them if we already had the new images there. I also feel it's somewhat strange to have a picture of a B&W G3 at the top, when the Beige was the first one anyway... and is directly followed by text on it. Dan 01:23, 14 March 2006 (UTC)

I think an article should have a lead image, and I don't think a side view of the B&W G3 (with the large lettering) was unreasonable. Having pictures of all three case styles for the beige G3 is relevant as well (couldn't find an AIO G3 yet). However, I'm thinking of splitting the article anyway, since the two have little to do with each other apart from the CPU. Any thoughts on this? -- grm_wnr Esc 01:45, 14 March 2006 (UTC)

Well, that's the problem. They have a great deal to do with each other; initially when I came upon the article, I was wondering why on Earth they were grouped together... of course, it became rather apparent that it would be tough to do it otherwise. They're both "Power Macintosh G3s". Sure, one is the Beige and one is the B&W, but PowerMac G4 and PowerMac G5 both bring up the same, uniform thing, by and large.

I think what is "tricking" us into thinking that they should have separate pages is the enormous aesthetic difference; after all, compare the original "Graphite" Power Macintosh G4 (Yikes! based, preferably), to the last iteration of the G4... the Mirrored-Drive Doors model. If you look at the iMac and iBook pages, they're also all grouped together.

Plus, what would we call the page? It would be sort of hard to do it and allow people to immediately get the page they want.

If we can figure something out, that'd be great, but if not, it should stay.

Btw, I really don't think we should add any more pictures, just because it starts to make the page unwieldy, as well as increases load times for dial-up users Dan 04:54, 14 March 2006 (UTC)

Standard disambiguation rules say that the two pages should be called Power Macintosh G3 (Beige) and Power Macintosh G3 (Blue & White). Note that this page does not even have full tech specs yet, if we add them, the page will soon become unwieldy. These are really radically different machines on the inside, and the Power Mac G4s are more closely related to the G3 than the beige is to the B&W, New World ROM and all (especially the aforementioned Yikes!, which is a B&W with a different CPU). They just happen to use the same CPU, which gives them the same name, but that's no compelling reason these must share an article. Not surprisingly I would advocate splitting the other articles too once they get enough content (Power Macintosh G4, in particular, is still much too short for a split, but iMac and iBook look promising).
As for the images, I guess the B&W side view isn't really needed, and I'll leave out the G3 minitower until I find a better (larger) picture that really shows what the tower looks like. Deal? -- grm_wnr Esc 09:25, 14 March 2006 (UTC)
P.S.: An interesting metric when considering a split is whether incoming links to this combined page actually mean both or only one of the models. My count is 2 (both) / 8 (one) / 3 (unknown, could be one way or the other). I didn't count links coming from the "Post-1998 Apple products" template (which would logically only link the B&W). Now, someone counting differently (and with a different bias) will get slightly different results, but is clear that the ratio is heavily in favor of one of the two models, suggesting that the split would make sense. -- grm_wnr Esc 11:08, 14 March 2006 (UTC)
P.P.S.: Actually, what I wrote above is not strictly correct, there are two options for disambiguation:
I prefer the latter (with a bit more explanation than the usual disambiguation page), though I'm not passionate about it. -- grm_wnr Esc 11:45, 14 March 2006 (UTC)
I see you've thrown out the other image now... *grumble* Okay, I'll leave it for now. If no comments against the split materialise, I'm going to go ahead with it soon, and then both articles can have two images again. -- grm_wnr Esc 14:15, 20 March 2006 (UTC)


I actually quite like the idea of two different pages, but like I said it might be a bit annoying for people to type in "Power Mac G3" and get two pages. I think, though, that the page names you listed are definitely the right ones to choose. And you're right, they're in many ways different computers (similar in more ways than you might think, though, they both used iterations of the Grackle chip), and their CPUs were actually different. Both G3s, sure, and obviously they ran at different bus speeds (66 vs. 100MHz), but they were based upon different manufacturing processes... but then, you probably knew that. The New World ROM, IMO, was a dumb idea, since large ROMs are cools :P ...lol. anyway >_>

Yep.. the PCI-based G4 was a B&W G3 with a PPC7400 chip, and the ADB port hacked off. Specs... that'll be incredibly easy. We can add all the stuff that no one cares about too... Heathrow, Grackle, CUDA, SWIM 3, MESH, etc.

I would also really like to see (and hey, help with) the iMacs being split up too. Maybe since they're so diverse (so many different revisions), "iMac" should still direct to iMac, and then there could be separate links for each different one... obviously Rev. A and B should be grouped together since they're basically the same thing... Dan 21:45, 20 March 2006 (UTC)

I've gone ahead and made the split now - it's easily reverted, if somebody doesn't like it at all. The new page (Power Macintosh G3 (Blue & White)) can be moved to a better name if anyone can think of one, without much hassle (only one redirect points to it). -- grm_wnr Esc 16:31, 21 March 2006 (UTC)
I like it. I did swap pictures (minitower with dt for the template), but they're both still on the page. As for the B&W page, I linked a few thinks which previously hadn't been linked because they were already on the page when it was one article. And I screwed one of the links up. :) Dan 21:42, 22 March 2006 (UTC)

WP:MOSNUM[edit]

I feel it's pointless to include Gibibits for several reasons.. though, mainly because it's just sort of ridiculous. And all that ;nbsp just sort of makes things confusing. Angelic Wraith 18:37, 16 March 2006

My argument is based on Wikipedia guidelines (WP:MOSNUM). Putting nbsp between the value and the unit is unambiguously specified by WP:MOSNUM. Using binary multipliers is specifically recommended. Indeed, there is a specific guideline that revisions to binary multipliers should be accepted where the value is a binary-based value. I believe that in this case there is no reason to go outside the Wikipedia guidelines. If you don't like the guidelines, you are free to contribute and work towards having them revised. Duckbill 19:42, 16 March 2006 (UTC)

Umm... I'm not sure what you're asking for. "Binary Prefixes?" That's just what we have, isn't it? 228 512 byte blocks is just that, right?

As for the nbsp... I haven't the slightest idea what the reasoning behind that is, but it makes it harder for people who actually do edit this article, not the people who come by and put "nbsp" in everywhere. Believe me, if I had any idea how to combat this, I would work to get the spaces removed altogether... "333 MHz" and "768 MB" looks incredibly stupid.

If this sort of thing continues, I'm just going to stop editing any page that has anything to do with computers altogether, and remove anything I've added, which is considerable, whether it's considered vandalism or not. Dan 21:15, 16 March 2006 (UTC)


Dear Angelic Wraith, I am pleased that you are enuthusiastic on having this article be the best it can be. I had a sniff around and had a read on the articles Wikipedia:Assume good faith, Wikipedia:No angry mastodons, and Wikipedia:Resolving disputes, and found them to be helpful for me.

I have been working on the assumption that Wikipedia guidelines outweigh individual opinions. Do you think that idea makes sense, and if so do you think it might apply here? Duckbill 22:29, 16 March 2006 (UTC)

You wrote As for the nbsp... I haven't the slightest idea what the reasoning behind that is. I believe the reasoning for that is a consensus has been reached that numbers with units are best with a space between the value and the unit, and that a consistent style is better than having each individual contributor use their own style. Also, I believe that it has been felt that breaking the line between the value and the unit looks sufficiently bad that a non-breaking space should be used to keep them together, even if this makes the wiki markup marginally less easy to follow.

You wrote Believe me, if I had any idea how to combat this, I would work to get the spaces removed altogether... "333 MHz" and "768 MB" looks incredibly stupid. If you wish to 'combat this', you could start a discussion thread on Wikipedia talk:Manual of Style (dates and numbers). Whilst I would wish you every success, I would not be wanting to place a wager that the guidelines would be changed as a result.

You wrote I'm not sure what you're asking for. "Binary Prefixes?". The size limit for LBA(28) addressing is a function of the size of the LBA address space (which is a function of the representation of the LBA as a 28-bit value, resulting in an address space of 228 blocks) and the size of a hard disk block (512 or 29 bytes). This results in a total size of 228 × 29 bytes or 228+9 bytes or 237 bytes. So everything in the derivation of this value was a power of two. As a result, this is a prime candidate for representation using one of the binary multiplier prefixes. The appropriate prefix is the one equivalent to 230 which is gigabinary. So we end up with 27 or 128 as the value before the prefixed unit, resulting in a representation of 128 GiB.

Conventionally, the resultant figure is more familiarly known by an approximation using the power-of-ten multipliers, 137 gigabytes. So it's appropriate to mention that value as the main value. If we are to allude to a power-of-two significance by using an alternative representation, we can just use the compact form 128 GiB. If the reader is in any doubt as to what GiB means, then they can just follow the link which will either be there, or it will have been linked earlier in the article.

The issue has already been discussed at length, consensus has been reached, and the result is embodied Wikipedia guidelines. There is little to be gained by individuals flying in the face of the guidelines in individual articles. If there is a weight of opinion against the guidelines, this should result in a change of consensus, and a change in the guidelines.

Mentioning the 128 GiB is probably superfluous. Hard disks are normally measured in gigabytes. This isn't an article about the 28-bit-LBA limitation. We can write the whole thing more succintly and avoid gibibytes altogether. Duckbill 23:41, 16 March 2006 (UTC)


Thank you for responding candidly and cordially. That's more than I can say for myself- at least the cordial part :)

No, you're right, as a rule the guidelines usually outweigh personal opinion. It's just that I can't see any reasoning behind them whatsoever, and I feel that following rules for their own sake isn't necessarily the thing to do, especially if it irritates some people (let's just use the name "Dan"), and in no way increases (or even decreases) the effectiveness of the page.

When I look at a webpage and it says something along the lines of "The Power Mac G3 was introduced at 233, 266 and 300 MHz," I almost immediately leave, because it looks like it was written by someone who doesn't know what they're talking about; it's sort of like reading an essay full of spelling and grammar errors, and by grammar I mean not only the stuff you normally see, but misused tenses and plurals. It just irritates me; it comes across as ignorance, and on pages I've been working on for months, it greatly annoys me. Especially when I used to get issues of MacWorld and read up about Apple's new computers (I was what, twelve at the time when G3s were new?).

"The 300 MHz model shipped with 1 MiB L2 Cache" makes me want to rip my hair out. Why on earth are there discussions about Apple's own Manual of Style... regarding things like possessiveness (you shouldn't write "G3's"), when we go against the very way they present their products? MiB is also very likely to confuse anyone not used to Wikipedia articles... I've been reading about computers for years, and I never once stumbled across "MiB." At least a byte is by itself, and there's no BiB :P

You wrote I believe the reasoning for that is a consensus has been reached that numbers with units are best with a space between them, and that a consistent style is better than having each individual contributor use their own style.

That's not reasoning, that's an opinion for which I have yet to see any substantiation. I asked why it's considered better.

Also, I believe that it has been felt that breaking the line between the value and the unit looks sufficiently bad that a non-breaking space should be used to keep them together, even if this makes the wiki markup marginally less easy to follow.

You lost me with this.

The size limit for LBA(28) addressing is a function of the size of the LBA address space (which is a function of the representation of the LBA as a 28-bit value, resulting in an address space of 228 blocks) and the size of a hard disk block (512 or 29 bytes). This results in a total size of 228 × 29 bytes or 228+9 bytes or 237 bytes. So everything in the derivation of this value was a power of two. As a result, this is a prime candidate for representation using one of the binary multiplier prefixes. The appropriate prefix is the one equivalent to 230 which is gigabinary. So we end up with 27 or 128 as the value before the prefixed unit, resulting in a representation of 128 GiB.

Here I'm completely lost. I don't mean to be rude, but this seems really circuitious and tautalogical, to the point where I'm just not understanding you.

Mentioning the 128 GiB is probably superfluous. Hard disks are normally measured in gigabytes. This isn't an article about the 28-bit-LBA limitation. We can write the whole thing more succinctly and avoid gibibytes altogether.

I would normally agree with you, except there is no such thing as either a 128GB or 137GB HDD. Maybe you're right, though.

Anyway, as it is, I don't see the need to change the units. If on a personal level you don't care, then why bother? Dan 00:21, 17 March 2006 (UTC)

USB 1.1 full-speed data rate[edit]

The article was giving the USB 1.1 full-speed data rate in both Mbit/s and MB/s (12 and 1.5).

Really we only need to say that it is a USB 1.1 port, and quickly, as an aid to memory mention what the data rate was, in the way that it is usually specified, which is 12 Mbit/s. Supplying a conversion of 1.5 MB/s is a bit over-the-top here. Confusingly, it's also the same number as the bit rate for low-speed USB, which the port would also have supported, but which doesn't get a mention. Duckbill 23:51, 16 March 2006 (UTC)

USB 1.1 is low-speed USB. What are you talking about? >_< I'm having trouble understanding what you're trying to say... especially in the first sentence. Anyway, the reason I put both, is because USB (and FireWire) is always advertised in Mbps, whereas everthing else (ATA, SATA, SCSI) is advertised in MBps. A lot of people see "12Mbps" and think it's 12MBps. Ok? I mean, think about it. Even if they do know that a bit is one eighth of a bite, they shouldn't have to do the conversion in their heads- sure it's easy for us, but a lot of people have trouble with it. Dan 00:00, 17 March 2006 (UTC)

No, USB 1.1 is a serial bus standard encompassing two data rates:

  • Full speed - 12 Mbit/s
  • Low speed - 1.5 Mbit/s

"Full speed" is 12 Mbit/s.

On your second point re confusion of Mbit/s and MB/s, that is precisely why the recommended terms are Mbit/s and MB/s. The term "Mbit/s" (which is what I wrote in the article) includes the entire word "bit", and as such means that it cannot feasbly be confused with a bytes (or if it is, the reader is having serious comprehension problems, in which case all bets are off anyway). Duckbill 00:12, 17 March 2006 (UTC)

Notice Mbps and MBps [3] Note no space between any units and numbers on Apple's page. [4]. Dan 00:27, 17 March 2006

That's fine. Those are Apple documents on Apple's web site. Apple are perfectly at liberty to use whatever style they see fit. But we are not reporting text verbatim. We are reporting the information accurately, and in Wikipedia that information should be represented according to Wikipedia guidelines. So Apple may well write "2.5GHz", but in Wikipedia that quantity is represented as "2.5 GHz" (2.5&nbsp;GHz). Duckbill 18:06, 20 March 2006 (UTC)

This same discussion took place on Apple Macintosh. I was thinking about it today and look at this reasoning for the space. 1) Which looks correct to you: 1MB or 1 MB, 1foot 1 foot. It is a unit of measure, numbers and their unit should be separated because it looks better IMO, much easier to do mathmatical calculations with the space, and the manual of style has it this way. This is wikipedia, guidelines have been constructed over the years and you have to respect that. Removing anything you have added from wikipedia is vandalism as well, since it's not yours, clicking save means "You agree to license your contributions under the GFDL". Why do you need to keep fighting this fight? Mike (T C) Star of life2.svg 06:00, 21 March 2006 (UTC)

Well, I suppose it's a "fight" on both sides if you "fight back." Anyway. . . I'm not really sure why you feel it's "easier for calculations"... it almost makes it harder for me... but it's not much of a point. As for "1MB" and "1foot".. I would write it as "1MB," (or 1 megabyte) and 1 foot (or perhaps one foot). I do think 1MB looks a bit stupid, but any number other than "1" with KB/MB/GB etc (i.e. 128MB). looks fine, simply because the "1" is too narrow. Unfortunately, it boils down to opinion, and obviously we differ there.

On Board SCSI controller[edit]

The article claims an "NCR 53C96 SCSI controller". This controller can neither be found in Apple's specification (http://developer.apple.com/documentation/Hardware/Developer_Notes/Macintosh_CPUs-G3/PowerMac_G3/PowerMac_G3.pdf) nor on this site: http://www.heise.de/ct/97/13/208/ nor on any other site in the internet. Even a detailed picture (http://www.heise.de/ct/97/13/208/gossamer.jpg) of the Gossamer mainboard does not show any dedicated SCSI chip. Instead, both Apple's document and the computer magazine c't clearly state that the I/O controller Heathrow handles SCSI (among other interfaces). This solution is called MESH.

Of course it could be possible that NCR's controller is included as a part/module of Heathrow, but this is nowhere documented.

Thus there are a lot of proofs that there is no dedicated SCSI controller chip, and there is not a single proof that there is one. So either the article must be corrected, or a new photo of a Gossamer motherboard must be added that clearly shows the NCR chip.

This article is meant to be objective and should not reflect a single subjective idea or opinion or memory of a single person.


MESH is Macintosh Enhanced SCSI Hardware. I'm also familiar with that picture.. and with the insides of the Beige G3.. I have three of them. The 53c96 isn't a chip, at least not in the G3s; I'm fairly certain it's incorporated into the MESH custom IC... which includes the SWIM3 controller (floppy drive), heathrow (IDE) etc, so you wouldn't see anything labeled "NCR 53C96". Software, such as TechTool Pro, even reports the presence of the 53c96 SCSI chip.. as well as the VIA1, SCC, etc. Dan 00:08, 31 March 2006 (UTC)
Update mmm, nope, you were right. Though, if it had used the 53c96 chip, it wouldn't have been visible... it still would have been part of the Heathrow; it's MESH-based. I was under the impression that the 53c96 was part of MESH. My apologies :) Dan 19:19, 31 March 2006 (UTC)

SCSI Error[edit]

In addition to the above problem the article states that the computer has "...Fast SCSI chain (up to 5 MB/s..." This is an error Fast SCSI has a 10MB/s speed. The original SCSI standard was 5 MB/s. Either this machine has Fast SCSI @ 10MB/s or else SCSI @ 5MB/s. Will a knowledgable party please correct it?

The article was right in alternate cases. They had 2 SCSI buses - internal & external. The internal bus was fast scsi, the external was normal scsi —Preceding unsigned comment added by Pantergraph (talkcontribs) 21:43, 2 December 2010 (UTC)

Compatible optical drives[edit]

Given the section on optical drive compatibility, it would probably be beneficial to compile a list of what is known to work. I have found that a BTC CDRW5224 reads and burns correctly on a Revision A G3 with Mac OS 9.2.2, and supports booting. Removing the front panel of the tray will allow the drive tray to open when the case is put back on. Dashwarts (talk) 18:57, 8 June 2010 (UTC), boot section updated Dashwarts (talk) 05:25, 9 June 2010 (UTC)

PCI[edit]

Where the slots 32 or 64-bit wide? The article is ambiguous in this respect. Someone not using his real name (talk) 06:49, 17 December 2013 (UTC)

If this photo is depicting what it claims to (and it seems to), then they were 32-bit, unlike in the successor Power Macintosh G3 (Blue & White). Someone not using his real name (talk) 06:51, 17 December 2013 (UTC)