More on Intel CPU Sockets
After reading the lists in CPU
Competition and Nitpick on Eric
McCann's Comments, John Cate adds:
One small inaccuracy in the list....
Socket 4 was the socket used in the Pentium 60 and 66 MHz systems.
It was a 273-pin socket that came in both LIF and ZIF formats. These
were the only CPUs it was used for, although a 120-133 MHz OverDrive
chip later came out as an upgrade path, and PowerLeap created an
adapter to allow Pentium MMX and K6 chips to work, as well.
Socket 5 was a 320-pin socket for the second-generation Pentium
(75-200 MHz). It was virtually identical to Socket 7 except that it
lacked support for split-level voltages. Adapters also exist to allow
MMX and K6 chips to run in it.
Socket 7 was a 321-pin socket for Pentium MMX, K6-x and Cyrix 6x86
and MII chips. Super Socket 7, you covered already. Although I would
add that most Super-7 processors spanked Intel processors of the same
speed on integer operations. The K6-III of 1999 was the fastest x86 in
the world at its introduction, except in the FPU department, and I
still have a machine in daily use that runs one.
Don't get me started on Socket 8. That was an Intel rip-off from the
beginning. Sockets 8 and 370 are virtually identical in the way they
function, and the only reason for the change was to deny Pentium Pro
owners a reasonable upgrade path.
The other sockets and slots, you hit right on the head.
All this talk of CPUs and slots and sockets on the Windows side
sure makes me glad I'm a Mac user.
Low End Mac Saved Me Money
Following a lead on our Week's Best Power Mac G4
Deals, Chris Borchert writes:
Thanks to your site I just got a great deal on a new G4. You might
want to mention that for PowerMax they too charge a installation fee
for the "free" RAM they provide. It is $20, and I know because I just
purchased my G4 from them.
Thanks for the note. We're always trying to help our readers find
good values. It's pretty standard procedure for companies to charge an
installation fee, but I've never seen it noted on the PowerMax website.
I'll make sure future "Best Deals" pages include this fact.
Overclocking an iMac
Recent discussion of overclocking the beige G3 inspired Alvin to
write:
Thank you for your time. I have an iMac 350. I checked Bekkoame's
site on overclocking the iMac
DV. It says it can go up to 1 GHz with x10 multiplier. Is it
possible to reach 1 GHz if I added a good fan like GlobalWin
(while looking for a volt in the iMac's board that can power it) or
sets of fan in it (with intake fan, outlet fan, paste, good sets of
heatsinks, etc.)?
Is it possible to increase the CPU's voltage, or is it locked? When
one overclocks, will the other parts become overclocked, too, and will
they require cooling, too? Any links on overclocking the hard disk,
video card, and memory for iMac 350 and up?
Why does the iMac have up to x10 multiplier, does it mean it can be
overclocked that many times?
First, quoting from the Bekkoame site, "Please note that any
modifications you make to your Macintosh are made at your own risk." If
you overclock things too far, you can damage your CPU and other
components.
The different multipliers made it easy for Apple to change jumpers on
the beige G3 or make some resistor changes on the slot-loading iMac and
change the processor speed. By planning ahead, the iMac motherboard
supported CPUs up to about 500 MHz before Apple adopted the next
generation G3 processor. This saves the expense of designing a new
motherboard every time you increase processor speed; it doesn't mean
Apple was ever able to actually test a 1 GHz G3 in an iMac - it's
just a theoretical possibility.
I've been inside early iMacs, but never one of the slot-loading models,
so I don't know if the CPU in these models can be removed. If the
processor can't be removed, you may be able to change the multiplier
from 3.5 to 4 and get your 350 MHz iMac running at 400 MHz. I suspect
you would run into problems running it at 450 MHz or faster - even if
you add a big heat sink and additional fans.
When IBM or Motorola offers the same chip at three or four speeds (350,
400, and 450 MHz for instance), the slower ones are usually the ones
that failed quality control tests at the higher speed(s). If they had
passed, the chips would be sold as more expensive chips due to the
higher rating.
If the CPU in your iMac is socketed, you may well be able to push
things to 1 GHz, assuming you can find a fast enough G3 on the
open market. Should is the operative word, as I don't know of
anyone who has done this.
Also note that because you are adding and/or removing resistors, there
is a greater risk of damaging the computer's motherboard than there is
in jumpered models like the beige G3. For maybe a 15% improvement in
CPU speed, I wouldn't advise doing this kind of surgery.
Hard drives are limited by their rotation speed. I have never heard of
overclocking a hard drive, which would generate considerable excess
heat. If you iMac's hard drive is too slow, consider replacing it with
a newer, faster 7200 rpm hard drive with a nice big 2 MB or
8 MB buffer.
It may be possible to overclock Rage Pro 128 circuitry on the
motherboard, but I've never heard of it being done.
Dropping a G4 into a Beige G3
In the same vein, Dan Brown wonders:
I have a G4/450 Sawtooth Mac, and a beige G3/366 (wife's Mac) and
have considered upgrading the G4 to an 800 MHz or 1 GHz clock
speed.
Is there a possibility I can take the G4/450 ZIF out of my G4 and
put it in the beige G3 after replacing the G4's ZIF with a higher speed
ZIF?
To the best of my knowledge, this will not work. The "Yikes!" G4
used the same ZIF socket as the beige G3 and blue & white G3; the
AGP models had an improved memory bus that the G3 processor didn't
support. Because of this, I don't think it will work.
If any readers can shed more light on the subject, I'm all ears.
More on Overclocking the Beige G3
I've been reading up on your pages about speed bumping the beige G3
(I have a 266 MHz machine) but I'm still not sure what physically has
to be done. What does a jumper actually consist of? Should I be looking
for something that can be purchased at my local computer store to plug
into the J16 jumper bank or will a piece of bent wire suffice?
The local Mac people want nothing to to do with my experimenting so
I can't get any info there.
How do I actually do this?
I explained the steps yesterday, and I really don't want to take
apart the beige G3 in my son's bedroom for photos, so I googled and
found Willy's Wierd
G3 Desktop. This page has a photo of the top and bottom of the
jumper block, which may be white or black. (For the record, what makes
Willy's G3 weird is that one of the jumpers seems to be shorted on the
motherboard.)
MacGurus also has a nice
graphic showing the layout of the G3 motherboard, which can
help you in locating the jumper block.
Refer to Clocking the Power Mac G3 for
details on what the various jumper positions indicate. I've updated the
page with the above links.
Fighting Spam
After reading my
thoughts on Mail's spam identification, Tom Mulhern writes:
I read toady's column and thought I would share an option with you
for dealing with spam.
Since lowendmac.com is run on FreeBSD, you can use Spambouncer or
Spam Assassin. I have Spambouncer installed, and I am enjoying it
immensely, since my spam is reduced dramatically. A few slip
through, but so far spam is reduced by 96-9% and Mail.app catches
those. See <http://www.spambouncer.org>.
I suggest that if you use Mail.app, you use one of the scripts that
adds SpamCop reporting. You
can get a free SpamCop or a very inexpensive SpamCop account. There are
2 scripts [linked
on MacUpdate] that will add this to Mail.
I have to recommend that you look into server side filtering. I
don't know how your *nix abilities are, but mine are a hair past
remedial, and I was able to do it. You or someone that is Unix savvy
can install this in about 5 minutes, and you can take back your inbox.
It is a nice feeling, and I encourage you to do it. Despite how well
Mail.app works, I find myself working remotely lately, and sometimes I
am on a dialup or a cellular connection and don't have the time or
desire to download spam at all, especially when I am paying for access
with my mobile phone (and Bluetooth - very cool)
One note about Bounce to Sender: It doesn't work for spammers. If
you want to bounce a message from your girlfriend, mother-in-law, etc.
to get out of some social engagement. (What? No I didn't know about the
[insert undesirable social event here]. I didn't get your email.) As
for spammers, no they will not get your bounce. Why? They fake their
return addresses. Look at the return address in some of your spam. It
is all fake. That is where you would be sending it to, a fake address.
Then what happens? You get the bounce back in many cases. So Bounce To
Sender is useless for spam.
I also, like many others, use a separate email to keep my email
address private. I use despammed.com, a free service, and they
even forward email!
Once I installed it, I couldn't believe that I didn't have to live
with spam. If you only use mac.com, then you'll have to rely on Apple
filtering your email. If you use a @lowendmac.com , then you could
effectively use this. I just thought I'd share this as I am so pleased
at having put such a major dent in my spam problem.
Last thing, false positives, 0. YMMV.
My count is 3 false positives so far - one from a mailing list and
two from the same source. I'm still in training mode, though, so Mail
hasn't yet come close to catching all the incoming spam.
There are pros and cons to server side filtering. The pro is that you
don't have to download the spam it catches. The con, and it's a big
one, is that you don't know when it throws out the baby with the bath
water. Some webmasters (including
MacSlash) have never received domain renewal notices because of
server side filtering. One false positive can have a very negative
impact.
At my last job, I ran the mail server. We used the MAPS RBL (Realtime Blackhole
List) to reduce spam. I chose MAPS because they have a very explicit
policy of only blocking servers proved to relay spam - and they would
remove servers once they cleaned up their act. It reduced spam but
didn't eliminate it.
That was the good thing. The bad thing was a client who couldn't email
us because his ISP was blacklisted. Is it worth losing a client (or
several) just so you can avoid downloading some spam and deleting it
yourself?
I have used SpamCop a few times and been reported to SpamCop a few
times. I don't like getting called on the carpet when someone who
subscribed to one of the LEM email
lists reports list messages as spam (every one of our lists
requires both an initial subscription request and email confirmation) -
and it has happened.
The AppleScript I have on Claris Emailer doesn't bounce spam, and I
agree that bouncing may not be the most effective tool out there.
Instead, the script sends an email message to the postmaster of every
mail server the message was relayed through, letting them know that I
received an email which I consider spam that was sent through their
system. Several have thanked me for reporting a problem they hadn't
been aware of.
My mac.com email is pretty low on spam, which is one reason I use it
for site-related email. I do have several other email addresses, some
on filtered servers and some on unfiltered servers. I don't run my own
servers, and I prefer that any filtering done on the server limit
itself to known relays. Too many black hole lists cast too wide a net,
which leads to false positives.
I prefilter my email using the free version of POPmonitor. If
it's not addressed explicitly to me, I'll probably delete it from the
server. If there's no subject or the sender's name appears with all
sorts of weird characters, I usually delete it. If I recognize the
subject as spam, it's gone. And if I wonder about it, I can preview the
first few lines of the email. POPmonitor lets me delete messages before
I check my mail. I use it every morning. Maybe someday I'll ante up for
the full version, which can be trained to automatically delete messages
base on rules you specify.
I'd much rather be in charge of filtering my email than have someone
else decide what gets blocked and why. I've been involved on all sides
of the issue - well, not as a spammer, but having been reported as one
due to email list - and that's where I've settled down on the
issue.
More on Radeons, Beige G3s, and OS X
Responding to the ongoing discussion of the ATI Radeon cards,
OS X, and beige G3s, Chris Ryan writes:
Thought I would add some info to this discussion. I have a Radeon
Mac Edition in my wife's Beige MT. This is a PCI Radeon with 32 MB DDR.
While I was at MWSF, the ATI rep that I was speaking with confirmed
that the Radeon Mac Edition is the better of the two cards and would
have been called the Radeon 7200 had it followed the naming of their
subsequent releases. Its availability is a bit limited (I purchased it
at MWSF 2001). Anyway, it has no problems with OS X that were
described in the article and works quite well with the PCI-Extreme hack
for OpenGL support.
Will Faster Ethernet Help?
Dan, enjoyed your article on upgrade a beige G3. I have one I've
upgraded with a faster hard drive, more RAM, ATI graphic card, USB
card, CD-RW. I'm running OS 9.2 and was considering upgrading to a
10/100 ethernet card, but I didn't know if I'd see any significant
speed gains surfing the Web with my cable modem, given that I still am
using the stock processor.
Any thoughts, either on speed gains or the best card to use?
Thanks.
In a word, no. Most cable modems provide 3 Mbps throughput or
less - sometimes less than 1 Mbps. 10Base-T ethernet is already
faster than that, so the only thing you would gain from moving to
10/100 ethernet would be faster file swapping with other computers
using 100Base-T ethernet.
Further Thoughts on Windows and Viruses
Following up my response to his
email on Windows, viruses, worms, and browser integration, Peter da
Silva writes:
I don't know the ins and outs of Windows (in)security, but I know
that Word and Excel macros, Visual Basic, handling of email
attachments, Outlook, and Outlook Express are all among the ways
viruses propagate in the Windows world. I was unaware that the tight
integration of Explorer with Windows also contributed to the
problem.
You know enough to have covered pretty much all the email-related
virus classes, anyway. :)
When you analyse them further, there are really only two classes:
macro viruses and viruses that take advantage of the browser-desktop
integration.
Both of them could be stopped dead fairly easily by Microsoft
reducing the integration between applications and the OS. They just
haven't chosen to do so. Why this hasn't been better documented and
publicised, I don't know . . . it seemed fairly obvious to me
back when it was first announced.
It's why I banned OE and IE several years ago at work, and why I
tried (and failed) to ban the general use of Microsoft Office. The
result has been that the only real virus infestation we've had has been
a Word macro virus.
I don't think Safari will ever integrate with the OS like IE does;
we already have Aqua to handle local and networked file services.
That's not what I meant when I talked about integration; I meant the
way Safari is tightly coupled to the underlying OS, optimized for
OS X, and cannot exist without it.
I'm not sure that these two concepts are related. Any application
written to a specific OS API, whether that's Apple's or Microsoft's,
cannot exist without it. Opera can get that level of integration
simply by writing a Windows-specific version . . . they don't
need any special knowledge. And on OS X, where the underlying OS
is itself an open-source project, it's not really clear that Apple can
isolate enough of the kind of "special knowledge" that would make a
difference.
Tight coupling to the OS really does imply the kind of reflexive
relationship that's been causing so much trouble in the Windows
world.
I suspect that you're right, and Apple will stay further back from
the edge than Microsoft, at least I hope so. Still, you do have to see
where that phrase raised a big red flag for me, no?
Yes. I'm so glad I don't have to live in the world of
Windows.
I do try to keep in touch with viruses and worms, though, as they are
perhaps the biggest Mac advantage in the era of cheap hardware and
pretty stable versions of Windows.
Like you, I'm surprised Microsoft hasn't done more to bulletproof
Windows. And I'm still wondering when we'll see the first worm or virus
written specifically for Mac OS X....
OS X, Unix, and Security
Regarding Peter da Silva's
concerns about security issues surrounding Browser/OS integration,
Peter Hummers writes:
Browsers in Unix-type OSes are not integrated with the OS,
but they can read file-heirarchies and html and text files. It's
common, especially on a command-line system, to use a Lynx browser as a
file-explorer; when you come to a text or html file that you might want
to edit, Lynx, through its preferences, will open the file in your
favorite editor (although you can't delete or move files this way).
Since the files and folders have names like / and /dev, it's easy to
point any browser at a file-tree.
This is not a security problem, as ordinary Unix-type OS users have
limitations on what they may or may not do to the system itself, and
the programs they run are under the same limitations. On today's
single-user desktop Unices like Linux, they may su - "switch users" -
to temporarily become an administrator, by giving the administrator's
password.
OS X (a Unix variant) users need to learn to use their computers as
unprivileged users; from the little time I've had on OS X, I see
there is a setting that will force users to enter a password to get to
the system settings, but it seems to require deliberate setting up.
People who are used to single-user systems, like classic Mac and that
other popular one, will probably not do so.
This is basic - and very important - UN*X security, for the
browser-integration reason, among others. You are often connected to
the Internet. People or scripts that hack into a user's account can go
no farther.
About Rick Barham's comments in the same Mailbag on the complexity
of OS X, I have to agree. Unix-type systems are quite simple once
you have mastered a few concepts, which are applicable to most Unices,
from Linux to BSD. Mac OS 9 and earlier were simple, too, if not
very stable. OS X seems to be a jumble of quirks and protocols,
neither fish nor fowl, as it were. I use console-FreeBSD and Mac 9 at
home, and I'm confused - and annoyed - in OS X. (OS X
is even based on FreeBSD.)
Also BSD and Linux take very little memory; it's obscene what
OS X demands, mainly to run its gooey GUI. (Yes, it's pretty, but
enough is enough, especially if you have to work on the computer ~40
hours a week.)
You know what's a nice, simple OS? Palm OS. It's beautiful; when
you're done in an application, just leave it. Your work is saved, and
when you come back you pick up where you left off. I'd like to see that
on a simple desktop machine or keyboarded notebook!
The Palm OS is more like Macintosh System 6 than any computer OS
since. Simple. Elegant. Fast. Does one thing at a time.
Mac OS X is a Macintosh OS designed for users, not a server OS. It's a
different paradigm, even though Apple does have BSD at the core of
OS X. And that's why X lets users automatically log into their
computers - while also protecting them by pretty much insisting that
they never work with root privileges. The normal user setting on
OS X is underprivileged compared with the admin
setting.
As for OS X bloat, you're right on the money. It would be so nice if we
could simply turn Aqua off and use a faster, simpler, less pretty
interface. There's a lot of potential flexibility that Apple keeps from
the users by insisting that we use Aqua in exactly the way Steve Jobs
insists is just right. Sounds more like a Big Brother move than
something from the company that celebrates diversity and
individuality.
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.
