1.1. Purpose and History

The purpose of this document is to give you the background information you need to be a savvy buyer of Intel hardware for running Unix. It is aimed especially at hackers and others with the technical skills and confidence to go to the mail-order channel, but contains plenty of useful advice for people buying store-front retail.

This document is maintained and periodically updated as a service to the net by Eric S. Raymond, who began it for the very best self-interested reason that he was in the market and didn't believe in plonking down several grand without doing his homework first (no, I don't get paid for this, though I have had a bunch of free software and hardware dumped on me as a result of it!). Corrections, updates, and all pertinent information are welcomed at esr@snark.thyrsus.com. The editorial `we' reflects the generous contributions of many savvy Internetters.

If you email me questions that address gaps in the FAQ material, you will probably get a reply that says ``Sorry, everything I know about this topic is in the HOWTO''. If you find out the answer to such a question, please share it with me for the HOWTO, so everyone can benefit.

If you end up buying something based on information from this HOWTO, please do yourself and the net a favor; make a point of telling the vendor ``The HOWTO sent me'' or some equivalent. If we can show vendors that this HOWTO influences a lot of purchasing decisions, we get leverage to change some things that need changing.

Note that in December 1996 I published an introductory article on building and tuning Linux systems summarizing much of the material in this HOWTO. It's available here.

This document actually dates back to 1992, when it was known as the ``PC-Clone Unix Hardware Buyer's Guide''; this was before Linux took over my world :-). Before that, portions of it were part of a Unix Buyer's Guide that I maintained back in the 1980s on USENET.

1.2. New versions of this document

New versions of the Unix Hardware Buyer HOWTO will be periodically posted to comp.os.linux.help and news:comp.os.linux.announce and news.answers. They will also be uploaded to various Linux WWW and FTP sites, including the LDP home page.

You can view the latest version of this on the World Wide Web via the URL http://www.linuxdoc.org/HOWTO/Unix-Hardware-Buyer-HOWTO/.

1.3. Feedback and corrections

If you have questions or comments about this document, please feel free to mail Eric S. Raymond, at esr@thyrsus.com. I welcome any suggestions or criticisms. If you find a mistake with this document, please let me know so I can correct it in the next version. Thanks.

1.4. Related resources

You may also want to look at the read the Hardware-HOWTO. It lists hardware known to be compatible with Linux, and hardware known to be incompatible.

3.1. Getting Down to Cases

I used to say that cases are just bent metal, and that it doesn't much matter who makes those, as long as they're above an easy minimum quality (on some really cheap ones, cards fail to line up nicely with the slots, drive bays don't align with the access cutouts, or the motherboard is ill-supported and can ground out against the chassis).

Unfortunately, this isn't true any more. Processors run so hot these days that fans and airflow are a serious concern. They need to be well designed for proper airflow throughout.

If you're fussy about RFI (Radio-Frequency Interference), it's worth finding out whether the plastic parts of the case have conductive coating on the inside; that will cut down emissions significantly, but a few cheap cases omit it.

Look for the following quality features:

• Aluminum rather than steel. It's lighter and conducts heat better.

• Unobstructed air intake with at least one fan each (in addition to the power supply and processor fans)

• No sharp metal edges. You doon't want to shred your hands when you're tinkering with things.

• There shouldn't be any hot spots (poor air flow).

• Sturdy card clips. Some poorly-designed cases allow cards to wiggle out of their slots.

• Effective and easy to use mechanisms for attaching hard drives, CD-ROM, CD-R/W, DVD, tape drive etc.

Should you buy a desktop or tower case? Our advice is go with tower unless you're building a no-expansions personal system and expect to be using the floppies a lot. Many vendors charge nothing extra for a tower case and the absolute maximum premium I've seen is $100. What you get for that is less desktop clutter, more and bigger bays for expansion, and often (perhaps most importantly) a beefed-up power-supply and fan. Putting the box and its fan under a table is good for maybe 5db off the effective noise level, too. Airflow is also an issue; if the peripheral bays are less cramped, you get better cooling. Be prepared to buy extension cables for your keyboard and monitor, though; vendors almost never include enough flex.

The airflow thing is a good argument for a full- or mid-tower rather than the `baby tower' cases some vendors offer. However, smaller towers are getting more attractive as boards and devices shrink and more functions migrate onto the motherboard. A state of the art system, with all 3" disks, 300W power supply, half-size motherboard, on-board IDE and 64meg of RAM sockets, and half-sized expansion cards, will fit into a baby or midized tower with ample room for expansion; and the whole thing will fit under a desk and make less noise than a classic tower.

For users with really heavy expansibility requirements, rackmount PC cases do exist (ask prospective vendors). Typically a rackmount case will have pretty much the same functionality as an ordinary PC case. But, you can then buy drive racks (complete with power supply), etc. to expand into. Also, you can buy passive backplanes with up to 20 or so slots. You can either put a CPU card in one of the slots, or connect it to an ordinary motherboard through one of the slots.

I've recently become a big fan of Antec cases. They're inexpensive, sturdy, and thoughtfully engineered.

3.2. Power Supplies and Fans

Power supplies can matter but quality is cheap; give preference to those with a Underwriter's Laboratories rating. There's some controversy over optimum wattage level. On the one hand, you want enough wattage for expansion. On the other, big supplies are noisier, and if you draw too little current for the rating the delivered voltage can become unstable. And the expected wattage load from peripherals is dropping steadily. On the other hand, processors and their cooling fans eat a lot more power than they used to.

The choice is generally between 200W and 300W. After some years of deprecating 300W-and-up supplies, I'm now persuaded it's time to go back to them; a modern processor can consume 50-75W by itself, and for the newer dual-processor board the power supply needs to be rated 450W or up.

Quality brands include PC Power and Cooling, Antec and Sparkle. A lot of no-name power supplies are out of spec; this is an area where buying a name brand is good tactics (the price premium won't be large).

About that annoying fan noise, ask if the power-supply fan on a target system has a variable speed motor with thermostatic control; this will cut down on noise tremendously. However, be aware that a thermostatic sensor basically measures the temperature at the sensor (typically within the power supply box) and makes sure there is enough airflow to keep the power supply from overheating. The sensor does not know a thing about the temperature in certain hot spots likely to develop in a PC case (CPU, between SIMMs, between drives mounted in vertically adjacent bays).

This can be a problem, because in garden variety tower cases there often isn't enough airflow to cool all components effectively even if a single fan is going at full speed. This is especially true if your computer has lots of add-on cards or hard disks (not much airflow between cards or between drives). Note that the fan in the power supply was basically designed to cool the power supply, not the components in the case. Not providing additional fans is a sign of cheapness. On tower PCs with "expensive" engineering (e.g. HP Vectra, Compaq) one will find one to two extra fans besides the one in the power supply.

So the bottom line is, use thermostatic controls if you can to cut noise. But if you want high reliability, use two or more fans. Modern designs normally also have a small auxilliary fan mounted right over the chip.

The noise produced by a fan is not just a function of the speed with which it turns. It also depends on the nature of the airflow produced by the fan blades and the bearings of the rotor. If the blades causes lots of turbulent airflow, the fan produces lots of noise. One brand of fans that is much more silent than most others even if going at full throttle is Papst.

3.3. Motherboards

Provided you exercise a little prudence and stay out of the price basement, motherboards and BIOS chips don't vary much in quality. There are only six or so major brands of motherboard inside all those cases and they're pretty much interchangeable; brand premiums are low to nonexistent and cost is strictly tied to maximum speed and bus type. Unless you're buying from a "name" outfit like Compaq, Dell, or AST that rolls its own motherboards and BIOSes, there are only four major brands of BIOS chip (AMI, Phoenix, Mylex, Award) and not much to choose between 'em but the look of the self-test screens. One advantage Unix buyers have is that Unixes are built not to rely on the BIOS code (because it can't be used in protected mode without more pain than than it's worth). If your BIOS will boot properly, you're usually going to be OK.

The old ``AT'' and ``Baby AT'' motherboard form factors are obsolete. It's an ATX world now.

There are still a few potential gotchas to beware of, especially in the cheaper off-brand boards. One is ``shadow RAM'', a trick some boards use for speeding up DOS by copying the ROM contents into RAM at startup. It should be possible to disable this. Also, on a cacheing motherboard, you need to be able to disable cacheing in the memory areas used by expansion cards. Some cheap motherboards fail to pass bus-mastering tests and so are useless for use with a good SCSI interface; on others, the bus gets flaky when its turbo (high-speed) mode is on. Fortunately, these problems aren't common and are becoming less so.

You can avoid both dangerously fossilized hardware and these little gotchas by sticking with a system or motherboard design that's been tested with Unix.

Some other good features to look for in a motherboard include:

• Gold-plated contacts in the expansion slots and RAM sockets. Base-metal contacts tend to grow an oxidation layer which can cause intermittent connection faults that look like bad RAM chips or boards. (This is why, if your hardware starts flaking out, one of the first things to do is jiggle or remove the boards and reseat them, and press down on the RAM chips to reseat them as well -- this may break up the oxidation layer. If this doesn't work, rubbing what contacts you can reach with a soft eraser is a good fast way to remove the oxidation film. Beware, some hard erasers, including many pencil erasers, can strip off the plating, too!)

• The board should be speed-rated as high as your processor, of course. It's good if it's rated higher, so upgrade to a faster processor is just a matter of dropping in the chip and a new crystal.

• Voltage, temperature and fan speed monitoring hardware. This is now common on motherboards based on recent iterations of the Intel support chips, especially those designed for server use. Linux supports drivers that can read this hardware, and monitoring can help you spot incipient board failures.

If you're changing a motherboard, see the Installing a Motherboard page first. This one even has a Linux note.

3.4. Memory

All current PC designs include a cacheing memory controller and some fast on-chip cache that combine to produce higher effective speeds. Judging the cache design used to be one of the trickiest parts of the motherboard, but that stuff is all baked into the processor itself now.

Current motherboards use PC100 or PC133 SDRAM, which comes packaged on 168 pin DIMM modules. SDRAM is a large step forward in memory speed, at 10ns. SDRAM does not need to be installed in pairs. The key words you want to see on the spec sheet are:

• SDRAM DIMM -- the physical module type.

• ECC -- support for error correcting memory, important for reliability; Unix makes more efficient use of hardware and thus beats on your memory harder than Windows does. Be sure your motherboard supports ECC and note that "ECC compatible" is marketing manure for a board that alllows you to use ECC memory but doesn't use the ECC.

• PC100 -- this is the memory-module specification for current motherboards with a 100MHz Memory Bus support. This is minimum; PC133 or PC266 is better.

• 16x72 -- if it says 16x64, the extra 8 bits per 64 needed for ECC are not present.

For current motherboards with 133MHz Memory Bus support, PC133 should be used instead of PC100; it gives 33% greater memory bandwidth at very little additional cost. DDR-SDRAM and RDRAM are faster memory types that retrieve data in chunks and give you faster throughput. So-called `PC266' memory is designed for motherboards that transfer at 133 but double the width of the front-side bus connecting processor and memory.

As the throughput of processor-to-memory buses rises, memory latency (bus cycles required for the first fetch in a chunk) is becoming a more important statistic. Lower numbers are better.

For more technical stuff on memory architectures, see The Ultimate Memory Guide maintained by Kingston Technologies.

3.5. Buying a Video Card

Video controllers translate byte values deposited in their video memory by your GUI (usually an X server under Linux) into an analog RGB signal which drives your monitor. The simplest kinds treat their video memory as one big frame buffer, requiring the CPU to do all dot-painting. More sophisticated ``accelerated'' cards offer operations such as BitBlt so your X server can hack the video memory algorithmically. These days almost all cards even at the low end actually have some acceleration features.

Cards are rated by the maximum number of analog signal changes they can produce per second (video bandwidth). Video bandwidth can be used to buy varying combinations of screen resolution and refresh speed, depending on your monitor's capabilities.

Another important variable of video cards is the size of their on-board video RAM. Increased memory lets you run more colors at higher resolutions. For instance, a 1MB card usually will only allow 256 colors at 1024x768 while a 2MB card usually allows at least 16-bit color (a palette of about 65,000 colors). You'll need 4MB of video memory to use 24-bit or ``true'' color (16 million colors) at 1024x768.

The card's video RAM size has no effect on its speed. What does affect speed is the type of memory on board. VRAM (Video Random Access Memory) is fast but more expensive; it features a dual-ported design allowing two devices (the CRT controller and the CPU) to access the memory at the same time. DRAM (Dynamic Random Access Memory) is is similar to the RAM used in main memories. It is cheaper, more common, and slower (because the CRT controller and the CPU must take turns accessing the video buffer).

A quick review of monitor standards:

The Horizontal and Vertical Frequency columns refer to the monitor scan frequencies.

The vertical frequency is the upper limit of the monitor's flicker rate; 60Hz is minimal for ergonomic comfort, 72Hz is VESA-recommended, and 80Hz is cutting-edge. At resolutions above VGA, horizontal scans take long enough that the monitor may never reach anywhere near the vertical-frequency maximum; how close it gets is a function of the horizontal-scan frequency (higher is better).

For more information on how to avoid the evil screen flicker, see the XFree86 Video Timings HOWTO, a tutorial written by your humble editor and included with the XFree86 distribution.

It's no longer possible to find MDA video boards and monitors out there. Anyway, prices for SVGA have collapsed so totally that it's wouldn't be worth bothering.

XGA is an IBM-proprietary included for completeness, but is vanishingly rare in the clone market. 8514/A is another IBM standard supported by a few graphics accelerator cards. It is interlaced, and thus has a tendency to flicker. The VESA 1024x768 standard makes XGA and 8514/A obsolete.

SVGA or `Super VGA' strictly refers only to 800x600 resolution, but is widely used for 1024x768 and even 1280x1024 resolutions. Standards above 1024x768 are weak and somewhat confused.

These days, most vendors bundle a 15" or 17" monitor and super-VGA card with 1024x768 resolution in with their systems. Details to watch are the amount of RAM included (which will affect how much of that maximum resolution and how many colors you actually get), and whether the memory is dual-ported VRAM (slightly more expensive but much faster).

More information on video standards is available at the Video Display Standards page of the PC Guide site.

3.6. Selecting a Monitor

3.7. Keyboards

It's important to get a high quality keyboard with good key feel. See the typing-injury FAQ from sci.med.occupational to see what happens if you don't. Carpal tunnel syndrome is no fun for anyone, but it hits hackers particularly hard. Don't be a victim!

Hal Snyder of Mark Williams, Co sent us the following caveat: "We find that about 10% of cheap no-name keyboards do not work in scan code set 3. We are interested in scan code set 3 because only there can you reprogram the keyboard on a per-key basis as to whether keys are make-only, make-break, or autorepeat. It is a big win for international support and for X."

He continues: "Keytronic, Cherry, and Honeywell keyboards, as well as a large number of imports, work fine. My advice is to either buy a respected brand of keyboard, or deal with a vendor who will allow you to return an incompatible keyboard without charge."

Allen Heim <ajh@netscape.com> writes: I'm sold on NMB keyboards (http://www.nmbtech.com/), available from Global Computer (http://www.globalcomputer.com/), or at (800) 8-GLOBAL. Their line of mechanical-switch keyboards now have a lifetime warranty, and I've just ordered my second RT-8200 unit. I don't see it listed on their website currently, but I do see their "Windows 95-enhanced" model, the RT-8200W. It's the same thing, but with extra keys (which may be programmable; think of the Emacs "meta" key--could be useful).

Following the wave of mid-1990s publicity about repetitive-strain injuries, ergonomic keyboards have become increasingly common. One that looks promising to your editor (though I haven't yet used it) is the Marquardt MiniErgo MF2, from Marquardt Switches, Inc.; 2711 Route 20 East, Cazanovia NY 13035, phone (315)-655-8050; suggested list price $170, AT-compatible interface). Michael Scott Shappe <mshappe@jeeves.net> sent me a rave review of the Marquardt after having used it about six months.

The MF2 features a conventional QWERTY layout, but with the right and left halves split apart and rotated about 30 degrees towards each other in a shallow V shape. The theory is that being able to angle your arms inward and your elbows out produces a less stressful typing position.

The MF2 has no keypad, but it does have the standard 12 function keys across the top and arrow keys at the point of the V (meant to be thumb-operated).

I have seen and used a device called the Maltron ergonomic keyboard. This keyboard splits the keys into two main groups, each arranged in a dished hollow. Each hand also has easy access to separate thumb pads of nine keys each; there is one numeric and arrow-key pad in the center of the unit, between the right and left-hand groups. Also, the keys in each group do *not* have the alternate-row staggering of the conventional (Scholes) keyboard; this subtle change reduces torsional stress on the fingers and wrists tremendously.

I found the Maltron easy to use, and regretted having to give it back.

For more details on many ergonomic keyboards and typing-injury issues in general, see Dan Wallach's FAQ on repetitive strain injuries and ergonomic input devices, published monthly in news.answers

3.8. Floppy Drives

There's not much to be said about floppy drives. They're cheap, they're generic, and the rise of CD-ROM drives as a cheap distribution medium has made them much less important than formerly. You only ever see the 3.5-inch `hard-shell' floppies with 1.44MB capacity anymore.

You'll probably never use floppies for anything but first boot of a new operating system. Bootable CD-ROMs, standard of most PCs these days, eliminate even that use. So go ahead and settle for cheap Mitsumi and Teac floppy drives. There are no `premium' floppy drives anymore. Nobody bothers.

3.9. Printers

There's one huge gotcha about printers: Don't Get a `GDI' printer!

3.10. Power Protection

3.11. Radio Frequency Interference

(Thanks to Robert Corbett <Robert.Corbett@Eng.Sun.COM> for contributing much of this section)

Radio Frequency Interference (RFI) is a growing problem with PC-class machines. Today's processor speeds (above 33MHz) are such that the electromagnetic noise generated by a PC's circuitry in normal operation can degrade or jam radio and TV reception in the neighborhood. Such noise is called Radio Frequency Interference (RFI). Computers, as transmitting devices, are regulated by the Federal Communications Commission (FCC).

FCC regulations recognize two classes of computer:

If a PC is to be used in a home or apartment, it must be certified to be FCC class B. If it is not, neighbors have a legal right to prevent its use. FCC class A equipment is allowed in industrial environments.

Many systems are not FCC class B. Some manufacturers build boxes that are class B and then ship them with class A monitors or external disk drives. Even the cables can be a source of RFI.

It pays to be cautious. For example, the Mag MX17F is FCC class B. There are less expensive versions of the MX17 that are not. The Mag MX17 is a great monitor (I wish I had one). It would be painful to own one and not be allowed to use it.

An upgradeable system poses special problems. A system that is FCC class B with a 33 MHz CPU might not be when the CPU is upgraded to a 50 or 66 MHz CPU. Some upgrades require knockouts in the case to be removed. If a knockout is larger than whatever replaces it, RFI can leak out through the gap. Grounded metal shims can eliminate the leaks.

Even Class B systems don't mix well with wireless phonesets (not cellular phones, but the kind with a base station and antennaed headset). You'll often find a wireless phone hard to use withing 20 feet of a Class B machine.

To cut down on RFI, get a good metal case with tight joints, or at least make sure any plastic one you buy has a conductive lining. You can also strip the painted metal-to-metal contacting parts of paint so that there's good conductive metal contact. Paint's a poor conductor in most cases, so you can get some benefit from this.

4.1. How To Pick Your Processor

Right now (mid-2001), the chips to consider for running Unix are the Pentium IIs and Pentium IIIs and their clone equivalents from AMD or Cyrix. The Pentium IV is something of a dog (very poor price-performance, actually slower than a III on some benchmarks), and the Itanium isn't out of the starting gate.

Brands don't matter much, so don't feel you need to pay Intel's premiums if you see an attractive Cyrix, AMD or other chip-clone system offered.

To compare the performance of different Intel-based systems with each other and with machines from other manufacturers, you can take a look at the SPECmark Table at ftp://ftp.cdf.toronto.edu/pub/spectable. That document recommends (and I do too) that you read the SPEC FAQ at http://www.specbench.org/spec/specfaq.html to get background before browsing the table.

Good current advice about chipsets can be found at The Cheap /Linux/ Box.

4.2. Of Memory In...

Buy lots of RAM, it's the cheapest way to improve real performance on any virtual-memory system. 64MB now comes standard on most clone configurations. This is good enough for X.

Tuning is simple. Watch your job mix with top(1) and add memory until you're not swapping to disk any more.

4.3. Bus Wars

This is yet another area in which progress has simplified your choices a lot. There used to be no fewer than four competing bus standards out there (ISA, EISA, VESA/VLB, PCI, and PCMCIA). Now there are effectively just two -- PCI for desktop/tower machines and PCMCIA for laptops.

4.4. Disk Wars: IDE vs. SCSI

 ———-  ———   ————-   ———
 |Internal|--|Adaptec|—|New Quantum|—|UMAX   |
 |Devices |  |1542CF | ^ |  Disk     | ^ |Scanner|
 ———-  ——--  | ————- | ——— 
                       |               |
                   New $16 cable   Cable that came
                                     with scanner.

4.5. Other Disk Decisions

Look at seek times and transfer rates for your disk; under Unix disk speed and throughput are so important that a 1-millisecond difference in average seek time can be quite noticeable.

4.6. Tuning Your I/O Subsystem

(This section comes to us courtesy of Perry The Cynic, <perry@sutr.cynic.org>; it was written in 1998. My own experience agrees pretty completely with his. I have revised the numbers in it since to reflect more recent developments.)

Building a good I/O subsystem boils down to two major points: pick matched components so you don't over-build any piece without benefit, and construct the whole pipe such that it can feed what your OS/application combo needs.

It's important to recognize that ``balance'' is with respect to not only a particular processor/memory subsystem, but also to a particular OS and application mix. A Unix server machine running the whole TCP/IP server suite has radically different I/O requirements than a video-editing workstation. For the ``big boys'' a good consultant will sample the I/O mix (by reading existing system performance logs or taking new measurements) and figure out how big the I/O system needs to be to satisfy that app mix. This is not something your typical Linux buyer will want to do; for one, the application mix is not static and will change over time. So what you'll do instead is design an I/O subsystem that is internally matched and provides maximum potential I/O performance for the money you're willing to spend. Then you look at the price points and compare them with those for the memory subsystem. That's the most important trade-off inside the box.

So the job now is to design and buy an I/O subsystem that is well matched to provide the best bang for your buck. The two major performance numbers for disk I/O are latency and bandwidth. Latency is how long a program has to wait to get a little piece of random data it asked for. Bandwidth is how much contiguous data can be sent to/from the disk once you've done the first piece. Latency is measured in milliseconds (ms); bandwidth in megabytes per second (MB/s). Obviously, a third number of interest is how big all of your disks are together (how much storage you've got), in Gigabytes (GB).

Within a rather big envelope, minimizing latency is the cat's meow. Every millisecond you shave off effective latency will make your system feel significantly faster. Bandwidth, on the other hand, only helps you if you suck a big chunk of contiguous data off the disk, which happens rarely to most programs. You have to keep bandwidth in mind to avoid mis-matching pieces, because (obviously) the lowest usable bandwidth in a pipe constrains everything else.

I'm going to ignore IDE. IDE is no good for multi-processing systems, period. You may use an IDE CD-ROM if you don't care about its performance, but if you care about your I/O performance, go SCSI.

Let's look at the disks first. Whenever you seriously look at a disk, get its data sheet. Every reputable manufacturer has them on their website; just read off the product code and follow the bouncing lights. Beware of numbers (`<12ms fast!') you may see in ads; these folks often look for the lowest/highest numbers on the data sheet and stick them into the ad copy. Not dishonest (usually), but ignorant.

What you need to find out for a disk is:

1. What kind of SCSI interface does it have? Look for "fast", "ultra", and "wide". Ignore disks that say "fiber" (this is a specialty physical layer not appropriate for the insides of small computers). Note that you'll often find the same disk with different interfaces.

2. What is the "typical seek" time (ms)? Make sure you get "typical", not "track-to-track" or "maximum" or some other measure (these don't relate in obvious ways, due to things like head-settling time).

3. What is the rotational speed? This is typically 4500, 5400, 7200, 10000, or 15000 rpm (rotations per minute). Also look for "rotational latency" (in ms). (In a pinch, average rotational latency is approx. 30000/rpm in milliseconds.)

4. What is the `media transfer rate' or speed (in MB/s)? Many disks will have a range of numbers (say, 7.2-10.8MB/s). Don't confuse this with the "interface transfer rate" which is always a round number (10 or 20 or 40MB/s) and is the speed of the SCSI bus itself.

These numbers will let you do apple-with-apples comparisons of disks. Beware that they will differ on different-size models of the same disk; typically, bigger disks have slower seek times.

Now what does it all mean? Bandwidth first: the `media transfer rate' is how much data you can, under ideal conditions, get off the disk per second. This is a function mostly of rotation speed; the faster the disk rotates, the more data passes under the heads per time unit. This constrains the sustained bandwidth of this disk.

More interestingly, your effective latency is the sum of typical seek time and rotational latency. So for a disk with 8.5ms seek time and 4ms rotational latency, you can expect to spend about 12.5ms between the moment the disk `wants' to read your data and the moment when it actually starts reading it. This is the one number you are trying to make small. Thus, you're looking for a disk with low seek times and high rotation (RPM) rates.

For comparison purposes, the first hard drive I ever bought was a 20MB drive with 65ms seek time and about 3000RPM rotation. A floppy drive has about 100-200ms seek time. A CD-ROM drive can be anywhere between 120ms (fast) and 400ms (slow). The best IDE harddrives have about 10-12ms and 5400 rpm. The best SCSI harddrive I know (the Seagate Cheetah) runs 7.8ms/10000rpm.

Fast, big drives are expensive. Really big drives are very expensive. Really fast drives are pretty expensive. On the other end, really slow, small drives are cheap but not cost effective, because it doesn't cost any less to make the cases, ship the drives, and sell them.

In between is a `sweet spot' where moving in either direction (cheaper or more expensive) will cost you more than you get out of it. The sweet spot moves (towards better value) with time. Right now (July 2001), it's about at 18GB drives, 6ms, 10000rpm, ultra2 SCSI. If you can make the effort, go to your local computer superstore and write down a dozen or so drives they sell `naked'. (If they don't sell at least a dozen hard drives naked, find yourself a better store. Use the Web, Luke!) Plot cost against size, seek and rotational speed, and it will usually become pretty obvious which ones to get for your budget.

Do look for specials in stores; many superstores buy overstock from manufacturers. If this is near the `sweet spot', it's often surprisingly cheaper than comparable drives. Just make sure you understand the warranty procedures.

Note that if you need a lot of capacity, you may be better off with two (or more) drives than a single, bigger one. Not only can it be cheaper but you end up with two separate head assemblies that move independently, which can cut down on latency quite a bit (see below).

Once you've decided which kind of drive(s) you want, you must decide how to distribute them over one or more SCSI buses. Yes, you may want more than one SCSI bus. (My current desktop machine has three.) Essentially, the trick is to make sure that all the disks on one bus, talking at the same time, don't exceed the capacity of that bus. At this time, I can't recommend anything but an Ultra/Wide SCSI controller. This means that the attached SCSI bus can transfer data at up to 40MB/s for an Ultra/Wide disk, 20MB/s for an Ultra/narrow disk, and 10MB/s for a `fast SCSI' disk. These numbers allow you do do your math: an 8MB/s disk will eat an entire bus on its own if it's `fast' (10MB/s). Three 6MB/s ultra/narrow disks fit onto one bus (3x6=18MB/s<20MB/s), but just barely. Two ultra/wide Cheetahs (12.8MB/s) will share an (ultra/wide) bus (25.6<40), but they would collide on an ultra/narrow bus, and any one Cheetah would be bandwidth constrained on a (non-ultra) `fast' bus (12.8 > 10).

If you find that you need two SCSI buses, you can go with `dual channel' versions of many popular SCSI controller cards (including the Adaptec). These are simply two controllers on one card (thus taking only one PCI slot). This is cheaper and more compact than two cards; however, on some motherboards with more than 3 PCI slots, using two cards may be somewhat faster (ask me what a PCI bridge is :-).

How do you deal with slow SCSI devices - CD-ROMS, scanners, tape drives, etc.? If you stick these onto a SCSI bus with fast disks, they will slow down things a bit. You can either accept that (as in ``I hardly ever use my scanner anyway''), or stick them onto a separate SCSI bus off a cheap controller card. Or you can (try to) get an ATA version to stick onto that inevitable IDE interface on your motherboard. The same logic applies to disks you won't normally use, such as removables for data exchange.

If you find yourself at the high end of the bandwidth game, be aware that the theoretical maximum of the PCI bus itself is 132MB/s. That means that a dual ultra/wide SCSI controller (2x40MB/s) can fill more than half of the PCI bus's bandwidth, and it is not advised to add another fast controller to that mix. As it is, your device driver better be well written, or your entire system will melt down (figuratively speaking).

Incidentally, all of the numbers I used are `optimal' bandwidth numbers. The real scoop is usually somewhere between 50-70% of nominal, but things tend to cancel out - the drives don't quite transfer as fast as they might, but the SCSI bus has overhead too, as does the controller card.

Whether you have a single disk or multiple ones, on one or several SCSI buses, you should give careful thought to their partition layout. Given a set of disks and controllers, this is the most crucial performance decision you'll make.

A partition is a contiguous group of sectors on the disk. Partitioning typically starts at the outside and proceeds inwards. All partitions on one disk share a single head assembly. That means that if you try to overlap I/O on the first and last partition of a disk, the heads must move full stroke back and forth over the disk, which can radically increase seek time delays. A partition that is in the middle of a partition stack is likely to have best seek performance, since at worst the heads only have to move half-way to get there (and they're likely to be around the area anyway).

Whenever possible, split partitions that compete onto different disks. For example, /usr and the swap should be on different disks if at all possible (unless you have outrageous amounts of RAM).

Another wrinkle is that most modern disks use `zone sectoring'. The upshot is that outside partitions will have higher bandwidth than inner ones (there is more data under the heads per revolution). So if you need a work area for data streaming (say, a CD-R pre-image to record), it should go on an outside (early numbered) partition of a fast-rotating disk. Conversely, it's a good convention to put rarely-used, performance-uncritical partitions on the inside (last).

Another notes concerns SCSI mode pages. Each (modern) SCSI disk has a small part of its disk (or a dedicated EEPROM) reserved for persistent configuration information. These parameters are called `mode pages', for the mechanism (in the SCSI protocol) for accessing them. Mode page parameters determine, among others, how the disk will write-cache, what forms of error recovery it uses, how its RAM cache is organized, etc. Very few configuration utilities allow access to mode page parameters (I use FWB Toolkit on a Mac - it's simply the best tool I know for that task), and the settings are usually factory preset for, uh, Windows 95 environments with marginal hardware and single-user operation. Particularly the cache organization and disconnect/reconnect pages can make a tremendous difference in actual performance. Unfortunately there's really no easy lunch here - if you set mode page parameters wrong, you can screw up your data in ways you won't notice until months later, so this is definitely `no playing with the pushebuttons' territory.

Ah yes, caches. There are three major points where you could cache I/O buffers: the OS, the SCSI controller, and the on-disk controller. Intelligent OS caching is by far the biggest win, for many reasons. RAM caches on SCSI controller cards are pretty pointless these days; you shouldn't pay extra for them, and experiment with disabling them if you're into tinkering.

RAM caches on the drives themselves are a mixed bag. At moderate size (1-2MB), they are a potential big win for Windows 95/98, because Windows has stupid VM and I/O drivers. If you run a true multi-tasking OS like Linux, having unified RAM caches on the disk is a significant loss, since the overlapping I/O threads kick each other out of the cache, and the disk ends up performing work for nothing.

Most high-performance disks can be reconfigured (using mode page parameters, see above) to have `segmented' caches (sort of like a set-associative memory cache). With that configured properly, the RAM caches can be a moderate win, not because caching is so great on the disk (it's much better in the OS), but because it allows the disk controller more flexibility to reschedule its I/O request queue. You won't really notice it unless you routinely have >2 I/O requests pending at the SCSI level. The conventional wisdom (try it both ways) applies.

And finally I do have to make a disclaimer. Much of the stuff above is shameless simplification. In reality, high-performance SCSI disks are very complicated beasties. They run little mini-operating systems that are most comfortable if they have 10-20 I/O requests pending at the same time. Under those circumstances, the amortized global latencies are much reduced, though any single request may experience longer latencies than if it were the only one pending. The only really valid analysis are stochastic-process models, which we really don't want to get into here. :-)

4.7. Souping Up X Performance

If you care about X performance, be sure you get a graphics card with a dedicated blitter and a high-speed local-bus connection. If it says "AGP" you have this; AGP is a cross-vendor standard for a local bus optimized for graphics.

These cards speed up X in two ways. First, they offload some common screen-painting operations from the main processor onto specialized processors on the card itself. Secondly, by using a local bus, they make it possible to send commands to the card faster than the ISA bus could allow. The combined effect can be eye-poppingly fast screen updates even at very high resolutions.

There's no longer much reason to bother with any of the closed-source X servers like MetroLink or X/Inside. XFree86 now supports most of the high-end cards that used to be the special preserve of the closed-source X versions.

If you're feeling really flush, plump for a 17" or even 20" monitor. The larger size can make a major difference in viewing comfort. Also you'll be set for 1600x1200, which many cards can support these days. In the mean time, the bigger screen will allow you to use fonts in smaller pixel sizes so that your text windows can be larger, giving you a substantial part of the benefit you'd get from higher pixel resolutions.

6.1. Mouse Types

In the beginning, there was only the Mouse Systems 3-button serial mouse; this reported status to a serial port 30 times a second using a 5-byte serial packet encoding now called ``C'' protocol. The Logitech Series 7 and 9 mice were Mouse Systems-compatible. All Unixes that have any mouse support at all understand C-protocol serial mice.

Then Microsoft got into the act. They designed a two-button serial mouse which reports only deltas in a three-byte packet; that is, it sends changes in button status and motion reports only when the mouse is actually moving. This is called `M' protocol. Microsoft sold a lot of mice, so Logitech switched from `C' to `M' — but they added a third button, state changes for which show up in an optional fourth byte. Thus, `M+' protocol, upward-compatible with Microsoft's `M'. Most Unix vendors add support for M+ mice, but it's wise to check (Linux's XFree86 has it).

Bus mice are divided into 8255 and InPort types. These report info continuously at 30 or 60 Hz (though InPort mice have an option for reporting deltas only), and you get interrupts on events and then have to poll hardware ports for details. Bus mice are no longer widely available and there is no good reason to bother with them.

In addition to serial mice and bus mice, there are ``keyboard mice''. On PS/2s there are two identical-looking keyboard ports, labeled (with icons) ``mouse'' & ``keyboard''. Both are 6-pin mini-DINs that look like the regular PC keyboard port only smaller. Physically, the connector goes to the keyboard processor (often an 8042); electrically, it provides TTL levels (serial clock, serial data, ground and +5V); logically, it uses the same protocol as normal serial mice. The same keyboard processor that decodes the keyboard decodes the mouse. PS/2s have this port, many newer motherboards do as well. Confusingly, XFree86 thinks these keyboard mice are bus mice.

All things considered, Unix users are probably best off going with a serial mouse if their motherboard isn't one of the newer ones with a built-in PS/2 mouse. Most current clone motherboards give you two serial ports, so you can still dedicate one to this and still have one for the all-important modem. Not only are the compatibility issues less daunting, but a serial mouse loads the multitasking system less due to interrupt frequency.

6.2. Buying a Mouse

Beware that most clone vendors, being DOS oriented, bundle two-button M-type mice. Thus, you may have to buy your own three-button mouse. Ignore the adspeak about dpi and pick a mouse or trackball that feels good to your hand.

Your humble editor really, really likes the Logitech TrackMarble, an optical trackball that eliminates the chronic roller-fouling problems of the older TrackMan. They're well-supported by XFree86 (type MouseMan), so any Linux or BSD will accept them.

7.1. The Simple Answer

The modem market has stabilized, with a clear leader at a reasonable price. If you can afford it ($270 in August 2001) go buy a U.S. Robotics V.Everything external. You can then know that you've got the best and skip the rest of this section.

7.2. Overview of the Modem Market

The modem market is like consumer electronics (and unlike the computer market as a whole) in that price is a very poor predictor of performance. For ordinary file transfers, some $50 modems are better than some $150 modems. Paying top dollar mainly buys you better tolerance of poor connections and better performance at heavy-duty bi-directional transfers (such as you would generate, for exmaple, using SLIP or PPP over a leased line to an Internet provider).

In today's market, the typical modem does a nominal 56kbps -- V.90 plus V.29 or V.17 fax transmission and reception (over plain old phone lines you won't get more than 53K of that). You don't see much in the way of slow/cheap to fast/expensive product ranges within a single brand, because competition is fierce and for many modem board designs (those featuring DSP (Digital Signal Processor) chips run by a program in ROM) adding a new protocol is basically a software change.

For much more information on high-speed modems for Unix, see The Linux/Modem Compatibility Knowledge Base. The page is aimed at Linux users, but the advice is general to any non-Windows OS.

7.3. Internal vs. External

Most modems come in two packagings: internal, designed to fit in a PC card slot, and external, with its own case, power supply, and front-panel lights. Typically you'll pay $20 to $30 more for an external modem than you will for the internal equivalent. You'll also need a serial port to connect your external modem to.

Pay that premium — being able to see the blinkenlights on the external ones will help you understand and recover from pathological situations. For example, if your Unix system is prone to ``screaming-tty'' syndrome, you'll quickly learn to recognize the pattern of flickers that goes with it. Punch the hangup/reset button on an external modem and you're done — whereas with an internal modem, you have to go root and flounder around killing processes and maybe cold-boot the machine just to reset the card.

See Rick's Rants for extended discussion of this point.

7.4. Pitfalls to Avoid

If the abbreviation ``RPI'' occurs anywhere on the box, don't even consider buying the modem. RPI (Rockwell Protocol Interface) is a proprietary ``standard'' that allows modem makers to save a few bucks at your expense by using a cheap-jack Rockwell chipset that doesn't do error correction. Instead, it hands the job off to a modem driver which (on a Unix machine) you will not have.

Also avoid anything called a ``Windows Modem'' or ``WinModem'', ``HCF, or ``HSP''; these lobotomized pieces of crap require a Windows DLL to run. They will eat up to 25% of your processor clocks during transfers, and hog high-priority interrupts (causing your machine to stall under Windows even if your processor still has spare cycles).

Multi-user Unix eats enough processor clocks that you want to be sure of good hardware buffering in your UART — that is, enough of it to avoid losing characters between modem and PC if the OS is a bit slow responding to an interrupt (V.42bis in hardware won't detect this!). This means you want a 16550A or equivalent UART. If you're using an external modem, this is an issue about your serial-port board(s). If you're using an internal modem, the UART is on the modem card itself. So, when buying internal modems, ask what the UART type is. If the vendor says 16540, lose them.

Many fax modems come with bundled MS-DOS fax software that is at best useless under Unix, and at worst a software kluge to cover inadequate hardware. Avoid these bundles and buy a bare modem — it's cheaper, and lowers the likelihood that something vital to your communications needs has been left out of the hardware.

Avoid ``Class 1'' and ``Class 2'' modems. Look for ``Class 2.0'' for the full EIA-standard command set.

7.5. What to Buy

What you want, these days, is a V.90 modem. V.90, formally, is the ITU-T recommendation for asymmetric data signalling rates of up to 56Kbps in the direction of a digitally connected server to a capable client, and up to 33.6Kbps in the direction of the client to the server.

The technology is based on eliminating restrictions imposed by the conversion of analog signals to digital form in the downstream data path (server -> client). Data flow in the server to client direction does not occur in the form of a modulated carrier, it is instead sent as binary numbers representative of 256 possible voltage levels. The reason for the asymmetrical send/receive rates is because in the direction from the client to the server it is not possible to use a digital coding scheme and make it work as well as V.34 does, thus V.34 is used instead. It isn't possible because the telco's line card has a codec that is a much better digital level changer for the transmit direction than it is for the receive direction. The codec used in the customer's modem is, in that respect, somewhat more sophisticated and was designed to work as a fairly good level changer in the receive direction (which the telco's codec was not designed to do).

Note: Achievable bit rates are limited to less than 56kbps in the United States by FCC regulations that limit power input to the network.

Beyond the V series, the telcos are pushing a technology called ADSL (Asymmetric Digital Subscriber Line). This method converts existing twisted-pair telephone lines into access paths for multimedia and high speed data communications. ADSL transmits more than 6Mbps to a subscriber, and as much as 640 kbps more in both directions.

An ADSL circuit connects an ADSL modem on each end of a twisted-pair phone line, creating three information channels; a high speed downstream channel, a medium speed duplex channel, and a POTS (Plain Old Telephone Service) channel. The POTS channel is split off from the digital modem by filters, thus guaranteeing uninterrupted POTS, even if ADSL fails. The high speed channel ranges from 1.5 to 6.1 Mbps, while duplex rates range from 16 to 640 kbps. Each channel can be sub-multiplexed to form multiple, lower rate channels.

7.6. Fax Modems

Fax capability is included with most all modems these days; it's cheap for manufacturers, being basically a pure software add-on. The CCITT also sets fax protocol standards. Terms to know:

V.17 isn't common yet, but it doesn't usually cost extra over V.29 when you find it.

There's a separate series of standards for software control of fax modems over the serial line maintained by the Electronics Industry Association and friends. These are:

Class 1 — base EIA standard for fax control as extensions to the Hayes AT command set.

Class 2.0 — enhanced EIA standard including compression, error correction, station ID and other features.

Class 2 — marketroidian for anything between Class 1 and Class 2.0. Different ``Class 2'' modems implement different draft subsets of the 2.0 standard, so ``Class 2'' fax software won't necessarily drive any given ``Class 2'' modem.

There's also a proprietary Intel "standard" called CAS, Communicating Applications Specification. Ignore it; only Intel products support it.

The GNU toolset includes a freeware fax transmission and reception toolset, Netfax. Look for it at prep.ai.mit.edu:pub/gnu/fax-*. It says it requires a modem conforming to the ``Class 2'' control standard, but you'd be safest getting a 2.0-conformant modem for reasons explained above. Netfax also requires GNU Ghostscript to do Postscript handling for it.

8.1. CD-ROM Drives

Standard CD-ROMs hold about 650 megabytes of read-only data in a format called ISO-9660 (formerly ``High Sierra''). All current Unixes now support these devices. In fact, most Unix and Linux software is now distributed on ISO-9660 CD-ROM, a cheaper and better method than the QIC tapes we used to use.

CD-ROM drives may be driven through SCSI, through enhanced IDE (ATAPI), or through a proprietary interface card (like the Mitsumi and Sony interfaces). Unix support for CD-ROMs is usually through SCSI drivers (exception: BSDI/386 supports the Mitsumi interface; Linux supports the Mitsumi and Sony interfaces). I recommend that you avoid the proprietary cards; they will effectively cost you money when you need to upgrade. Besides, street prices for SCSI CD-ROMS have dropped below $200.

(A few external CD-ROMs come with a parallel-port interface. Avoid these; they tend to have very slow transfer rates.)

Any CD-ROM you buy should be at least a ``double-spin'' drive meeting the MPC2 (Multimedia PC) standard of a 300K/sec transfer rate when reading. digital data. The older single-speed drives, which only supported the 150K/sec rate Red Book standard for audio CDs, are obsolete. The lowest speed you can buy these days is 4X (600K/sec). 6X, 8X, 10X, 12X, 24X, 32X, 40X, and 56X are available.

The next level up in CD hardware standards is CD-ROM XA. So far, drives that support XA are few and expensive. It's not yet in wide use in the DOS/Windows world, and I don't know of any Unix support for it, either in closed- or open-source code.

CD-ROM access times about 280ms for high-end double-speed drives (to put this in perspective, it's about 30 times slower than a typical 9ms hard disk, but considerably faster than a tape). Accordingly, modern 32X drives are about half the speed of a hard drive.

Most CD-ROMS will include a headphone jack so you can play audio CDs on them. Better-quality ones will also include two RCA jacks for use with speakers. Another feature to look for is a drive door or seal that protects the drive head from dust.

CD-ROM formats are still an area of some confusion. A slight enhancement of the original ``High Sierra'' CD-ROM filesystem format (designed for use with DOS, and limited to DOS's 8+3 file-naming convention) has been standardized as ISO-9660.

There is a de-facto Unix standard called `Rock Ridge' pioneered by the Sun User's Group shareware CD-ROMs. This is a way of putting an extra layer of indirection on an ISO-9660 layout that preserves Unix's long dual-case filenames. Some Unixes (notably Linux, netBSD, freeBSD and BSD/OS) can mount Rock Ridge filesystems.

More much more detail on CD-ROMs, CD-ROM standards and how to buy drives is available in the alt.cdrom FAQ, available for FTP as cdrom.com:/cdrom/faq. It is also archived in the news.answers tree at rtfm. This FAQ includes comparison tables tables of numerous drive types, CD-ROM sources, and ordering information.

8.2. Sound Cards and Speakers

Look for the following features as a minimum in your sound card:

• 16-bit sampling (for 65536 dynamic levels rather than 256).

• Mono and stereo support.

• Full-duplex mode.

• Sampling rates ranging fron 8K/sec (voice-quality) through 11KHz (AM-radio quality), 22KHz (FM-radio quality) and standard audio (44.1KHz).

• MIDI interface via a standard 15-pin D-shell connector.

• RCA output jacks for headphones or speakers.

• A microphone jack for sound input.

The most important feature is your sound card is what type of MIDI synthesis it has. All current sound cards have some sort of MIDI (musical instrument digital interface) compliance, but if nothing else, make sure your card supports the General MIDI standard. MIDI is just a set of commands issued by the application that tells the sound card which instrument to play, at what note, and for what duration.

Older and cheaper cards use FM synthesis. This synthesis uses a combination of sine waves to imitate the sounds of the different instruments. The result is like the sound tracks of most computer games sold a few years ago; imitation music with an arcade-like sound.

The method used by most modern sound boards is called wave table synthesis. In this method, digitized samples of actual instrument sounds are used as templates for the tones generated by the MIDI commands. Wave table cards vary in the quantity and quality of samples; one figure of merit often quoted is the wave table ROM size (often 4MB or 8MB). Also some boards have wavetable RAM that can store samples loaded from a disk.

Soundcards with DSP (Digital Signal Processing) can perform synthesis effects on board, relieving the CPU for other tasks. Some DSP chips are even software-programmable. Some high-end cards even include 3D sound effects. Whether the system used is SRS (Sound Retrieval System), Q-Sound, or Spatializer, it is designed to improve the perceived stereo effect of your speakers. These 3D effects work by delaying the timing of certain portions of the audio signal so that different frequencies hit your ear at slightly different times. The downside is that some of the cards equipped with 3D sound add a noticeable amount of noise to the card's output.

If you play a lot of computer games, you'll need to pay attention to compatibility. DOS games are written almost exclusively for the Creative Labs specification; you will need a card that is 100% Sound Blaster compatible. Many vendors do not license the Creative Labs specification but claim that their cards are 100% game compatible. This means that the sound will work, but not all sounds that you hear will be the ones that the game programmers intended. If you play many DOS games, it would be best to buy a Sound Blaster and save yourself a migraine.

Lastly, try to avoid sound cards with built-in amplifiers that are more powerful than 4 watts/channel. Sound cards that have more powerful amplifiers are said to have the problem of adding noise to the card's output. Use powered speakers with a 4 watt/channel card to solve this problem. Most cards are equipped with 4 watts/channel anyway. Wavetable cards are so inexpensive these days that it's almost worth their additional cost over a regular FM synthesis card. If you decide to settle for an FM card, make sure that there is a daughterboard made for the card that will let you upgrade to wavetable synthesis. In some cases, however, the wavetable card is cheaper than buying an FM card and then deciding that you want the wavetable upgrade. If you do decide on the wavetable as your card of choice, PC Magazine rated the best MIDI wavetables (MIDI being the most important feature in my opinion) the Media Vision Premium 3-D, Media Vision Pro 3-D, Creative Labs Sound Blaster AWE32, and the Turtle Beach Monterey (although there are value editions of the Sound Blaster 32 that have fewer ROM instrument samples but maintain the superior MIDI wavetable synthesis).

Linux, includes drivers for the SoundBlaster series and many other boards including the PAS Adlib, the Gravis Ultrasound, and the ProAudioSpectrum. Many of these are included in the 2.0 and newer kernels. See the OSS/Free home page for details

Warning: some sound cards require a specific CD-ROM type! Avoid these.

In speakers, look for a magnetically-shielded enclosure with volume, bass and treble controls. Some speakers run off the card's 4-watt signal; others are ``self-powered'', using batteries or a separate power supply. Your major buying choice is which one of these options to pursue.

Generally speaking, you'll get better value if you buy your sound card, CD-ROM, and speakers not separately but as a bundled ``multimedia upgrade kit'' (even though Unix users will usually have to throw out the bundled software).

One final, important tip: that audio cable from your CD-ROM back to the sound card is used only when you play audio CD-ROMs through your speakers. Software-generated sound goes through the system bus, so you can play ``Doom'' with sound even if your sound board won't accept the audio cable connector.

10.1. When to Buy

It used to be that good configurations for Unix were what the market called `server' machines, with beefed-up I/O subsystems and fast buses. No longer; today's `servers' are monster boxes with multiple power supplies and processors, gigabytes of memory, and industrial-grade air cooling -- they're not really suitable as personal machines. A typical SCSI desktop workstation is as much as you'll need.

Prices keep dropping, so there's a temptation to wait forever to buy. A good way to cope with this is to configure your system on paper, get a couple of initial estimates, then set a trigger price, below the lowest one, at what you're willing to pay. Then watch and wait. When the configuration cost hits your trigger price, place your order.

The advantage of this method is that it requires you to settle in your mind, well in advance, what you're willing to pay for what you're getting. That way, you'll buy at the earliest time you should, and won't stress too much out afterwards as it depreciates.

Before you shop, do your homework. Publications like "Computer Shopper" (and their web site at http://www.computershopper.com) are invaluable for helping you get a feel for prices and what clonemakers are doing. Another excellent site is ComputerESP.

10.2. Where to Buy

The most important where-to-buy advice is negative. Do not go to a traditional, business-oriented storefront dealership. Their overheads are high. So are their prices.

Especially, run -- do not walk -- away from any outfit that trumpets `business solutions'. This is marketing code for the kind of place that will justify a heavy price premium by promising after-sale service and training which, nine times out of ten, will turn out to be nonexistent or incompetent. Sure, they'll give you plush carpeting and a firm handshake from a guy with too many teeth and an expensive watch -- but did you really want to pay for that?

There are two major alternatives to storefront dealerships and one minor one. The major ones are mail order and computer superstores. The minor one is computer fairs.

10.3. Computer Fairs

I used to be a big fan of hole-in-the-wall stores run by immigrants from the other side of the International Date Line, but most of those places have been driven out of the regular retail game by the superstores. The only place you find diaspora Chinese and Indians selling cheap PCs over the counter anymore is at computer fairs. (Usually they're doing it to publicize a mail-order business.)

You can find good `loss-leader' deals on individual parts at these fairs (they're especially good places to buy disk drives cheap). But I call them a minor alternative because it's hard to get a custom SCSI-based configuration tuned for Unix built for you at a fair. So you end up, effectively, back in the mail-order or Web channel.

10.4. Mail Order

Direct-mail or Web buying makes a lot of sense today for anyone with more technical savvy than J. Random Luser in a suit. Even from no-name mail-order houses, parts and system quality tend to be high and consistent, so conventional dealerships don't really have much more to offer than a warm fuzzy feeling. Furthermore, competition has become so intense that even mail-order vendors today have to offer not just lower prices than ever before but warranty and support policies of a depth that would have seemed incredible a few years back. For example, many bundle a year of on-site hardware support with their medium- and high-end "business" configurations for a very low premium over the bare hardware.

Note, however, that assembling a system yourself out of mail-order parts is not likely to save you money over dealing with the mail-order systems houses. You can't buy parts at the volume they do; the discounts they command are bigger than the premiums reflected in their prices. The lack of any system warranty or support can also be a problem even if you're expert enough to do the integration yourself — because you also assume all the risk of defective parts and integration problems.

Cruise through `Computer Shopper' and similar monthly ad compendia. Even if you decide to go with a conventional dealer, this will tell you what their premiums look like. These days most mail-order outfits with enough cash to advertise on glossy paper are good risks.

Watch out for dealers (Spectrum Trading for one) who charge ridiculous shipping fees. One of our spies reports he bought a hotswappable hard disc drive tray that weighed about 3 lbs. and cost $250 and they charged $25 to ship it UPS groud.

Don't forget that (most places) you can avoid sales tax by buying from an out-of-state mail-order outfit, and save yourself 6-8% depending on where you live. If you live near a state line, buying from a local outfit you can often win, quite legally, by having the stuff shipped to a friend or relative just over it. Best of all is a buddy with a state-registered dealer number; these aren't very hard to get and confer not just exemption from sales tax but (often) whopping discounts from the vendors. Hand him a dollar afterwards to make it legal.

(Note: I have been advised that you shouldn't try the latter tactic in Florida -- they are notoriously tough on "resale license" holders).

(Note II: The Supreme Court has ruled that states may not tax out-of-state businesses under existing law, but left the way open for Congress to pass enabling legislation. Let's hope the mail-order industry has good lobbyists.)

10.5. Computer Superstores

Big chain superstores like CompUSA and Circuit City give you a reasonable alternative to mail order. And there are good reasons to explore it -- these stores buy and sell at volumes that allow them to offer prices not far above mail-order. (They make back a lot of their margin on computer games and small accessories like mouse pads, cables, and floppy disks.)

One thing you should not buy mail-order if you can avoid it is a monitor. Monitors are subject to substantial quality variations even within the same make and model. Also, one good bump during shipping can twist the yoke on a monitor so the image is tilted with respect to the bezel. So buy your monitor face-to-face, picking the best out of three or four.

Another good argument for buying at a superstore is that you may have to pay return postage if you ship a mail-order system back. On a big, heavy system, this can eat your initial price savings.

The only major problem with superstores is that the salespeople who staff them aren't very bright or very clueful (it's a sort of Darwinian reverse-selection effect; these are the guys who are fascinated by computer technology but not smart enough to be techies). Most of them don't know from Linux and are likely to push things like two-button mice and (worse!) controllerless modems, that you can't use. Use caution and check your system manifest.

But if you shop carefully and don't fall for one of their name-brand "prestige" systems, you can get prices comparable to mail order with the comfort of knowing there's a trouble desk you can drive back to in a pinch. (Also, you can see your monitor before you buy!)

10.6. Other Buying Tips

You can often get out of paying tax just by paying cash, especially at computer shows. You can always say you're going to ship the equipment out of the state.

A lot of vendors bundle DOS or Windows and variable amounts of apps with their hardware. If you tell them to lose all this useless cruft they may shave $50 or $100 off the system price.

11.1. Minimum Warranty Provisions

The weakest guarantee you should settle for in the mail-order market should include:

• 72-hour burn-in to avoid that sudden infant death syndrome. (Also, try to find out if they do a power-cycling test and how many repeats they do; this stresses the hardware much more than steady burn-in.)

• 30 day money-back guarantee. Watch out for fine print that weakens this with a restocking fee or limits it with exclusions.

• 1 year parts and labor guarantee (some vendors give 2 years).

• 1 year of 800 number tech support (many vendors give lifetime support).

Additionally, many vendors offer a year of on-site service free. You should find out who they contract the service to. Also be sure the free service coverage area includes your site; some unscrupulous vendors weasel their way out with "some locations pay extra", which translates roughly to "through the nose if you're further away than our parking lot".

If you're buying store-front, find out what they'll guarantee beyond the above. If the answer is "nothing", go somewhere else.

11.2. Documentation

Ask your potential suppliers what kind and volume of documentation they supply with your hardware. You should get, at minimum, operations manuals for the motherboard and each card or peripheral; also an IRQ list. Skimpiness in this area is a valuable clue that they may be using no-name parts from Upper Baluchistan, which is not necessarily a red flag in itself but should prompt you to ask more questions.

11.3. A System Quality Checklist

There are various cost-cutting tactics a vendor can use which bring down the system's overall quality. Here are some good questions to ask:

• If you're buying a factory-configured system, does it have FCC certification? While it's not necessarily the case that a non-certified system is going to spew a lot of radio-frequency interference, certification is legally required — and becoming more important as clock frequencies climb. Lack of that sticker may indicate a fly-by-night vendor, or at least one in danger of being raided and shut down! (For further discussion, see the section on Radio Frequency Interference above.)

• Are the internal cable connectors keyed, so they can't be put in upside down? This doesn't matter if you'll never, ever ever need to upgrade or service your system. Otherwise, it's pretty important; and, vendors who fluff this detail may be quietly cutting other corners.

12.1. Tricks and Traps in Mail-Order Warranties

Reading mail-order warranties is an art in itself. A few tips:

Beware the deadly modifier ``manufacturer's'' on a warranty; this means you have to go back to the equipment's original manufacturer in case of problems and can't get satisfaction from the mail-order house. Also, manufacturer's warranties run from the date they ship; by the time the mail-order house assembles and ships your system, it may have run out!

Watch for the equally deadly ``We do not guarantee compatibility''. This gotcha on a component vendor's ad means you may not be able to return, say, a video card that fails to work with your motherboard.

Another dangerous phrase is ``We reserve the right to substitute equivalent items''. This means that instead of getting the high-quality name-brand parts advertised in the configuration you just ordered, you may get those no-name parts from Upper Baluchistan — theoretically equivalent according to the spec sheets, but perhaps more likely to die the day after the warranty expires. Substitution can be interpreted as ``bait and switch'', so most vendors are scared of getting called on this. Very few will hold their position if you press the matter.

Another red flag: ``Only warranted in supported environments''. This may mean they won't honor a warranty on a non-DOS system at all, or it may mean they'll insist on installing the Unix on disk themselves.

One absolute show-stopper is the phrase ``All sales are final''. This means you have no options if a part doesn't work. Avoid any company with this policy.

12.2. Special Questions to Ask Mail-Order Vendors Before Buying

• Does the vendor have the part or system presently in stock? Mail order companies tend to run with very lean inventories; if they don't have your item in stock, delivery may take longer. Possibly much longer.

• Does the vendor pay for shipping? What's the delivery wait?

• If you need to return your system, is there a restocking fee? and will the vendor cover the return freight? Knowing the restocking fee can be particularly important, as they make keep you from getting real satisfaction on a bad major part. Avoid dealing with anyone who quotes more than a 15% restocking fee — and it's a good idea, if possible, to avoid any dealer who charges a restocking fee at all.

Warranties are tricky. There are companies whose warranties are invalidated by opening the case. Some of those companies sell upgradeable systems, but only authorized service centers can do upgrades without invalidating the warranty. Sometimes a system is purchased with the warranty already invalidated. There are vendors who buy minimal systems and upgrade them with cheap RAM and/or disk drives. If the vendor is not an authorized service center, the manufacturer's warranty is invalidated. The only recourse in case of a problem is the vendor's warranty. So beware!

12.3. Payment Method

It's a good idea to pay with AmEx or Visa or MasterCard; that way you can stop payment if you get a lemon, and may benefit from a buyer-protection plan using the credit card company's clout (not all cards offer buyer-protection plans, and some that do have restrictions which may be applicable). However, watch for phrases like ``Credit card surcharges apply'' or ``All prices reflect 3% cash discount'' which mean you're going to get socked extra if you pay by card.

Note that many credit-card companies have clauses in their standard contracts forbidding such surcharges. You can (and should) report such practices to your credit-card issuer. If you already paid the surcharge, they will usually see to it that it is returned to you. Credit-card companies will often stop dealing with businesses that repeat such behavior.

12.4. Which Clone Vendors to Talk To

16.1. Older Memory Types

The last-generation memory package was a SIMM (Single In-Line Memory Module). We include some information on these here for people upgrading old systems.

SIMMs have 9 bits wide (8 bits of data, one for parity) and either 1, 4, or 16 megabytes in size. The 16s are slightly cheaper per megabyte.

There are two physical SIMM sizes; 30-pin and 72-pin. The 30-pin size is long obsolete; most SIMM-using motherboards have exclusively 72-pin sockets.

Also, your memory may or may not have on-board parity check bits. Memory with parity is specified as ``x36'' (4 check bits per word); memory without is ``x32''. Some motherboards require parity memory; some (including those based on the Neptune chipset) can use parity if present; some (including those based on the Triton chipset) cannot use it. Parity memory is a good idea for Unix machines, which stress the hardware more than other OSs.

You'll hear a lot about ``EDO'' (Extended Data Out) RAM. This stuff is 5%-10% more expensive, but that pays for a significant performance boost in ``burst mode'' (that is, when fetching large contiguous sections of memory into the cache). Most current motherboards (including those based on the Triton chipset) can use EDO but don't require it.

You do not want to buy 1MB SIMMs any more, unless you're filling out memory in an existing system configured with 1MB SIMMs. The trouble with these (besides the fact that four of them are more expensive than a 4MB chip) is that, due to interleaving constraints, you'll probably have to throw them away if you ever add more memory.

These interleaving constraints differ from motherboard to motherboard. Typically, memory sockets are set up in banks of either 2 or 4, with a requirement that you fill all sockets of each bank with the same size of chip. Also, if you mix chip sizes in the banks, the sizes have to decrease smoothly (so there won't be gaps in the memory address space).

So your possible configurations look like the following table. All of these are possible on 4-bank, 16-slot boards; only the asterisked ones are possible on 2-bank, 8-slot boards. All the capacities above 64MB require that the board accept 16MB SIMMs. All current 2x8 boards accept these.

        * 4MB  = 1-1-1-1
        * 8MB  = 1-1-1-1 + 1-1-1-1
          12MB = 1-1-1-1 + 1-1-1-1 + 1-1-1-1
          16MB = 1-1-1-1 + 1-1-1-1 + 1-1-1-1 + 1-1-1-1
        
        * 16MB = 4-4-4-4
        * 20MB = 4-4-4-4 + 1-1-1-1
          24MB = 4-4-4-4 + 1-1-1-1 + 1-1-1-1
          28MB = 4-4-4-4 + 1-1-1-1 + 1-1-1-1 + 1-1-1-1
        
        * 32MB = 4-4-4-4 + 4-4-4-4
          36MB = 4-4-4-4 + 4-4-4-4 + 1-1-1-1
          40MB = 4-4-4-4 + 4-4-4-4 + 1-1-1-1 + 1-1-1-1
        
          48MB = 4-4-4-4 + 4-4-4-4 + 4-4-4-4
          52MB = 4-4-4-4 + 4-4-4-4 + 4-4-4-4 + 1-1-1-1
        
          64MB = 4-4-4-4 + 4-4-4-4 + 4-4-4-4 + 4-4-4-4
        
        * 64MB = 16-16-16-16
        * 68MB = 16-16-16-16 + 1-1-1-1
          72MB = 16-16-16-16 + 1-1-1-1 + 1-1-1-1
          76MB = 16-16-16-16 + 1-1-1-1 + 1-1-1-1 + 1-1-1-1
        
        * 80MB = 16-16-16-16 + 4-4-4-4
          84MB = 16-16-16-16 + 4-4-4-4 + 1-1-1-1
          88MB = 16-16-16-16 + 4-4-4-4 + 1-1-1-1 + 1-1-1-1
        
          100MB = 16-16-16-16 + 4-4-4-4 + 4-4-4-4 + 1-1-1-1
          112MB = 16-16-16-16 + 4-4-4-4 + 4-4-4-4 + 4-4-4-4
        
        * 128MB = 16-16-16-16 + 16-16-16-16
          132MB = 16-16-16-16 + 16-16-16-16 + 1-1-1-1
          136MB = 16-16-16-16 + 16-16-16-16 + 1-1-1-1 + 1-1-1-1
        
          144MB = 16-16-16-16 + 16-16-16-16 + 4-4-4-4
          148MB = 16-16-16-16 + 16-16-16-16 + 4-4-4-4 + 1-1-1-1
          160MB = 16-16-16-16 + 16-16-16-16 + 4-4-4-4 + 4-4-4-4
        
          192MB = 16-16-16-16 + 16-16-16-16 + 16-16-16-16
          196MB = 16-16-16-16 + 16-16-16-16 + 16-16-16-16 + 1-1-1-1
        
          208MB = 16-16-16-16 + 16-16-16-16 + 16-16-16-16 + 4-4-4-4
          256MB = 16-16-16-16 + 16-16-16-16 + 16-16-16-16 + 16-16-16-16

As you can see from the table, most configurations with 1MB chips in them can't be upgraded without throwing away chips.

Here's some more about memory packaging.

There are also adapter cards that will accept 30-pin SIMM modules and plug into a 72-pin socket; make sure you've the room for these if you intend to try them, as they protrude quite a bit above the board. There are also companies that will, for modest fees, desolder the memory chips from 30-pin SIMMs and wave solder them onto 72-pin boards.

(Thanks to Dave Ihnat for the glossary.)