01/27/2001

More reviews of reviews. I liked how yesterday went, and I found some really good stuff yesterday, so let’s continue on and see what’s good and why.

2001 Upgrade Guide (Ace’s Hardware)

This is an outstanding upgrade guide, working from the assumption that you have an older system (a K6-2 or Celeron with a TNT2 board, which is a pretty common setup), then they test a number of upgrades so you can see what makes a difference. Unfortunately these upgrade candidates already have a modern hard disk and sound card, so they don’t closely simulate a real-world system, but they do isolate the components, so while these upgraded systems will outperform yours, you can see precisely what effect upgrading the video card will have.

For example, you can see right away from their graphs that replacing a K6-2’s TNT2 video card with a GeForce 2 GTS will only improve Half-Life frame rates slightly (up to 25.5 from 22.1), while trading up to a Duron 850 while keeping all the same peripherals increases rates to 51.8 from 22.1. How valuable is that information? I found a GTS card for $229. The same place has a Duron 850/Gigabyte 7ZX-1 bundle for $222. The upgrades cost the same amount, yet one of them increases performance significantly while the other just barely helps. It’s the difference between throwing away $240 and spending $235 wisely (after shipping).

The other great thing about this guide is that it tests more than just first-person shooters. For FPS, DDR gives marginal improvements indeed, but for other types of games, its improvement can be immense. Mercedes-Benz Truck Racing and Formula One 2000, for instance, are faster with a DDR-equipped Duron 850 than it is with a PC133-equipped Athlon 1100.

This guide shows when a GHz+ CPU and new memory technology makes sense, and when it doesn’t, letting you decide when it makes sense to buy the latest and greatest.

Overall: great methodology, nice balance of real-world tests (assuming gaming’s your thang, which it probably is if you read this stuff, since you won’t see much difference between a Celeron 667 and a 1.2 GHz Athlon for office apps). A lot of work goes into guides like this, but it’s worth it. Maybe someday articles like this will be the norm on the hardware sites, rather than the exception. Hey, I can dream, can’t I?

VIA Apollo Pro 266 (THG)

This is an analysis piece combined with a preview of VIA’s Apollo Pro 266 chipset. Good explanation of PC architecture for one who doesn’t understand what the north bridge and south bridge are, plus the benchmarks are using boards you can actually buy, rather than reference designs.

Tom Pabst takes his usual swipes at Rambus, and points out that the Pentium III isn’t really able to take advantage of DDR, as evidenced by its similar performance to Rambus- and PC133-equipped systems. Pabst concludes with an assertion that a DDR Pentium 4 chipset would prove how terrible Rambus really is, since the bottleneck with DDR seems to be the CPU, rather than the memory itself. Unfortuantely, he doesn’t provide anything at all to back up this claim, so he comes off as an anti-Rambus bigot. Has he seen a P4 run with DDR? Maybe he’s under NDA, but if he is, he can at least say, “I can’t tell you why I know this, but DDR chipsets for the P4 will prove how worthless Rambus is,” and it would be better than what he wrote. But his speculation of DDR performance with the P4 and how it will compare is no more valuable than yours.

This article does give the useful information that DDR on the Pentium III probably isn’t worth the bother.

Value Biz PC Guide (Sharky Extreme)

Unusual for hardware sites, good focus on what’s necessary for business. No benchmarks; I’d have liked to have seen illustrations of why CPU speed isn’t as important as, say, disk speed, for business apps. Hardware recommendations are solid, and I’m happy to see they don’t assume businesses overclock. They don’t. I disagree with the $100 CD-R recommendation; you’re better off with a Plextor drive with Burn-Proof, especially since such a drive will allow you to multitask. Since time is money, businesses can’t afford to waste time burning coasters. If a slower, cheaper CPU is necessary in order to afford a better CD-R, then so be it.

Some discussion of when SCSI would be appropriate on the desktop also would have been nice, as SCSI does have its place in the office.

But overall, this is a solid guide. By blindly following its advice, you’ll build a better PC than you’ll get from many of the direct PC vendors.

Internet Connection Sharing (Dan’s Data)

Nice, down-to-earth, and pretty thorough overview of what it takes to share an Internet connection whose primary target is people who are less ambitious than me–an old 386 or 486 running Linux isn’t among the options he presents. I guess he could have titled it “ICS for the Rest of Us.”

This is thorough without getting too bogged down in particulars, and it’s cross-referenced with an outstanding Networking 101 piece by the same author, and weird jargon is cross-referenced with an online dictionary. Some reviews of the various options would be nice, but he gives a good thumbnail sketch of each option’s advantages and drawbacks. The author, Dan Rutter, is a mainstream computer journalist in Australia who seems to have a very high standard for his work.

Definitely bookmark his networking piece, http://www.dansdata.com/network.htm , and if you keep a notebook, print out a copy to put there as well, as it’s an outstanding overview that answers most of the common networking questions like the difference between a hub and a switch. You may find yourself referring back to this one as well, but it’s more specialized and as such, not as generally useful.

His other stuff is useful, well-written, and downright entertaining. Few computer writers are fun to read. Dan Rutter usually is. Many people consider Ace’s Hardware the best of the hardware sites, but I really think Dan’s Data gives Ace’s a big-time run for the money.

01/23/2001

Mailbag:

More Networking

What’s going on with memory prices? Every time I say they’re stable, they drop again. I’m not going to say anything about current prices, except they’re low. Face it: I remember five years ago, paying $48 for an 8-meg stick, and I felt like I was stealing it. Kingston memory for $6 a meg! Unbelievable!

I told Dan Bowman on Sunday that you can get a 128-meg PC133 Kingston module at Outpost.com for $59 with a $20 mail-in rebate. Then yesterday he sends me word that I can get a 128-meg PNY PC133 stick from globalcomputer.com for $49. No rebate hassles whatsoever, and plenty of stock. So $6/meg has become $.31/meg. Prices may stabilize there, or they may free-fall some more.

What happened? Overproduction. Millions of chips were produced for millions of computers that didn’t sell over Christmas, which is supposed to be the heaviest buying period of the year. Not a whole lot of upgrades were bought either. And now, with demand for Rambus increasing a little and DDR looming overhead like the Enola Gay, they’re stuck with a bunch of inventory that’s living on borrowed time. Gotta move it, because demand’s moving elsewhere. There’ll be demand for SDRAM for many years to come (just as there’s still some demand for EDO DRAM today), but its days as the memory everybody wants are about to come to a close.

So as long as you have some use for SDRAM, this is a great time to buy. But keep in mind that the stuff you buy now probably won’t move with you to your next PC. A current PC with 384 MB of PC133 SDRAM will be useful for many years to come, true, but next year when you buy a motherboard that takes DDR or Rambus, you’ll have to buy new memory again, so it makes absolutely no sense to hoard this stuff.

So should you buy? Windows 9x sees diminishing returns beyond 128 MB of RAM, unless you’re playing with RAM disks. Windows 2000 really likes 256 MB of RAM, but for the things most people do, there’s little point in going past that. Of all the OSs I use right now, Linux does the best job of finding a use for such a large amount of memory. So if you’re below any of those thresholds, sure, buy. But if you’re there already, you’re better off banking that money until the time comes for your next major upgrade.

But if you are buying, let me reiterate: Get the good stuff. I had a conversation with someone on a message board today. He asked why, if 95% of all memory chips are fine, it makes sense to pay more for a brand name. I pointed out to him that with 8-16 chips per module, a 95% rate means you have a 25-50 percent chance of a bad module, since it just takes one bad cell in one chip to make the module unreliable. It’s much better to get A-grade chips, which have a .1% defect rate, and buy from a name brand vendor, who will in all likelihood do their own testing and lower the defect rate another order of magnitude. To me, knowing that I won’t have problems attributable to bad memory is definitely worth the few bucks. Even the bottom-feeders aren’t beating that Kingston price by much, and the shipping will make the cheap, nearly worthless memory cost more than the good stuff.

Tracking down memory problems is a real pain, unless you’ve got a professional-quality memory tester. I do. Still, verifying a memory problem and then isolating it to a single stick can take hours. I have all the facilities necessary to let me get away with buying the cheap stuff and I won’t do it. That should tell you something. Buying generic memory isn’t like buying generic socks or generic spaghetti. In memory, brand is a lot more than status.

Partition Magic. I tried unsuccessfully last night to track down a copy of Partition Magic 6 so I can revise the article on multi-booting Windows 98 and Windows Me that won’t go in the March issue of Computer Shopper UK. It’ll be in the April issue instead. I also had to deal with some personal issues. It’s not like my whole world’s upside down–it’s not–but a pretty important part of it is right now.

Mailbag:

More Networking

Boot multiple operating systems for free

~Mail follows today’s post~

XOSL doesn’t seem to like my Promise Ultra66 controller. At least not all the time. I don’t like that. I also don’t like how XOSL installs itself in the root directory–my poor root ballooned to over 40 entries after installing it. That’ll cause some system slowdowns. I don’t like having any more than 16 entries in there if I can avoid it.

Fortunately you can install XOSL to a dedicated partition, and that looks to be the better method.

But when XOSL works, it seems to work well. It’s slick and versatile and gives you a great deal of freedom over how and where you install your OSs, as well as how many you can install (and let’s face it, with 30-gig drives selling for $99 at CompUSA, running multiple operating systems is going to get common).

And I see from Brian Bilbrey’s site that patents may accomplish what the RIAA could not. Makes me wonder why one of the RIAA members didn’t just buy Fraunhofer Institut (who owns the applicable patents on MP3) and start charging outrageous royalties immediately. That’ll kill new technologies faster than anything — just ask Rambus.

~~~~~~~~~~

From: “Dustin D. Cook” <dcook32p@nospam.htcomp.net>
Subject: Windows Me
Dave,

Here’s my two cents on Windows Me.

I have been testing this operating system for some time now before I begin pre-installing it on new computers. We’ve run the gamut of stress tests, benchmarks, and usability tests, and we have some interesting results.

All tests were run on multiple machines with a minimum system being an AMD K6-2 500 with 64MB RAM and a Voodoo3 3000 graphics card. The best system tested was an AMD Athlon 1.0 GHz with 512 MB Mushkin PC133 2.0 memory and
a GeForce2 GTS 32MB DDR. We used the same HDD for each machine. It is a Maxtor DiamondMax 45 Plus (ATA/100, 7,200 RPM).

Windows Me should take home a gold medal for speed. It booted quickly, it loaded programs at blistering speeds, and it performed very well in our 3D tests. All-in-all, Windows Me is about one percent faster than Windows 98 SE. This came as something of a surprise to me. I was expecting slightly degraded performance due to the additional system overhead of Internet Explorer 5.5 and the new features of Windows Me. Either Microsoft did some serious “tweaking” to their code, or I’m missing something entirely about this operating system.

Stress tests were a different story. Occasionally, Windows Me would lock-up on us for no apparent reason. The same computer running Windows 98 SE would never falter during our tests. Actually, sometimes Windows Me would lock-up when we were not even running the tests! We replaced some hardware in the machine, but it did this on all of the test PCs. This was a big problem for us. We still haven’t officially tracked down the killer, but we think it involves the new version of Internet Explorer. We had already completed our tests before the new service pack was released, so I don’t have any data from that version. The stress tests involved opening a 25 MB Excel 2000 spreadsheet and minimizing it; open eight browser windows and loading miscellaneous things like Flash movies, several animated GIFs and PNGs, and several Java applications; having The Matrix DVD-ROM’s menu playing in WinDVD 2000; and running Unreal Tournament at 1280x1024x16bpp with our custom “movie”. Windows 98 SE performed admirably, but, as I had mentioned earlier, Windows Me couldn’t do it.

In the usability tests, we had some elderly people try out each computer. This isn’t really a test that can be easily replicated, but overall Windows Me seemed easier for them to use.

What’s my opinion on Windows Me? I think Microsoft made a fairly good product. I’m not very impressed by the lack of native DOS support. I frequently use that to diagnose customer’s computers. What do I do if I have forgotten my boot diskette? I return to the shop and grab one instead of making one right there. The stability issue is a big concern of mine. I’ll try to reproduce those results after downloading the new Internet Explorer service pack, and I’ll write back to you with those results. The speed is commendable. I appreciate the extra “oomph” that Windows Me appears to have behind it. The boot time is quite impressive!

My prize goes to Windows 98 SE. Speed is a very good thing, but when it comes at the cost of stability…we have a problem. My customers don’t want their machine freezing every time they try to open http://thesiliconunderground.editthispage.com/ . 😉

Sincerely,

Dustin D. Cook,
A+ Campus Computers
Stephenville, Texas – USA
~~~~~

Thanks for the info!

You can add DOS support back in with a utility available at www.geocities.com/mfd4life_2000 — that was one of the first things I did after installing WinMe. As for stability, IE5.5SP1 might help. Running 98lite (www.98lite.net) to remove IE 5.5, then replacing it with IE 5.01 (or not at all) could help. I’m not at all impressed with IE5.5, so I’m inclined to speculate the blame lies at its clumsy feet.

I’ll keep experimenting with it myself. And I’m hoping my page is simple enough that it won’t crash any browsers. 🙂

AMD’s turnaround

AMD just turned their fourth consecutive profitable quarter, and they say they expect to sell out of Athlons this year. This exposes AMD’s prime weakness: Even though this year Intel has repeatedly failed to execute while AMD has had smooth sailing, Intel has the tremendous advantage of capacity. AMD has two fabs. I don’t remember how many Intel has. Eight?
I was talking with someone before church Wednesday night about new PCs, and he said, “I hear AMD is actually outperforming Intel these days.” AMD, of course, has always been known as the budget chip maker. In 1992, if you wanted a bargain PC, you got an AMD 386DX/40. In the mid 1990s, for a bargain you bought AMD 486s. In the late 90s, you bought AMD K6s. Suddenly, AMD’s not doing much in the low end. They’ve stopped taking orders on K6-2s and they’ll ship their final one this month. The Duron’s a great chip, but they’re not making them in huge quantities. They cite lack of an inexpensive integrated chipset, which is a perfectly valid reason, but there’s another reason. Why should they produce large numbers of Durons? They’re selling every Athlon they can make, so why sacrifice high-margin chips to make lower-margin chips?

It’s been a very interesting year. AMD bet the company on its new fab, knowing that if they made one mistake there was every possibility they were toast. Going into 2000, Intel looked like a company that could do no wrong. But Intel made tons of mistakes this year (the i810 and 1.13 GHz PIII recalls, other high-speed chips that you could read about but not buy, the lack of a suitable replacement for the venerable 440BX chipset, delays on the P4), having a year that made some of AMD’s bad years look good or at least acceptable. Meanwhile, AMD executed. Unlike some, I was fairly confident that AMD would find some way to survive, but survival was about all I expected from them, and about all anyone had any right to expect, given their track record and financial condition.

The guy at church asked if he should sell his Intel stock and get some AMD. I told him I didn’t think so. If AMD can turn around, so can Intel. As for whether AMD is a good investment, I don’t know. They’d look a whole lot better to me if they had a couple more fabs. They’re doing great on CPUs and flash memory, but they need enough capacity to be able to afford to flood the market with Durons, and they need enough capacity to be able to manufacture chipsets if they so choose, rather than developing chipsets as stopgap solutions until VIA and SiS and ALi can step in and then phasing them out. Intel learned that building both chipsets and CPUs allows you to, to a great degree, control your own destiny, not to mention make an extra $20-$30 per computer sold. Intel’s mistake was guessing incorrectly (or perhaps not caring) what consumers wanted, betting on Rambus, and then delivering a bunch of technologies that it was quickly forced to recall.

I don’t know that AMD can turn itself into a giant like Intel, but a strong AMD is good for all of us. It keeps Intel honest.