Apple for Consumers

For consumers, I think it is fairly easy to make a compelling case for buying an Apple computer. First, since the hardware and the software are fully under the control of one company, stuff just works (mostly). Second, since Apple has only a small share of the market and OS X has a better security model than Windows, as long as you don’t run Windows on your Mac, you don’t have to worry about spyware and viruses. As Apple gains momentum (and the user-base becomes ever more smug) this will change, but for now, OS X is safe. This eliminates a major headache for most consumers and extends the life of the machine. (Believe it or not, many users buy a new computer to eliminate problems caused by spyware and viruses.) Third, with the switch to Intel and the new pricing, Apple computers are now just as powerful as their PC counterparts. They are also competitively priced when compared to similar quality equipment (e.g., MacBooks vs. Thinkpads). All these features make a compelling case for first-time computer buyers to purchase a Mac. When my wife needed a computer, I suggested she by a MacBook, even though I am a longtime PC user.

For longtime PC users, the decision to switch is a bit harder. OS X is unfamiliar, and these users may have a myriad applications that are not available on the Mac. However, with the release of the Intel Macs and Parallels Desktop, this all changed. It is now possible to buy a Mac, run OS X to get all its benefits, and still run your old Windows applications. Yes, this does make that copy of Windows running in Parallels vulnerable to spyware, trojans, and viruses, but probably less so than a native Windows install. That’s because you most likely won’t be reading email and surfing the web under that Windows install. Furthermore, if you can use Microsoft Office for the Mac, you avoid many problems with Word macro viruses as well. Oh, and since Apple laptops now allow you to use two-finger taps on the trackpad to simulate a right click, interacting with Windows is no problem at all. Apart from inertia, and price (on the low end), I think that there is little reason not to buy a Mac even for PC users (unless you play games, but hopefully that will change in the near future).

All these factors combine to make Apple computers a compelling choice. Furthermore, as more users adopt Apple, there will be less and less need to use that copy of Parallels Desktop since there will be a native Mac app. Even game developers will come around if Apple captures enough of the consumer market. With 12% of all new laptops sales in the U.S. going to Apple, maybe we’ve already reached the tipping point, beyond which many application developers can no longer afford to ignore Apple. Microsoft will likely be the last hold out.

Apple for the Business World

Unfortunately, I think the consumer market is the low-hanging fruit for Apple. Don’t get me wrong, there is absolutely nothing wrong in tackling the easy stuff first. However, I think that Apple will have a much harder sell in the business environment. In fact, I think Apple will have to radically rethink its business model to compete in this space.

Businesses do not like to be beholden to a single vendor for anything business critical. Such a vendor essentially owns your company. They can raise prices, jerk you around, force you to update (if licenses expire), etc. and there is nothing you can do about it. Microsoft is in this position with many businesses already, and I’m sure businesses don’t like it. They continue to tolerate it because they see little choice.

The problem for Apple is that, from this perspective, they are worse than Microsoft. They control both the software platform and the hardware. One might argue that Apple will be forced to compete with PC vendors, since companies can always switch back, but that is an expensive proposition. Consider that a company’s employee base is trained on a specific system, with a specific set of, potentially custom, software. What if Apple has production problems and cannot ship machines in the quantity your business needs? What if your 3 new salesmen have to wait 8 weeks for a computer? There is no one else to turn to in a pinch.

In fact, the inertia, within a company, behind an existing platform makes it that much harder to convince a business to switch to Apple. Compound that by saying, “Oh yeah, and if you make the switch, Apple owns you in terms of both hardware and software.” and it makes a switch to Apple almost unthinkable. For my personal computer (since I maintain it myself), I’ll gladly switch to Apple as soon as those Merom laptops come out. But if I ran a company, I don’t want to be in a position where I must trust Apple to competitively price its hardware, to be fair in negotiating volume purchasing discounts, and never run short of supply.

Why is it that when it comes to computers and technology, people want to immediately relate one problem with another. Even if there is virtually no correlation, users decide to create one.

He goes on to posit that there is some strange aspect of technology that causes people to make correlations between unrelated problems, driving “computer experts” nuts.

While I’m glad to see I’m not the only person who’s noticed related-blame syndrome, I disagree with the idea that it is a mysterious affliction limited to technical problems. Dan emphasizes that this is a tech specific problem by analogizing the type of correlation I described above with the following non-tech example:

“Remember how last week at lunch your soup was cold? Well now my fridge doesn’t work.”

But I think this is a false analogy. Although the soup was cold and fridges make food cold, everyone understands the two are unrelated. A better analogy is when someone says:

“The other day, you fixed the radiator in my car and now my car stereo doesn’t work.”

I’ve definitely seen this before, and I’ve said similar things as well. And when I’ve said them, I’ve been convinced it was true. The mechanic will deny it of course, but it is often of little value. I am not a car expert but I am distrustful of automechanics. I know enough to know, maybe when he was fiddling around fixing my hypothetical radiator, maybe he jiggled loose the power connector for the radio. I don’t know if that’s even possible, but it sure seems suspicious that my radio stopped working shortly after the mechanic did something I don’t understand to my car.

As with cars, people really don’t understand computers. So, when you fiddle around with iPhoto to help them import pictures, they find it suspicious that their Internet connection no longer works only two days later. In both the automotive case and the computer case, the only reasoning going on is: Some expert did something mysterious to help me the other day and boy was it complicated. They must have screwed something up doing that complicated thing that has led to my most recent problem.

I will say that the auto-mechanic does have it a little better than the computer expert though. People think a bit before blaming the automechanic, so the mechanic doesn’t get “You fixed my car stereo the other day and now I have a flat tire!” as frequently as computer expert gets “Hey, remember how you fixed my mouse settings, well now my Internet doesn’t work.” Perhaps I should write a short rant on how it seems that computers emit some strange electromagnetic fields that disconnect people’s higher brain function as soon as they sit in front of one. Sounds like a good first article for the Rants category.

loose (verb)

trans. To let loose, set free; to release (a person, an animal, or their limbs) from bonds or physical restraint.

lose (verb)

Of a thing: To be deprived of or part with (a portion of itself, a quality, or appurtenance)

There are other definitions for each word, of course, but none are similar. The problem lies in the pronunciation of lose. The “o” in lose sounds like an “oo”; lose is pronounced looze. This leads to the “loose” spelling of the intended “lose”. The fact that loose is also a word, albeit with a very different meaning, means that spell-checkers can’t catch the error. To get a feel for the pervasive nature of this problem, keep an eye out for it when reading articles online, especially blogs and other “amateur” content.

With so many people making this error, it’s only a matter of time before some dictionary acknowledges that l-o-o-s-e is a valid spelling for lose. And then, the distinction between lose and loose will be lost, apart from the context in which the word appears. There is precedent for such things. The word peruse is one such example. This is what OED now has to say of the word peruse.

3. trans. To examine in detail; to scrutinize, inspect, survey, oversee; to consider, to take heed of. Now also (influenced by sense 4c): to look over briefly or superficially; to browse.

Yes, you read that correctly. The word means either to examine in detail or to look over superficially. In essence the word now simply means to examine. Apart from that, the word is meaningless, since it is often hard to know from context which sense is intended. In fact, the slippery tongued can use that ambiguity to give the wrong impression. Fortunately, the loose/lose distinction is much clearer from context.

Oh, and just in case you think this is a common typo, I’ve asked a number of people who write “loose” for lose if they see a problem with their spelling, and often they don’t see anything. In other words, they no longer make the distinction between the two spellings. Once enough people migrate to that camp (and I’ve met quite a few), the metamorphosis will be complete and probably irreversible. In short, we’ll see lots of insane folks setting their minds free instead of simply being “deprived of” or forced to “part with” it.

My goodness it’s slow. My wife asked what was wrong with the scroll bar and after a quick look, it was just the machine being really slow. Firefox takes some time to load, as do most other non-Apple applications. The machine isn’t exactly what I would call responsive. My wife’s MacBook is lightening fast in comparison. Applications load instantly and are quite responsive. Even Microsoft Office, which is running under Rosetta (i.e., a PowerPC app running on a translator) is more responsive than Firefox on the iMac.

Apple isn’t kidding when they say their new machines are 5x faster than the previous generation. They just don’t tell you it’s because their old machines were super slow. My Dell Latitude is of the same vintage as the iMac (circa 2003) and is noticeably faster. I can now understand the excitement amongst Mac users regarding the performance of their new machines. But instead of being happy about the performance boost, perhaps they should be angry that Apple had been overcharging them for slower hardware until the switch to Intel. At least now users are getting what they’re paying for with their Mac hardware.

A classic example is the Ford Pinto. I don’t think I can say it much better than Wikipedia:

Through early production of the model, it became a focus of a major scandal when it was alleged that the car’s design allowed its fuel tank to be easily damaged in the event of a rear-end collision which sometimes resulted in deadly fires and explosions. Critics argued that the vehicle’s lack of a true rear bumper as well as any reinforcing structure between the rear panel and the tank, meant that in certain collisions, the tank would be thrust forward into the differential, which had a number of protruding bolts that could puncture the tank. This, and the fact that the doors could potentially jam during an accident (due to poor reinforcing) made the car a potential deathtrap.

Ford was allegedly aware of this design flaw but refused to pay what was characterized as the minimal expense of a redesign. Instead, it was argued, Ford decided it would be cheaper to pay off possible lawsuits for resulting deaths. Mother Jones magazine obtained the cost-benefit analysis Ford had used to compare the cost of an $11 repair against the cost of paying off potential law suits. The characterization of Ford’s design decision as gross disregard for human lives in favor of profits led to major lawsuits, inconclusive criminal charges, and a costly recall of all affected Pintos. Ford lost several million dollars and gained a reputation for manufacturing “the barbecue that seats four.”

The scandal surrounding the Pinto led to some landmark judgements against Ford and an initial jury verdict of$125 million in punitive damages (quite a sum for that time). There is some controversy regarding how bad the Pinto problem really was (see Wikipedia), but it is still a notorious case of management disregarding safety in favor of small profits.

Recently, Dell and Sony users have experienced catastrophic battery failures in which the laptop actually explodes! Although there have been reports of computers catching fire in the past, I’ve never heard of an explosion like the one linked above. Basically, Dell has been shipping consumers time-bombs. Literally.

What makes this oh-so Pinto like is that, Dell and Sony apparently knew about the problem 10 months earlier. Sony and Dell discussed the issue of excess metal particles in their batteries puncturing delicate dividers, which could potentially lead to fires or explosions. Instead of recalling the product, Dell and Sony decided to wait and see if the laptops would actually catch fire or explode in practice before issuing the recall! In short, Sony and Dell knew that they could be shipping viable bombs to their customers, but decided to wait and see if the detonator would actually engage before notifying customers and recalling the product. To me, this seems exactly like the Pinto; Dell and Sony likely estimated that the expected cost of any liability was less than the cost of the recall, so they crossed their fingers and waited. The companies are lucky that no one has been seriously injured or killed because of this issue. Imagine if a laptop exploded in the passenger cabin of a commercial jet! At least we now know that the computer industry has no more of a conscience or concern for consumer safety than the 1970’s car industry.

The morning after the MacBook first started shutting down, I decided to see if the problem could be fixed at home. I googled around for some help and found that resetting the PMU might fix the problem. Unfortunately this also failed. I wanted to also reset the PRAM but the machine wouldn’t stay on long enough to allow me to do so. I finally did discover that holding the power button down for a while would allow the computer to boot after a long beep. So, at least the machine wasn’t a paper weight.

Since my attempts at a home remedy had failed, it was time to call Apple Support. Here was the first (and only, so far) strike for Apple Support. They had me on hold for over 25 minutes. While I was listening to the hold music, I thought back to a week earlier when my wife had come back to a machine that was mysteriously off even though she had only put it to sleep the night before. While strange, I didn’t think much of it. Was that the first of the random power downs? I was jarred back to the present by two DTMF tones, and then a ring. Finally, I thought, I was through, but I was simply put on hold again, this time with a female voice saying, “We’re sorry, all representatives are busy now…” instead of the previous male voice.

After finally getting through to a representative, I explained the problem and what I had done. The rep on the phone was very polite and friendly and asked if could hold again for 5 minutes. Ugh. This time, though, the hold was only five minutes, and he came back with some suggestions. The first was to do a PMU reset, but then he said to skip that since I’d already tried it. Next, he wanted me to re-seat the RAM. Embarrassingly, I didn’t have a screw driver small enough to remove the cover for the hard disk and RAM, and communicated this to the rep. He promptly stated that that was fine, and that I should take the machine in to an Apple store if possible, since the next step, if the memory re-seat failed, was to test with a new battery and power supply. So, I made an appointment to go to the Apple Store where we bought the computer for that same afternoon.

Upon arriving at the store, we had to wait ten minutes because they were a bit swamped at the Genius Bar. (Who the heck calls tech support people geniuses, anyway? Seriously, their official title is Genius. It said so on their t-shirts. My experience with other companies’ technical support suggests that the name is a bit of a misnomer for tech support folk. In Apple’s defense, the support folks were very knowledgeable, though we didn’t interact enough for me to decide if they were really geniuses.) Anyway, after we spoke to someone, they quickly looked up the purchase of the machine, noted that it was only 18 days old, and immediately decided to replace the machine with a brand new one. They even swapped the hard-drive for us. So, after a few hours, my wife had a working MacBook. I’d asked what they would have done if the machine was older, to which the Genius responded that they would have let us hold on to the MacBook until a logic board arrived for replacement, which usually takes a few days. In any case, the in-store replacement was awesome. 5 hours from store open to working laptop. That’s hard to beat.

Overall, I was extremely happy with Apple Support, except for the hold time issue. At no time did anyone make me go through a ridiculous set of irrelevant steps. The reps/Geniuses seemed happy to skip a test that I said I’d already done, and when all else failed, they promptly replaced the machine. The only thing more I could ask for is that the machine never failed in the first place. But hey, every company has their disasters. (Though every rev A Apple seems to have more than its fair share of problems, but that is for a different article.) There was a time when all 7 Thinkpads in the research group I was working for had to have the hard drive replaced. And even IBM’s customer service tried to jerk me around when they were unable to get a replacement drive for my laptop because they were back-ordered on the 60GB model. After waiting two weeks for a replacement to come in, I had to spend a day on the phone with them until they finally agreed to upgrade me to an 80GB drive which they did have in stock. So no one is perfect. If you have to use the support, it doesn’t get better than the experience I had with Apple. The only thing left to see is if they properly update the AppleCare contract with the new MacBook’s serial number.

Postscript
The really disturbing thing, which I can’t blame on Apple Support, but do blame on Apple, is not knowing if the new laptop will actually fix the problem. Is there a new logic board for the MacBook that fixes this problem or do they just roll the dice and hope the new boards/machines don’t have the problem? It seems to be the latter but I haven’t been able to find out. More importantly, when there is a revision (if there isn’t one already) that fixes the problem, can users that do not currently experience the problem get the upgrade? A company should not deny a fix to a user for an issue that turns the machine into a useless brick, even if the user isn’t currently experiencing the problem but might in the future.

Apple’s Macintosh platform has held a strong appeal for me ever since they released MacOS X. (Prior to that, I made fun of Macs saying that MacOS 9 finally caught them up to Windows 95’s preemptive multitasking.) As a hardcore Linux user at the time, the idea that I could run a UN*X derivative on a laptop without fighting with drivers to make things work, have the ability to seamlessly run Microsoft Office, and have a slick, refined, GUI was very appealing.

The problem was that I simply refused to pay 50-100% more for inferior (i.e., much slower) hardware. However, today, the situation is different. MacOS X has gotten better, Apple’s hardware has gotten better, and their prices have gotten a lot lower. Taken together, this made Apple’s machines worth a second look. As far as I can tell, the Macbook and Macbook Pro are actually competitively priced when compared to similar laptops. What’s more, since Apple now uses Intel processors, one can actually run Windows at reasonable speeds.

As a long-time and mostly satisfied Thinkpad user, I wanted to see how the various Apple models stacked up, especially compared to my IBM Thinkpad T42. Looking at the Apple line-up, there are three things that differ from the T-series I currently use: the display (both physical size and resolution), the keyboard, and the pointing device.

The Pointing Device

One of the things I love about my Thinkpad is the trackpoint input device and its 3, that’s right 3, buttons. If you’ve ever used xfig, that middle mouse button is almost essential. Compare this to the Apple with one mouse button. My experience with the trackpad on the Dell Latitude I briefly used was also not pleasant. Fortunately, my wife recently purchased a MacBook and so I could experiment with the trackpad and single mouse button. Much to my surprise, the Apple trackpad is great. I still like the trackpoint better, but only marginally. I could get used to the Apple trackpad. I also learned that with a two finger tap, you can right click, so that is effectively 2 mouse buttons. I can live with 2 mouse buttons, though I’d prefer three. As an aside, the two finger scroll thing is pretty darn nifty as well.

The Keyboard

The keyboard on the Thinkpad is fantastic. Its so good, I prefer it to a full-size keyboard. However, I didn’t have extensive experience with the Macbook Pro keyboard and the Macbook keyboard looks darn right uncomfortable. The best I could do for the Pro was try it out at the local Apple store. I must say that the keyboard on the Pro feels great. Almost as comfortable as my Thinkpad. The keyboard on the MacBook is different but still good. After using my wife’s MacBook for several hours, I stopped noticing the keyboard, which is a good thing.

The Display

With both the pointing device and keyboard on both the MacBook and MacBook Pro passing muster, let’s turn our attention to the difficult issue of the the Apple display. Just looking at the specs gave me pause. My Thinkpad has a 14.1 inch, 1440×1050, display. A nice resolution at a reasonable size that still allows for portability. (I have no idea who carries around these 17″ notebooks, they are huge, heavy, and unusable on aircraft.) For their size, it looked like the MacBook and MacBook Pro were rather low resolution. As someone mentioned to me, if you have good eyes, effective screen real-estate may be better measured in pixels as opposed to actual area since that controls the minimum font size. I’ve been accused of using microscopic fonts, and as someone who does a lot of programming and writing, effective screen real-estate matters. I need to know exactly what I would be sacrificing screen-wise to move to the Mac (and as I’ll discuss later, what I get in return).

To find out I compared the stats of my current Thinkpad screen to those of the MacBook and MacBook Pro. The results are in the table below.

This means that with the 15″ MacBook Pro (MBP) you actually gain 13% physical real-estate, but lose resolution. Remember, that if you use small fonts, resolution controls the smallest usable font size, and thus controls effective real-estate. From this perspective, the MBP has 15% less effective screen real-estate. The MacBook (MB) has 15% less physical area and a whopping 30% less effective real-estate. The final difference in the display department is under the hood. The 15″ MBP has a dedicated ATI Radeon X1600 while the MB has the integrated Intel GMA 950. This certainly makes it seem like the MBP is a much better replacement for a Thinkpad T42, but remember, the real choice is between a new Thinkpad and a MacBook or MacBook Pro. So the question is, “does the MB or MBP offer enough in exchange for the poorer screen to make the switch worthwhile?”

MacBook vs. MacBook Pro

As we saw above, the MacBook (MB) has a considerably smaller screen compared to the Thinkpad T-series measured in both pixels and physical area. The MacBook Pro (MBP) has a larger screen but fewer pixels. Since I am a small font guy, the fewer pixels matters more at the given screen sizes than the smaller physical area. Furthermore, 75% of the time I use the computer in my office or home office. If these things are not true for you, then the analysis in this section might differ a bit. On to what is offered by the Apple in exchange for a lower resolution/smaller display.

The first thing to consider is battery life. The reports for the 15″ MBP’s roughly 3.5 hour battery life seem in line with that of my 2 year old (when it was new) and friends’ new Thinkpads (I’ve never gotten 4 hours out of my Thinkpad battery). The reports for the MacBook battery life are much more impressive, and my experience with by wife’s MacBook confirms the reports. As opposed to feeling lucky when the battery lasts 4 hours, it seems that 4 hours is about average with 5 hours not unusual if you are careful. I’ve never seen the 5:40 minutes the Ars Technica article reports though. I also don’t understand why people say that turning off wireless extends battery life significantly. I can’t tell that much of a difference. Anyway, on the battery front, the MacBook offers quite a bit, the MacBook Pro, not so much, when compared to the Thinkpad.

The next thing to consider is size. I don’t have a good photo (yet), but the MBP is rather large compared to my Thinkpad T42 (though not thicker) and slightly heavier. The MB is slightly lighter but about the same size. The 15″ MBP is almost too large in my opinion. The MB is just about right. I won’t mention the 17″ MBP except to say that it is ridiculously large.

As for the graphics subsystems in the MB and MB Pro, I also don’t mind the GMA 950 graphics. I don’t play 3D games and the stats for the GMA950 seem about the same as that for my T42’s ATI Radeon 9600; I’ve never found the T42 wanting in the graphics department.

Conclusions, Pricing, and Reliability

Based on how I use my computer, the MacBook is really appealing. With the MacBook I get a nicely sized machine with great battery life. The screen is smaller, but for the difference in price between the MacBook and the Pro I can almost buy 2 20″ Dell wide-screen LCDs to make up for it during my office/home office usage (75% of overall usage). Besides the screen, going from the MBP to the MB means one has to give up some faster processor options, the dedicated 3D graphics, and the ExpressCard slot. To me this isn’t a big sacrifice. The graphics on the MB are similar to what I have now and I’m fine with that. I only use the PCMCIA in my laptop for a CF Card adapter and so I feel that the ExpressCard is pretty much worthless. Finally, I don’t think the 2.16 GHz is worth the extra money (Over $300). However, by giving up these options, I get extra battery life and save almost $800. I suspect that for most non-gaming users this is also the right trade-off, making the MacBook (not Pro) the right choice if going Apple is a forgone conclusion. For those of us for whom Apple is one of many options, the question is “how does the Apple stack up against a new Thinkpad T60 (with the low-res screen) price wise?”

All machines have Bluetooth, wireless, etc. The Z61t comparison is straight-forward. The T60 above is about as equal as I could get it. The problem is that you can’t get the T60 with the low-res screen and the 2.0GHz Core Duo, so I had to settle for the 1.83 GHz. The T60 also has a slightly larger hard-drive (because of a special promotion that makes the 100GB cheaper than an 80GB), and a better optical drive. Let’s say that makes up for the 0.17 GHz in the CPU. The MacBook still has a better display (which beats the T60 screen by more than the specs imply, I’ve seen both screens), which, when combined with the benefits of OS X outlined earlier, I would say is worth $100. (A similar comparison of the MacBook Pro and the T60 with the better screen shows similar results, though in that case, the Thinkpad has the bigger, though not better, screen.) Some might say that the X-series Thinkpad is a better comparison to the MacBook, but it has a smaller screen (physically) and no built-in optical drive which is a show-stopper for me. Oh and no matter what you do, don’t buy more memory or harddisk than listed above from Apple. The prices are offensive. For example, you can buy 2GB of laptop memory from Crucial for about $250, while Apple will charge you $500!

The last issue is one of reliability and service. It is possible to get cheaper machines with similar specs, but not from a company I would trust to build a solid laptop and give excellent service when (not if, but when) there is a problem. IBM has had unbelievably good service. I’ve called support to report a failure and the Airborne Express guy has been at the office with a box in under 4 hours. I had the computer back the next afternoon. Thinkpads are also darn tough, though they do have their issues (hence my experience with the service department). Apples also seem darn tough. I’ve seen 3+ year old heavily used iBooks and PowerBooks still in service and working well. Contrast this to my Dell Latitude where the CD-RW drive falls out, so it is taped in! Apple service isn’t quite as good as IBM, but then whose is? And that includes Lenovo; while their service might be adequate, I doubt it will stay up to IBM standards over the next 3 years.

As for Apple’s recent troubles with the MacBook and the MacBook Pro, here are my thoughts. There do seem to be some Quality Control issues on both machines. However, the Pro seems to be suffering more than the MacBook. The main remaining MacBook issue seems to be heat. While an issue, my wife’s MacBook seems fine. Don’t get me wrong, it is super hot when it runs, but I don’t feel it is much hotter than other new laptops. Overall, I think that the Thinkpads are among a few laptops that are in the same class as the Apple in terms of build quality (though it probably should be stated the other way around, Apple is one of the few machines in the same quality class as the Thinkpad). As for the 3000 series, I don’t believe that it’s in the same class as the Apple and Thinkpad. Lenovo makes both the 3000 and the Thinkpad, so what is different between the models? I suspect quality and ruggedness is a big difference.

Pulling the Trigger

Despite all this I haven’t yet bought the MacBook (and 20″ Dell Screen). The main reason is that I’m not pressed to replace my Thinkpad and I would like to get a Merom powered MacBook. If you need the computer now, though, it’s probably not worth waiting. Rumor has it that both the MacBook and the MacBook Pro will be upgraded together early in September, so its worth a few weeks wait to me.

In the review, MacAskill says that his company is primarily interested in minimizing the cost, per application performance unit, per watt, or as he says on his blog $/CPU/Watt (a common criteria for web companies, including Google if I recall correctly). The primary question he asks is that in terms of this metric, is the T1000 (CoolThreads) better than the Quad-core Opterons (Olde Faithful, machines with dual dual-core CPUs) they currently use. To spoil the review, the conclusion was an emphatic no. It wasn’t even close. In response, readers clamored for a better comparison, suggesting ways to tune performance, citing Linux’s poor thread scheduling compared to Solaris, and calling for Solaris results for the T1000. Since it was such a blowout in favor of Olde Faithful, the engineer in me has to ask, is there even a point in doing a Solaris test? Was there any point in doing the Linux test in the first place?

CoolThreads just too Expensive?

The real reason that I suspect that CoolThreads is doomed to lose is cited on the last page of the article. MacAskill says:

Olde Faithful costs $3,595. At the “sweet spot”, she uses 157.2 watts of power, which costs us $26.20/month in power and cooling costs, or $314.40 per year. Total cost for 1 year? $3,909.40. … CoolThreads costs $8,395 for a similar model with 4GB … Total cost for 1 year? $8546.52.

In other words, CoolThreads costs more than twice as much as Olde Faithful. So what does this mean? Going back to the original evaluation metric, $/CPU/Watt, let’s say that Olde Faithful consumes wo Watts of power, generates po performance units (however you want to measure them), and costs co dollars. (And you thought algebra was useless.) Then its evaluation metric, call it Mo, is Mo=co*wo/po. Call CoolThread’s metric Wc, and let it consume wc watts of power and deliver po performance units. We know that CoolThreads has a cost of 2*co. Knowing nothing about power consumption (wc) and performance (pc), we still know that for CoolThreads to win (i.e., for Mc < Mo), (2*co*wc/pc) < (co*wo/po ). Cancelling the co from both sides we see that 2*wc/pc < wo/po. After some manipulation we get that pc/wc > 2*po/wo. In other words, Coolthreads has to deliver more than twice as much performance per watt in order to win.

In my experience, this seems really unlikely for real applications. Let’s say that since Olde Faithful has only 4 processors, and CoolThreads has 8 processors, it handily delivers twice the performance. This is a very generous assumption, by the way, since each CoolThreads processor is not as fast as the Opterons in Olde Faithful. If each Niagra core in CoolThreads consumes just as much power as a core in the Opteron, then we also double power consumption, pc/wc remains unchanged, and thus Olde Faithful is better by a factor of 2! In order for CoolThreads to just tie in this scenario, each core would have to deliver the same performance as the Opteron in half the power. This seems very unlikely, since if it were possible and Sun knew how to do it, they could just release a laptop x86 and make a killing (since I also doubt the ISA makes much of a difference, as Intel proved in the RISC/CISC performance battle).

Of course, it is possible that the application smugmug is running is not CPU intensive, and therefore the processor performance is irrelevant, and thus the simplicity of the Niagra cores in Coolthreads could dramatically reduce power and have it come out as a winner. However, with power management technology, it seems unlikely that this would give CoolThreads a 2x advantage. In short, I just don’t see Sun developing an architecture that generates double the performance per watt as AMD. Even though Sun typically has much better I/O than x86 systems, x86 boxen have been improving on this front and I don’t think the gap is close to the needed 100% for most applications. I vaguely recall that the better I/O might get you 30% if you are I/O bound, but I could be wrong on this. To get the 2x with only a 1.3x improvement in I/O performance, you’d need a corresponding .65x drop in power consumption (a 35% reduction) in the I/O subsystem alone, which also seems unlikely.

So what about Solaris? This is more promising, since it is possible to deliver 2x or more performance improvements with little extra power consumption in software. In processor hardware you are lucky to get 3% with a single technique (hence the nickname the 3%-club for people doing processor architecture research). System architecture is a better bet, but software can give really big wins. However, to get 2x, it usually means that the original implemenation had some serious bottlenecks that have gone unfixed. These do exist, and I’ve seen some research in the OS-design-for-web-serving space that highlights these issues, but I suspect that Linux and its scheduler aren’t even close to 2x worse than Solaris in performance/power, performance alone, or power alone. Which means that Olde Faithful is going to win again.

Conclusion

In short, I can come to the same conclusion as MacAskill without running any numbers. If I had to budget my time, I’d forget about Solaris and forget about the CoolThreads box. Unless those machines get cheaper, they just won’t make sense if you care about $/CPU/Watt. Now, if you are an AMD user, it is definitely worth investigating the new Intel Woodcrest and Conroe based machines. Those machines essentially take the best of AMD’s architecture and Intel’s fab and deliver a great product in terms of price performance and performance/watt. I was astounded by the initial performance numbers reported for the Conroe, and am even more shocked to see that those numbers hold up to independent scrutiny. Now, if only the Xserve from Apple had the same compelling price point as the Mac Pro…

Postscript - Added August, 18, 2006

Looks like Sun will be helping with Solaris numbers and some performance tuning after MacAskill’s original post got Dugg at ton. I’ll be curious to see if the tuned Ubuntu or Solaris on CoolThreads can beat Olde Faithful. It will be most impressive if they succeed. Hopefully, there is a detailed analysis as to why they could double the performance per watt of the T1000 if they do.

1. If the domain is less than four letters long, it’s taken.
2. If the domain is an english word in common use, it’s taken.
3. If the domain is two common english words pasted together, it’s probably taken.
4. If it sounds good, its probably taken. (ok, maybe this is a bit cynical.)

It seems I was close to right in my rules of thumb, except that the situation is even worse than I thought. Odds are you can’t get a firstnamelastname domain name either.

Now, the above is probably common knowledge for folks who deal with DNS and domain registration on a regular basis. What really shocked me was the sheer number of domains that were registered solely to collect ad revenue until they are sold. I’m not entirely sure how the economics of this works since I haven’t kept up with how domains are registered after domain registration became commercial, but I’d be curious to know what kind of return these domain “squatters” are generating on their electronic real-estate.