Akemi's Anime World

Akemi’s Anime Blog AAW Blog

A Supercomputer In Your Pocket

A supercomputing expert recently made some comments about the computing power of an iPad 2 relative to a Cray-2. It was a funny coincidence, because I had just, completely at random, been goofing around with some similar calculations myself.

So, for the heck of it, here are my very rough and mostly unscientific comparisons of classic supercomputers to one pocketable device and one luggable one: A current iPod and my new laptop, a MacBook Pro quad-core 2.2GHz Sandy Bridge i7.

These two devices are arbitrary; I picked an iPod because it’s cheap, popular, and lacks unnecessary (for this comparison) cellular hardware, and the laptop because it’s portable, pretty beefy, and I just bought one.  The numbers will be about the same for an iPhone 4 or any similarly high-end smartphone (or much higher, in those that use an iPad 2 class dual-core CPU), or any relatively high end laptop.


One thing I’m ignoring in my attempt to compare stuff you can buy in 2011 to decades-old supercomputers is GPUs. Modern GPUs are incredibly powerful, even in high-end handhelds, but they are also very specialized, so while you can find GFLOPS ratings for many GPUs, and some general-purpose computing tasks can be run on them, you can’t run the Linpack benchmark on most, and it’s hard to make a one-to-one comparison of the computing power of a GPU versus a general-purpose number-cruncher, be it modern CPU or classic supercomputer.

The other caveat is the standard one about synthetic benchmarks across very different systems; you can only really compare some specific task. The Linpack benchmark suite has been used since the early ’90s to benchmark massive supercomputers, and it’s easy enough to download and run yourself, plus it’s the standard used by the Top500 fastest supercomputer list. Since it also produces results in FLOPS (floating point operations per second), which have been used for a very long time to measure computing power, I’m using it as the measure of power. It’s actually a fairly specialized (and not necessarily very useful) set of math routines, but it’s as universal a benchmark as any.

Vector calculations also complicate things; being more specialized, classic vector computers often had disproportionately high performance at the cost of less versatility. Of course, many CPUs (including all modern ones) also have a vector processing unit (Apple based a whole advertising campaign on the power of the AltiVec vector unit in their G4 CPUs). These usually produce much higher FLOPS ratings for a subset of appropriate tasks, making comparisons somewhat harder.

All that said, off we go.

A 35-year-old Cray In Your Pocket

Bottom line first: a 4th Gen iPod Touch (or an iPhone 4) is significantly more powerful than a Cray-1 from the late ’70s.

iPod Touch equals Cray1

The Cray-1 was sort of the start of the classic supercomputers; it was a big, cool-looking, appliance-sized thing that sold well and debuted in 1976, a couple years before I was born (and just before Apple started selling the Apple I).


Assuming Wikipedia is correct, the Cray-1 generally performed at about 136 MFLOPS, but could max out at 250 MFLOPS with carefully-tuned code that made use of its vector capabilities.

However, Linpack, which presumably uses more complicated routines, isn’t so kind to the Cray-1’s capabilities.  The Linpack FAQ actually has some figures for the machine (verified in the huge list of Linpack results linked from this page, down at the bottom); with the software available at the time of its release, it could apparently manage about 3.4 MFLOPS. A few years later, with better compilers, it turned in a much more respectable 12 MFLOPS.

As for the iPod Touch, there is, conveniently, a Linpack app in Apple’s App Store; the current (4th Gen) iPod tests at a little over 38 MFLOPS with it (my 3rd-gen Touch manages about 30 MFLOPS). Its ARM CPU (like the Cray) isn’t the strongest in that particular sort of calculation; in more optimized sorts of calculations using the right compiler and routines, it can apparently manage around 260 MFLOPS, which would be similar to what the Cray-1 could do under ideal circumstances.

Cray-1 Hardware Specs

The Cray-1 had a single, hand-wired, 64-bit, 80MHz CPU with vector capabilities, 1M 64-bit words of RAM (8MB in modern terms), one or more external 300MB hard drive units (which could be combined up to 4.8 GB, if I understand the manual correctly), and cost in the range of $5-8 million. It was freon-cooled and used (again, according to the manual) 115kW of electricity before you factored in storage and auxiliary hardware. Upgrades over the next few years offered versions with up to 32MB of RAM and 256MB of solid state storage.

Cray-1 CPU, at EPFL, Switzerland, photographed by Rama

A Cray-1 on display at EPFL in Switzerland (photographed by Rama)

Also, in addition to the main CPU unit (the tower-shaped thing above), which was about 9 feet (2.5m) wide and 6.5 feet (2m) high and weighed 5 and a quarter tons (4700kg) fully equipped, the Cray-1 also required two coolant condensing units, a power cabinet, two 150kW generators, and a smaller computer (closer to a modern desktop in size) that served as the user interface. The disk units were separate and rather large, as well.

iPod Touch Hardware Specs

A maxed-out iPod Touch (4th gen) uses a 32-bit, 800MHz A4 CPU (ARM Cortex-A8 class) with floating point and vector units, has 256MB of RAM, 64GB of solid state storage, and costs $400.  The iPhone 4, if you prefer, has the same CPU, 512MB RAM, up to 32GB storage, and costs $700 unsubsidized. The iPod weighs 3.6 oz (101g). The iPhone gets 2.75 hours out of its 5.2 Watt-hour battery running full-bore, meaning it uses a little under 2W (some significant portion of which is the screen backlight, GPU, and wireless hardware); the iPod Touch should be about the same.


Which is to say, in ballpark terms, you can walk out and buy a $400 iPod that has somewhat better performance than a Cray-1 from 35 years ago that would have cost you $8 million.  At 1/20,000th the cost it has 8x more RAM, 13x more storage, weighs 50,000x less not counting all the accessories, and uses at least 60,000x less electricity.


A 20-year-old Supercomputer On Your Lap

The Cray-1 was a hand-built bundle of circuit boards and wads of wires, and I’m comparing it to a pocketable media player/gaming doodad (or phone); I’m a lot more interested in a comparison between more modern, refined supercomputers from, say, when I was in high school, and a modern supercomputer-on-your lap, the high-end laptop I’m typing this on.

The early ’90s were an interesting time for supercomputers, so I like using 20 years ago as the comparison; there was a boom in the construction of computers that used hundreds of off-the-shelf CPUs in parallel instead of fancier custom hardware, and there was a corresponding rapid increase in the speed of such systems. It was, effectively, the start of the “modern,” massively-parallel supercomputer era.

So, my 2011 MacBook Pro 17″ versus 1991, give or take a couple of years.

The bottom line is that the CM-5—a top-tier supercomputer from 1993—is very similar in specs to my laptop, and if you go back an even 20 years to 1991, this thing I use everyday would probably have been the fastest computer on earth by a modest margin.

A MacBook Pro equals a Thinking Machines CM5 Supercomputer


To start with, I downloaded a copy of Linpack from Intel and ran it myself; I got about 38 GFLOPS. (In benchmarks that make more use of vector math, for reference, I’ve seen as much as 59 GFLOPS. The GPU is theoretically capable of a ridiculous 500-700 GFLOPS, if you believe AMD, but that’s even more specialized sorts of calculations with little to do with everyday, non-3D-graphics computing.)

For comparison, we’ll use the Top500 List’s Rmax values, which are measured real Linpack throughput. In 1990, pre-list, the fastest supercomputer in the world was apparently the NEC SX3/44R, which clocked at 23.2 GFLOPS. It has the same 4 cores, interestingly, as my laptop, although they were 400MHz vector processors, so not quite comparable.

In mid-1993, when the Top500 list debuted, the top public computer was a Thinking Machines CM-5/1024, clocked at 59.7 GFLOPS (the NSA had a somewhat faster CM-5 called Frostburg with 512 better processors and a lot more RAM, but it was classified at the time); the #2 that year, a 544-core version of the same CM-5, managed 30.4GFLOPS, and 512-core versions took the 3 and 4 spots.

I like the CM-5 as a comparison; it wasn’t necessarily the best supercomputer in absolute terms, and the company was horribly mismanaged, but it was promoted heavily at the time (here’s a promo video) and was more or less the first of the “mass market” (if such a term applies to less than a dozen sales) supercomputers. Plus, it was the first system to sit atop the Top500 list.

The CM-5 also looks rather sexy, even earning a cameo in Jurassic Park; it’s the big black series of towers with a bunch of scary-looking flashing red lights running the park (the lights were mostly for show, but did actually show the status of the processors).

The Frostburg CM-5 Supercomputer

The NSA's CM-5 "Frostburg," looking suitably imposing (photo by Austin Mills)

CM-5 Specs

The CM-5 used a variable number (up to 1024 in practice) of Sun SPARC processors; 512 was the most common (not that very many were sold) configuration.

I found plenty of info about the NCSA’s CM-5/512 (ranked #3 in June 1993, retired in 1997), so I’ll use that as the sample system.

It had 512 SuperSPARC I 32 MHz CPUs, 16GB of RAM (cheaper configurations had 8GB), and 140GB of disk storage.

The storage array was composed of 1.2GB 3.5″ SCSI disks with throughput of 2MB/s each, which were effectively in a sort of RAID-0 array; a paper I found (PDF) said that each 56-disk-set had a theoretical throughput of about 110MB/s, but in practice only managed 32MB/s reading and a much better 90MB/s writing, so presumably that 140GB array managed about 75MB/s write, 200MB/s read.

The CPUs alone would have drawn 7kW of power; I can’t find a number anywhere for the whole system, but probably at least 20kW, if not much, much higher; pricing was a weird, scandalous subsidy thing with DARPA, but somewhere over a million bucks. The individual cabinets were about 8 feet (2.5m) tall (this page has some photos with people for scale), and took up more or less a whole room; I couldn’t find any weight figures, but presumably several tons.

2011 MacBook Pro Specs

In comparison, my MacBook Pro has a four-core 2.2GHz CPU, 8GB of RAM (supports up to 16GB, if you’re rich), and a 120GB SSD (which I added; OWC 6G, SandForce 2200-based; about twice the performance of the stock SSD Apple offers) that can transfer data at around 500MB/s in both directions, plus a slower, old-fashioned spinning drive of 750GB that can manage around 90MB/s. The Sandy Bridge i7 CPU draws 45W at 2.2GHz, while the whole computer uses under 85W including screen, GPU, and battery charging. It cost me around $3000 including the high-end SSD; a more stock configuration with a smaller screen (but otherwise identical specs) would cost closer to $2500.


Which, when you put it together, means that my high-end but not-particularly-unusual laptop from 2011 is about 18% faster than a 1993 CM-5/512, has the same amount of RAM as the lower-end configuration, a similar amount of storage that can transfer data at about twice the speed (or, if you go with stock hardware, 5 times as much at half the speed), and uses at worst 1/200th the power, if not much less than that, for maybe 1/500th the cost.

A very close match to one of the best room-sized computers a few million dollars would buy you in 1993, and you can carry it in a briefcase and run it off of an internal battery for several hours.

Or, alternately, it would have been the fastest (publicly-known) computer on earth 20 years ago. If we compare it to our Cray-1 era of 35 years ago, it might well be faster than every general-purpose computer on the planet combined.

My Predictions About the iPad And The Future Of Computing

While I’m a fairly hardcore tech geek, AAW isn’t (currently) that kind of site at all. Still, I wanted to put some thoughts down in writing mainly so I can refer back to a publicly published opinion in a few years as proof that I either accurately predicted things, or was spectacularly wrong. On account of the thematic mismatch, I’ve intentionally backdated this post so it doesn’t get in the way of the anime goodness.

I also note that nothing here is original—it’s all been said by other people many times. I’m just reiterating the opinions I think are right.

Anyway: The iPad, iPhone, and where computing is heading.

First, just to note (and I had this opinion back when I bought my father an iPad a week after it launched), the people who were saying that it’s just a big iPod/iPhone and the ones who were saying it was revolutionary and was going to change everything were, of course, both right. It is just a big iPod, and that’s exactly why it’s revolutionary—Apple decided, I believe correctly, that the iPhone/iPod Touch is all the computer a substantial majority of everyday folk need or want. Its only limitation was the tiny screen, which inherently limits what you can do with it, so they put exactly the same device behind a bigger screen, and bingo, you have the portable computer for the proverbial everyman.

I’m a geek. I do tech support for a lot of people. And if there’s anything I’ve learned, it’s that most people do not know how to use their computer. Not because they’re stupid, it just takes a lot more thinking and understanding of the thing than they’re willing to put in, because it’s a tool, not a goal in and of itself. To use the long-since-beaten-to-death car analogy, I don’t know how to rebuild the engine of my car, but that doesn’t mean I’m stupid or shouldn’t be driving it. It just means that it’s a complicated thing with a simple user interface. So is the inside of your TV—you don’t know how it works, but you know how to turn it on, change the channel, and adjust the volume.

Computers aren’t like that. They’re supposed to be easy, but they’re not. They’re complicated things designed to give far more options and freedom than the average user needs—or, more importantly, wants.

The iPad strips away the unnecessary options and complexity and interface abstractions to let the average person do what they want—check their email, surf the web, watch some stuff on YouTube, play some games, and download some special purpose apps for whatever it is they  personally want their computer to do. It doesn’t require understanding much of anything, and—very importantly—it doesn’t break (in the software sense). I get paid good money to go to people’s houses and un-screw-up their computers, because keeping a modern all-purpose computer running smoothly is a dark art. I don’t care what you say about how easy or dumbed-down the OS is, or how bulletproof the software is, it just barely works, and people are afraid of breaking it by looking at it wrong. The iPad takes some of that away.

The real analogy is console gaming versus PC gaming. Leaving aside the update-creep in recent-generation console games, it breaks down pretty clearly, and has for the past roughly two and a half decades: If you buy a console, it does nothing but play games, and it isn’t going to look as fancy as an expensive gaming computer, but you’re pretty sure that when you stick the cartridge/card/disc into the thing and push the on button, the game will work. It doesn’t cost all that much, gives a good experience, and is bulletproof. To top it off, most software is heavily vetted (Nintendo Seal of Approval, anyone?), so while there is less of it than in a free-for-all, you’re pretty sure what is there will do more or less what it claims to.

Doesn’t mean that console games are better or worse than PC games, just that there are legitimate tradeoffs, tradeoffs that many—the large majority, depending on how you count—people are willing to make. Like me—I’d much rather spend a couple hundred bucks on a console that I know will be good for several years, and that will just work when I stuff a game in it than a couple of grand on a gaming rig that will be outdated in six months and will require regular updates, patches, driver adjustment, etc to keep in tip-top shape. Other people go for that, and more power to them. Just like other people go for turbocharged, tricked-out Honda Civics instead of reliable, un-tricked-out Civics or even stock, “easy” sports cars like a Corvette or Eclipse .

The point here is that, to date, the only option has been the equivalent of gaming computers—they do far more and are drastically more powerful than the average person cares about. The problem was that there wasn’t a general purpose console computer as an alternative. Until now.

Case in point: My dad. He’s in his mid-80s, and is a very smart guy. He taught me to use a computer, and is quite capable with his desktop computer. But when I got him an iPad—to use as a large-type e-reader only—he took to it immediately, and uses it more than his desktop. I don’t get asked technical questions or to fix things anymore, because it just works. Almost no learning curve at all, to boot.

Does this mean that “traditional” computing is dead? No. It just means that what we’ve treated as a “real” computer to date is, in fact, the souped-up, professional-grade monster that, fact of the matter is, the average person doesn’t actually need. Apple sees this, and that’s why they’re ahead of the curve and Microsoft—who as a company seems to still be convinced that the average user actually wants Windows—is flailing.

My prediction: Apple will, within a year or two probably, release a desktop iMac style computer that runs iOS. People will mock it as a toy, just like they did with everything else Apple has released in the last several years, but it will sell increasingly well. Within ten years—2020, which is a nice convenient Cyberpunk number—the majority of computing devices sold, desktop or portable, will run iOS, Android, or a similarly simplified, console-style OS.

Windows, MacOS, and Linux in their traditional form will still be around, but as “pro” machines for professionals, geeks, hobbyists, and people who actually need that kind of horsepower and flexibility. Photoshop jockeys, gamers, number crunchers, me. But everybody else—including “pros” when they’re not working—will use a console computer. Car analogy, “real” computers (which is a stupid term—“traditional” is what they really are) are big trucks, construction vehicles, and exotic sports cars; console computers are everything else you see on the way to work.

Second prediction: Nobody “wins.” I’m sick and tired of reading how Android or iPhone are “winning” or “losing” the smartphone war. It’s a war, yes, but only in that they’re competitors; there is no reason whatsoever that there be only one “winner.” It happened, more or less, with old-school desktop OSes in the ’90s, but this is not the ’90s. These days if it does email, the web, and maybe sudoku, it’s good enough for a lot of people—that’s all that matters. Document cross-compatability is dying, and along with it the idea that if your platform can’t run Microsoft Office it isn’t worth using. Microsoft probably realized this a long time ago, and that’s why they tried so hard to stall the web with IE6, ActiveX, etc—they could see it would eventually render their monopoly irrelevant, and that’s exactly what has happened. The barrier to entry is far, far lower now.

I don’t think Apple wants, in as much as a company can want anything, to own 100% of the computing market. I think they will be quite happy to own, say, 50-70% of the top third of that market, give or take, plus some portion of the middle tier. I’ll bet that the iPhone/iPad/iPod Touch/iWhateverIsNext ends up with, say, 20-30% of the new console computing/phone computing market in five years. And I think they’ll be perfectly happy with that, making snazzy devices and piles of money (don’t forget that Apple is making more profit from 3% of phone handsets than more or less the entire rest of the industry combined—I don’t think they’re too disappointed with that kind of “loss”).

Flavors of Android will probably have a substantial share, including a bunch of the bottom, as will, in all likelihood, the successor to Win Phone 7 (or whatever they’re calling it this week), and maybe HP’s WebOS and other new platforms as well. Nobody will “win” any more than anybody “won” the TV market, or car market, or refrigerator market, or any other appliance market.

We will, of course, see.

To repeat: Console computing is the future, but traditional computers won’t go away entirely either, and nobody is going to win the smartphone war.

AAW 3.0

So, after 16 months and 2 days of work (between the initial draft and today) AAW3.0 is finally live.  The final tally was 35 initial drafts and 10 additional final candidate refinement versions between concept and the design you see now.

Among the fun stuff the new look and feel ads is a full mobile version of the site, selectable (from a desktop browser, if you’re so inclined) via the On This Page stuff over on the left of each page. It should default to that automatically, unless you prefer the full version, if you’re using an iPhone, iPod, Android phone, Blackberry, or Symbian handset, although it’s only been tested with iOS hardware currently (let us know if you get the wrong one on your phone so we can work on that!).

Some of the things that people who’ve been to the site before might not notice have changed: I’ve completely rewritten and seriously expanded the company profiles, rearranged much of the older writing on the site, added a couple of additional old photo albums from Japan that I’d prepared before but never actually published, updated a few entries in the glossary, added a new history page documenting the evolution of AAW over the past 12 years (complete with screenshots and some links to the WayBack machine, so you can actually try the old versions for yourself if you’re into nostalgia and/or making fun of my crappy 1999-era design skills).

We’re in the process of re-recording and expanding on the audio samples that accompany the Japanese lessons; we have much better recording hardware now, so the sound should be considerably clearer, and I wanted to start from scratch so it all sounds as good as possible. Currently there’s just the beginning re-done, but now that the incredibly time-consuming rebuild is complete, I’ll be working on more as time permits. They are best heard as HTML5 audio in a decent browser (meaning Safari or Chrome 4+, Firefox 3.5+, and eventually Opera and Internet Explorer 9 ); it might work with QuickTime in older browsers, but no guarantees—do yourself a favor and upgrade anyway.

Old-time readers might be glad to note the resurrection of the screencap galleries from the days of yore; we’ll see how that goes, but small galleries, including commentary on the images, have been added to about a dozen shows, with more to be added steadily over time.

We’ve also added large-sized box art images to everything. Not just cheesy “grab what Amazon’s got,” either—Akemi personally scanned every DVD reviewed here that I own. That means over 200 extra-high-quality box art images for those who like such things (the other 200 are made up of the best images I could find; if you have a copy of one of the handful of ancient VHS tapes we couldn’t come up with anything decent for, and are willing to scan it for us, we’d be eternally grateful).  Note that you can click the box art on any review to get a larger version.

Our big new feature, in terms of “fun ways to find more stuff to watch,” is an analogy added to every single anime reviewed here, such that you can get a pithy answer to the question “What else is it like?” (and let me tell you, writing anything at all for 400 anime isn’t easy).

In the “new but not quite ready for prime time” feature department, we’re also experimenting with an amusingly unscientific graph of quality versus time for TV series; the first such experiment is on the Kanokon review:

Amusing quality graph for Kanokon

Here you can see in geeky detail exactly how Kanokon is lame, and where it gets particularly abysmal.

I plan on adding more of these to other TV series over time.

Along with that is our first experiment in side-by-side comparisons of Blu-ray discs versus the upscaled DVD version of the same title, as can be seen toward the bottom of the Ponyo image gallery.  I sort of enjoy comparisons like that, so as time permits I will add more.

Finally, in addition to the two newly posted reviews, I’ve updated about 50 older reviews with adjustments ranging from minor grammar fixes and availability corrections up through significant re-writes of about a dozen others. Some of these will be showcased in the new Pick Of The Week rotation (now fully-automated from a hand-selected queue, so if I’m asleep at the wheel you’ll still get the next pick right on time).

We’ve also decided to completely ditch a few features that hadn’t been properly updated in years, and have removed a few things that were going to hold up converting to the new look due to needing significant re-writes.  These orphan pages (mostly old song lyrics that needed serious updates on the translations) will eventually sneak back in as I revise them.

Of course, some stuff is no doubt messed up; if you find such a thing, tell us and get a chance to win this month’s contest!  (Actually, it’s technically next month’s contest, but you just get an extra week to find stuff that way).  I hope people actually enter, because I have a whole shelf full of DVDs here to pick from.

A few additional comments for the geekier types who might be curious:

Everything on the site is, or should be, valid HTML5 according to the current state of the draft spec, and all CSS is valid CSS3 with the exception of some -moz and -webkit additions to add support for rounded corners, shadows, and alpha channels to older versions of Safari, Firefox, and Chrome. We’re using just a bit of @font-face fun, with the attractive Helvetica Neue Ultralight clone Lane as a fallback for the page subheadings for people who don’t have the real thing. The site has been tested to be at least usable in Opera 6-10, Safari 2-5, Chrome 4-5, Firefox 1.5-3.6, Internet Explorer 5-8, Camino 1.6-2, Mobile Safari in iOS3 and iOS4 (which just launched today, but that was enough time to install it and test!), the default Android 1.5 browser, the Wii and PS3 browsers, plus some archaic oddities, for fun (Netscape 4, IE5 Mac, Firebird, Lynx).  It uses progressive enhancement, meaning that it’s totally usable even with no styling whatsoever on Netscape 4, all the way up to a bunch of little touches like rounded corners and drop shadows in modern Webkit or Gecko browsers. It looks pretty much perfect in good new browsers (last two versions of Opera, Safari, Firefox, and Chrome), and fine but without pretty touches in IE8. IE6 and IE7 get a somewhat dumbed-down version that looks similar custom-built for them, and while old versions of Opera and Firefox screw some stuff up they work and look pretty good. The only real disappointment is that the PS3 browser chokes somewhat (ugly but usable; the Wii amusingly is perfect, since it uses Opera instead of the embarrassingly underfeatured NetFront mobile browser Sony licensed) and IE5 Mac looks relatively bad (but also usable); not enough people use either for me to care much. (That said, c’mon, Sony, fix the browser already! It doesn’t even work on your own forums. Heck, any Android or Apple phone from the last year or two makes it look primitive.)

Interesting statistic: IE6 is now down to only 5% of AAW page views, which is a huge change from when IE6 was dominant and IE5 still had about 10% share when we did our previous design—it’s amazingly freeing to just design for good browsers and decide not to waste time on anything more than a simple version for broken old versions of IE.

Almost everything has been optimized for speed—code organization and optimization, minimizing HTTP requests, server-side compression, hand-tuned content caching, and more. On a fast connection page loads should rarely take more than a second—average over DSL after the big stuff has cached, on my 4-year-old laptop running Safari 5, is about half a second until you can start reading.  Even at dialup speeds (I checked!), after the initial 45-second delay to load the stylesheet and javascripts you should be able to start reading a new page within a second or so of clicking a link.

Of course, if you’re the geeky type and you think we’ve done something wrong, by all means, give me a hard time. That’s how you learn.

And that about sums it up—hope you enjoy looking around the new site!