danger: technical stuff about the future!

An article posted only a couple hours ago on Ars Technica regarding memory bandwidth vs. many-core computing made me raise their eyebrows. The gist of the article is, as computers are now, adding more cores to the processor will not continue to be beneficial since there is a limit to how fast information can be read or written to the RAM. As you add more CPU cores, each one gets less and less memory bandwidth; the number of CPU cores has been increasing faster than memory bandwidth lately.

So, the article raises a valid point: the number of cores per socket is effectively limited in a practical sense by the total rate of memory access. I believe the problem with this argument—and the reason that we will be seeing vastly more than 16 cores in the future—is that you can always just split everything up.

You can put fast and compact DRAM on-die, right there on the processor die, as another level of cache (with maybe 128MB of memory per core? That’s not too hard). You can have multiple sockets: as massively multicore becomes the order of the day, more powerful computers will have appropriately more CPU sockets; the space and importance assigned to card slots may well lessen, as well.

And just giving an individual memory space to each processor isn’t such a bad idea either. So, say you have 128 processor cores, each one with maybe 128 or 512MB of dedicated, independent memory (perhaps on-chip)—you can then just have another 32GB or so of core memory shared between all of them and everything works out great. You can assign large chunks of work to each core, and the total memory bandwidth is off the charts.

There should also be a way to have data read from the core memory be uploaded simultaneously to an arbitrary set of CPU core caches, to make synchronizing data sets between cores easy as pie.

The point I’m making is, the article isn’t wrong in what it’s saying, but I believe it has an improper focus. It should focus not on the naysaying, pointing out all the reasons why technology can’t move on. (Don’t be silly, technology can always move on.) The article should focus on what should be changed about current computing architecture to adapt in the future.

edit: Basically, don’t freak out because current computing architecture has limitations. There is always some major aspect of the way computers are built that has to change next. If we’d done it perfect the first time, we would all be demi-gods ascended above the pithy material plane by now.

tl;dr – Don’t say that computers are limited because there are bottlenecks. Just make a whole ton of little computers and put them in the same box all wired together. Only, you know, with advanced technology. So that they go fast. And stuff.

gauss chariot

I think it would be fun to make a map in Hammer (the Source engine map editing kit) where the people who work on it just add a few things per day, slowly expanding the contents of the map until there is a gigantic building filling up the entire 0.5×0.5×0.25 mile space. Then we could go around trying to finish it off and compile it (haha, it would probably take about a week to compile on a regular machine). I figure we could just rule out moving parts and map triggers, since stuff like that would just overwhelm the engine, plus they’re one of the hardest things to do when you’re mapping. I think getting the cubemaps right would be overwhelming enough, if it’s even possible.

Either way, I think I’d be willing to work on something like that.

Important things I left out of the last post:

Someone actually developed a volumetric 3d display that displays dots on thin air. I must be frank, I totally did not see this coming. It can display about a hundred dots per second so far, and can actually display two to three meters away from the projector. That’s right. Just the projector, and empty air.

You might laugh when you hear how they do it though. It’s kind of like a humorous mockery of Ye Olde Starre Warse holograms, because it actually superheats tiny pockets of air with a focused infrared laser, creating a flash of light and a small plasma explosion. (Here’s an article about it, and the official press release translated into english). So it totally works and everything, but it’s really loud.

Something that will delight fans of technology and newer user interfaces: Johnny Lee has done a number of really neat projects which save hundreds and hundreds of dollars off of expensive equipment by merely using a wii-mote to do the same thing. (Specifically: using the wiimote for 3d head tracking, multipoint interface with your fingertips in the air rather like Minority Report, a digital white-board you can put pretty much anywhere…) It’s extremely impressive, and you need to check it out.

Among all the weird things that happen to city wildlife, I think this particular link in the food chain ranks rather high up on the scale of the weird and unexpected. I mean, pelicans eat FISH, not… certainly not that.

My room mate signed me up for a 440 gloss page magazine called Bridal Guide… I got in my mailbox and had no idea what to think. Then he put up bald-faced lies about it for almost a day before casually bringing up that it was totally his fault and on purpose. I encourage you to do this as well, it’s a great prank; you can get free magazine subscriptions right here, with no commitment. Seriously, it does work. Just… well, this is better for doing in dorms where you’re not going to be living there the next ten or twenty years; if you’re too concerned about causing them bulk mail problems, you could always just double check to make sure that they don’t distribute the addresses to anyone.

One of the more silly brainchildren of the half of my brain devoted entirely to ludicrous engineering (the part that will forever be deeply enamored with the Space Fountain) is the house heated by bulk mail. The idea is, get yourself signed up for as many cupon newspapers, thick magazines, and free, worthless publications as you can, then stockpile them and use them to heat your house via a clean paper-burning stove, which should also include a catalytic converter in the chimney stack to cut down on any nasty chemical emissions due to weird inks or plastics that get burned along with it. (Mind you, I’m not really overly concerned about emissions in general, but seriously, it’s coming out right over your house.) If it worked, you’d basically be heating your house to free, with fuel being delivered to you sporadically by the national mail service.

That shit would be so cache.

ray tracing is a good thing

Imagine if you will, a laptop computer for which plugging in is optional, which is powered by fuel. Or a laptop that charges in three minutes and lasts for days at full power. Or one of those sonicare vibrating toothbrushes, the size of a regular toothbrush, that still last a full week without recharging. Feasible electric cars, capable of going as fast or faster than the cars we have right now. A camera that you don’t have to keep recharging every time you use it. James Bond-like equipment, like his magnetic climbing grapple; robots such as you might see in Ghost in the Shell or some other science-fiction.

All these things can and, if I may, will happen, soon after we make some real breakthroughs in portable power sources. A good marking point would be when you can power an SUV for a mile over flat ground with a device the size of a nalgene bottle. When we get to that (actually modest) goal, then we can start doing some cool things.

Imagine, if you will. Imagine a screen that sees where you are and projects the screen differently into each of your eyes. Like a window into another world, such a screen could show you anything, in beautiful, nearly flawless 3d. Or perhaps even a similar screen for which each pixel could change its appearance from every angle, as a tiny projector. Too much screen data to display state-of-the-art moving graphics from every angle, such a screen would still be perfect for static or slowly-changing scenery, sublimely beautiful from every angle. Given good enough eye targeting and a high degree of resolution and accuracy, such a screen could also display different images to each eye of a large number of people simultaneously. (Think of that, no more screen-looking!)

Retinal laser-projection displays, on the other hand, have even more potential. While they’re probably not practical for viewing by multiple people, the resolution and fine detail achievable is simply staggering.
We’re talking about the possibility of a resolution comparable to the actual native resolution of your retina itself. Not only that, but get this: you could actually detect the focus of the eye and change the displayed image accordingly. So essentially, it’s possible to construct a display nearly indistinguishable from normal real-life vision.

So that’s the science I was talking about. With display technology in mind, ray tracing is the future of graphics. Since it’s actual light simulation, for all practical purposes there is no limit to the amount of processing power you can pour into it before it stops improving.

And… do my eyes decieve me, or does Wikipedia have RSS feeds for every single page now? Actually, it appears to be an RSS feed for all of Wikipedia. I wonder what uses that could have.

edit: They do indeed have individual feeds for every single article, and several other things. See Wikipedia:Syndication for details.

how to dry a klein bottle

I am actually pleasantly surprised at my readership here; it’s pretty impressive. So far I’ve gotten traffic from Google for “Bill Nye” and “carbon nanotube ultracapacitors” (once each). Evidently I’m getting most of the traffic from somewhere else; maybe one of my readers just keeps opening the same page a bunch of times. So seriously, if you’ve been reading these posts and haven’t commented, just drop a small one somewhere; if you just dropped by from some random link, posting a comment wouldn’t hurt; if you’re disgruntled with this boring load of videos and commentary because I should be writing about something else, just say so here and I’ll try my best.

(Bleh, WordPress updated their whole dashboard/options/writing pages since I joined just a couple days ago. I don’t think I like it as much.)

Forward brave soldiers! I smell manflesh ahead!

This talk that this guy (Clifford Stoll) does is brilliant. The man is a genius, it’d be really neat to actually hang around and talk to him, he’s done so much. He also has a company that makes and sells Klein bottles. Just browse around in the site and read, his sense of humor comes out in every area. For instance, how do you dry a Klein bottle?

And this video, this ADVERTISEMENT, which I saw about a week ago, made me melt and explode in hopeless, impotent anger. For you see, the video itself is mediocre, and spends almost no time actually showing the details of the car itself, and 80% of the time gloating and puffing themselves up about how scientific and futuristic they are. I’m sure it’s a nice car and all, but seriously now.

And THEN, starting at 2:30, the guy winds into this “someday, think of the possibilities” spiel about how, you know, you use electricity to compress air, and they’ve also made an electric generator that runs on compressed air… and some day… this car… will run FOREVER, for FREE, on perpetual motion. It makes me hurt inside.

This conversation ensued when I showed the video to chafez:

chafez: compressed air as a source of energy?
chafez: hmm
chafez: apparently someone found air that will compress itself…

chafez: whoaw
widdershins: what
chafez: sure air takes energy to compress
widdershins: you finally heard him say it didn’t you
chafez: this is supposed to be a joke right?
widdershins: no, dude. it’s an AD

widdershins: ok i just watched the perpetual motion thing again
widdershins: …
widdershins: and i want to KILL HIM
chafez: lets do it
chafez: in such a way that we cannot be blamed
widdershins: yes
chafez: such as
chafez: put him in a cart
chafez: on a cable
chafez: over a cliff
chafez: with the only thing holding him up
chafez: being a motor
chafez: powered by compressed air
widdershins: generating electricity
chafez: compressed by a generator
widdershins: for an air compressor
widdershins: yes
chafez: and we will tell him that he is perfectly safe
widdershins: yeeeesss
chafez: because prepetual motion is holding him up
widdershins: get in, sir, you said it yourself

the unreal winter

I had a long, unusually coherent, and unsettling dream wherein Seattle was hit by an extremely high-yield nuclear blast. It was very… well, unsettling.


Speaking of unsettling, I did not remember them making magnets this large (I mean, look at at that). Check out the supermagnets down below. For reference, have you ever played around with those rare earth magnets before? They look kind of like these, you can buy them at ThinkGeek. And they’re really really hard to take apart when they stick together. Yeah, well those are the lower quality ones. These are HIGH quality, and large enough to kill you dead. The force of impact between two magnets of this size would be enough to completely shatter both magnets, as well as probably send magnetized metal shards everywhere, injuring you. Walk through an area with metal things in it, even ten feet away, and if you’re carrying one of these magnets they will fly off the table and break your bones.

I wonder how they took those pictures? It’d be hard to do without damaging the camera.

Last time I remember looking at United Nuclear, they wouldn’t sell magnets approaching that size unless you had a research grant, because they were afraid of lawsuits in the likely event of grievous injury. Evidently that’s all been ironed out now, because anyone harboring two or three hundred bucks and the irrepressible desire to be in the possession of a life-threatening magnet can acquire one.

That said, I would just LOVE to someday make a cube magnet six inches on a side out of the 2x2x3″ block magnet. The equipment necessary to assemble the dang thing would be expensive, the process of actually doing it would be incredibly difficult, the 18 component magnets themselves cost $175 each plus whatever large shipping and handling costs they levy, but the resulting object would be a solid-state, everlasting, lurking menace worthy of anyone who’s taken over the world. You could just mount it in the center of a large, empty room, and… have ceremonies in there, or just play around or something.


Flixxy is a pretty neat site. From what I can tell, they just link to videos that are really cool, eliminating much of the dross from YouTube and several other prominent video sites. Many, many fascinating things can be found just by browsing around inside. Not only that, but they eliminate video comments and channel subscriptions, which I am all for. I mean, they don’t actually host the videos themselves anyway, so I guess it makes sense.


In the field of holographic displays, various research projects have popped up the past five years or so. One particular Scientific American article introduced several, one involving projecting an image onto a spinning disc, one involving a screen with two images interlaced vertically and a thinly barred screen (like a diffraction grating only larger) revealing each image to only one eye. Another actually projected frames onto multiple screen surfaces at different depths, which when used together create a very realistic impression of a smooth 3d image. The interlaced screen is little more than a very complicated system so that you don’t need to actually wear 2-screen googles (it did require head-tracking). The other two appeal to my sense of completeness better, especially the spinning disc one.

There are basically three categories that I can think of for 3d displays.

The first merely displays different images to each eye, like with old-school 3d goggles, the interlaced screen, and the one method that will eventually win for efficiency, awesomeness, and quality in personal display devices: retinal laser projection, which excites me particularly.

The second category traces light patterns into the air, unfortunately on a moving surface or some substrate; holographic projectors like you see in sci-fi movies barely have any practical implementation at all. The advantage and disadvantage of this method is that while it is tracing a true 3d image in the air, visible from every angle you can get a vantage point from, projected images don’t occlude anything behind them; that is, they don’t appear to be solid, you can see right through them. This means that you can see the internal workings of devices if they’re shown, but this method would not be suitable for showing three-dimensional movies.

The third category projects a different image at every angle. From what I’ve seen so far, it’s only practical to do this in a horizontal fashion, so that the image does not change when viewed higher or lower, only distorts some. With camera- or head-mounted tracking, you can in fact make it appear correctly when viewed from closer to above it, but that doesn’t really matter usually. It works great, because as you can see, projected objects appear to be occluded.

Another, more interesting and by all indications undemonstrated category of device (which I shall refer to as category 3a) would be a SURFACE that appears different from every angle. Such a device, if made practical, could be made into something the size and shape of modern LCD monitors. The key would probably lie in making microscopic projectors and lenses less than a quarter of a millimeter square on the end, then forming them into a gigantic array ten, twelve, thirty inches in diagonal. Such a device could be used as a regular flat boring 2d monitor for normal applications, then periodically switch to a deep, near-realistic window into another world. Like the Wizard of Oz, seeing such a transition would be no less breathtaking than a transition from washed out sepia to full, vibrant color.

Category 3a devices needn’t be flat, though. You could also form them into a cylinder, or perhaps a sphere, for full 360° viewing pleasure.

When you consider the big picture, the first and third categories will probably endure for a long time, while the second is slightly more limited in its practical uses. The more information you display, the harder it is to make out. For practical purposes, the category 3 devices can be made to do everything category 2 devices can do, with a little more computational work involved.

It occurs to me I should consider lasting practicality. Category 1 devices are of course already plenty practical; categories 2 and 3 may win or lose based on the necessary size of the device itself. Both kinds could hurt you if you touched them, since they’re spinning so fast. The real issue of size lies in the ability to project onto the device from below. This is a simple matter with category 2 devices, but with the spinning mirror in a category 3 device this is a more difficult proposition. The bulkier projection equipment in the demo I linked to is actually mounted above the mirror itself, making the design large and impractical for home use. If you inverted the mirror, then, you could easily project from the bottom, but it would only be viewable from so high up. Angling the mirror nearer the vertical would help too, but this would result in a wider base. So all things considered, spinning-mirror model category 3 devices need some revision in form to be ready for home application, but should be sufficient once that is accomplished.

With all that said, I’m still waiting for the vast improvements needed in the color selections and brightness/contrast of modern 2d display monitors – namely, exponent-curved 48 bit brightness levels and 4 or 5 frequency displays to more accurately represent actual vision. Until they get that right, 3d displays should stay in museums.

Edit: More specific information on retinal laser displays. There’s a few other articles as well on in-depth design, which I can’t seem to find because I read them in some magazine in Canada, and can’t find them again.

oh and about recently

I was going to say something terribly interesting, I thought, but it turns out that it was just this: that I went and tried to turn in my programming assignment yesterday (Thursday, I work on a subjective-day system). I mailed it in like it seemed apparently asking for, but it turned out it’s not due until NEXT week. Nonetheless, I’m glad I did it, and the professor seems happy about it. She was somewhat confused (Oh, that was you), but when you get down to it, it took about five minutes to write the actual code. And it was fun, since it’s reversing sound files and stuff. So yeah.

I like Daft Punk.

And something that I just ran into. On some sites, specifically Microsoft now, I actually have to concentrate sometimes to read CAPTCHA messages (try refreshing that a few times, you might see what I mean). If they aren’t just keeping ahead of the curve, and they’re in fact actually doing this to keep automatic bots out of their email system, then it spells the very first sign of flattening wavy seas signaling the slow rising of the true AI behemoth rising from the freezing watery depths. It seems like every few days recently, I notice something that makes me think just how close to the truth Shirow Masamune may have been. (Mr. Shirow wrote both Ghost in the Shell and Appleseed, futuristic mangas which were eventually made into television and/or movie form. He’s rather a visionary as far as future technology goes.)

atomic batteries to power

After catching up on Brendan‘s recent bloggings, I was pleasantly blown off my feet by recent progress in the field of exoskeletons. However, this serves to hilight the one area that will almost undoubtedly be the most important field of technological advances in coming years: power sources. The main problem with these exoskeletons and other autonomous devices is the lack of a suitably lasting and quiet power source. The noise you most often hear on a robotic prototype such as these is a constant BZZZZZZ noise — this is coming from a small, internal-combustion-engine powered generator. While these are certainly powerful enough, they are highly inefficient and more importantly, very noisy, which makes them highly impractical for many military applications.

Recent developments in batteries and other such direct energy storage devices is promising. Improvements in recent years allowed the construction of battery-powered rappelling devices, such as you’ve probably seen in James Bond, just much bulkier and not quite as sexy.

One particular type of battery that is particularly exciting is the carbon-nanotube ultracapacitor, which has made significant research progress in recent years. Essentially, using carbon nanotubes can increase the surface area of electrodes in a capacitor dramatically, allowing for sturdy, compact capacitors with an energy density equivalent to that of standard chemical batteries today. The main difference is, they last many times as long (since they undergo no actual chemical changes) and can be recharged and discharged in an extremely short period of time, and be capable of extremely powerful bursts of energy.

Other possibilities for quiet, portable high-powered energy sources might be something like fuel cells, which have been in the technological background for some time now. Remaining far too experimental and expensive for consumer use, they’ve dropped out of the public view for the past few years.

In a rapid change of topics, I recently purchased a gas-powered airsoft blowback pistol – a KWA Glock 18C, semi-automatic or full-automatic selectable. It’s very hard to have more fun with an airsoft gun than you can with a blowback gas pistol.

And on the anime front, Kimagure Orange Road is pretty dang good. Rest assured that this torrent download is 100% legal in the USA now that the license has run out. And don’t forget to watch the two movies, which finish off the series nicely (I’m taking Connway‘s word on this one).