The Human Mind... boggles.

31 August 2007 — dan

Last night I watched, or at least attempted to watch, an episode of The Human Mind (subtitled "And How To Make The Most Of It"; this debut episode is reviewed here by someone less annoyed than me).

The Human Mind managed the remarkable feat of being staggeringly dumbed down, yet also, frequently, incomprehensible.

Robert Winston's made some great documentaries, but this sure as hell wasn't one.

For me, the high point was a guy who can flawlessly remember ten consecutive shuffled packs of cards. We were told that he did so by walking around London, looking at landmarks, associating mental images of things like teddy bears and cakes with suits and numbers, and then associating, say, a teddy bear eating a cake with Tower Bridge in order to be able to remember that this point in his walk was the Jack of Diamonds.

Just do that 519 more times, and you've got it!

It's just that simple!

Yes, that really was all the explanation we got. Perhaps something that'd make sense of it got left on the cutting-room floor.

As it stood, though, I found this part of the show very much like watching Look Around You, but without the humour.

The episode also featured a fireman, whose story was told over about three hours of brightly coloured stock footage of fire and explosions and men with big hoses, without which the audience was presumably expected to go and watch the football instead, or just drool until we all died of dehydration.

This fireman once saved a bunch of other firemen by ordering them to leave a burning building where, a mysterious intuition told him, something awful was about to happen. Which it did.

After eight or nine more hours of stock footage - and interview footage of the fireman, who was interviewed in a slightly smoky room, to make sure we didn't absent-mindedly start thinking he was a pastrychef - we were told that he'd actually seen very clear evidence that a backdraft situation was developing. And then he just got a bit of a hunch before he added it all up consciously.

This doesn't sound like a very big deal to me.

But apparently it was worth a third of the episode, all by itself.

Oh, and the beginning of the episode sang the praises of the Durham fish oil trial, in which omega-3 oils apparently made kids smarter.

Except that study is complete bollocks [latest update here!]. There is no reason whatsoever to suppose that fish oil supplementation does anything for brain development in otherwise well-nourished children.

I suppose Winston's just phoning this one in from the voice-over booth and trousering the proceeds.

You wouldn't think he'd need the money, but I don't know why else anyone'd want to put their name on crap like this.

Posted in MiniReviews, Science. 9 Comments »

Overlord update

1 July 2007 — dan

I've played enough of Overlord now to get a proper feel for the game (one of the seven bosses dead, three of the four minion colours collected). I am continuing to like it.

The PC control system works pretty well. The console version of the game uses an analogue stick to let you tell your minions where to go, on the occasions when you're not just saying "go in the direction I'm pointing", but need to steer them around the map. The PC version lets you do this by holding both mouse buttons and moving the mouse. This usually does not result in half of your minions drowning.

The different flavours of minion are also pretty easy to manage, because the game deliberately limits you to ordering one type around at a time, or telling them all to move at once. I presume they were tempted to include some kind of RTS-type grouping so you could order two or three flavours around in a group; I'm glad they didn't, if only because that would have further tempted them to make fights you could only win by doing that.

The level design is also good. The levels so far look like the kind of "natural layout" game levels in which you're forever wandering around places you've already been trying to figure out where the hell you need to go next, but they are not in fact that kind of level. Which is good, because there's no map display.

The level structure - move this to access that, get a shortcut back to the start when you get to the end, all that stuff - is also competently done.

I've seen a couple of bugs - the game locked up once, and there's other occasional oddness like minions that're carrying something getting stuck on an obstacle even after you've removed it. The bugs are easy to work around, though.

Back in the real world, I keep feeling the urge to order our cats to charge out, kill something, and bring back treasure.

The first part's probably quite doable, but the cats unfortunately do not share my opinions regarding what constitutes "treasure".

Posted in Games, MiniReviews. 6 Comments »

It's good to be bad

30 June 2007 — dan

Yes, as I anticipated, Overlord is fun. And it runs fine at decent resolution on my GeForce 7900 GT, once I disabled a couple of pretty-features in the config.

It's hard not to love a game where you come upon what is obviously Bag End, but cannot be bothered to stoop to enter the silly little round door.

Instead, you just send your minions to bash that door down, flood inside, smash and kill everything in sight, and then bring anything of value out to you.

(Ideally, you'd be picking your nose while you waited for them to return. Perhaps in Overlord II.)

Posted in Games, MiniReviews, Nerdery. Leave a Comment »

The continuing quest for a decent USB drive box

3 May 2007 — dan

As I've mentioned before, ordinary everyday external USB hard disk boxes do not ever tell their drive to go into sleep mode.

How much harm this actually does is questionable, because spinning up hard drives causes motor electronics and bearing wear, just like running the drive all the time. There's doubtless some point on the duty cycle graph below which using sleep mode does more harm than good.

But since external drives are very seldom the main drive for a computer, they usually don't need to be spinning for a very large fraction of their lives. So they probably will last quite a lot longer, not to mention use less electricity and make less noise, if they're spun down when they're not needed.

But virtually no external boxes have that feature. They use cheap USB-to-ATA bridge hardware that can't do spin-down at all. There's no standard way to even send a spin-down message via USB (you can do it via FireWire) - but you could still use extra software, or just a little switch on the box, or something. But nobody does. You have to buy a NAS box if you want a sleep feature, which is overkill if all you need is a plain external drive.

If you want your external hard drive to stop spinning, you've got to turn off the box. For a lot of the cheap ones that means unplugging the power. Then the DC lead falls down behind the desk.

M'verygoodfriends at Aus PC Market, though, now have a cheap stopgap solution.

Noontec drive box.

This box comes from international megabrand Noontec/BlueEye, whose Web site is as I write this not responding to hails.

(Readers outside Australia may find the same product being sold under yet another weird name, by a company that may even have a working Web site!)

The box has a few points in its favour.

One: It's cheap. $AU77 without a drive, including delivery anywhere in Australia.

Two: It accepts SATA drives, up to 500Gb in capacity. No good if you want to use an old PATA drive, but new SATA drives are now often cheaper than new PATA ones.

Three: It turns off when it loses USB signal, either because the host computer has shut down or because its data cable has been unplugged.

The box also has a simple power button on the front. That'd be a selling point all by itself, since it means you don't have to fumble around the back of the thing to shut it down.

The automatic power-down function is not, regrettably, matched by an automatic power-up function when you turn your computer back on again. You have to power the box up manually. Just poking the power button is not a huge chore, though, and for my money it definitely beats coming back to your computer after a weekend away and discovering that while the PC's been off all that time, the bloody external drive you forgot to unplug has been spinning for sixty completely pointless hours.

There's a second button on the front of the box, that runs some automatic backup software of unknown quality. You may find that useful. If you don't (or aren't running Windows), just don't install the software and the button will be harmless.

The Noontec box is made out of aluminium, so it ought to get decent convective cooling if you set it up vertically in the supplied stand. Any 7200RPM drive should be OK in it, if you don't live in a tropical jungle.

If this box cost $AU150 or something, I'd find it hard to recommend. For $AU77 delivered, though, its convenience features make it a winner. It even comes with a padded bag!

Australian shoppers can click here to order it from Aus PC Market.

Posted in MiniReviews. 5 Comments »

Mouse of champions

12 April 2007 — dan

Microsoft's IntelliMouse Explorer 3.0 was, arguably, the best mouse a right-handed user could get when I reviewed it - and compared it with Logitech's show-off dual-pickup MouseMan Dual Optical - back in 2001.

Explorers old and new

And, as you may have heard, it has returned.

The final decision back in '01 came down to what shape you preferred, and left-handers were left out in the cold then, as they still pretty much are now.

But the Explorer 3.0 felt good to little old right-handed me, and it worked well. The two side thumb buttons are in just the right place, and the mouse feels neither too small nor too big. I like a teeny mouse for use with a laptop, when you often have a crummy wrist angle and need to hold the mouse with your fingertips from above to avoid strain injuries. But with an ergonomically correct desk setup, a big, though sculpted, mouse like this works for me. Perhaps I'm getting set in my ways in my old age, but I've kind of settled on the Explorer 3.0.

(I don't want a cordless mouse; yes, they work perfectly well these days, but I prefer a bit less weight and zero battery concerns. I keep my mouse cord organised with a simple weight, which you can readily make yourself; that may have something to do with the disappearance of the WireWeights company. More elaborate cord management contraptions are still on sale!)

Explorer 3s seem to last pretty well, too. My mouse gets a whole lot of use, but I can count on at least three years of service from an Explorer 3.0 before the cable goes flaky or the wheel starts mis-counting.

Microsoft were left behind in the feature-chart race, though. So they retired the 3.0, and created a new and awful version 4.0 of the Explorer.

This whole post is very much the outside scoop for gamers, of course, but the Explorer 4.0's suckage centred around its new and allegedly fantastic "tilt wheel", which you could not only roll up and down and click for the button-3 function, but could also tilt left and right for horizontal scrolling.

The tilt function made the click function hard to use, and they deleted the clicky detents in the wheel rotation that're essential when gamers want to accurately select a weapon.

So people who liked the old 3.0 started paying premiums for new old stock on eBay. Microsoft eventually noticed this, and reintroduced the older model.

Apparently the new Explorer 3 has a faster sensor chip in it, or something, but the change isn't significant enough that Microsoft bothered calling the new-old-mouse the Explorer 3.1. It is, for all practical intents and purposes, the same as the good old 3.0.

Except now, as you can see, it's dark slate-grey, with only slightly cheesy matte silver side buttons.

Here in Australia, m'verygoodfriends at Aus PC Market sell the new Explorer 3 for $AU69.30 including delivery anywhere in the country. Australian shoppers can click here to order one.

That's really not a bad price at all. Microsoft's fancy-pants Razer-collaboration Habu costs twice as much, and Logitech's flagship corded mouse, the G5, is not a lot cheaper.

Microsoft now seem to be calling the Explorer 4.0 just the "IntelliMouse Explorer", and OEM (no-fancy-box) versions of it can be had in the States for quite a bit less than the price of a new 3.0. At that price it's a perfectly OK desktop mouse, but it's still no good for many games. Aus PC Market have given up selling it.

Interestingly, Microsoft's main list of mice doesn't include the 3.0 at all any more. Look under "gaming products", though, and you can find it, next to the Habu.

The Explorer 4.0 tilt wheel also lives on in some even swoopier products. I'm not itching to try any of them, though.

Viva 2001!

Posted in MiniReviews, Nerdery. 13 Comments »

The storage appliance, not the guitar

11 April 2007 — dan

David, from Western Australia, writes:

During my daily trawl across the more interesting places on the net, I encountered this little device, as I’m sure you have.

Naturally, I was interested... until I saw the bloody price. I could build my own happy little 1.5Tb RAID box for the cost of that thing and still have change left over.

But it did bring something to light. Its storage mechanism seems intriguing, for as much I like RAID, I hate sacrificing to the parity god, especially if I’m using 500Gb drives.

Which got me thinking. This Drobo thingummy seems to be an expansion of JBOD with some kind of parity calculator. Now granted, the implementation of this in a shiny box is reason you pay for this thing, however my question as it were is this:

Have you ever encountered a method of implementing this on a PC? I’ve done a bit of hunting and come up with squat, but I do remember you writing about JBOD related things on several occasions, so I turn to you as a font of knowledge, oh mighty Dan.

The Drobo box does indeed look like an interesting little thing, and certainly seems to be a step toward the home mega-storage Nirvana I've written about before. But it has its limitations, chief among which is exactly the same Parity God sacrifice you'd make with a do-it-yourself RAID rig.

If you only read the glossy main Web site, all you'll find is that the manufacturers allege that their storage scheme uses "advanced storage concepts such as virtualization, but it is not a derivative of RAID". Well, who knows what the heck it is, then, but it's clearly doing something analogous to parity RAID, otherwise it wouldn't be possible to yank a drive and upgrade it any time you liked. So, obviously, all of the data on any given disk must be reconstructable from the content of the other disks, and the amount of capacity offered up to the Parity God must be at least as much as the size of the biggest disk.

Drobo aren't really hiding this, though; their knowledge base confirms it. I suppose the thing could theoretically offer more capacity if it did real-time compression, but that'd make it hilariously slow and not gain much for the kinds of files that many Drobos are probably going to end up containing, anyway.

So if you add a 1000Gb drive to a Drobo that already has three 500Gb drives in it (ignoring real-gigabytes versus drive-manufacturer-gigabytes for now), you'll take your aggregate capacity from 1000Gb to 1500Gb. If you've got four 500Gb drives in your Drobo and swap one of them for a larger one, the actual capacity won't increase at all!

So I suppose it's basically working like RAID 4, but with support for dissimilar disks.

As with normal RAID, if you change the drive setup your Drobo may be churning away for hours. And, just like a rebuilding RAID array, a disk loss during this period will poleaxe the entire array. The Drobo does, however, have battery backup to prevent a mere mains interruption from clobbering your data. So you should factor the price of a UPS into your equivalent-PC calculations.

In answer to your actual question, no, I don't know of any remotely user-friendly way of doing this same sort of thing on a PC. FreeNAS could be a thing if you don't want to take the more traditional route of pirating a really expensive version of Windows, but plug-and-go it ain't.

Drobo really are quite straightforward about these capacity issues, though, including the powers-of-two versus powers-of-ten capacity rip-off. The knowledge base makes clear that four "500Gb" drives will only give you 1397 formatted gigabytes.

(It's also a neat hack that the thing reports 2Tb capacity no matter what drives you actually install in it.)

In answer to another of my immediate questions about the thing, it is also apparently possible to swap drives from a dead Drobo into a new one. But there is of course no way to read Drobo disks on anything else. Software RAID (and, in theory, quality hardware RAID controllers) gets around this vendor lock-in problem, but for home users it's not that big a deal, if of course you make backups. Which home users don't.

Oh, and as someone else has noticed, the Drobo site tagline currently says "whose" where it ought to say "who's". That's the kind of attention to detail you love to see from a storage vendor, isn't it?

Posted in MiniReviews, Nerdery. 8 Comments »

An extraordinary coincidence

22 February 2007 — dan

My productivity has just dropped to zero.

I cannot imagine what connection that might have with the copy of Supreme Commander the nice postman brought me.

I only dabbled with the widely-pirated beta version, so the retail version is pretty much new to me.

Except it isn't, because everybody who tells you that Supreme Commander is Total Annihilation on steroids is exactly right. A TA player will feel very much at home.

It's certainly taken me back to 1997, when I was playing Total Annihilation on my K6-200 (with crazy-fast Tseng ET6000 graphics card and useless-for-TA Monster 3D Voodoo Graphics accelerator) at 1024 by 768 and winding the game speed down as soon as battle proper commenced, to keep the frame rate out of slideshow territory.

Now here I am again, doing exactly the same thing ten years later on my 20-times-faster PC.

Admittedly, I am now playing with one monitor at 1600 by 1200 and the other one (which doesn't seem to be as useful as you might think, but is so cool that I cannot countenance disabling it) at 1024 by 768.

And it's all pretty and 3D accelerated.

And everything's larger in scale, more like real military units in size-to-weapon-range terms. SupCom can also support bodaciously hyper-gigantic maps, though there's not much point trying to play on one today unless you've got a PC that fell through a time warp from a thousand years in the future.

By and large, though, SupCom still feels awfully TA-ish.

All this same-old-same-old stuff does not mean that SupCom is not a fine game. It appears to be one, from what little I've seen so far, even if those timeless unit pathfinding problems are still there. And very noticeable to dorks like me who insist on making 200-unit armies on the early tutorial levels.

(Pathfinding problems are a big obstacle when you start playing with the nifty formation and coordinated attack features. I suppose the developers could have smoothed it over a bit by allowing re-formationing units to cheat and walk through each other, but I'm sure some munchkin would immediately figure out how to use that to make 40% of his units invulnerable at any given time.)

(Oh, and I can't say I'm a huge fan of frickin' SecuROM copy protection, either, but presumably that'll be turned off a couple of patches down the track, as usual. And it only stopped the game from starting the first time I tried to run it. Fingers crossed.)

So SupCom is not just TA warmed over. It's a cool modern RTS that does stuff that nobody else's RTS games can do. It's just that a lot of the stuff that it does was already done by TA, because TA was so very far ahead of its time.

Command queuing, smart unit selection hotkeys (yes, control-Z to select all units the same as the ones currently selected still works, though it doesn't seem to be mentioned in the manual; does anybody know how to set map bookmarks?), the ability to issue commands to factories to affect the units they produce... all TA stuff, and all beefed up in SupCom.

(The perfect example being telling a factory to send its ground units over the hills and far away, then setting one or more air transports to assist the factory, which will cause the transports to ferry the units to their destination automagically. I don't think transports-assisting-other-transports works right yet, though.)

And SupCom, like TA, is still a RTS game for people who hate micromanagement. I don't think micro is bad; I'm just not into it. So Blizzard-y games with lots of unit abilities that you have to play like a piano if you want to do even slightly well leave me cold.

(Yes, I'm aware that high-level TA degenerated into an evil clickfest, as people discovered that vast crowds of missile trucks were unanswerable early on, while giant flocks of stealth fighters, carefully managed, were just as invincible later.)

The Gamereplays Supreme Commander section looks like the best site to soak up info on the game at the moment. Almost all of the replays they have online are for beta versions and fail amusingly...

Supreme Commander replay error

...on the retail release, but that'll change.

I'm pleased to see, as I peruse the replays-I-can't-play list, the irrepressible Gnugs mixing it up in SupCom. Now we old-timers need only see a gigantic Swedish Yankspankers sign rotating over the SupCom battlefield to feel perfectly at home.

(Although SY might, of course, be a little busy.)

The Yankspankers were the people responsible for the TA Demo Recorder, which allowed games to be recorded and played back via a sort of benign man-in-the-middle attack. SupCom has its own record/playback system built in, of course.

If you don't have a sufficiently bitchin' computer to play SupCom, I strongly recommend you pick up a copy of TA - and the Core Contingency and Battle Tactics expansions as well, even if the hovercraft were all useless. Going back to TA today is not like going back to the original Command and Conquer; the 256 colour graphics look distinctly dated, but TA's gameplay is still great.

(Kingdoms was kind of interesting, and prettier, but the original is better.)

Because TA doesn't use 3D acceleration at all (zillions of tiny polygons were un-acceleratable by 3D cards of TA's era), it's also an excellent game for computers with crappy 3D adapters, including boring business boxes and your Aunt Mabel's dreadful Dell.

Any current CPU will push TA along at warp speed at as high a resolution as you can fit on your monitor, and it's a young enough game that you can play online using nice normal TCP/IP, rather than having to do some bizarre tunnelling trick with IPX/SPX or something.

If your PC is large and veiny enough for SupCom, though - the minimum requirements are not completely laughable, but more of everything is a very good idea - forget its little brother.

Get in on the ground floor of the connoisseur's RTS for the next ten years.

Posted in Games, MiniReviews, Nerdery. 9 Comments »

3.8 trees per day

29 November 2006 — dan

Because I Just Don't Learn about responding to people's news-announcement e-mails, please find herein my review of Uniblue System's "newsworthy and highly topical" LocalCooling software.

LocalCooling interface

(Mmm, alpha transparency. Picture of woogle not included in standard install.)

LocalCooling is software for Windows PC power saving. According to the press release, "if enough people join in, the Local Cooling Community could seriously cut CO2 emissions from computers".

Not that computers themselves actually emit carbon dioxide, of course, but the power stations that run them usually do.

Global warming, LocalCooling. Geddit?

LocalCooling is from the makers of the harmless but largely useless SpeedUpMyPC, which is one of those apps with "273 Five Star Reviews!" because it's got an affiliate program and all the reviews are from penny-ante sites (like mine) that want to cash in (a temptation I've managed to resist. Feel free to throw me a tip in appreciation of my shining incorruptibility).

If you want software to clean your Windows registry, disable useless things in system startup and so on, there's plenty of excellent freeware you can use. Commercial system-speedup utilities often do more than the freeware does, but they very seldom do anything extra that actually needs to be done. Voodoo memory management twiddlers, for instance, are extremely popular among grunty dads and other Power Users, but (a) Windows' memory management has actually been very good in every version since Win2000 and (b) memory's so cheap these days that you'd do better to put the $US30 you're considering spending on memory-optimisation software towards the price of another whole gigabyte of RAM, which'll set you back less than $US100.

Aaanyway, LocalCooling restricts itself to power management. It's got a slick and easy to use interface, but the single functional difference between LocalCooling and the normal Windows Power Options is that LocalCooling lets you set your PC to shut down if it isn't used for X hours (or even minutes, which could be a good way to annoy someone).

To make that setting more useful, you can tell LocalCooling to not shut down before a particular time (presumably resetting at midnight or something; I was never clear on when you were allowed to start feeding mogwai again, either). You can also tell LocalCooling to never shut the computer down if some particular single application is still running.

If you're in the habit of forgetting that you've left your computer on, then I suppose this could be of some use. But it's not as if the standard Windows power saving features don't already have standby/hibernation options. A stood-by PC only draws a few watts, and a hibernated PC draws no more power than it would if it were "off".

Like many modern appliances, PCs are never really "off" while they're still connected to a live power socket. The "vampire" draw of all those little red standby lights all over your house, not to mention the watts used by old heavy linear plugpack power supplies and the smaller but still non-zero number of watts drawn by newer lightweight switchmode plugpacks, can add up to a significant amount of juice.

But I don't have a problem with "off" PCs keeping that one little LED lit on the motherboard and waiting for things like wake-on-LAN signals that will never come, because that extra power also maintains the system BIOS configuration data. Loss of the "CMOS" config is not nearly as big a deal today as it was back in days of yore when losing it meant your computer forgot about all of its hard drives, but it can still be a pain. Modern PCs also have easily replaced lithium coin cell battery backup for CMOS data, instead of the nasty rechargeable or soldered-on lithium batteries that a lot of old PCs had, but that battery will last effectively forever {like, ten years plus} if the PC's got wall power all the time.

Anyway, only if your PC is unable to hibernate is turning it off really a better option. There's no power drain difference at all.

And... that's about it for LocalCooling's actual functions. It lets you set monitor and hard disk sleep timers as well, but so does the standard Power Options.

LocalCooling also has a neat-o tracker that shows you how many "Trees", "Gallons" (of oil) and "kilowatt-hours" you've saved by using it, but this tracker suffers the most common failing of bad science - arbitrary measurements. Not entirely arbitrary, since it's calibrated in watts, but it's got arbitrary fudge factors for the number of watts your PC consumes.

(And also for how those watts convert to trees and gallons, but that doesn't matter much.)

LocalCooling identifies your CPU, monitor, hard drive(s) and graphics card, but it can only actually have a clue about the real power consumption of the CPU and graphics card. Even then it is basically guessing, because it can't tell how hard either is working.

It evaluates my Manchester-core Athlon 64 X2 as being good for 59 watts of power consumption, which is a reasonable enough average figure for a well-used computer. But there's a big difference between idle and working-hard power figures; an easy 50 watts for something like my computer.

Since, like a lot of nerds, I run distributed computing software 24/7, my CPU usage is pegged at 100% all the time and that extra fifty watts ought to be added on.

This is something that a future version of the LocalCooling software could do (this one is only v1.03), since CPU usage is easy to measure and map with moderate accuracy against processor type. But it doesn't do it yet.

The same applies to the video card. Modern high powered 3D cards really are high powered when they're doing complicated stuff in 3D mode, but not when I'm sitting here typing text.

LocalCooling estimates only 20 watts for my GeForce 7800 GT. That actually seems to be something like its idle power consumption, with full-blast 3D mode consumption up around 60 watts (Nvidia specify the 7800 GT as an 85 watt card).

LocalCooling could build a proper database of monitor power consumption, too. They have to depend on Windows' opinion of what monitors are connected being correct, but beyond that they currently just take a wild-ass guess about how much power each monitor consumes. For my 21 inch Samsung CRT they reckon 70 watts, which is OK for a 21 inch LCD, but is a significant under-count for my monitor (which is specified at 125W, but probably more like 100W most of the time). 70 watts would be way off if the "21-22 inch Screen" they detected was actually some old three-foot-deep NEC behemoth.

Oh, and LocalCooling doesn't yet seem to know if you've got multiple monitors. It only notices the primary. That's bad.

LocalCooling makes a couple of less objectionable guesses, too. It assumes each of your hard drives draws eight watts, which may or may not be an overestimate but doesn't matter a whole lot, since the difference won't add up to much unless you've got a bunch of 10,000RPM drives or something.

LocalCooling also tacks on an extra 15 watts for "everything else" - system fans, motherboard chipset power, expansion cards and so on. That's fine; few enough users have a wind tunnel PC full of 30 watt Delta fans.

LocalCooling also doesn't seem to have any integration with the low power modes that various modern processors have. They can step down to a lower clock speed or even voltage on demand. Of course, you should be able to make that stuff happen anyway, possibly with integrated Windows drivers and maybe with extra software from Intel or AMD, but LocalCooling is happy to duplicate the basic Power Options features, so it ought to do this too.

One day we may be running PCs with hardware monitoring that can actually track the current being drawn from the power supply, perhaps even with some granularity so you can see how much juice each subsystem requires.

We're not doing that now, though, so it's impossible for software to really accurately track system power.

LocalCooling could definitely do a considerably better job of it than it does, though. I hope it does in future versions. At the moment all it does is let you join a "community" of people all vying to see how much power they can save. Which is good. But the actual power saving numbers are pretty much random, and LocalCooling does not, currently, actually do anything of importance that you can't do with plain unvarnished Windows.

Posted in Electricity, MiniReviews, Software, Windows. 5 Comments »
« Older posts
Newer posts »