Meet you back here in half an hour.
What are you going to do?
What I always do - stay out of trouble... Badly.
Saturday, January 24
On Being The Wrong Size
There's a problem with the
Lord of the Rings and
Hobbit movies: The scale doesn't make sense. Little things, like the fact that the town of Dale is about as big as Minas Tirith. And big things like the dwarves' attempt to drown Smaug in molten gold.
If we assume that the volume of molten gold is about the same as an Olympic swimming pool (and frankly, it looks larger), then we're talking about 2.5 million litres of gold. That would weigh close to 50 thousand tons - 50 billion grams. Gold in our world is currently runs about $40 per gram.
That's two trillion dollars worth of gold right there. And never mind the hoard itself, which is much larger.
It's possible that gold is more common in Middle Earth than on Earth, but that just means it's less valuable, since it has little practical use in a pre-industrial economy. (It doesn't corrode, which is good, but it's soft and very heavy.)
And the dwarves pay Bard in silver, not in gold, and yet that is enough for him to risk his life to smuggle them into Esgaroth. Either the values of silver and gold are inverted - in which case the dwarves wouldn't be hoarding gold - or the economy of Middle Earth is bigger than Earth's - which isn't possible; they have nothing we'd even recognise as a city in the modern sense.
Of course, this particular part is Peter Jackson, not Tolkien, and The Hobbit is a children's story (and canonically, is told by Bilbo, who is not an entirely reliable narrator).
Still... Where is everybody in the Lord of the Rings movies? The world seems to be utterly depopulated. It's the movie equivalent of a Bioware game.
Posted by: Pixy Misa at
04:48 PM
| Comments (8)
| Add Comment
| Trackbacks (Suck)
Post contains 297 words, total size 2 kb.
1
The gold hoard is actually a Dwarven mine tailings pit?
Posted by: Mauser at Saturday, January 24 2015 05:47 PM (TJ7ih)
2
They're mining for iron and dammit, more of that useless yellow crap?
Posted by: Pixy Misa at Saturday, January 24 2015 11:56 PM (PiXy!)
3
The fight with Smaug really pissed me off, Jackson just pulled that one out of someplace dark and stinky. Made no sense at all, either technically, or in terms of the book's plot.
As for where everybody is, in the Lord of the Rings, I think it's worth remembering that the Fellowship were actively avoiding populated areas, because the spies of Mordor were known to be all over the place, and the last thing they wanted was for Sauron to get wind of what they were up to. That was quite explicit in the books.
Posted by: Brett Bellmore at Sunday, January 25 2015 05:18 AM (L5yWw)
4
A good point. Also, the area between Minas Tirith and Mordor
would have been depopulated; Minas Ithil was taken and Osgiliath destroyed, so there wasn't much reason to hang around. Looking at the map (for the first time in a long while), almost all of Gondor lay south and east of the capital.
So Jackson got that one pretty much right.
Posted by: Pixy Misa at Sunday, January 25 2015 04:24 PM (2yngH)
5
South and
west, that is. Mordor is to the east.
Posted by: Pixy Misa at Sunday, January 25 2015 04:42 PM (2yngH)
6
In the books, Middle Earth is fairly depopulated, partly due to the previous war against Sauron and partly the whole raids by goblins and orks and trolls and 'evil men' thing.
The other thing is that the whole continent survives on medieval era tech which doesn't allow that many big cities, there isn't enough food. Which raises the question, how many elves are there in Rivendel and where do they get their food from? Imports from Bree and the Shire?
Posted by: Rikto at Monday, January 26 2015 09:36 AM (zDlKl)
7
Also I hate 80-90% of the 2nd hobbit film. I'm glad I haven't had to watch the 3rd.
Posted by: Riktol at Monday, January 26 2015 09:37 AM (zDlKl)
8
Pixy, bug report. The HTML editor works fine in Firefox 34.0.5 but it doesn't work in IE 11 (11.0.9600.17501).
Posted by: Steven Den Beste at Wednesday, February 04 2015 02:38 PM (+rSRq)
Hide Comments
| Add Comment
Wednesday, January 21
Samsung U28D590D Review
TL;DR: Oh my God, it's full of stars.
Tech Specs
Resolution: 3840x2160
Input: 1 x DisplayPort 1.2, 2 x HDMI 1.4
Output: 1/8" audio jack
Power: External power brick
The
Samsung U28D590D is a 1.5th generation 4K monitor. Early 4K screens were targeted squarely at the professional market, with prices upwards of $3000. Second generation 4K screens, showing up now, are priced under $1000, still use high-quality IPS panels, but forego some high-end features like colour calibration.
The U28 sits below those, as a high-end consumer model, for consuming rather than producing high-resolution content. And it's priced appropriately; I paid A$499 for mine on sale; regular online prices range upwards from $549 to $699.
Out of the box it takes a few minutes to attache the stand (you'll need a large-bladed screwdriver, either plus or minus) and get it plugged in. The default settings are retina-searingly bright; I have the brightness and contrast turned down to 60 currently and it's still on the bright side.
This is a subjective review; I have no measurement equipment. But there are no static colour or brightness inconsistencies significant enough to notice (and there were on my old Dell U2711, a professional monitor), and no visible dead or stuck pixels (though at 4K they might not be easy to find).
Colours are vibrant and text is very sharp. It's not quite perfect - I'm judging it against my 2560x1600 Nexus 10 - but it's very good indeed.
The stand isn't as solid as it could be, particularly compared to the Dell, which is rock steady. I'd feel comfortable leaning on the Dell if I needed to climb on the desk to change the light bulb (and have); I'd never do that with this monitor.
This is a TN panel, so there is going to be some colour shift if you view it from an angle. Good news is that the horizontal viewing angle is as good as any monitor I've seen, including expensive IPS screens. Bad news is that if you stand up and look down at a 45 degree angle, white turns to blue-grey and other colours take on a distinct blue shift.
Both my desktop with its Radeon 7950 and my notebook with its Intel integrated graphics (Haswell CPU) recognised the monitor immediately and worked flawlessly over HDMI, albeit at 30Hz. What didn't work so well was connecting my 7950 over DisplayPort.
On one of the two mini-DisplayPort outputs on the card, the display shows graphical glitches on random horizontal bands every 10-30 seconds. On the other port, the whole screen goes black and then restores itself every 20-60 seconds.
Adjusting the resolution to 2560x1440 stops the problem, but that doesn't look particularly great. On the HDMI port at 30Hz, the display is rock solid. I suspect this is an issue with my DisplayPort cable - my card has mini-DP, and the included cable is full-size, so I picked up a mini-DP to DP cable from the corner computer store. I suspect it might be an older cable only rated for DP 1.1, and so not able to reliably carry the 4K@60Hz signal. I'll order another cable online and test that again.
Right now I'm running quite happily at 30Hz; you can notice the difference, but for work, watching movies, and light gaming it's not a really problem.
The remaining question is Windows 8's scaling. It's a bit of a mixed bag. I'm having more luck (for some reason) on my desktop than on my notebook (a 1080p 13.3" screen, so basically the same DPI as this). I'll report back on that aspect in a few days.
The on-screen display is driven by a little joystick on the back of the monitor; you can easily reach it from the front - it's just behind the lower right corner. It works quite well, and has all the usual amenities. I haven't yet tried features like picture-in-picture or picture-by-picture, but I have little use for them anyway.
Overall, it's a good monitor for its target audience, and at this price, a good buy. The big selling point is not this monitor in particular but 4K and high-resolution displays generally. : It's like washing the mud off your screen; everything is suddenly so clear. I'd find it hard to go back, and I've only had it a day.
Posted by: Pixy Misa at
11:08 AM
| Comments (3)
| Add Comment
| Trackbacks (Suck)
Post contains 726 words, total size 5 kb.
1
It's too bad that Windows and Mac OS GUIs didn't embrace scalable graphics until so late, which surely was a major reason that higher resolution displays failed to catch on until recently. I recall a few ~300 dpi grayscale CRTs back in the mid 80s, but likely cost, widespread color, and software problems prevented any significant adoption.
Posted by: Kayle at Friday, January 23 2015 07:05 AM (W8clb)
2
Yes, there were always a few high-res displays, for medical imaging and engineering and similar uses. The IBM T220 was a 4K LCD back in 2002 - but it cost the best part of $20,000.
Posted by: Pixy Misa at Friday, January 23 2015 03:55 PM (2yngH)
3
The Tektronix 4014 was 4K*3K in 1974. (I worked on it as my first job out of college.)
Posted by: Steven Den Beste at Saturday, January 24 2015 12:26 AM (+rSRq)
Hide Comments
| Add Comment
Tuesday, January 20
4K Or Not 4K
Guess the answer is 4K. Dick Smith has the
Samsung U28D590D on sale for $499 right now, and even had it in stock at my local store, so I trotted over there at lunchtime and got one.
This is one I considered and rejected before, because it's a TN rather than an IPS panel. Having now had a chance to look at one up close... Frankly, if they'd said it was IPS I probably would have believed them. It doesn't have the characteristic colour shift of TN when viewed at a sharp angle, though there is a brightness/contrast shift. I think it will do fine.
Also, at $499 for a 28" 4K monitor, I'm willing to forgive a few minor foibles. Just 15 months ago the only 4K monitor for sale here cost $4200.
The stand does wobble if you poke it, which is the first thing every review comments on. Samsung, spend another $5 on the stand for next year's model, okay?
Review later once I finish work for the day and have a chance to set it all up.
Update: The display glitches when I run it on DisplayPort at the full 4k@60Hz. It doesn't glitch on HDMI running 4k@30Hz. I'll try it with a different DisplayPort cable as soon as I can find one.
Running at 30Hz isn't a killer, but you do notice, so I'm really hoping that it will be happy at 60Hz with a cable swap.
Update: Some other people have reported these issues with this monitor and Radeon video cards, and the solution is to find a better cable. So I'll do that. Meanwhile:
Also on the plus side, I plugged it into my new notebook, and it worked immediately. Only 4k@30Hz, but many notebooks have trouble running higher than 1080p for an external display, so I'm happy with that.
Posted by: Pixy Misa at
02:25 PM
| Comments (2)
| Add Comment
| Trackbacks (Suck)
Post contains 311 words, total size 2 kb.
1
What do "TN" and "IPS" mean?
Posted by: Steven Den Beste at Tuesday, January 20 2015 03:21 PM (+rSRq)
2
TN is "twisted nematic". IPS is "in-plane switching".
I'm a little fuzzy on exactly how this works at the molecular level, but as I understand it, in a TN panel the liquid crystals are perpendicular to the screen, where in IPS they're parallel - in the plane of the screen, hence the name.
The result is that IPS panels look much the same from any angle, where TN panels can have a pronounced colour shift at sharp angles - in cheap screens, the colours can be completely inverted if you view the monitor from a 45 degree angle above. This monitor showed almost none of that.
Posted by: Pixy Misa at Tuesday, January 20 2015 03:57 PM (PiXy!)
Hide Comments
| Add Comment
Monday, January 05
Next Problem
Most of my remaining errors were being caused by Windows Error Reporting.
This thing is awful if you're trying to run older software. Turn it off and suddenly things work again.
Posted by: Pixy Misa at
03:56 AM
| No Comments
| Add Comment
| Trackbacks (Suck)
Post contains 33 words, total size 1 kb.
Sunday, January 04
Ticket #12350 (New Defect)
Hardware Virtualisation support not detected if Hyper-V installed.
Which means I can't create 64-bit virtual machines in VirtualBox, which was kind of the point. (And CentOS 7 is 64-bit only.)
It's something of a relief, though, as my first thought was a hardware or BIOS issue.
Update: So I turned off Hyper-V (which required two reboots, but was easy enough) and now it works.
And that's the hardest problem I've had with this install so far.
Why haven't I run into this before? Because Hyper-V is only included in Windows 8 Pro, and I've only used Unpro before now.
Posted by: Pixy Misa at
10:47 PM
| No Comments
| Add Comment
| Trackbacks (Suck)
Post contains 103 words, total size 1 kb.
How Did We Cope?
Steam.
Library.
Click. Scroll scroll scroll. Shift-click.
Right-click.
Install.
Disk Space Required: 406849 MB
Disk Space Available: 3131066 MB
OK.
more...
Posted by: Pixy Misa at
06:31 PM
| No Comments
| Add Comment
| Trackbacks (Suck)
Post contains 31 words, total size 1 kb.
Saturday, January 03
Kei
So, my New Year's Resolution was to set up Kei, my shiny new Windows PC that has been waiting to be assembled and installed for... More than two years.
And done. Not perfect, but done.* Finished about 3AM today.
Since I had a lot of parts to play with, even though many are two years old, the result is pretty good: 8-core 3.6GHz CPU, Radeon 7950, 32GB RAM, 2 x 960GB SSDs, 5 x 2TB disks, and a Blu-Ray burner for burning all my Blu-Rays. 22 USB ports (8 x USB 3 and 14 x USB 2), and 12 x 1/8" audio ports for that perfect 16.2 mix.
Currently Kei is hooked up to a spare 1600x1200 monitor and I'm sharing the keyboard and mouse via
Synergy, which works perfectly except that the Windows
Do you want to allow this program to do stuff? dialog flicks me to the primary monitor on Nagi every time. (Update: The solution for this is the obvious one - plug the keyboard into Kei and share in the opposite direction. Works great now.)
Now I have a bit of an install party going on. I built Nagi late in 2008 when the original Kei went flaky due to faulty memory. So Nagi has everything installed. Despite a few hiccups over the years, that adds up to 841 applications.
I'm hoping to whittle that down a bit. Really, I just need my JetBrains IDEs, Microsoft Office, Adobe Creative Cloud, Visual Studio, Sony's creative software, Notepad++, Cygwin, VirtualBox, Chrome, Firefox, Firefox DE, iTunes, Clementine, Steam, Origin, Xshell, Xftp, Zoom Player, Python and Ruby and about a hundred libraries for each, uTorrent, Bryce, Carrara, Hexagon, Poser, virtual machines running CentOS 7 and Ubuntu 14.04...
Hmm. This might take a while.
The other thing I need to do is decide on a new monitor; my Dell U2711 is visibly declining. Right now, my short list is either the Philips 40" 3840x2160 monitor or a couple of Dell's new U2515H 2560x1440 screens. The Dell is about half the price of the Philips and has about half the pixel count, so it all balances out. The only real problem with the Philips monitor is that it has a static television-type stand, where the Dell is height adjustable and rotates on all three axes. Well, that and 40" is huge.
* As usual, the disk layout isn't what I wanted. Windows 8.1 denies the existence of my motherboard's RAID functions, and it also denies the existence of either my Adaptec RAID card or my cheap dual-port SATA card, so after much messing around I'm left with RAID-0 for the local drives and one drive not doing anything at all. Plus the SSDs aren't on the first two SATA ports, so it seems that Windows put a boot partition on one of the disk drives, so if that ever dies Kei won't boot any more.
But those minor issues aside, the dual SSDs, eight cores, and 32GB of RAM really makes a difference.
Posted by: Pixy Misa at
01:56 PM
| Comments (5)
| Add Comment
| Trackbacks (Suck)
Post contains 502 words, total size 3 kb.
1
I don't think I have 22 USB devices, let alone need 22 ports for them!
Great googly moogly, Pixy's created Skynet.
Posted by: Wonderduck at Sunday, January 04 2015 04:04 PM (jGQR+)
2
I'm pretty sure I have 22 USB devices if I gather everything together. But this is a slightly older motherboard, so it has
lots of USB 2 but only a couple of USB 3 ports (14 and 2 respectively). I added a 2-port USB 3 card, then I found a cheap 4-port USB 3 card and added that as well.
Posted by: Pixy Misa at Sunday, January 04 2015 04:44 PM (PiXy!)
3
And I've installed 399 programs so far, so I'm on track to exceed Nagi's count later this week...
Posted by: Pixy Misa at Sunday, January 04 2015 04:45 PM (PiXy!)
4
Can you uninstall the "French Spam" attraction plugin for The Pond? :-(
Posted by: Wonderduck at Wednesday, January 07 2015 05:37 AM (jGQR+)
5
Ooh, now I'm getting NFL football spam, too!
Posted by: Wonderduck at Wednesday, January 07 2015 01:11 PM (jGQR+)
Hide Comments
| Add Comment
Thursday, January 01
The Landauer Bound
We all know that while Moore's Law is still holding - just - Dennard Scaling failed in the early 2000s. Clock speeds have remained unchanged for more than a decade.
What most don't realise is how fast we are approaching the thermodynamic limits of computational efficiency. If Koomey's Law holds, we will reach perfect efficiency - the dreaded Landauer Bound - by 2048.
Mark that date in your calendars, boys and girls! 2048 is the year it all ends.
Posted by: Pixy Misa at
11:19 PM
| Comments (6)
| Add Comment
| Trackbacks (Suck)
Post contains 82 words, total size 1 kb.
1
The reason clock rates leveled off was that we ran screeching into Planck's Constant.
Posted by: Steven Den Beste at Thursday, January 01 2015 11:57 PM (+rSRq)
2
Or at least, Planck's Constant's Little Friends. As I understand it (I don't even play a quantum physicist on television) it's essentially a long list of second- and third-order effects of quantum uncertainty.
So, for example, as transistors get smaller, you have to reduce the switching voltage so that you don't damage them over time. A lower voltage also means less current leakage with narrow insulators. But a lower switching voltage directly leads to slower switching. So although the smaller transistor can switch faster, if you actually
do that, all your painfully-earned gains go out the window.
There's a review up on Anandtech right now where they overclock an 8-core CPU. Getting it from the default clock (which varies between 3 and 3.5GHz depending on core temperature) up to 4GHz uses an extra 70W of power and raises the core temperature from 54C to 65C, which is into the realm of diminishing returns but not awful. But getting it to 4.4GHz - just another 10% - requires a further 120W of power (on a 140W CPU) and pushes the temperature to a scorching 93C.
Posted by: Pixy Misa at Friday, January 02 2015 12:41 AM (PiXy!)
3
Thing is, if the transistor is smaller and the voltages are lower, then you get more quantum leakage ("tunneling") across the junction. At a certain point there's as much leakage when the transister is "off" as current flow when it is "on", and the device isn't useful any longer.
The only solution is to increase the voltage, with all the bad things that leads to.
Posted by: Steven Den Beste at Friday, January 02 2015 03:21 AM (+rSRq)
4
I don't even play a quantum physicist on television
If need be, I can get
a real one for you. Assuming he's not jetting to some overseas conference or another, that is. Or swing dancing. Or both
Posted by: Wonderduck at Friday, January 02 2015 11:18 AM (jGQR+)
5
Another limit we're coming close to is the size of a silicon atom. As strange as it might seem, FETs these days are so small that you can count the atoms in the junction.
Posted by: Steven Den Beste at Friday, January 02 2015 05:01 PM (+rSRq)
6
Maybe
reversible computing will arrive in time to save us. (I recall reading a paper about these ideas back in the late 90s, hmm... I think it was something by Tom Knight, which he called adiabatic circuits)
Posted by: Kayle at Saturday, January 03 2015 06:53 AM (W8clb)
Hide Comments
| Add Comment
76kb generated in CPU 0.0264, elapsed 0.6474 seconds.
55 queries taking 0.6327 seconds, 371 records returned.
Powered by Minx 1.1.6c-pink.