I've been on more or less pure ice before, and it was difficult, but not THIS bad. I managed to navigate around without ramming into every damn thing. My only guess is that this happened somewhere where they don't usually get this sort of weather, and so nobody had proper tires. (A good set of all-weather tires can get you over the ice as long as you take it slow and don't do anything stupid.)
Other than that, I can't imagine why it would be this bad.
Posted by: Shamus at Wednesday, January 24 2007 02:13 AM (GDT1x)
There's a comment on Ace's site about the location (in Portland):
I recognize that as the intersection of SW 20th Ave. and Salmon St. It doesn't come across well in the video, but both of those streets have very steep grades. Steep enough, actually, that one would have to be a real dope to try and drive down either of them under those conditions.
I know those streets well. That's right next to the Multnomah Athletic Club. Portland gets really wicked ice storms sometimes, and those street always rack up a big score when it's icy out.
The Columbia River Gorge cuts a hole right through the middle of the Cascade Mountain range. It's colder east of the Cascades in winter, and sometimes a cold east wind down the gorge meets a warm front right there. The cold air goes under, on the surface, and the warm air goes over, and starts to rain. One time it put two inches of ice on top of everything in the city. It's just about the most slippery surface you can imagine, because it's ice covered with water.
That slope next to the Mac club is rather steep.
Long time residents know better than to attempt it after an ice storm.
The person driving the minivan was not actually a SoCal immigrant ... he was a retired Pendleton Woolen Mills CEO. But he was also 86 years old, and it looks like he has about a one second reaction time. By the time he hit either the gas, it was time to stop. By the time he hit the brakes, it was a bad idea ...
Posted by: kristopher at Wednesday, January 24 2007 12:48 PM (jcvPd)
OMG! Someone did it! My friends and I were emailing the same thing to each other when we saw it "it should be set to music - the Blue Danube works".
Posted by: Teresa at Wednesday, January 24 2007 09:08 PM (gsbs5)
Okay, water on ice - that would explain it. In the video it looks like snow on ice, which isn't too bad. Water on ice is something that I've seen, but never tried to DRIVE OVER the stuff. Yikes already.
Posted by: Shamus at Wednesday, January 24 2007 10:19 PM (GDT1x)
Mr. Pixy, are you in control of Ace's blog? If so, would you consider making his RSS feed open pages with 'new comment thingy' instead of the old-style comments? New comment thingy recognizes me correctly, but the other one seems unwilling to do so, hence my comments go uncredited because I'm too lazy to fill in the name every time.
This is particularly annoying because on two separate occasions I said something funny, but now I have no proof. Life is unfair without new comment thingy.
Posted by: Kevin at Thursday, January 25 2007 12:21 AM (H826O)
Actually, that particular video was probably of snow on ice. But the camera angle doesn't really make clear that the street there is tilted about 20 degrees. Once you lose control and start sliding, you won't stop until you reach the bottom or you hit something.
Not exactly. I host it, and do any technical support needed, but I don't control the content or design. So I could suggest to Ace that he change his feed, and I could make the change if he asked me to do it, but I can't just go and change stuff.
But there is a solution on the horizon: New Comments Thingy is part of the new blogging system I'll be installing at mu.nu next month. Once Ace switches to the new system, everything will work perfectly.
Trust me. ;)
Posted by: Pixy Misa at Thursday, January 25 2007 02:23 AM (GaSFI)
12Once you lose control and start sliding, you won't stop until you reach the bottom or you hit something.
Posted by: Pixy Misa at Thursday, January 25 2007 02:24 AM (GaSFI)
Processing 0.05, elapsed 0.0557 seconds.
42 queries taking 0.0192 seconds, 138 records returned.
Page size 98 kb.
Powered by Minx 0.81.19a.
Processing 0.09, elapsed 0.1028 seconds.
42 queries taking 0.0383 seconds, 138 records returned.
Page size 98 kb.
Powered by Minx 0.81.19a.
So, not quite twice as fast, and hence not quite back to the speeds of Old Minx on Nabiki.
I do have a solution for this: Replace the template interpreter with a compiler that custom-builds all the database queries. And I might even do that at some point, but not today. 55 milliseconds is something I can live with.
Oh, and the perennial favourite:
Retrieved from cache, processing 0.0, elapsed 0.0019 seconds.
Page size 98 kb.
Powered by Minx 0.81.19a.
This is updated from my previous post. I now have three new servers to play with, so let's see how they go.
Kasumi, Ukyo and Shampoo are Xeon 3060s: dual-core 64-bit 2.4GHz
Akane and Nabiki are Opteron 170s: dual-core 64-bit 2.0GHz.
Ranma is an Athlon XP 3000+: single-core 32-bit 2.16GHz.
Martina is an Athlon XP 2800+: single-core 32-bit 2.08GHz.
Naga is an Athlon 64 3200+: single-core 64-bit 2.0GHz.
Namo was a P4 Celeron: single-core 32-bit 1.7GHz
Lina: Pentium 4 2.6GHz
Amelia: Core Duo 1.66GHz
Haruhi: Core 2 Duo 2.4GHz
Core 2 Duo
I'm not sure where the difference between the Linux and Windows versions comes from; I'm guessing that Psyco would still be using Python's string libraries, and they're compiled using a different (better) compiler on Windows, perhaps Intel's. I'm still using GCC 3.4.6 (which is what CentOS installs); I might be able to do better with GCC 4.1, and I'll probably try that at some point.
Posted by: Pixy Misa at Tuesday, January 23 2007 02:16 AM (GaSFI)
One thing I noticed - the string optimisations that make Python so effective for this sort of thing only appeared in 2.4. 2.3 exhibits the same sort of behaviour as IronPython - the "String" benchmark is about 600 times slower.
Posted by: Pixy Misa at Tuesday, January 23 2007 02:23 AM (GaSFI)
SoftLayer are running a sale this weekend, and it looks like I'll be picking up three Xeon 3060 servers, with a total of 10GB of memory and 2.5TB of disk.
That's a whole lot of fast. It's equivalent in terms of CPU, memory and disk alike to five more of the Opteron servers we have right now.
I'll be bringing forward the commissioning of Kasumi by two months, but the deal I'm getting works out equal to two months free over the first year, so effectively I get a development and beta-test box for two months at no extra charge.
The page you see here takes around 40ms to produce.
The same page on my test server, on essentially the same version of Minx, takes around 60ms.
The same page on the new version of Minx takes around 95ms.
One interesting point there is that although Nabiki (the server) and Martina (my test box) deliver the same results in trivial Python benchmarks, the actual application runs significantly faster on Nabiki. Those same trivial benchmarks indicate that the Xeon 3060 servers I'm planning to deploy on will be 60% faster than Nabiki. I'm very curious now to see how that translates to the real application.
On a more depressing note, jumping from 23 fields per entry to 118 (I think those counts are correct) has a real and noticeable impact on performance. And I still have five tables to add to that join...
The performance hit doesn't seem to be on the MySQL side. At least, once something is in the query cache, it coughs up the data more or less instantly. Rather, it's the Python DB library handling all the fields. It takes time. On my test box, Python can retrieve from MySQL and stash into native data structures about 300,000 fields per second. That puts a hard limit on how fast Minx can run; the time taken to run the queries on New Minx on Martina is about 85% of the total time taken to generate the page on Old Minx on Nabiki. I'm going to keep on working to improve performance, but it looks like I'm fighting Amdahl's Law here.
My goal was to deliver complex pages in under 100ms and simple ones in 10; that's still achievable unless I really break something, but I'm going to have to borrow some of that performance from the new hardware.
Posted by: Andrew at Monday, January 22 2007 10:26 PM (7Q1Mu)
Yeah, it started fine with Minx and then dropped dead on the first page request.
It's almost certainly because I'm running with Psyco, which is a JIT compiler. But if I turn Psyco off, all the timings change. Psyco speeds up loops and numeric processing by as much as 100x, and string processing only about 2x, so it makes a huge difference when you're working out what to optimise.
Posted by: Pixy Misa at Monday, January 22 2007 10:39 PM (GaSFI)
Just to comment on what Steven has written about P2P:
First, I'll be getting my anime OP/ED site back up shortly, but initially it will be direct download. I have the bandwidth to spare, and I can manage the interface nicely with the current version of Minx, but I don't have BT support working yet.
Second, on BitTorrent itself. BT is different from most P2P applications (or at least the older ones) in that it's a file-swarming protocol rather than a file-sharing system. If you look at the standard BT client (which is open-source, and written in Python), it downloads one file at a time, and shares only that file with others, and only while you leave the program running. You can certainly end up uploading more than you download, particularly with the way I had the site set up before. With 50 or so torrents active, I could only allocate a limited amount of bandwidth to each, so downloaders would preferentially get data from other users (if there were any).
If you use a program like Azureus or μTorrent, it looks more like a traditional P2P app, but still has a difference: Only the torrents you specifically have active are shared. You can't share your whole anime directory; the programs just don't support doing that.
The wide screen is essential. Just about every show is widescreen these days. Crescent Love is widescreen, and it won the worst animation honors with the cantalope (or did it?). So it's the killer feature. The rest is rather unimportant for me, and of course I regret paying for the telephone handset electronics which I am not going to use ever. But it's not all that much premium.
Posted by: Pete Zaitcev at Wednesday, January 10 2007 05:17 PM (9imyF)
I meant the iPhone, and by the way, its resolution is far higher than that of any Pilot, at least in DPI figures. But it was brought to my attention that the memory capacity is miniscule on the device (4GB). I'd need to put all anime on SD cards. This makes it suck harder than PSP, which can do the same today, costs 2 times less, and has a bigger screen (physically bigger, not in pixels).
Posted by: Pete Zaitcev at Thursday, January 11 2007 12:33 AM (9imyF)
7I meant the iPhone, and by the way, its resolution is far higher than that of any Pilot, at least in DPI figures.
I double-checked. Both are 320x480. My Palm Pilot has a 3.9" screen, as against 3.5" for the iPhone, so the iPhone is 10% better on DPI. That's higher, but not far higher.
Posted by: Pixy Misa at Thursday, January 11 2007 01:17 AM (GaSFI)
Posted by: Pixy Misa at Thursday, January 11 2007 08:23 PM (GaSFI)
If you don't want the phone, but want the other goodies, just be patient. I'd guess not much more than a year before the next gen ipod launches.
In the keynote, Jobs made a big deal about the three devices in one (iPod, Phone, internet communicator) I don't doubt that one day there will be an iPod that is also an internet communicator that is not a phone.
I still haven't figured why I got a phone with a camera in it when that was the Next Big Thing. Maybe it was the horror in the eyes of the Radio Shack guys when I told them I won't buy a new phone more often than once every 4 years.
GPS is the frst useful feature I've wanted in a cellphone since the belt clip.
Posted by: TallDave at Thursday, January 18 2007 10:12 PM (odS+4)
Suggested retail price of $399; the Seagate 750GB drive sells for around $340, so that's a pretty good deal. Seagate launched the 750GB back in April at a price of $559.
This is good news, because I need a bunch of these. And SoftLayer do use Hitachi drives. (I just checked: Akane has Hitachi drives; Nabiki has Western Digital. Ranma, which is hosted at LayeredTech, also has Western Digital.)
The new drives are 5-platter models, so they don't actually represent an advance in areal density, but they do represent an advance in the amount of storage you can jam into a 1U server.
I was --><-- this close last night to pulling the trigger on buying a handful of 500 gig drives last night to stick in a media server. Glad I waited.
(not that I'd buy a handful of these right off the bat -- I just want them to drive down the price of the 500 gigers!)
In a 1U server do you just mirror drives? On that note -- out of curiosity, what do you do for backups? I retired all our tape drives in favor of just copying two redundant copies of all our data on all the different servers on two JBOD boxen, but I can't help feeling that my way is somewhat schleppy.
Posted by: bkw at Saturday, January 06 2007 01:55 PM (bRLba)
RAID-5 for files, RAID-10 for databases. (The 1U servers I use have four drive bays.)
And yeah, I back up my files by copying everything onto another computer. (Or archive it onto DVD-R if it's static data.)
Posted by: Pixy Misa at Saturday, January 06 2007 08:51 PM (GaSFI)