If you can make your lab/company fork for it, man, don't hold back!!!
To complement your nice Tyan motherboard, get one or two (XXX check for physical sizes) of them realizm 800 from 3dlabs. They are the only 16 lanes PCIexpress videocards I know of. Not sure about GPGPU, but at 3840 x 2400, solitaire is bound to look amazing... especially if you can get some nice 9.2Mpixel displays as well: High end videocard without a matching display, what would be the point? Check for instance the IBM T221).
Interesting idea! If you can work out a floating point location and colour value for each grain (AFM anyone ?), you could map these characteristics onto the center of your output image pixels, using a function of the distance between grain location and pixel location. Resulting values in whatever you want (8 bit unsigned, integers...)
Depending on your output image resolution (which you choose), the number of grains contributing to a single pixel would vary. Also, grains too far away from a pixel center don't contribute. Measurement errors are averaged out. Interpolation could also go temporal when suited (across frames).
Keyword: scattered data interpolation, RBFs, shepard.
I used shepard on some 3-D scattered datasets (confocal microscopy). Reconstruction works well, although the algorithm is quite hungry. RBFs may be a better option.
Who knows, when the technology catches up (disk space, microscopy, algorithms) this may be the norm in film restoration... At that point, rendering the film on your cheap 128 bit per RGBE channel, multi giga pixel handheld display would also be done in real time...
I am booting linux 2.4 via syslinux and a usb flash smartmedia (16 megs SMs and readers were incredibly cheap)
I got a nice logo up using syslinux. The booting goes like:
displaying the logo
displaying loading linux.....
displaying loading init_rd.....
the screen is cleared and the words Uncompressing linux... Ok, booting the kernel.
I found the relevent bit of code in linux/arch/i386/boot/compressed/misc.c
I also found a reset_terminal() in linux/drivers/char/console.c but tweaking it doesn't seem to help.
I could go for the bootsplash patch, but that adds nearly 200K to the (compressed) kernel.
I don't really want to as I'd be happy to just have the Uncompressing linux... line under my syslinux logo (640x480x16 colours only adds 10K to the distribution). After that, all the kernel messages are sent to/dev/ttyS1 so they don't appear on the screen.
So, anybody knows where in the kernel source (preferably 2.4) is the screen cleared? Any clues I could follow?
I checked the specs and saw the interconnect they used for the cluster is SCI provided by a company called dolphin. 64 bit cards, works with linux. Kind of expensive, though... SCI reminded me of an old project aimed at transporting TCP/IP packets over SCSI... There's even a page on sourceforge now and some benchmarks.
Is SCSI P2P used in real world clusters though? How does it compare to SCI or gigabit ethernet? Price? Performance? Status of the project? No idea...
I tasted some green beans that guy was roasting in a food exhibition (he was selling bread ovens with a rotating coffee roasting bit inside) and I just couldn't get over how nice the smell was (both bread and coffee). Coffee being my favourite drug, I'm really envious you manage to grow your own, even if it's only a bush or 2!
Has anyone checked if the game gear will take 2300mAh NiMH batteries? It's pretty much the same problem with digital cameras: put new alcalines in... and cry
Others actually have internal 10 bit ramdacs but don't allow driver access to them. Such a pity.
The matrox parhelia is such a beast.... I mean matrox does provide gigacolor as either standalone or as a plugin for a limited number of applications, but maybe an easier way to play with that resolution is through full-screen opengl where 32 bit pixels can be used as 10+10+10+2 bit alpha.
I'm thinking wrapping glui in python, but haven't got round doing that yet. Don't other cards also provide 10 bit resolution through opengl? might be worth a look. That's why I thought doing some opengl mojo is still way more compatible than using proprietary plugins.
Skype's GUI is QT
+ they already have a linux version. So I'm pretty confident they will have an OS X version... if they had enough money to buy the QT TrioPack license, that is!
I may be way off with my assumptions, but bear with me:
I presume you are americam;
I presume you don't value bottled water.
Fair enough, you can buy shit with sugar in for half the price of bottled water. Maybe all this sugar is also what keeps you going. Again, fair enough.
The simple point about bottled water is that it tastes actually quite nice, and certainly much better than tap water. It has other uses too:
If you want to make yourself a decent espresso (illy maybe?) and use tap water, then you've just ruined it. This is also true for tea although I'm out of depth, here.
Use waters with neutral tastes such as evian or volvic (those you can find pretty much everywhere in Europe, I presume in the US as well) but avoid contrex... unless you like it of course.
RS-232 is good, no denying it. Have a quick look though, at FTDI for an ASIC that give you USB to a 8 bit bi-directional// or serial bus. Linux/macos/bsd drivers? sure!
Also, many projects rely on the cypress EZ-USB too, some of which even ended-up on sourceforge.
How do you link the stepper shaft to the camera?
on
70 Megapixel Webcam
·
· Score: 4, Interesting
So far, this is what I have:
Hugin is getting really good as a frontend for panotools. It'll be really great when alpha layers become available too!
Getting the camera to take remote piccies is possible as well (although getting access to all the manual parameters maybe a problem -- no luck there with my canon a40)
A stepper motor and its RS-232 interface is not that expensive or hard to find anymore (50 quid at Milford instruments).
Or... you can build your own out of a floppy drive connected to the parallel port. Maybe a better solution, the milford stuff is getting pretty hot after a while and requires 9-15V
The question remains: How do you attact the stepper shaft to the camera? (I mean other than duct tape or lego) It would be nice to have a 90degree bent bracket as well to take piccies vertically.
Has anybody built a tripod like this? What did you guys use?
If you ever get the chance try Manchester Oxford Road station and get yourself a ticket to Liverpool Lime Street (once you're there, try and find Dr Duncans and order yourself a nice pint of Cains Dark Mild... mhhhh)
Anyway, just listen to the announcements once you're in Oxford Road station: "we are sorry that the blahblahblah is running approximately 20 minutes late... We apologise for the delay and for the inconvenience this might cause you". The keyword of course being approximately. Just wait another couple of minutes and the announcement is updated... "running approximately 23 minutes late..."
Chances are that the Liverpool train will be running anything between 10 minutes and 40 minutes late (which means one of the trains was cancelled and the next one is also running 10 minutes late). It _is_ a fsckin joke. Once I even was told (although that was a train from Victoria) that the driver went missing... result? another train cancelled.
The alternative is to take a coach, and all things considered (leaving on time and arriving on time, ticket is cheaper) nearly makes up for the reduced comfort (that is compared to the ideal situation whereby you manage to seat your arse somewhere on the fsckin train). Only problem is that the last coach back to Manc leaves 1 hour earlier than the last train. You just can't win.
Point taken. Having some limited experience working in hospitals (yeah, this is/. after all, anything is possible) I have this feeling that sometimes doctors aren't totally honest with their patients. So they promise some kind of miraculous cure (even sometimes believing themselves it would work) which more than likely the patient (or as I said their relatives) will want to try out. So there they go, enduring the most aggressive form of therapy... and to what result? well, maybe helping those patients die decently would've better for them. For doctors of course, it's more like... well we tried everything that was humanly (humanely?) possible and here's the bill and gee won't that look great in my CV when I publish my results (whatever they are).
Yes, I know I am being sarcastic since doctors often don't even have a choice anymore. Whatever treatment they will prescribe, they'll probably get sued anyway by said relatives of the deceased.
Hey, in my days you had to take philosophy whether you wanted or not. Helps balance that scientist's urge to try things, I guess. That's what I wanted to say;-)
From the website: Memtest86 is released under the terms of the Gnu Public License (GPL).
That's what I meant!
I guess there's no incompatibility between commercial and GPL. it's just that GPL projects are usually on.org and not.com appart maybe from that one ^_^
Re:Knoppix as a diagnostic tool
on
Knoppix 3.3 Is Out
·
· Score: 2, Informative
Check this one out: memtest86 (don't be fooled by the.com it's gnu software). Make sure it's not a memory problem. Cut yourself a CD which you boot from to run the tests.
I couldn't even run knoppix (let alone XP) on that computer (a SN41G2 shuttle) until I had the one stick swapped for another one.
If you get any errors on memtest86 (say after running it for 8 hours continuous) then you probably won't be able to install any OS (and/or experience random crashes).
Also, make sure you plug in a shitty pci videocard if you have a system which uses main memory for as display memory (like the nforce chipset). That segment of memory wouldn't get tested by memtest86. Setting that memory to 0 in the bios wouldn't help either. Hello? who turned off the lights;-)
Re:This is the kind of research I like to see.
on
The Oldest Mouse Contest
·
· Score: 4, Insightful
Keywords are palliative treatment, quality of life.
Need to be balanced with patient's choice (or their relatives?) but if the prognosis is bleak then maybe it's more important to spend quality time with your loved ones rather than enduring agressive treatment that's not going to be effective anyway.
Oh yeah... trials on terminal patients. Maybe people like grandparent (and those who modded him up) don't see the ethical issues involved. Sad it came from somebody involved in bioinformatics. Don't you guys have any philosophy lectures anymore? even basic stuff?
Simply put, what next? you are given permission to start trials on patients who are going to die after all, then what? trials on prisoners? soldiers? random population sample? Thinking about it, it's not like this hasn't been tried before...
Slashdot Mirror
To complement your nice Tyan motherboard, get one or two (XXX check for physical sizes) of them realizm 800 from 3dlabs. They are the only 16 lanes PCIexpress videocards I know of. Not sure about GPGPU, but at 3840 x 2400, solitaire is bound to look amazing... especially if you can get some nice 9.2Mpixel displays as well: High end videocard without a matching display, what would be the point? Check for instance the IBM T221).
Anandtech reviewed the Realizm 800 here.
Mhhh... If you wanted the machine to be a server of some sorts, then I just wasted 10 minutes typing all this!
Interesting idea! If you can work out a floating point location and colour value for each grain (AFM anyone ?), you could map these characteristics onto the center of your output image pixels, using a function of the distance between grain location and pixel location. Resulting values in whatever you want (8 bit unsigned, integers...)
Depending on your output image resolution (which you choose), the number of grains contributing to a single pixel would vary. Also, grains too far away from a pixel center don't contribute. Measurement errors are averaged out. Interpolation could also go temporal when suited (across frames).
Keyword: scattered data interpolation, RBFs, shepard.
I used shepard on some 3-D scattered datasets (confocal microscopy). Reconstruction works well, although the algorithm is quite hungry. RBFs may be a better option.
Who knows, when the technology catches up (disk space, microscopy, algorithms) this may be the norm in film restoration... At that point, rendering the film on your cheap 128 bit per RGBE channel, multi giga pixel handheld display would also be done in real time...
- displaying the logo
- displaying loading linux.....
- displaying loading init_rd.....
- the screen is cleared and the words Uncompressing linux... Ok, booting the kernel.
I found the relevent bit of code in linux/arch/i386/boot/compressed/misc.cI also found a reset_terminal() in linux/drivers/char/console.c but tweaking it doesn't seem to help.
I could go for the bootsplash patch, but that adds nearly 200K to the (compressed) kernel.
I don't really want to as I'd be happy to just have the Uncompressing linux... line under my syslinux logo (640x480x16 colours only adds 10K to the distribution). After that, all the kernel messages are sent to /dev/ttyS1 so they don't appear on the screen.
So, anybody knows where in the kernel source (preferably 2.4) is the screen cleared? Any clues I could follow?
Is SCSI P2P used in real world clusters though? How does it compare to SCI or gigabit ethernet? Price? Performance? Status of the project? No idea...
I tasted some green beans that guy was roasting in a food exhibition (he was selling bread ovens with a rotating coffee roasting bit inside) and I just couldn't get over how nice the smell was (both bread and coffee). Coffee being my favourite drug, I'm really envious you manage to grow your own, even if it's only a bush or 2!
Who does it remind me of... Eric Serra? Jarre?
Has anyone checked if the game gear will take 2300mAh NiMH batteries? It's pretty much the same problem with digital cameras: put new alcalines in... and cry
I'm thinking wrapping glui in python, but haven't got round doing that yet. Don't other cards also provide 10 bit resolution through opengl? might be worth a look. That's why I thought doing some opengl mojo is still way more compatible than using proprietary plugins.
Skype's GUI is QT + they already have a linux version. So I'm pretty confident they will have an OS X version... if they had enough money to buy the QT TrioPack license, that is!
I hope it'll be smaller than my quantum bigfoot too :-)
- I presume you are americam;
- I presume you don't value bottled water.
Fair enough, you can buy shit with sugar in for half the price of bottled water. Maybe all this sugar is also what keeps you going. Again, fair enough.The simple point about bottled water is that it tastes actually quite nice, and certainly much better than tap water. It has other uses too:
If you want to make yourself a decent espresso (illy maybe?) and use tap water, then you've just ruined it. This is also true for tea although I'm out of depth, here.
Use waters with neutral tastes such as evian or volvic (those you can find pretty much everywhere in Europe, I presume in the US as well) but avoid contrex... unless you like it of course.
Talking of the 80s, looks like you can even play your old NES cartridges on it... I think.
Could the jukebox in this picture possibly be the one on this site?
Also, many projects rely on the cypress EZ-USB too, some of which even ended-up on sourceforge.
- Hugin is getting really good as a frontend for panotools. It'll be really great when alpha layers become available too!
- Getting the camera to take remote piccies is possible as well (although getting access to all the manual parameters maybe a problem -- no luck there with my canon a40)
- A stepper motor and its RS-232 interface is not that expensive or hard to find anymore (50 quid at Milford instruments).
- Or... you can build your own out of a floppy drive connected to the parallel port. Maybe a better solution, the milford stuff is getting pretty hot after a while and requires 9-15V
The question remains: How do you attact the stepper shaft to the camera? (I mean other than duct tape or lego)It would be nice to have a 90degree bent bracket as well to take piccies vertically.
Has anybody built a tripod like this? What did you guys use?
If you ever get the chance try Manchester Oxford Road station and get yourself a ticket to Liverpool Lime Street (once you're there, try and find Dr Duncans and order yourself a nice pint of Cains Dark Mild... mhhhh)
Anyway, just listen to the announcements once you're in Oxford Road station: "we are sorry that the blahblahblah is running approximately 20 minutes late... We apologise for the delay and for the inconvenience this might cause you". The keyword of course being approximately. Just wait another couple of minutes and the announcement is updated... "running approximately 23 minutes late..."
Chances are that the Liverpool train will be running anything between 10 minutes and 40 minutes late (which means one of the trains was cancelled and the next one is also running 10 minutes late). It _is_ a fsckin joke. Once I even was told (although that was a train from Victoria) that the driver went missing... result? another train cancelled.
The alternative is to take a coach, and all things considered (leaving on time and arriving on time, ticket is cheaper) nearly makes up for the reduced comfort (that is compared to the ideal situation whereby you manage to seat your arse somewhere on the fsckin train). Only problem is that the last coach back to Manc leaves 1 hour earlier than the last train. You just can't win.
Yes, I know I am being sarcastic since doctors often don't even have a choice anymore. Whatever treatment they will prescribe, they'll probably get sued anyway by said relatives of the deceased.
Hey, in my days you had to take philosophy whether you wanted or not. Helps balance that scientist's urge to try things, I guess. That's what I wanted to say ;-)
From the website: Memtest86 is released under the terms of the Gnu Public License (GPL).
That's what I meant!
I guess there's no incompatibility between commercial and GPL. it's just that GPL projects are usually on .org and not .com appart maybe from that one ^_^
I couldn't even run knoppix (let alone XP) on that computer (a SN41G2 shuttle) until I had the one stick swapped for another one.
If you get any errors on memtest86 (say after running it for 8 hours continuous) then you probably won't be able to install any OS (and/or experience random crashes).
Also, make sure you plug in a shitty pci videocard if you have a system which uses main memory for as display memory (like the nforce chipset). That segment of memory wouldn't get tested by memtest86. Setting that memory to 0 in the bios wouldn't help either. Hello? who turned off the lights ;-)
Need to be balanced with patient's choice (or their relatives?) but if the prognosis is bleak then maybe it's more important to spend quality time with your loved ones rather than enduring agressive treatment that's not going to be effective anyway.
Oh yeah... trials on terminal patients. Maybe people like grandparent (and those who modded him up) don't see the ethical issues involved. Sad it came from somebody involved in bioinformatics. Don't you guys have any philosophy lectures anymore? even basic stuff?
Simply put, what next? you are given permission to start trials on patients who are going to die after all, then what? trials on prisoners? soldiers? random population sample? Thinking about it, it's not like this hasn't been tried before...
Slashdot had an article on the subject some time ago (original salon article).
Has been up for years (even if it doesn look very active) and always has good info on.