Domain: tcd.ie
Stories and comments across the archive that link to tcd.ie.
Comments · 114
-
Re:video stuff
here is a paper, though not very recent, describing the use of GPUs to aid in realtime motion estimation. I believe the latest hardware could do miracles here, especially with a fast back channel like the PCI successor.
-
Interesting work that raises some questions...
Having done a similar work for my final year project this year, I have some experience attempting general purpose computation on a GPU. The results that I recieved when comparing the CPU with the GPU were very different with many of the applications coming in at 7-15 times slower on the GPU. Further, I discovered some problems which I mention below:
! Matrix results
As in mentioned earlier in the report, the graphics pipeline does not support a branch instruction. So with a limitied number of assembly instructions that can be executed in each stage of the pipeline (either 128 or 256 in current cards), how is it possible for them to perform a calculation on a 1500x1500 matrix multiplication. To calculate a single result 1500 multiplications would need to take place and if they are really clever about how they encode the data into texture s to optimise access, they would need two texture accesses for even 4 multiplications. By my calculations that is 1875 instructions, where you can only do 128 or 256.
My tests found that using the Cg compiler provided by NVidia, that a matrix of size 26x26 could be multiplied before the unrolling of the for loop exceed the 256 limitation.
One aspect that my evaluation did not get to examine was the possiblity of reading partial results back from the framebuffer to the texture memory along with loading a slightly modified program to generate the next partial result. They don't mention if they used this strategy so I assume that they don't.
! Inclusion of a branch instruction
Even if a branch instruction were to be included into the vertex and fragment stages of the pipeline, it would cause serious timing issues. As student of Computer Science, I have been taught that the pipeline operates at the speed of the slowest stage and from designing simple pipelined ALUs, I see the logic behind it. However, if a branch instruction is included then the fragment processing stage could become the slowest as the pipeline stalls waiting for the fragment processor to output its information into the framebuffer. I believe it for this reason that the GPU designers specifically did not include a branch instruction.
! Accuracy
My work also found a serious accuracy issue with attempting compuation on the GPU. Firstly, the GPU hardware represents all number in the pipeline as floating point values. As many of you can probably guess, this brings up the ever present problem of 'floating point error'. The interface between GPU and CPU are traditionally 8-bit values. Once they are imported into the 32-bit floating point pipeline the representation has them falling between 0 and 1, meaning that these numbers must be scaled up to their intended representations (integers between 0 and 255 for example) before computation can begin. Combine these two necessary operations and what I saw was a serious accuracy issue where five of my nine results(in the 3x3 matrix) were one integer value out.
While I don't claim to be an expert on these matters, I do think there is the possiblity of using commodity graphics cards for general purpose computation. However, using hardware that is not designed for this purpose holds some serious constraints in my opinion. Anyone who cares to look at my work can find it here -
Re:Nail on the Head
It's not always trivial, but by complex math, do you mean implementing complex math (i = sqrt(-1)) or complex as in difficulty? Judging by your "I'm an engineer," I'm guessing the former, probably electrical or computer, which require complex math to calculate things such as inductance. Complex math may be restricted by the computer language - C and Fortran contain complex (imaginary) math libraries, but some languages don't (BASIC and Pascal didn't when I learned them on the Apple ][) which would require the rules to be implemented by hand. Still, many complex math libraries in common use need to be implemented by the programmer - quaternions, for example. These are not simple to understand - take a look at this page and tell me it's straightforward and easy to understand. I've written the things several times for 3d programs and still struggle through the equations when I try to figure them out (most people just skip the figuring out and go straight to the implementation, tho).
Complex equations (as in difficulty), such as solving an integration on a computer from, say, n to infinity, technically would take infinite time on a computer, so artificial bounds are necessary - but what is a good infinity? This varies depending on the accuracy required, what n starts at, sometimes the type of curve/function, etc., so the programmer needs to know approximations of certain characteristics and then estimate the values. On computers, these points are estimated using several learned techniques for estimation. I've learned a few of them in school - Newton-Raphson, Runge-Kutta, Euler's method - I personally haven't used any of them in years, so I don't remember which applies to this particular case, but I know they exist if I ever need them and could probably find the one I need fairly quickly (I know the books I would look in, at the very least). Half of the problem is knowing a where to look for the solution. -
Shut up, bloody Vikings!Manually babelfishing (I'm I mere Norwegian, so bear with me, Swedish brothers
;-)
Oh, I'm sorry - I'm in the wrong skit. I meant to say thank you for the translation.
No offence meant to the nice Slashdotter from Norwegia. It's a Monty Python Skit. Spam, spam, spam, spam! -
Use Kile
Kile, the KDE Integrated LaTeX Environment, is a pretty powerful LaTeX IDE for KDE. It has toolbars and menus that will be of major assistance, and has commands that automate certain things, like creating a skeletal structure for your document.
Of course, you'll need more than that--an IDE can't teach you everything. I particularly like this LaTeX tutorial--it was the first one I found when I first started learning LaTeX a little over a year ago, and it's pretty good for starting one off. -
Impression
Walking home from the lab I found it looks quite good. 9MB *.mov at tcd.ie(if someone could mirror it, might be useful)
Next step would be to take some consecutive slots and make some kind of movie.
Have fun -
Re:Novell:Mormons::The Enquirer:Scientologists
If that's truly a South Park episode, then they ripped it off of Rowan Atkinson, the British comedian known for his Mr. Bean show. Except in his version, the jews were right: Atkinson's "Hell" sketch
-
Re:Don't be so hard on this man
Here is an even better example of someone with a similar affliction - though note that I don't particularly buy the DSM categorisations that lump these people together, but there's clearly something wrong with her given that her life is ruled by this stuff - not so sure about Soto, who appears to have it pretty much under control and just have a messy living room (as do I). Would have been interesting to see some photos of the Soto's house for comparison.
-
Great scope tutorial
-
Re:yep
In fact, the Xbox is the only popular console in recent memory that has an x86 chip. The Nintendo 64 and Playstation series (as mentioned), have all sported MIPS architectures. The Dreamcast and previous SEGA consoles were (Hitachi SH)RISC.
MIPS is used in routers (especially high-end, like from Cisco), broadband/wireless adaptors, TV's, and TV set-tops.
MIPS is commonly taught to CS / programming students to introduce them to assembly and applied architecture concepts. A common and short book for the beginner is "Introduction to RISC Assembly Language Programming" by John Waldron. It uses the SPIM emulator.
-
Re:yep
In fact, the Xbox is the only popular console in recent memory that has an x86 chip. The Nintendo 64 and Playstation series (as mentioned), have all sported MIPS architectures. The Dreamcast and previous SEGA consoles were (Hitachi SH)RISC.
MIPS is used in routers (especially high-end, like from Cisco), broadband/wireless adaptors, TV's, and TV set-tops.
MIPS is commonly taught to CS / programming students to introduce them to assembly and applied architecture concepts. A common and short book for the beginner is "Introduction to RISC Assembly Language Programming" by John Waldron. It uses the SPIM emulator.
-
Re:This won't work
Yes, freenet and friends exist and are under active developement.
And there are more promising concepts waiting to be explored.
Seems like the *IAA can only lose if they intend to enter an arms race with academia. -
Re:So what about web+P2P?
As you're interested in P2P and how its done, here's a good read for you and everybody else who wants to take a challenge..
(found in the dusty bookmarks) -
Re:About LaTeX..Couple questions, I thought I read on one site that you can only go 4 levels down on sections/subsections.
Another poster has answered this below..
Is this true? (Hopefully using the right term...I mean itemized lists with roman numerials, numbers, letters for each part)
If you mean "itemized" or "enumerated" lists then yes there is a limit it appears you can go 5 deep.
The following will give a "Too deeply nested" error. Due to the "sub sub sub sub sub sub item"
N.B. It it not very pretty due to having to get past the "comment compression filter"...
\documentclass{article} \begin{document} \begin{itemize} \item Item \begin{itemize} \item Sub item \begin{itemize} \item Sub sub item \begin{itemize} \item sub sub sub item \begin{itemize} \item sub sub sub sub item \begin{itemize} \item sub sub sub sub sub item \end{itemize} \end{itemize} \end{itemize} \end{itemize} \end{itemize} \end{itemize} \end{document}
As with many aspects of LaTeX however if you find it doesn't do something it probably means it's not prudent (from a structural perspective) to do it anyway. For example if you really need that level of deep reference you may well be better off with part,chapter,section, subsection,
... . . .,itemize etc... Ironically I tried posting this reply with some deep nesting, slashdot posts are limited to three levels deep! ;-) Of course if you wish to you can always override the builtins with your own "super list" or something.Also, can ya'll post some good links to a newbie learning LaTex..and some good reference sites that have all the tags layed out with good explanations?
Sure, below are a list places I would reccomend starting, you havn't said if you use Windows, *nix or Mac so i've added both (sorry if you are a Mac man you'll have to Google yourself).
- Editing:
- *nix If you are a *nix user I would reccomend the following editing combination.
- XEmacs
- AucTeX. A sophisticated editing mode for LaTeX
- preview-latex. Places the rendered equations and images directly in the editor window making "equation tuning" and other tasks a snip.
- Windows
- WinEdt. A very sophisticated text editor for Windows. Its forte is LaTeX. It is not free, but well worth the money.
- Learning resources:
- Other random stuff
- dvipdfm. For converting the output of LaTeX into PDF (highly recommended)
- Prof. Knuth's home page(The author of TeX).
- CTANThe Comprehensive TeX Archive Network. Here you will be able to download packages, utilities and tools that do not come by default in your LaTeX distribution.
-ed
-
Better links
Here's the official smart couch page:
http://www.dsg.cs.tcd.ie/index.php?category_id=350
There are links to more aricles, and to two RealMedia streams of when it appeared on BBC and RTE news.
Here's a longer article about the smart couch which also shows pictures of the smart sword I've been working on:
http://news.bbc.co.uk/2/hi/technology/3107746.stm
And this is the page for my smart sword project:
http://www.dsg.cs.tcd.ie/index.php?category_id=351 -
Better links
Here's the official smart couch page:
http://www.dsg.cs.tcd.ie/index.php?category_id=350
There are links to more aricles, and to two RealMedia streams of when it appeared on BBC and RTE news.
Here's a longer article about the smart couch which also shows pictures of the smart sword I've been working on:
http://news.bbc.co.uk/2/hi/technology/3107746.stm
And this is the page for my smart sword project:
http://www.dsg.cs.tcd.ie/index.php?category_id=351 -
Someone who did some work with it
You could get in touch with this guy. Some of the screens look pretty good for a hobbyist's efforts at PS1. Also he's a old school C64 geek so he's gotta be ok.
-
Re:Actually thats the recommended approachI believe you are thinking of Shocking Nonsense
I remeber seeing this first from the UNIX System Administration Handbook.
-
Re:So Much For *BSD Is Living
AT&T filed suit against Berkeley Software, claiming that proprietary code agreements had been haphazardly violated. In the same year, BSD filed countersuit, reciprocating bad intentions and fueling internal rivalry
Oh and Linux has no problems with lawsuits.Theo de Raadt
You've got RMS, don't talk about Theo.BSD's filesystem implementation to be "very poorly performing."
Obsolete refernce, doesn't mention softupdates.rapid, decentralized development models are inherently superior to slow, centralized ones in software development.
The Linux kernel development has to push everything through Linus so you've got bottlenecks too.corporations and coders alike to reap profits without reciprocating the goodwill of open-source
You've got LindowsOS so don't talk about corporate greed.The failure of BSD culminated in the resignation of Jordan Hubbard and Michael Smith from the FreeBSD core team.
So all the current success is not happening, huh?
Is it just me or does the *BSD is Dying troll seem like he has real problems -
ISO HTML
If you want to be standards-compliant, and don't need the extras that XHTML provides, check out ISO/IEC 15445:2000 and the associated User's Guide.
-
ISO HTML
If you want to be standards-compliant, and don't need the extras that XHTML provides, check out ISO/IEC 15445:2000 and the associated User's Guide.
-
Ethics of Free SoftwareA few years ago, Bertrand Meyer penned a fascinating article, "The Ethics of Free Software". (Printed in Software Development magazine (reg req'd), but mirrored many places including here and here.)
Meyer criticizes the self-assumed ethical superiority of ESR, RMS, and others, and in particular notes the "gun evangelism" ESR intertwines with his open-source evangelism.
This thoughtful article should be required reading for all overly-strident geeks.
-
Re:Why
These are pretty basic questions that have already been figured out. A quick Google search brings up this little FAQ that you might find interesting: http://ntrg.cs.tcd.ie/mepeirce/Project/Mlists/min
i faq.html -
Re:My SigSome form of a Spring festival has been celebrated since the bronze age,
" The word Easter comes from the name of Ostrae, the Anglo-Saxon goddess of the dawn. As with the Easter festival and that of Dionysus, Ostrae's feast was celebrated around the time of the Spring equinox. Although there are no myths which speak of a death and resurrection of Ostrae, her role as goddess of the Spring indicates her connection to those gods who participated in the cycle of the seasons by just such an event, namely Osiris and Dionysus. Ostrae's festival, naturally enough, involved dawn rites - an important feature in the Christian celebration of the Resurrection. Indeed, many aspects of how the death and resurrection of Christ is celebrated by Christians today have their origins in the pagan past."
"Lent, for example, began as a form of sympathetic magic. People fasted as Spring began so that more energy would be available for the growing seeds. Lent takes its name from the Old English for "lengthening", with reference to the change in daylight hours as Spring approaches. At the equinox, when there begins to be more day than night, birds start to lay their eggs. The brightly coloured eggs of forest birds were traditionally collected either in their nests or in baskets made to resemble them. In the Christian tradition the egg is used in the iconography of Easter as a symbol of death and rebirth, a symbolism which existed long before the time of Christ. For the Anglo-Saxons the egg also stood for fertility and the sun. The sun played a large role in the festival of Ostrae and was symbolically rekindled as a bonfire, also a significant part of today's Easter liturgy."
A paper by Mr. John Duffy delivered to the College Theological Society at the Inaugural Meeting of the 168th Session on December 1st 1997.
I am happy to take a nice sip of wine and enjoy the show. Seriously, I look forward to continuing this (and other debates) debate later.
-
Re:AS long as thay have anonomous cash
David Chaum has been concerning himself with these issues for years. If you read some of his writings you will find that he shares your concerns.
Rather than badly paraphrase his thinking, I'll just quote the introduction to "Security without Identification":
Computerization is robbing individuals of the ability to monitor and control the ways information about them is used. Already, public and private sector organizations acquire extensive personal information and exchange it amongst themselves. Individuals have no way of knowing if this information is inaccurate, outdated, or otherwise inappropriate, and may only find out when they are accused falsely or denied access to services. New and more serious dangers derive from computerized pattern recognition techniques: even a small group using these and tapping into data gathered in everyday consumer transactions could secretly conduct mass surveillance, inferring individuals' lifestyles, activities, and associations. The automation of payment and other consumer transactions is expanding these dangers to an unprecedented extent.
Organizations, on the other hand, are attracted to the efficiency and cost-cutting opportunities of such automation. Moreover, they too are vulnerable, as when cash, checks, consumer credit, insurance, or social services are abused by individuals. The obvious solution for organizations is to computerize in ways that use more pervasive and interlinked records, perhaps in combination with national identity cards or even fingerprints. But the resulting potential for misuse of data would have a chilling effect on individuals. Nevertheless, this is essentially the approach of the electronic payment and other automated systems now being tried. Although these systems will require massive investment and years to complete, their underlying architecture is already quietly being decided and their institutional momentum is growing.
This momentum is driving us toward a seemingly irreconcilable conflict, between organizations' need for security and the benefits of automation on one side, and individuals' need for ensured privacy and other protections on the other. But this conflict may be avoided by early adoption of a fundamentally different approach to automating transaction systems. This new approach is mutually advantageous: it actually increases organizations' benefits from automating, including improved security, while it frees individuals from the surveillance potential of data linking and other dangers of unchecked record keeping. Its more advanced techniques offer not only wider use at reduced cost, but also greater consumer convenience and protection. In the long run, it holds promise for enhancing economic freedom, the democratic process, and informational rights.
Of course the technology Chaum advocates is not the only way to conduct monetary (and other) transactions. You can be sure that there are powerful forces that would like nothing better than to have improved access into people's private business. At the very least, people should realize there are other options. -
Interplanetary Internet
Some interesting links:
Future:
Interplanetary Internet (IPN)
IP in Space
Now:
SCPS gateway runs on FreeBSD
SCPS: Space Communications Protocol Standards -
Re:Standards . . . HAH!
The word standard implies that it is unwaivering and uniform. XHTML is most definitely NOT that!
I don't believe the w3c has ever claimed xhtml as a standard. Or any of their html specifications either. I think the only "standard" html is ISO-HTML.
-
Hype, hype, hype...
This is amazing technology, and it will revolutionize digital cameras if/when it comes down in price. HOWEVER, this is not how the human optic system works. Even in our optics, we have seperate receptors for red, green, and blue, and our brains do the interpolating. As most will remember from basic elementary biology, our eyes detect light through rods and cones. All quotes are from this link. "The retina has ~126 million photo receptors, 120 million rods and 6 million cones." Rods gather any light they can, and compile the data together to show the best possible image in the dimmest light; therefore, rods will display a black and white image. This is why the darker it gets, the harder it is to differentiate yellow from white: you are depending more and more on the rods.
HERE is where it gets interesting, and where I get to my point. Cones are what we use to see color. An individual cone cannot see red green and blue as this marketing hype would lead us to believe. "The cones come in three types: Red (60%), Green (30%) and Blue (10%). The red and green cones are randomly distributed in the center of the fovea and the blue cones form an annulus around the outside." So in effect this camera will actually surpass the human eye.
As a side note, the link goes to a very interesting document that states how "126 million photoreceptors must be transmitted to the brain via 1 million fibers in the optic nerve [while] [t]he overall compression ratio of 126:1 is not evenly distributed." Check it out. -
Huh?From the article:
Instead of trying to make thousands of transactions a day totaling only a few pennies or less (which is what ads result in if you're lucky), I propose a simpler system; a small yearly fee (less than $10USD) which works out to a micropayment per day. Memberships and subscriptions tend to cost more money for a smaller period of time.
So how is this different from regular subscription websites? It's cheaper?
I think micropayments are definitely the Right Idea for the web, but I don't see how they could be properly implemented using current payment systems. Off the top of my head, I think a payment system suitable for micropayments would need (at the minimum) the following properties:
- implicit (yet secure) payments. The user should be able to configure their (trusted) web browser to automatically make requested per-page micropayments to a server if those payments are below a threshold (e.g., $0.001). The browser can prompt the user for permission to make larger micropayments. It's very important that the user does not need to intervene in the micropayment process every time they request a document. Since a user can not read through the whole source of the browser and anything else that might need to make payments, perhaps the browser and other programs should call an external program to make the payments, the user's `payment agent'. This would be a small program that makes payments while following the user's policies and restrictions.
- extremely low (or non-existent) per-transaction fees. If the provider(s) of the payment system are charging $0.10 to the payment receiver for each micropayment, it obviously won't work. This essentially implies the next requirement.
- contact with payment system provider(s) not required for every transaction. If the server collecting micropayments must contact the provider(s) every time a payment is collected, the system will not be feasible. The server should be able to store up many micropayments and redeem them with the provider all at once every day/week/month.
Creating a digital cash system that has all the properties we'd like is a damn hard problem that hasn't be solved yet. However, cryptographic tools such as one way functions and PKI are very powerful. I don't think we've fully exploited their possibilities yet, so I'm still hopeful that a true digital cash scheme will one day be created. -
Re:I am stunned and amazed....
Forget Jesus, Santa, Mithra, and other mythology. December 25th is Sir Issac Newton's Birthday,and he definitely did exist! What would science be without him?
-
Re:What about next time?
(They didn't have images in their records for the last 2000 years;
Funny, I thought the Book of Kells was produced c.800 CE (well within the last 2000 years) and was illustrated. The illustrations being a major part in the significance of the work (because they're an insight into the artistic heritage of the Celtic world).
frankly, if something's really So Important That It Must Be Saved, it can be done in the good queen's English.)
Oh, and it wasn't in English either.
-
Aliens have warchalked them!!!!
Alien visitors picking up the signal have warchalked the area just outside of San Diego. Aerial observers have taken a photo of this this.
-
GL ripped all of SW from Trinity...
Wow, isn't that the Death Star in the top right corner?! long room page
-
Re:other architecture, too.
To be fair, though, the Vatican library is just a vaulted room, wheras - coincidence or not - this and this (the original post) are strikingly similar, even down to the far wall, the frequency of alcoves and the shape of the moulding between floors.
-
Re:Nothing new there
a strange coincidence???
(from the Long Room site) -
Re:Debt?
the link given by Hemos is merely a red herring.... he's really talking about the original death star plans which coincidentally appear on the Long Room page. Very controversial move, H.
-
Re:Debt?
the link given by Hemos is merely a red herring.... he's really talking about the original death star plans which coincidentally appear on the Long Room page. Very controversial move, H.
-
Deathstar
Well, it should be Trinity College who should be crediting LucasFilm for this blatant rippoff "Deathstar" from RTOTJ:
http://www.tcd.ie/Library/Images/sphblue.jpg -
Re:It's ok...
Here it is, in all it's original glory (This is book 1, section 1): http://www.maths.tcd.ie/pub/HistMath/People/Newto
n /Principia/Bk1Sect1/PrL1S1.pdf
And, oh, by the way, having a mastery of Latin will aid your understanding quite a bit... -
Re:It's ok...
You can read it here.
-
Re:Debate is getting old
Ya know, as much as I love to see Hillary Rosen gagging on her own foot, this isn't really news. She went up against an audience of students--people who typically have very little money and are hostile toward big, greedy corporations
You know, I was thinking something similar. I go to Trinity College Dublin (a sister school of Trinity Cambridge and Oxford) and we have some pretty ancient debating societies.
I've been to some of these and know that the debates are a breed unto their own. If you didn't know what you were up for, you'd be eaten alive. You really need some quick wit to do this and the heckling can get serious. -
Author giving talk tonight in TCD, Ireland
Anyone planning to go to this probably knows already, but Brian King is giving a talk on Mozilla in Trinity College Dublin tonight. More details at http://netsoc.tcd.ie/events/0203/mozilla.php
-
Any questions for the author?
I know this is short notice, but I'm going for dinner with one of the the authors tonight in two hours time, and then going to his talk with the Internet Society in Trinity College Dublin (Ireland). The details of the talk are here.
-
Any questions for the author?
I know this is short notice, but I'm going for dinner with one of the the authors tonight in two hours time, and then going to his talk with the Internet Society in Trinity College Dublin (Ireland). The details of the talk are here.
-
Any questions for the author?
I know this is short notice, but I'm going for dinner with one of the the authors tonight in two hours time, and then going to his talk with the Internet Society in Trinity College Dublin (Ireland). The details of the talk are here.
-
Re:IANALWA, but this can't be all bad
-
Digicash
There's a company : eCash Technologies Inc having a viable solution for micropayment, but I don't think it will become a standard before all patent claims on this are lifted (a cryptographer, David Chaum owns most of the patents on digital cash), like RSA usage took off when we approached the end of the patent and explodes now.
Another technology that won't took off before patents expire...
See also http://www.aci.net/kalliste/dcguide.htm.
-
Re:sTUpId cARboN rOd ... DOH !You forgot two:
THE CLASSICS: DEEP SPACE HOMER
Episode Guide
BTW, I see some prick mod'd you down for injecting a little bit of ON TOPIC humor (must be some type 'A'sshole who combs the bristles to his toothbrush after flossing for an hour).
Well, here are some ON TOPIC links that shows this entire article is OLD hat (relatively speaking).
Scientific America : STRANGE ATTRACTORS - Chemists make magnets without metal
Cryostat Modeling for the Superconducting Interaction Region Magnets: CESR Phase III
All of these articles are circa 1997-98. -
Re:Verification vs anonymity
I just posted this in response to somebody else, but a while ago my friend pointed me to this page on digital cash systems that seem to describe and solve exactly the problems you state. If the government provided a public list of tokens used to vote, anyone could verify that the total is correct, and any individual could verify that his or her token is in the list, although nobody can match any token but his own to anybody else. It seems to me like a system like this could work really well.
-
Re:Are Anonymity and Authenticity both possible?
A little. It sounds exactly like the problems in a digital cash system. Try this page, especially the section on blind signatures and double spending.
Under a system like this, it would be possible to issue a blind-signed token to everyone who wants to vote. You provide ID and a blinded token, and the government verifies the ID and signs the token. You get it back and unblind it, and you now have a provably valid token, even though nobody but you knows what you originally sent to the gov't. Then, you take it to a polling place and vote with it. Trying to vote with the same token twice is exactly the same as trying to double-spend a digital cash token - you have a 1 in 2^k (I think) chance of it working, where k is the "security parameter" discussed by Chaum in his paper.