"The conversation you are about to engage in may be monitored by outside parties. Please press '1' to confirm you permit this monitoring. To deny permission, and disable all monitoring, please press '2', or simply stay on the line."
In that case there might as well be a 3rd option:
"To quickly schedule a one-way trip to Mexico, please press '3'"
Each posted message requires several page views (login/preview/confirm,etc)
Articles with high message counts seem likely to attract readers (the count gives a rough measure of how interesting/controversial the topic is)..
More readers means more page views... (have to click on messages to view threads/etc)
In short, it is in SlashDot's FINANCIAL interest to periodically let an imflammatory 'article' "slip by". I bet a ton (long/short/metric/etc) of money is made just from advertising / pay-per-click revenue...
Does anyone else see the irony in the parent post?
While this sound may be a "nonlethal weapon", fast-food burgers are anything but. While I don't normally condone what the article is proposing, anything that causes young kids to associate something negative with fast-food places is truly a blessing.
If you haven't already, I highly recommend you check out the "Supersize Me!" video.. (Best $4 at Blockbuster that you will ever spend).
If the doctor reports aren't convincing, then make sure you see the very end... All fries and burgers are not created equal!
you'd find a genuine and deep-seated desire to do what is best.
These people can be a powerful & highly destructive weapon against the American people... All it takes is one manipulative person, *somewhere* in the chain of command (not necessarily at the top) to manipulate the good intentions of others to match his desires.
Add a dash of lack of critical thinking, and you have a large group of earnest people, working extremely hard (since they believe it in the greater "good") for truly nefarious purposes.
And given their size, they might not be able to see the whole picture until it is too late.
additions. To get 100% of this speed one needs to have a sophisticated handling of cache memory, among other requirements.
One problem with "optimization" is that often the wrong thing is optimized.
For example, Someone might be able to implement a convolution (dot product) using the regular O(N^2) formula... For a vector of length N, this will result in O(N^2) floating-point adds and multipications being performed... Loop-unrolling could be used to keep all the units on the P4 busy...
However, the same result could be computed using a FFT, which is O( N Log2 N ) in # of adds & mults. Thus, this method requires significantly less computation [picture what happens when N=128], and a tad bit more memory. This implementation will not fully utilize a P4's arithemtic units although it would likely complete in far less time than the first method.
Thus, if someone is trying to show off how fast their new processor/supercomputer/cluster can be, they may not use the fastest implementation!
I recall seeing a few "supercomputer" specs recently that advertised "TFLOP/s" rates that were calculated based upon the maximum instruction bandwidth... This is deceptive advertising since even highly compute-intensive programs will not have an instructions stream composed solely of floating-point arithmetic... (And if a huge number of floating pointer operations are being done without any loads/stores, then perhaps there is a way to simplify the calculation!)
People usually refer to Amdahl's law for parallelizability... But a similar concept exists when taking about instruction mixes and functional units.
The artificial way of being an extrovert is to pretend to care, pretend to listen to people, pretend to....
Oh, you mean the typical American...
Disclaimer: "typical" means pop culture in this context, since there really is no such thing as "typical American".
I am not trying to start a flame war...
Before modding this or the parent, consider this...
The "doctor" that wrote this so-called "book" is a medical student...(although nearly done)
http://www.rojakpot.com/default.aspx?location=7&va r1=2
Details on the MBBS program:
http://www.manipal.edu/kmc/institutions/twinningpr ogram-aua.htm
Also consider this: When does a PhD in CS/ECE write a book on BIOS settings? That wouldn't even qualify as a MS thesis topic unless someone's entire committee didn't understand it.
If the "Editors" didn't even give this half a though, then why are the EDITORS???
This has been happening ever since I first started reading Slasdot 2-3 years ago...
I believe the editors let an inflammatory article "slip" by one or twice a week go increase ad revenue:
This "article" went up on SUNDAY night... Which means most./'ers will be spending their WORKday tomorrow (and the rest of the week) posting rebuttals, reading others rants, etc... Since this is a fairly specialized topic (BIOS config), unless the editors allow the summary to be inflammatory, most people won't bother to click the "Read More..." link...
This will probably be one of those articles that hits +500 messages real quick... Each message means multiple pages were loaded (login/preview, compose message, post response)... and thats only for those who bother to comment! Many many people will read the comments just to see what other are saying (since they will see so many were posted).
I think it is just a quick way for CowboyNeal to get some extra $$$ real quick. They [the editors] could at least be honest about it....
With FPGA's I don't think anyone would seriously consider doing their own layout, cell by cell, for a design.
I have met at least one person who has done this...
Yes, an automatated program can try many different types of optimizations... But a human knows how to change the design to suit the implementation better. (e.g. exploit problem-specific aspects). There is often no need to consider many cases when it is clear which one will produce a very-near best result.
For example, TI supplies a hand-optimized convolution routine for its x6700 series DSPs... With the TI compiler and a well-written C program (remember, convolution is just a simple for-loop, much like a dot-product), the hand-optimized ASM code is 2-3x FASTER than the code with any combination of optimization settings (O1, O2, O3, others, etc). FYI this operation, is a critical computation performed in a lot of communications/image/signal processing applications. It is not a stupid, meaningless benchmark.
And that is just for a stikin' for loop that is a bout 4-5 lines of code in C!
interesting how Logitech comes up with such exclusive features.
Ahem! I have a feeling there are a few *cough* patents that are the root of such evil^H^H^H^Hexclusivity...
In many senses, computer mice are like toilet paper... There really much left in terms of *useful* innovation, but that's where marketing depts take over...
For example, the 3rd button/scrollwheel has its uses, but most things work quite well without. (I'd bet that most Linux users don't even realize that their scrollwheel may not work by default). And Mac users will probably be the first to say that even two buttons are overrated...
I don't think many HPC groups would consider this to be a problem... The really really large clusters are mainly diskless... Hard drives are a nightmare for both cooling and reliability... Plus, some of the really compute intensive applications like finite-element analysis require relatively little I/O. (I think they may just read some data and then crunch on it for a while...)
Re:Never Program- RFC822, workarounds
on
PHP 5.1.0 Released
·
· Score: 1
On a side note, the only way around this is to have a programming language that tags data as tainted if it came from or was derrived from user input and not allow it for input, and having functions that sanitize data and remove that flag. This is first off slower and second not optimal in many situations.
I think there may be a way to avoid the performance penalty with the use of strongly typed language...
If I were using C++, I would have several types for "strings". (WebQueryString, CommandArgString, SQL_QueryString, etc). Either built-in (or user specified) methods would santize/convert strings..
For example, if I wanted to use a web form value as a command-line argument (say for a filemanagement script), then the mere act of casting a WebQueryString to a CommandArgString would cause it to be properly escaped.
Similarly, passing a WebQueryString to a function that took a SQL_QueryString type would cause it to be properly escaped... (Well, I'm being a bit sloppy, queries would probably have to be split apart into many different types to represent the different components (noun, verb, etc))
Basically, a strong type system could be used to (either manually or automatically) force the programmer to santize the data... If designed & implemented well, this could result in faster development times and more secure software...
It really annoys me that just to properly write a simple set of PHP scripts (such as an address book with a MySQL backend), a great deal of effort is expended in just parsing and (more importantly) santizing data... Even with all the object-oriented and/or modular programming techniques, it is still all too easy to pass the wrong string (or call the wrong *_escape method) in a way that results in a security flaw...
The next article on this topic will be that the bees will be joining forces with the RIAA and issuing cease and decist orders to all icebergs and to every human with "ice", "berg", or "borg" in their name.
Re:I'm thinking of contributing to GCC...
on
GCC 4.1 Released
·
· Score: 1
In what sense do you mean that its inlining is bad? Inlining seems like a very simple optimization to implement...
I believe GCC uses heuristics (that can be modified with command-line parameters) to decide whether to inline a function or not.
Perhaps isn't not doing what you want because either
1) What you want isn't likely to improve performance...
2) Your not supplying the correct options to force it to do what you want...
3) what you want isn't really possible: (like to inline a virtual function call from a base pointer).
Wouldn't it be simpler just to have Firefox manage its own memory ? I think it might be possible without a complete overhaul. (But I'm far from being a firefox developer)
If the problem is that new windows tabs create a lot of objects that each get interspersed in memory, then why don't they just:
Overload "new" so that objects from each page are allocated from separate pools of memory. (using separate 4k memory pages for each tab/window wouldn't be a big deal)
Many times, applications do something similar so that they can create and delete many objects rapidly without the incurring the overhead of (implicitly) calling malloc and free each time....
I have been advocating that we should have a moderation system for the editors/submitters of articles...
Of course, somehow the editors don't seem too interested in the idea.
For example, the original summary of the Hot-Air Balloon article contained some wildly inaccurate metric-to-english conversions (off by almost 25%) that snubbed the true feat accomplished... I applaud the editors for removing the inaccuracies, but submitters should be held more "publicly accountable" for their choice of article selection and summary descriptions.
Especially the guy with the french-sounding name that posts links to his own block as stories...
"Stay here and FIX this bungled mess? Listen: 12 months ago you were all lambasting me for daring to speak against His Holiness Bush's Purity, the same guy who's now in the low 30's approval-ratings-wise. The day "should we start pre-emptive wars for the hell of it? (yes/no)" and "should we drop DRM? (OK/not sure yet)" and "should we ban anybody who makes more than $100,000/year from running for political office? (Hell, yes/hell yes)" show up on the ballot is the day I know I can vote and actually AFFECT something in this highjacked country. Picking between siamese-twin shit-for-brains (Bush AND Kerry) is not a choice. Here, I'll give America one last chance: "Hey, howzabouts we tear down all the football stadiums and build libraries in their place and pay the teachers the football player's salaries so we'll have an educated generation and have an OPTION besides outsourcing all our programming to Bangladesh? Anybody with me?" Crickets: *chirp* *chirp* *chirp*"
If you run for pres, I will vote for you and heartily encourage everyone to do the same...
Most Slashdoters should at least like the part about keeping their jobs...
Slashdot Mirror
It has actually been suggested that, in order to make Mars habitable, we should detonate nukes where we want to live???
Rigghht.
I'm surprised no-one has suggested the use of nukes to flatten the Rocky Mountains -- to facilitate cross-country transportation.
In that case there might as well be a 3rd option:
"To quickly schedule a one-way trip to Mexico, please press '3'"
At a recent IEEE conference, I noticed a large number of researchers' topics concerned voice recongition and voice synthesis.
Although I'm not sure for who they were working or from where the funding came. (Plus, it was an international conference).
Only if you like your scalp "extra crispy"....
Flame wars generate a lot of messages...
Each posted message requires several page views (login/preview/confirm,etc)
Articles with high message counts seem likely to attract readers (the count gives a rough measure of how interesting/controversial the topic is)..
More readers means more page views... (have to click on messages to view threads/etc)
In short, it is in SlashDot's FINANCIAL interest to periodically let an imflammatory 'article' "slip by". I bet a ton (long/short/metric/etc) of money is made just from advertising / pay-per-click revenue...
Such pieces of crap mean $$$$ for CowboyNeal.
Does anyone else see the irony in the parent post?
While this sound may be a "nonlethal weapon", fast-food burgers are anything but. While I don't normally condone what the article is proposing, anything that causes young kids to associate something negative with fast-food places is truly a blessing.
If you haven't already, I highly recommend you check out the "Supersize Me!" video.. (Best $4 at Blockbuster that you will ever spend).
If the doctor reports aren't convincing, then make sure you see the very end... All fries and burgers are not created equal!
These people can be a powerful & highly destructive weapon against the American people... All it takes is one manipulative person, *somewhere* in the chain of command (not necessarily at the top) to manipulate the good intentions of others to match his desires.
Add a dash of lack of critical thinking, and you have a large group of earnest people, working extremely hard (since they believe it in the greater "good") for truly nefarious purposes.
And given their size, they might not be able to see the whole picture until it is too late.
His main page says the server is being loaded... Well okay.. (slashdot has a tendancy to do that).
But he is nice, so he gives a screenshot...
Guess what folks, the links on his "screenshot page" appear to be live links!!! Well, I hope London is ready for this....
And traceroute shows:
(snip)
7: so-3-2-0-0.gar2.Dallas1.Level3.net (4.78.234.17) 26.187ms
8: ae-1-51.bbr1.Dallas1.Level3.net (4.68.122.1) 20.143ms
9: ae-1-0.bbr2.London1.Level3.net (212.187.128.57) 144.766ms
10: ge-10-0.ipcolo2.London1.Level3.net (212.187.131.8) 146.664ms
11: fe0-0.red1.mnet.net.uk (212.113.10.202) 144.729ms
12: cr1-fe20.rbs.uk.euroconnex.net (82.211.104.78) 155.163ms
13: cr1-fe01.gh1.uk.euroconnex.net (217.112.80.18) 150.020ms
14: wood.bocks.com (217.112.87.2) 155.446ms
One problem with "optimization" is that often the wrong thing is optimized.
For example, Someone might be able to implement a convolution (dot product) using the regular O(N^2) formula... For a vector of length N, this will result in O(N^2) floating-point adds and multipications being performed... Loop-unrolling could be used to keep all the units on the P4 busy...
However, the same result could be computed using a FFT, which is O( N Log2 N ) in # of adds & mults. Thus, this method requires significantly less computation [picture what happens when N=128], and a tad bit more memory. This implementation will not fully utilize a P4's arithemtic units although it would likely complete in far less time than the first method.
Thus, if someone is trying to show off how fast their new processor/supercomputer/cluster can be, they may not use the fastest implementation!
I recall seeing a few "supercomputer" specs recently that advertised "TFLOP/s" rates that were calculated based upon the maximum instruction bandwidth... This is deceptive advertising since even highly compute-intensive programs will not have an instructions stream composed solely of floating-point arithmetic... (And if a huge number of floating pointer operations are being done without any loads/stores, then perhaps there is a way to simplify the calculation!)
People usually refer to Amdahl's law for parallelizability... But a similar concept exists when taking about instruction mixes and functional units.
Yes, but we know what happens when you make a copy of a copy...
Easy
execute the following as root:
echo "127.0.0.1 slashdot.org www.slashdot.org" >>
After this one command, you will start making better use of your time.
I am not trying to start a flame war... Before modding this or the parent, consider this... The "doctor" that wrote this so-called "book" is a medical student...(although nearly done) http://www.rojakpot.com/default.aspx?location=7&va r1=2
Details on the MBBS program:
http://www.manipal.edu/kmc/institutions/twinningpr ogram-aua.htm
Also consider this: When does a PhD in CS/ECE write a book on BIOS settings? That wouldn't even qualify as a MS thesis topic unless someone's entire committee didn't understand it.
If the "Editors" didn't even give this half a though, then why are the EDITORS???
This has been happening ever since I first started reading Slasdot 2-3 years ago...
./'ers will be spending their WORKday tomorrow (and the rest of the week) posting rebuttals, reading others rants, etc... Since this is a fairly specialized topic (BIOS config), unless the editors allow the summary to be inflammatory, most people won't bother to click the "Read More..." link...
I believe the editors let an inflammatory article "slip" by one or twice a week go increase ad revenue:
This "article" went up on SUNDAY night... Which means most
This will probably be one of those articles that hits +500 messages real quick... Each message means multiple pages were loaded (login/preview, compose message, post response)... and thats only for those who bother to comment! Many many people will read the comments just to see what other are saying (since they will see so many were posted).
I think it is just a quick way for CowboyNeal to get some extra $$$ real quick. They [the editors] could at least be honest about it....
(emphasis mine)
Does the British Government know how to spell?
Any posion can be made to look like a cure if the disease it 'cures' doesn't kill you first...
I think there may be a way to avoid the performance penalty with the use of strongly typed language...
If I were using C++, I would have several types for "strings". (WebQueryString, CommandArgString, SQL_QueryString, etc). Either built-in (or user specified) methods would santize/convert strings..
For example, if I wanted to use a web form value as a command-line argument (say for a filemanagement script), then the mere act of casting a WebQueryString to a CommandArgString would cause it to be properly escaped.
Similarly, passing a WebQueryString to a function that took a SQL_QueryString type would cause it to be properly escaped... (Well, I'm being a bit sloppy, queries would probably have to be split apart into many different types to represent the different components (noun, verb, etc))
Basically, a strong type system could be used to (either manually or automatically) force the programmer to santize the data... If designed & implemented well, this could result in faster development times and more secure software...
It really annoys me that just to properly write a simple set of PHP scripts (such as an address book with a MySQL backend), a great deal of effort is expended in just parsing and (more importantly) santizing data... Even with all the object-oriented and/or modular programming techniques, it is still all too easy to pass the wrong string (or call the wrong *_escape method) in a way that results in a security flaw...
The next article on this topic will be that the bees will be joining forces with the RIAA and issuing cease and decist orders to all icebergs and to every human with "ice", "berg", or "borg" in their name.
In what sense do you mean that its inlining is bad? Inlining seems like a very simple optimization to implement... I believe GCC uses heuristics (that can be modified with command-line parameters) to decide whether to inline a function or not. Perhaps isn't not doing what you want because either 1) What you want isn't likely to improve performance... 2) Your not supplying the correct options to force it to do what you want... 3) what you want isn't really possible: (like to inline a virtual function call from a base pointer).
Wouldn't it be simpler just to have Firefox manage its own memory ? I think it might be possible without a complete overhaul. (But I'm far from being a firefox developer)
If the problem is that new windows tabs create a lot of objects that each get interspersed in memory, then why don't they just:
Overload "new" so that objects from each page are allocated from separate pools of memory. (using separate 4k memory pages for each tab/window wouldn't be a big deal)
Many times, applications do something similar so that they can create and delete many objects rapidly without the incurring the overhead of (implicitly) calling malloc and free each time....
I have been advocating that we should have a moderation system for the editors/submitters of articles...
Of course, somehow the editors don't seem too interested in the idea.
For example, the original summary of the Hot-Air Balloon article contained some wildly inaccurate metric-to-english conversions (off by almost 25%) that snubbed the true feat accomplished... I applaud the editors for removing the inaccuracies, but submitters should be held more "publicly accountable" for their choice of article selection and summary descriptions.
Especially the guy with the french-sounding name that posts links to his own block as stories...
If you run for pres, I will vote for you and heartily encourage everyone to do the same...
Most Slashdoters should at least like the part about keeping their jobs...