From the article they say that clusters max out at 64 machines, limiting their size - but also it's claimed that the cluster acts like a single machine, so my question is, why can't you cluster the clusters to use 4096 machines. Is it simply a case of (lack of) bandwidth linking the machines together?
It's probably a software limitation, but probably not a bad one. Large clusters get unwieldy quickly, and network latency and bandwidth is the bane of any parallel programmers existence. Communication between the nodes of a cluster is several orders of magnitude slower than referencing internal memory, and any no real parallel program has autonomous nodes. It's no coincidence that Donald Becker, a major contributor to Beowulf on Linux, also wrote huge chunks of the kernel networking, tons of network card drivers, and a network channel bonding implementation for Linux.
My point is that you can create a cluster with thousands of nodes, but doing so is an administrative and technological nightmare. For most parallel problems, it's much easier (and generally more efficient) to have a smaller number of more powerful nodes.
It's quite common in the Linux community, however, to distribute software in a manner which does not require the user to read anything, much less a license, to install and use it.
Agreed. But JC was extra careful with the quake source code (even more so than he was with the doom src code).
I wonder if the package required Slade to accept a license before accessing the code.
Not unless Slate found a way to delete the header of every source and header file, along with deliberately not read the file called readme.txt after uncompressing the archive.
Question: Do we know that small packets have less trouble gettinf through a cloged network pipe then large packets? This is what we need to do the encryption thing.
It doesn't matter. Bandwidth isn't _primary_ issue with modems -- it's latency (due to shitty buffering for hw codecs). And the answer to your question is that it depends on the commlunications protocol. PPP doesn't have priority packets, and neither does IPv4 (IPv6 does, though), so there's no way of doing what you're asking about with the current networking protocol.
Again, this is irrellevent. There isn't any small packet that a client can send verifying that the client hasn't been hacked to supt auto-aim, that the user hasn't re-written the ogl lib to have translucent walls, or modified theplayer models so they have lines indicating their trajectory.
He is violating his license by attempting to distribute a closed modification to a GPLed program under a more restricted license.
Therefore he is in violation of his license.
I really can't see why this is such a big deal. This guy is violating a software license, and breaking the law. Just because it's not written on a piece of shrinkwrapped cardboard or in a dialog box that you can't pass until you enter a 40 digit alphanumeric string doesn't make it any less binding.
Some dude said something like: Democracy is the worst type of governement, but we haven't found anything better yet.
AFAIK, it was Winston Churchill...
"No one pretends that democracy is perfect or all-wise. Indeed, it has been said that democracy is the worst form of Government except all those other forms that have been tried from time to time."
GnomeICU doesn't require panel, gnome-session or anything else. You need to have the gnome libraries present, but if you launch it with a -a (eg "gnomeicu -a") it won't try and dock to the panel.
Pretty damn convenient considering E16 handles both GNOME and KDE apps, and considering I despise panels, bars, or anything else that eats up 1/8 of my screen w/o providing any useful functionality. I get both KDE and GNOME functionality, a fucking awesome WM, and GnomeICU.
All banks and credit card companies insure that customers won't have to pay for fraud.
All companies offset all incurred costs to customers. That's the way business works.
On a side note, this could have been avoided had law permitted better/higher encryption on the CDUniverse site.
Wrong again. This has nothing to do with encryption and everything to do with improperly secured servers storing customer's CC numbers in databases. Like I said in a previous post, servers should _never_ store the actual credit card number. I think a one-way hash (via MD5 or equiv) would be suitable for e-commerce infrastructures (eg a large db of them is useless to script kiddiez, and it still allows the company to authenticate).
Alot of E-commerce companies put big efforts in making the "shopping experience" as easy and interresting for the user. Wonderful, the company stored your credit card number, you wont have to type it in again when you shop later!
This particularly irritating since both IE5 and Mozilla (soon to be NS5) offer to fill in forms at the click of a button. Also, if the company wants to store anything in their database it should be a one-way hash (MD5 or equiv) of the card number, _not_ the number itself.
That was one of the best Slashdot interviews I've seen since...I don't know when. It's good to see someone (or a group) speak thoroughly and clearly. And I'm sure that comes from years of dealing with aggressive press who don't want to listen.
Agreed. It's nice to see an educated and well articulated piece on/. for a change.
No one argues that hackers are mis-portrayed in the media.
I disagree. Supposedly reputable news establishments generally attribute report break-ins, defacements, and theft (eg _cracking_ behavior, or malicious hacking) to hackers. Unlike most hackers, I'm not particularly concerned about the "hacker" label. What I am concerned about is the implicit message the media is sending to the uninformed: that learning and privacy are analagous to criminal behavior merely because knowledge regarding either could be used to exploit badly designed or implemented security models. Correlation is not causation. Just because an apple is a fruit does not mean it's an orange.
Tell them what it means. It's a fine line between informative and over zealous
It is extremely difficult to convince most people to sacrifice convenience for security (witness the hundred of thousands of unprotected and unpatched Windows 9X and Windows NT systems accessible by any other machine via internet). Even though connecting a Windows 9X machine to internet is akin to hang gliding in a military no-fly zone, attempting to explain this to the masses will automatically place you in the "paranoid security nut" category. I'm not saying you shouldn't try (I've been trying to move my users from telnet to ssh for monthes), but noone should expect a chocolate coated, overnight change.
doesn't the way *nix in general handles file permissions pretty much prevent most malicious code such as worms from getting in and trashing things?
Nope. The god-like powers of root supercede file, memory, and device permissions. A lot has changed since 1988, but history has a way of repeating itself.
Remember, all it takes is a badly placed sprinf(), malloc(), or strcpy() in a SUID daemon to bring a so-called bulletproof security model to it's knees.
I think you are missing the distinction between emphasizing theming and providing customization options. Empasizing theming would mean providing theming capabilities at the expense of other features such as stability, usability, portability, etc. Noone can honestly tell me that the the most popular themable window managers (E, KWM, Sawmill, WM, FVWM, etc) do so at a the expense of other features (besides speed, that is. But cumulative speed is more a function of processor speed, user stupidity, and theme choices than window manager implementation).
As for Enlightenment, it seems silly for people to be complaining about something that is still nowhere near complete (don't believe me? check raster & mandrake's huge list on enlightenment.org). Also, E was completely rewritten betweem DR0.13 and DR0.14. So if raster & Co. were making any attempt to correlate version number incrementation with stability or completeness, then we would be at about DR0.02, not DR0.16.
Okay, now I'm rambling. I just saw the start of a big WM bashfest...
I wonder if hte following implementation would help: Set up a rotary multiplex with all pertinent client game stats (eg hp, velocity, weapons, ammo, etc) and appending, XORing, or checksumming them and appending the result to server-bound packets (ex for XOR ones, obviously). Set up a server thread dedicated to verifying the validity of clients specs, flagging them when they don't match, and booting a client when it exceeds the server's flag limit. Granted this probably wouldn't stop minor client hacks, but major ones (eg million hp, no damage, flying, etc etc) would be caught and logged. Dunno if this would add too much completity to the server... I have hte src but I haven't had time to poke around in it. ANyone more qualified care to comment?
-- odds of being killed by lighning and
Re:Possible additions to the h4x0rz guide
on
Guide to Slashdot
·
· Score: 1
use the 120 char max as a luser exploit by allowing it to partially mutilate an anchor tag. this should screw up slashdot for everyone, and make you look uber 31337.
-- rot13 my email address for the real thing
Possible additions to the h4x0rz guide
on
Guide to Slashdot
·
· Score: 5
Gotta have something about selecting all of the modular slashboxes for your user account, just so you can drag slashdot to a screaming halt and waste other lusers time.
Don't forget the HUGE cryptic geek blocks or arbitrarily complex self-decrypting perl code that prints "Hi Mom!" when executed. Use this as your signature.
Post everythign as html formatted, and leave an unterminated font modifier tag at the begining of the message.
use gratuitous
<BR>
tags to make your
document seem really long (and therefore important)
graduly hvae the spellnign and grammar deterate by arbitrly omitng vwls nd missplng wrdz.
use lots of cryptic abbreviateions and assume people understand what you're talking about.
don't justify any of your arguments. if someone responds asking you to support your statements, respond by posting 200 anonymous messages making derogatory comments about the poster's mother.
Besides the gratuitous spelling errors (wtf is Free BSD??), and the inaccurate licensing descriptions, the author also displays a complete lack of knowledge regarding the competition between DR-DOS and MS-DOS.
But DR-DOS couldn't cut it, and disappeared from the market. Frankenberg believes DR-DOS was driven out by a ruthless Bill Gates. Caldera released a lot of information a few monthes back regarding DR-DOS and the trial. Every version of DR-DOS had comperable or higher reviews than MS-DOS, and Microsoft refused to distribute Windows APIs to companies that worked with DR-DOS, and Microsoft deliberately put code in beta versions of Windows 3.1 so it behave erratically or crash with DR-DOS.
But hasn't Microsoft's monopoly power dwindled since the trail started? New competitors now are on the scene, including the Linux operating system. AFAIK a company is considered a monopoly if it controls 80% or more of a market. MS controls somewhere in the mid-90 percent range. In other words, millions of people could switch to Linux tomorrow and MS would still enjoy monopoly power.
In your current company, Encanto, you don't use Microsoft's software. You're using a Linux-like software, right? Ugh. Has linux really become so much of a cultural icon that all versions of UNIX and UNIX variants are assumed to be "Linux-like"? "Hey, you're using {MacOS,BeOS,GNOME,KDE,any windowed-gui interface}, that's Windows-like right?"
Why was it so tough for you to take on Microsoft? Did DR-DOS fail for the simple reason that Microsoft had better software?
With the possible exception of MS Office vs. Wordperfect Suite (where MS also used illegal bundling tactics), MS has never had _better_ software than any of it's competitors -- their products have always been marginally worse at best. DR-DOS was no exception -- it outperformed and outrated MS-DOS for years until MS tied Windows to DOS, and refused to deal with Novell while they owned DR-DOS.
In retrospect, isn't browser technology fairly insignificant? Would more browser competition change the technology market?
Yup, browsers sure are insignificant. URLs on billboards, _tons_ of ads on television for web pages -- I can't think of anything significant effect the _web_ or _web browsers_ have had on society or business. Whoops, actually internet (specificially the web) is changing everything! But I guess that's pretty insignificant. As for the second sentence.... Hey author, did you fail your fucking econ class in middle school? Ever heard of this little thing called market forces, or perhaps supply and demand?
"...we both wated and wated for Ismael because he was supposed to bring the ounce so we could get high..."
NO ONE expressed any sort of concern that this thirteen-year-old kid might be getting involved in drugs (a threat to himself). That a school-teacher could applaud such content without feeling some sort of concern for the child's perceptions and behaviour in regard to drugs is appalling to me.
Dude, if he's got an ounce of _any_ drug, he's not just "getting involved".
Do you think a win in the Bernstein case will deregulate encryption export, or will the result just be more political dodging? -- rot13 my email address for the real thing
The present invention overcomes the problems of the prior art and provides a microprocessor which is faster than microprocessors of the prior art, is capable of running all of the software for all of the operating systems which may be run by a large number of families of prior art microprocessors, yet is less expensive than prior art microprocessors.
Rather than using a microprocessor with more complicated hardware to accelerate its operation, the present invention combines an enhanced hardware processing portion (referred to as a "morph host" in this specification) which is much simpler than state of the art microprocessors and an emulating software portion (referred to as "code morphing software" in this specification) in a manner that the two portions function together as a microprocessor with more capabilities than any known competitive microprocessor. More particularly, a morph host is a processor which includes hardware enhancements to assist in having state of a target computer immediately at hand when an exception or error occurs, while code morphing software is software which translates the instructions of a target program to morph host instructions for the morph host and responds to exceptions and errors by replacing working state with correct target state when necessary so that correct retranslations occur. Code morphing software may also include various processes for enhancing the speed of processing. Rather than providing hardware to enhance the speed of processing as do all of the very fast prior art microprocessors, the present invention allows a large number of acceleration enhancement techniques to be carried out in selectable stages by the code morphing software. Providing the speed enhancement techniques in the code morphing software allows the morph host to be implemented using much less complicated hardware which is faster and substantially less expensive than the hardware of prior art microprocessors. As a comparison, one embodiment of the present invention designed to run all available X86 applications is implemented by a morph host including approximately one-quarter of the number of gates of the Pentium Pro microprocessor yet runs X86 applications substantially faster than does the Pentium Pro microprocessor or any other known microprocessor capable of processing these applications.
The code morphing software utilizes certain techniques which have previously been used only by programmers designing new software or emulating new hardware. The morph host includes hardware enhancements especially adapted to allow the acceleration techniques provided by the code morphing software to be utilized efficiently. These hardware enhancements allow the code morphing software to implement acceleration techniques over a broader range of instructions. These hardware enhancements also permit additional acceleration techniques to be practiced by the code morphing software which are unavailable in hardware processors and could not be implemented in those processors except at exorbitant cost. These techniques significantly increase the speed of the microprocessor of the present invention compared to the speeds of prior art microprocessors practicing the execution of native instruction sets.
...
Basically, they're describing a combo hardware/software dynamic OS emulation implementation. Almost like a JIT compiler for ASM on crack -- cool.
Beowulf clustered applications (along with COWs and many other distributed systems) are developed using mainly MPI or p-threads. There is almost no good reason to use PVM except to support legacy code.
And of the two, MPI is generally vastly superior to p-threads because (IIRC) MPI is a higher level implementation which provides communication routines that are optimized for each particular hardware implementation (depending on the version/implementation of MPI). For example, the actual implementation of MPI_Reduce() will vary depending on whether the nodes are in a shared/non-shared memory environment -- in a non-shared environment (eg. a Beowulf cluster), MPI uses a tree-method in order to distribute the data among the nodes in parallel.
Anyways, the point is that MPI is really just a communications specification (Message Passing Interface) with language bindings (C and Fortran). What you really want is a set of client/server daemons that _implement_ MPI.
Slashdot Mirror
From the article they say that clusters max out at 64 machines, limiting their size - but also it's claimed that the cluster acts like a single machine, so my question is, why can't you cluster the clusters to use 4096 machines. Is it simply a case of (lack of) bandwidth linking the machines together?
It's probably a software limitation, but probably not a bad one. Large clusters get unwieldy quickly, and network latency and bandwidth is the bane of any parallel programmers existence. Communication between the nodes of a cluster is several orders of magnitude slower than referencing internal memory, and any no real parallel program has autonomous nodes. It's no coincidence that Donald Becker, a major contributor to Beowulf on Linux, also wrote huge chunks of the kernel networking, tons of network card drivers, and a network channel bonding implementation for Linux.
My point is that you can create a cluster with thousands of nodes, but doing so is an administrative and technological nightmare. For most parallel problems, it's much easier (and generally more efficient) to have a smaller number of more powerful nodes.
--
odds of being killed by lighning and
Ouch, change to text formatted output.. :)
It's quite common in the Linux community, however, to distribute software in a manner which does not require the user to read anything, much less a license, to install and use it.
Agreed. But JC was extra careful with the quake source code (even more so than he was with the doom src code).
I wonder if the package required Slade to accept a license before accessing the code.
Not unless Slate found a way to delete the header of every source and header file, along with deliberately not read the file called readme.txt after uncompressing the archive.
--
odds of being killed by lighning and
Question: Do we know that small packets have less trouble gettinf through a cloged network pipe then large packets? This is what we need to do the encryption thing.
It doesn't matter. Bandwidth isn't _primary_ issue with modems -- it's latency (due to shitty buffering for hw codecs). And the answer to your question is that it depends on the commlunications protocol. PPP doesn't have priority packets, and neither does IPv4 (IPv6 does, though), so there's no way of doing what you're asking about with the current networking protocol.
Again, this is irrellevent. There isn't any small packet that a client can send verifying that the client hasn't been hacked to supt auto-aim, that the user hasn't re-written the ogl lib to have translucent walls, or modified theplayer models so they have lines indicating their trajectory.
--
odds of being killed by lighning and
He is violating his license by attempting to distribute a closed modification to a GPLed program under a more restricted license.
Therefore he is in violation of his license.
I really can't see why this is such a big deal. This guy is violating a software license, and breaking the law. Just because it's not written on a piece of shrinkwrapped cardboard or in a dialog box that you can't pass until you enter a 40 digit alphanumeric string doesn't make it any less binding.
--
odds of being killed by lighning and
This could only happen if _each_ contributing author for a project agreed to the license change.
:)
Otherwise the code written by the decenting author and hte derivative code has to be removed.
So, yeah this is technically possible, but hwat are the chances that every contributor to the Linux kernel would be willing to sell out?
Keep up the good work RMS.
--
odds of being killed by lighning and
Democracy is the worst type of governement, but we haven't found anything better yet.
AFAIK, it was Winston Churchill...
found this here...
--
odds of being killed by lighning and
GnomeICU doesn't require panel, gnome-session or anything else. You need to have the gnome libraries present, but if you launch it with a -a (eg "gnomeicu -a") it won't try and dock to the panel.
Pretty damn convenient considering E16 handles both GNOME and KDE apps, and considering I despise panels, bars, or anything else that eats up 1/8 of my screen w/o providing any useful functionality. I get both KDE and GNOME functionality, a fucking awesome WM, and GnomeICU.
--
odds of being killed by lighning and
All banks and credit card companies insure that customers won't have to pay for fraud.
All companies offset all incurred costs to customers. That's the way business works.
On a side note, this could have been avoided had law permitted better/higher encryption on the CDUniverse site.
Wrong again. This has nothing to do with encryption and everything to do with improperly secured servers storing customer's CC numbers in databases. Like I said in a previous post, servers should _never_ store the actual credit card number. I think a one-way hash (via MD5 or equiv) would be suitable for e-commerce infrastructures (eg a large db of them is useless to script kiddiez, and it still allows the company to authenticate).
--
odds of being killed by lighning and
Alot of E-commerce companies put big efforts in making the "shopping experience" as easy and interresting for the user. Wonderful, the company stored your credit card number, you wont have to type it in again when you shop later!
This particularly irritating since both IE5 and Mozilla (soon to be NS5) offer to fill in forms at the click of a button. Also, if the company wants to store anything in their database it should be a one-way hash (MD5 or equiv) of the card number, _not_ the number itself.
--
odds of being killed by lighning and
That was one of the best Slashdot interviews I've seen since...I don't know when. It's good to see someone (or a group) speak thoroughly and clearly. And I'm sure that comes from years of dealing with aggressive press who don't want to listen.
/. for a change.
Agreed. It's nice to see an educated and well articulated piece on
No one argues that hackers are mis-portrayed in the media.
I disagree. Supposedly reputable news establishments generally attribute report break-ins, defacements, and theft (eg _cracking_ behavior, or malicious hacking) to hackers. Unlike most hackers, I'm not particularly concerned about the "hacker" label. What I am concerned about is the implicit message the media is sending to the uninformed: that learning and privacy are analagous to criminal behavior merely because knowledge regarding either could be used to exploit badly designed or implemented security models. Correlation is not causation. Just because an apple is a fruit does not mean it's an orange.
Tell them what it means. It's a fine line between informative and over zealous
It is extremely difficult to convince most people to sacrifice convenience for security (witness the hundred of thousands of unprotected and unpatched Windows 9X and Windows NT systems accessible by any other machine via internet). Even though connecting a Windows 9X machine to internet is akin to hang gliding in a military no-fly zone, attempting to explain this to the masses will automatically place you in the "paranoid security nut" category. I'm not saying you shouldn't try (I've been trying to move my users from telnet to ssh for monthes), but noone should expect a chocolate coated, overnight change.
--
odds of being killed by lighning and
doesn't the way *nix in general handles file permissions pretty much prevent most malicious code such as worms from getting in and trashing things?
Nope. The god-like powers of root supercede file, memory, and device permissions. A lot has changed since 1988, but history has a way of repeating itself.
Remember, all it takes is a badly placed sprinf(), malloc(), or strcpy() in a SUID daemon to bring a so-called bulletproof security model to it's knees.
--
odds of being killed by lighning and
I think you are missing the distinction between emphasizing theming and providing customization options. Empasizing theming would mean providing theming capabilities at the expense of other features such as stability, usability, portability, etc. Noone can honestly tell me that the the most popular themable window managers (E, KWM, Sawmill, WM, FVWM, etc) do so at a the expense of other features (besides speed, that is. But cumulative speed is more a function of processor speed, user stupidity, and theme choices than window manager implementation).
As for Enlightenment, it seems silly for people to be complaining about something that is still nowhere near complete (don't believe me? check raster & mandrake's huge list on enlightenment.org). Also, E was completely rewritten betweem DR0.13 and DR0.14. So if raster & Co. were making any attempt to correlate version number incrementation with stability or completeness, then we would be at about DR0.02, not DR0.16.
Okay, now I'm rambling. I just saw the start of a big WM bashfest...
--
odds of being killed by lighning and
I like how the poster noted that IceWM "strives to be ... KDE compliant."
"It's like this. We're really trying hard to almost work correctly sometimes with KDE."
--
odds of being killed by lighning and
I agree.
I wonder if hte following implementation would help: Set up a rotary multiplex with all pertinent client game stats (eg hp, velocity, weapons, ammo, etc) and appending, XORing, or checksumming them and appending the result to server-bound packets (ex for XOR ones, obviously). Set up a server thread dedicated to verifying the validity of clients specs, flagging them when they don't match, and booting a client when it exceeds the server's flag limit. Granted this probably wouldn't stop minor client hacks, but major ones (eg million hp, no damage, flying, etc etc) would be caught and logged. Dunno if this would add too much completity to the server... I have hte src but I haven't had time to poke around in it. ANyone more qualified care to comment?
--
odds of being killed by lighning and
use the 120 char max as a luser exploit by allowing it to partially mutilate an anchor tag. this should screw up slashdot for everyone, and make you look uber 31337.
-- rot13 my email address for the real thing
<BR>
tags to make your
document seem really long (and therefore important)
-- rot13 my email address for the real thing
Besides the gratuitous spelling errors (wtf is Free BSD??), and the inaccurate licensing descriptions, the author also displays a complete lack of knowledge regarding the competition between DR-DOS and MS-DOS.
But DR-DOS couldn't cut it, and disappeared from the market. Frankenberg believes DR-DOS was driven out by a ruthless Bill Gates.
Caldera released a lot of information a few monthes back regarding DR-DOS and the trial. Every version of DR-DOS had comperable or higher reviews than MS-DOS, and Microsoft refused to distribute Windows APIs to companies that worked with DR-DOS, and Microsoft deliberately put code in beta versions of Windows 3.1 so it behave erratically or crash with DR-DOS.
But hasn't Microsoft's monopoly power dwindled since the trail started? New competitors now are on the scene, including the Linux operating system.
AFAIK a company is considered a monopoly if it controls 80% or more of a market. MS controls somewhere in the mid-90 percent range. In other words, millions of people could switch to Linux tomorrow and MS would still enjoy monopoly power.
In your current company, Encanto, you don't use Microsoft's software. You're using a Linux-like software, right?
Ugh. Has linux really become so much of a cultural icon that all versions of UNIX and UNIX variants are assumed to be "Linux-like"? "Hey, you're using {MacOS,BeOS,GNOME,KDE,any windowed-gui interface}, that's Windows-like right?"
Why was it so tough for you to take on Microsoft? Did DR-DOS fail for the simple reason that Microsoft had better software?
With the possible exception of MS Office vs. Wordperfect Suite (where MS also used illegal bundling tactics), MS has never had _better_ software than any of it's competitors -- their products have always been marginally worse at best. DR-DOS was no exception -- it outperformed and outrated MS-DOS for years until MS tied Windows to DOS, and refused to deal with Novell while they owned DR-DOS.
In retrospect, isn't browser technology fairly insignificant? Would more browser competition change the technology market?
Yup, browsers sure are insignificant. URLs on billboards, _tons_ of ads on television for web pages -- I can't think of anything significant effect the _web_ or _web browsers_ have had on society or business. Whoops, actually internet (specificially the web) is changing everything! But I guess that's pretty insignificant. As for the second sentence.... Hey author, did you fail your fucking econ class in middle school? Ever heard of this little thing called market forces, or perhaps supply and demand?
-- rot13 my email address for the real thing
"...we both wated and wated for Ismael because he was supposed to bring the ounce so we could get high..."
NO ONE expressed any sort of concern that this thirteen-year-old kid might be getting involved in drugs (a threat to himself). That a school-teacher could applaud such content without feeling some sort of concern for the child's perceptions and behaviour in regard to drugs is appalling to me.
Dude, if he's got an ounce of _any_ drug, he's not just "getting involved".
-- rot13 my email address for the real thing
Do you think a win in the Bernstein case will deregulate encryption export, or will the result just be more political dodging?
-- rot13 my email address for the real thing
...From the patent article...
...
Basically, they're describing a combo hardware/software dynamic OS emulation implementation.
Almost like a JIT compiler for ASM on crack -- cool.
...
-- rot13 my email address for the real thing
Beowulf clustered applications (along with COWs and many other distributed systems) are developed using mainly MPI or p-threads. There is almost no good reason to use PVM except to support legacy code.
And of the two, MPI is generally vastly superior to p-threads because (IIRC) MPI is a higher level implementation which provides communication routines that are optimized for each particular hardware implementation (depending on the version/implementation of MPI). For example,
the actual implementation of MPI_Reduce() will vary depending on whether the nodes are in a shared/non-shared memory environment -- in a non-shared environment (eg. a Beowulf cluster), MPI uses a tree-method in order to distribute the data among the nodes in parallel.
Anyways, the point is that MPI is really just a communications specification (Message Passing Interface) with language bindings (C and Fortran). What you really want is a set of client/server daemons that _implement_ MPI.
Okay, I'm done with the conch now.