It makes me feel conceited to think that I know better than wise and learned university staff.
That is utterly arrogant... but in any event, going from a 64-bit SPARC to a 32-bit x86 running linux might provide a significant advantage. Do you even have any clue of the circumstances that might have prompted such a decision by the University staff? No, none at all.
Usually, you'd send the MAC signal to any interface that has a physical connection to the NIC in question. If the machine is ATX and a NIC is powered, it will provide enough of a "connection" to your network, that it could receive some data on the hardware layer, make sense of it and issue a command to the system. Almost all boards with on-board NICs support the feature and others allow you to modify the MAC on the device.
WOL is really nice for system installations esp. WRT beowulf clusters, being the lazy admin that I am.
Or if you work for a University like I do, manage over 200 linux machines for multiple beowulf clusters, and most of all, live in the state of florida, where education budgets are slashed annually, paying the fees laid out by RedHat is impossible.
BTW, we've deployed CentOS-3 on all but one of our clusters with the lone cluster running CentOS-4.
Using the wrong data registers... ugh, i should have done this when I was sober. And watch the overflow on that long... It will take a while, but it'll overflow.
Were you correcting my spelling? Because I always make that mistake (myranet... it's myrinet damn it!). You know what I meant though;) It looks like BlueGene/L is using a hybrid backplane/hypertorus interconnect where a whole bunch of "machines" (more like system-on-a-chip) are connected via a backplane, then that case of "machines" is connected to another case in the same rack on some number of layers of interconnect. Then the racks are connected using some other protocol. Though you may not "be at liberty" to discuss this, the top500 site already disclosed an ample amount of information on the subject for any beowulfer to get the general idea of what type of interconnect topology/setup BlueGene/L is using.
And I quote:
The nodes are interconnected through multiple complementary high-speed low-latency networks, including a 3D torus network and a combining tree network. The physical machine architecture is targeted to be most closely tied to the 3D torus, a simple 3-dimensional nearest neighbor interconnect which is "wrapped" at the edges. An independent combining tree network provides for fast global operations, such as global max or global sum.
A slightly larger dose of logic would tell you that NASA has nothing to do with this cluster, that it belongs to the NNSA or the National Nuclear Security Agency. They are probably more interested in testing new reactor designes or running simulations to demonstrate the effects of an aircraft crashing into one of their reactor domes (though I honestly believe that no one really believes that will happen).
That's not how linpack works. Sure, increasing your number of nodes will give definite performance advantages to course-grained, embarassingly parallel applications, but Linpack is not one of these applications. As well, Linpack should not be used as a guide for raw floating point performance, but is much better suited to gauge throughput.
Linpack does its benchmarks using a more fine-grained algorithm, creating lots of communications for Message Passing to share segments of dense matrices for rather large linear systems. Not only is the number of nodes a factor, but so is the interconnect speed. If that cluster was using GigE for its interconnect, its Linpack benchmarks would not be nearly as impressive. Haven't RTFA but its likely that BlueGene/L is using Myranet or Infinband for its interconnect (or possibly a more proprietary backplane style interconnect, though that cluster is way too big for that).
These latest generations of high-speed interconnects (esp. Infinband) have brought clusters closer to the point of being near shared-memory performance and hence is more of a throughput test than anything else.
This description of the HPL benchmark (The "official" name for the Linpack benchmark) should provide some clarity as to how memory-dependent Linpack actually is:
The algorithm used by HPL can be summarized by the following keywords: Two-dimensional block-cyclic data distribution - Right-looking variant of the LU factorization with row partial pivoting featuring multiple look-ahead depths - Recursive panel factorization with pivot search and column broadcast combined - Various virtual panel broadcast topologies - bandwidth reducing swap-broadcast algorithm - backward substitution with look-ahead of depth 1.
They took a lot of time to get Linpack to be less shared-memory dependent, like adding the swap-broadcast algorithm (which i'm fairly certain was absent in the old mainframe version of Linpack), to make it more "fair" to run on a cluster versus a shared memory set up. However, on a typical cluster, Linpack can push your interconnect pretty hard, esp. if you are stuck on GigE. However, Linpack has _lots_ of settings and parameters to "tune" the benchmark for your particular cluster.
My point: Linpack/HPL is not an overall flops benchmark for a cluster. It measures the performance not only of double precision CPU performance, but also the performance of a cluster's interconnect.
Re:What it really does.
on
Firefox Hacks
·
· Score: 1
Easy DoS for his server... If you're on a fat pipe (Gigabit or some OC*_hello university students looking to get expelled_), you can probably use the java script libraries with mozilla (for parsing purposes) to whip up an app that floods his clipboard requests with an unlimited stream of data from/dev/urandom. Of course, you could do this on a smaller scale and just run up his bandwidth bill, say sending over a file to his server every minite or so. It's doable, right?
Yet I'm still compelled by the fact that two wrongs don't make a right. Or was it that two lefts don't make a right, but three do? Drinking is cool.
Imagine a beowulf cluster of those! No, really. All you'd need is a switch and some ethernet cables and you could have yourself a nifty 8 node cluster for under a grand... and bragging rights for probably having one of the worlds smallest clusters.
Users should not be able to set hard limits. You can, however, remove the -H and you should be able to reset the number of allowable processes (unless slackware has a default hard limit specified in/etc/security/limits.conf). It is up to the administrator to set reasonable softlimits (that users can change if need be) and reasonable hard limits to protect the server.
Your stack size is static, therefore it doesn't matter if you overflow the stack with and (n^n) algorithm or an (n^n)-1 algorithm as you will still exceed that same limit, hence the stack will be the exact same size, either way, when you crash.
I've recently been looking into running Linux on my primary desktop to record music with.
Don't even bother. The software that is available currently is utterly unacceptable for any type of music recording. Don't get me wrong, there are a few programs that have a lot of promise, but as of now there is nothing noteworthy. I've definitely got my share of Linux knowledge and I have had great difficulty getting the simplest of things to work correctly. Compiling and running is one thing... not having your entire plugin system crash on you is another. Windows and Macs have wonderful VST and DXi instruments with useful interfaces and very fast algorithms. Linux' LADSPA, while from a programmer's stand-point is quite impressive, is by no means a user-friendly setup.
In a nutshell: The features are not there, the stability is not there, the usability is not there, and the overall effectiveness of the applications is just not there.
Let it be directed by the UN or by like minded nations.
What libertarian party are you talking about? It is my knowledge that the libertarians would like nothing more that to lessen UN influence on any apsect of running this country.
Why did you even bring this up if you already knew that insider-trading, as described was unethical? Most people aren't aware of what ethics is and I pandered to them by providing a legal backing. The ethical argument I made was not about insider trading or hacking, it was about the scenario as a whole, ignoring the detriment of others for your own benefit. Both scenarios share this. Is it fair that these students get to know their acceptance status before anyone else, others that have to wait for their letters to arrive in the mail? Is it fair that no one else knew that the stock was going to tank and that the CEO was the only one warned?
Nervous, anxious wannabe Harvard type worries if he or she will be accepted.
Nervous, anxious wannabe Harvard type is told of way to find out early
Nervous, anxious wannabe Harvard type discovers acceptance letter and his or her anxiety/nervousness subsides.
Nervous, anxious Harvard type knew that it was wrong to do so, since any intelligent person knows that if Harvard hasn't told you personally, they don't want you to know (not to mention that some self-proclaimed hacker told you how to hack in to find out, ergo violating the school's acceptable use policy before even becoming a student), and is promptly made a disgrunteled, was-gonna-be Harvard type.
Then,
Nervous, anxious big-company-CEO worries if his company's stock, of which he owns 10 million shares, is about to tank.
Nervous, anxious big-company-CEO is told of a way to find out before said stock tanks.
Nervous, anxious big-company-CEO sells said stock before it tanks and his or her anxiety/nervousness subsides.
Nervous, anxious big-company-CEO knew that it was wrong to do so, since any business leader knows that there are strict laws prohibiting insider-trading and is promptly made someone's bitch in a federal prison...
I don't know about you, but I can see, although the magnitude of effect is quite different between the two scenarios, how the ethics of both are quite similar.
Don't be a fucking tool... its Musicmatch Jukebox. What is with these trolls and their apparent deafness when it comes to reading simple posts? It's starting to get a bit irritating... "I misread the article and though y instead of x". Its not funny anymore! jeesh
The IMF and the World Bank are the same thing. The World Bank is the money-force (American-based) behing the IMF (European-based). If she takes down the World Bank, the IMF will go too.
Slashdot Mirror
It makes me feel conceited to think that I know better than wise and learned university staff.
That is utterly arrogant... but in any event, going from a 64-bit SPARC to a 32-bit x86 running linux might provide a significant advantage. Do you even have any clue of the circumstances that might have prompted such a decision by the University staff? No, none at all.
Hehe, I remember that from a discrete math class, demonstrating the inferiority of spoken language with regards to ambiguity.
That's my new sig.
Usually, you'd send the MAC signal to any interface that has a physical connection to the NIC in question. If the machine is ATX and a NIC is powered, it will provide enough of a "connection" to your network, that it could receive some data on the hardware layer, make sense of it and issue a command to the system. Almost all boards with on-board NICs support the feature and others allow you to modify the MAC on the device.
WOL is really nice for system installations esp. WRT beowulf clusters, being the lazy admin that I am.
Or if you work for a University like I do, manage over 200 linux machines for multiple beowulf clusters, and most of all, live in the state of florida, where education budgets are slashed annually, paying the fees laid out by RedHat is impossible.
BTW, we've deployed CentOS-3 on all but one of our clusters with the lone cluster running CentOS-4.
Using the wrong data registers... ugh, i should have done this when I was sober. And watch the overflow on that long... It will take a while, but it'll overflow.
Better:
/* imaginary register for shock intensity */ /* imaginary register for shock enable */ /* imaginary register for shock status */
:"=r" (i) \
:"r" (k) \
:"eax") \ /* return status from register */
/* lets increase the dosage every time we shock */
/* we never get here */
/* fake, pseudo-asm */
#define _shock(mag) __asm__ volitile \
("movl %0,%%sri \
mov $1,%%sre \
movl %%srs,%eax \
movl %eax,%1" \
#define _unshock() __asm__ volitile \
("movw $0,%%sri")
int main(void){
long magnitude;
int status;
for (magnitude = 10; magnitude != 0; magnitude *= 2){
if (!(status = shock(magnitude))) return EXIT_FAILURE;
sleep(1);
if ((status = unshock())) return EXIT_FAILURE;
}
return EXIT_SUCCESS;
}
It might not work, but at least the idea is sadistic enough...
Were you correcting my spelling? Because I always make that mistake (myranet... it's myrinet damn it!). You know what I meant though ;) It looks like BlueGene/L is using a hybrid backplane/hypertorus interconnect where a whole bunch of "machines" (more like system-on-a-chip) are connected via a backplane, then that case of "machines" is connected to another case in the same rack on some number of layers of interconnect. Then the racks are connected using some other protocol. Though you may not "be at liberty" to discuss this, the top500 site already disclosed an ample amount of information on the subject for any beowulfer to get the general idea of what type of interconnect topology/setup BlueGene/L is using.
7 101
And I quote:
The nodes are interconnected through multiple complementary high-speed low-latency networks, including a 3D torus network and a combining tree network. The physical machine architecture is targeted to be most closely tied to the 3D torus, a simple 3-dimensional nearest neighbor interconnect which is "wrapped" at the edges. An independent combining tree network provides for fast global operations, such as global max or global sum.
http://www.top500.org/sublist/System.php?TB=2&id=
Enjoy.
A slightly larger dose of logic would tell you that NASA has nothing to do with this cluster, that it belongs to the NNSA or the National Nuclear Security Agency. They are probably more interested in testing new reactor designes or running simulations to demonstrate the effects of an aircraft crashing into one of their reactor domes (though I honestly believe that no one really believes that will happen).
That's not how linpack works. Sure, increasing your number of nodes will give definite performance advantages to course-grained, embarassingly parallel applications, but Linpack is not one of these applications. As well, Linpack should not be used as a guide for raw floating point performance, but is much better suited to gauge throughput.
Linpack does its benchmarks using a more fine-grained algorithm, creating lots of communications for Message Passing to share segments of dense matrices for rather large linear systems. Not only is the number of nodes a factor, but so is the interconnect speed. If that cluster was using GigE for its interconnect, its Linpack benchmarks would not be nearly as impressive. Haven't RTFA but its likely that BlueGene/L is using Myranet or Infinband for its interconnect (or possibly a more proprietary backplane style interconnect, though that cluster is way too big for that).
These latest generations of high-speed interconnects (esp. Infinband) have brought clusters closer to the point of being near shared-memory performance and hence is more of a throughput test than anything else.
This description of the HPL benchmark (The "official" name for the Linpack benchmark) should provide some clarity as to how memory-dependent Linpack actually is:
The algorithm used by HPL can be summarized by the following keywords: Two-dimensional block-cyclic data distribution - Right-looking variant of the LU factorization with row partial pivoting featuring multiple look-ahead depths - Recursive panel factorization with pivot search and column broadcast combined - Various virtual panel broadcast topologies - bandwidth reducing swap-broadcast algorithm - backward substitution with look-ahead of depth 1.
http://www.netlib.org/benchmark/hpl/
They took a lot of time to get Linpack to be less shared-memory dependent, like adding the swap-broadcast algorithm (which i'm fairly certain was absent in the old mainframe version of Linpack), to make it more "fair" to run on a cluster versus a shared memory set up. However, on a typical cluster, Linpack can push your interconnect pretty hard, esp. if you are stuck on GigE. However, Linpack has _lots_ of settings and parameters to "tune" the benchmark for your particular cluster.
My point: Linpack/HPL is not an overall flops benchmark for a cluster. It measures the performance not only of double precision CPU performance, but also the performance of a cluster's interconnect.
Goddamn, out of mod points... hahahaha
Easy DoS for his server... If you're on a fat pipe (Gigabit or some OC*_hello university students looking to get expelled_), you can probably use the java script libraries with mozilla (for parsing purposes) to whip up an app that floods his clipboard requests with an unlimited stream of data from /dev/urandom. Of course, you could do this on a smaller scale and just run up his bandwidth bill, say sending over a file to his server every minite or so. It's doable, right?
Yet I'm still compelled by the fact that two wrongs don't make a right. Or was it that two lefts don't make a right, but three do? Drinking is cool.
HAL sings the whole song but starts to _slow_ down to an almost inaudible frequency near "can't afford a carriage", if I recall correctly.
Imagine a beowulf cluster of those! No, really. All you'd need is a switch and some ethernet cables and you could have yourself a nifty 8 node cluster for under a grand... and bragging rights for probably having one of the worlds smallest clusters.
Moderation: +5 Funny, Insightful Troll
Don't act like you didn't know.
Nevermind... The process limit looks like it is protected by default (which is good) from user tampering.
Users should not be able to set hard limits. You can, however, remove the -H and you should be able to reset the number of allowable processes (unless slackware has a default hard limit specified in /etc/security/limits.conf). It is up to the administrator to set reasonable softlimits (that users can change if need be) and reasonable hard limits to protect the server.
Your stack size is static, therefore it doesn't matter if you overflow the stack with and (n^n) algorithm or an (n^n)-1 algorithm as you will still exceed that same limit, hence the stack will be the exact same size, either way, when you crash.
However, the joke is seen and that was funny.
I've recently been looking into running Linux on my primary desktop to record music with.
Don't even bother. The software that is available currently is utterly unacceptable for any type of music recording. Don't get me wrong, there are a few programs that have a lot of promise, but as of now there is nothing noteworthy. I've definitely got my share of Linux knowledge and I have had great difficulty getting the simplest of things to work correctly. Compiling and running is one thing... not having your entire plugin system crash on you is another. Windows and Macs have wonderful VST and DXi instruments with useful interfaces and very fast algorithms. Linux' LADSPA, while from a programmer's stand-point is quite impressive, is by no means a user-friendly setup.
In a nutshell: The features are not there, the stability is not there, the usability is not there, and the overall effectiveness of the applications is just not there.
um... you end up with an infinite loop that way too.
l drifle()... Stack overflow.
main()->readsite()->buildrifle()->readsite()->bui
Let it be directed by the UN or by like minded nations.
What libertarian party are you talking about? It is my knowledge that the libertarians would like nothing more that to lessen UN influence on any apsect of running this country.
Why did you even bring this up if you already knew that insider-trading, as described was unethical? Most people aren't aware of what ethics is and I pandered to them by providing a legal backing. The ethical argument I made was not about insider trading or hacking, it was about the scenario as a whole, ignoring the detriment of others for your own benefit. Both scenarios share this. Is it fair that these students get to know their acceptance status before anyone else, others that have to wait for their letters to arrive in the mail? Is it fair that no one else knew that the stock was going to tank and that the CEO was the only one warned?
Scenario:
Nervous, anxious wannabe Harvard type worries if he or she will be accepted.
Nervous, anxious wannabe Harvard type is told of way to find out early
Nervous, anxious wannabe Harvard type discovers acceptance letter and his or her anxiety/nervousness subsides.
Nervous, anxious Harvard type knew that it was wrong to do so, since any intelligent person knows that if Harvard hasn't told you personally, they don't want you to know (not to mention that some self-proclaimed hacker told you how to hack in to find out, ergo violating the school's acceptable use policy before even becoming a student), and is promptly made a disgrunteled, was-gonna-be Harvard type.
Then,
Nervous, anxious big-company-CEO worries if his company's stock, of which he owns 10 million shares, is about to tank.
Nervous, anxious big-company-CEO is told of a way to find out before said stock tanks.
Nervous, anxious big-company-CEO sells said stock before it tanks and his or her anxiety/nervousness subsides.
Nervous, anxious big-company-CEO knew that it was wrong to do so, since any business leader knows that there are strict laws prohibiting insider-trading and is promptly made someone's bitch in a federal prison...
I don't know about you, but I can see, although the magnitude of effect is quite different between the two scenarios, how the ethics of both are quite similar.
Don't be a fucking tool... its Musicmatch Jukebox. What is with these trolls and their apparent deafness when it comes to reading simple posts? It's starting to get a bit irritating... "I misread the article and though y instead of x". Its not funny anymore! jeesh
The IMF and the World Bank are the same thing. The World Bank is the money-force (American-based) behing the IMF (European-based). If she takes down the World Bank, the IMF will go too.