What the world needs now is love, sweet love It's the only thing that there's just too little of What the world needs now is love, sweet love, No not just for some but for everyone.
Yes - what a fantastic idea. Why not make a video about a technology relavant to Linux, *BSD and Solaris really difficult to access on Linux, *BSD and Solaris machines. Brilliant.
The (patent encumbered) middle ground would probably be MPEG - but for goodness sake, anything but WMV or AVI.
X(
OpenGL works just fine over a network. We use it at my work every day for running apps on our big cluster and exporting the display back to the users workstation. GLX is a "binding" (i'm not sure if that's the right word to use) of OpenGL to the X11 protocol. It need pretty good bandwidth though:)
I am also running a large compute cluster of v20z's and v40z's and have no such problems.
We had a couple of machines panic to begin with, but these were manufacturing problems rather than software and they have been replaced.
64bit Linux runs perfectly well on v20z's and v40z's. I suspect the "known stability problems in the kernel" that your Sun rep are talking about are most likely Sun FUD or the Sun rep's lack of knowledge about Linux x86_64 support.
And why not do Solaris/MySQL? That's Sun's point of view.
Besides, I doubt Sun care that much about small business. Sun doesn't make much money from shifting boxes, it does it through professional services (like IBM, just check their financials and you'll see).
Small businesses tend to not spend very much on professional services.
SUN is eyeing the entry into the larger mid to upper end business market thru Linux on Ultrasparc
From this I can only assume that you don't have a clue what you are talking about. Sun are already a player in the "mid to upper end business market" and have been for the past 20 years. Their current marketing strategy suggests that they aren't particularly interested in Linux either, having recently pushed out OpenSolaris (not that I don't wish they would start pushing Linux in some sort of useful way:( ).
That's NUMA - if your OS is NUMA aware it should try to place processes on the same processor as the memory that contains their data.
But yes, you're right, processes accessing memory on a different processor will suffer a latency (and to some extent bandwidth) hit. A well designed OS will help to mitigate it to some of the extent, but it's one of the reason that CPUs don't scale linearly.
Each Opteron has a dual bus memory controller on board (granted, only DDR400)... but as I increase the number of CPUs, the number of memory busses increases.
The 4 CPU boxes I'm working on have 8 independent DDR400 memory busses.
The only obvious way to get your memory bandwidth to scale is to have your memory controllers per CPU (or even better, on the CPU itself).
That's quite true. My background is in High Performance Computing so the majority of applications I see get a significant benefit from multiple CPUs/cores.
However, I believe we will see a change in development practices. Applications will become more multi-threaded as the number of CPUs per die increase. I think AMD have already announced that eventually they will only sell multi-core CPUs (in this market).
Of course, this means that programmers will have to learn how to write parallel code:)
I know this article is talking about Intel dual core chips, but for well-designed CPUs with integrated memory controllers (Power5, Ultrasparc IV, Opteron), the difference between a single dual-core CPU and two single-core CPUs is significant.
On chips with built in memory controllers, as you increase the number of cores on a chip the memory bandwidth per core decreases, however as you increase the number of chips in a system, the memory bandwidth per core remains the same and the number of cores increases.
That can amount to a big performance difference when running memory-intensive jobs.
Intel seem to be really losing the plot here at the moment. In multichip configurations, Intel's memory bandwidth already sucks compared to Opteron. Multicore per chip is only going to make it FAR worse.
Err, no. Incorrect - You need the PIN and *that* card, not just any card.
Keep your PIN (ie make sure no-one is watching you enter it) and card secure and nobody can clone your card (like they can with magstripe cards). They are far more secure than older cards.
Every TCO study I have seen into the cost benefits of Linux over Windows, and vice versa, seem to all be flawed. They are always paid for by someone with a vested interest in getting one "answer" or another. How can they be taken seriously... it's like going to Sun and IBM and saying "Whose hardware is better?"... I wonder what answer each company would give.
Surely version numbers are going up much faster now partly because a lot more work is being done on the kernels. There are far more people working on the kernel now than there were when 2.0 was about.
Speaking for myself, I've found 2.6 stable since I started using it when it hit 2.6pre. I've noticed no big changes in stability all the way from 2.0->2.6
Having said that, the idea of 2.6 progressing as a development tree certainly makes me feel uncomfortable.
As a University network administrator I thought I might answer this...
Universities already know that Bittorrent is not inherently bad. The problem is that there is a no way of distinguishing between a legitimate torrent (of say, a Linux distro) and a torrent of "unauthorised copyright material". If there were a way to easily differentiate between the two then I'm sure that many Universities would be quite willing to lift restrictions on bittorrents. Unfortunately, that's not the case.
I can guarantee, that if we altered our Bittorrent bandwidth restrictions to allow unfettered download/upload, our pipe would be saturated within a day.
The TGV is still the worlds fastest commercial train. Whilst it normally cruises around 300km/h, it has travelled at 515.3km/h on a REAL piece of track (i.e. not a test track). That top speed was set in 1990.
Slashdot Mirror
You're mistaken...
What the world needs now is love, sweet love
It's the only thing that there's just too little of
What the world needs now is love, sweet love,
No not just for some but for everyone.
You want a Linux designed to blow up and destroy ships?
It depends on the codec used inside the AVI file. I'm running Linux on PowerPC - so I can't use the Windows DLL hacks a lot of Linux on x86 folks use.
Yes - what a fantastic idea. Why not make a video about a technology relavant to Linux, *BSD and Solaris really difficult to access on Linux, *BSD and Solaris machines. Brilliant.
The (patent encumbered) middle ground would probably be MPEG - but for goodness sake, anything but WMV or AVI. X(OpenGL works just fine over a network. We use it at my work every day for running apps on our big cluster and exporting the display back to the users workstation. GLX is a "binding" (i'm not sure if that's the right word to use) of OpenGL to the X11 protocol. It need pretty good bandwidth though :)
apt-get update
apt-get upgrade
It doesn't get much easier than that.
No, that is a person seeking asylum. Refugees are people seeking refuge.
I am also running a large compute cluster of v20z's and v40z's and have no such problems.
We had a couple of machines panic to begin with, but these were manufacturing problems rather than software and they have been replaced.
64bit Linux runs perfectly well on v20z's and v40z's. I suspect the "known stability problems in the kernel" that your Sun rep are talking about are most likely Sun FUD or the Sun rep's lack of knowledge about Linux x86_64 support.
And why not do Solaris/MySQL? That's Sun's point of view.
Besides, I doubt Sun care that much about small business. Sun doesn't make much money from shifting boxes, it does it through professional services (like IBM, just check their financials and you'll see).
Small businesses tend to not spend very much on professional services.
That's NUMA - if your OS is NUMA aware it should try to place processes on the same processor as the memory that contains their data.
But yes, you're right, processes accessing memory on a different processor will suffer a latency (and to some extent bandwidth) hit. A well designed OS will help to mitigate it to some of the extent, but it's one of the reason that CPUs don't scale linearly.
Dual Bus memory still doesn't cut it.
Each Opteron has a dual bus memory controller on board (granted, only DDR400)... but as I increase the number of CPUs, the number of memory busses increases.
The 4 CPU boxes I'm working on have 8 independent DDR400 memory busses.
The only obvious way to get your memory bandwidth to scale is to have your memory controllers per CPU (or even better, on the CPU itself).
That's quite true. My background is in High Performance Computing so the majority of applications I see get a significant benefit from multiple CPUs/cores.
:)
However, I believe we will see a change in development practices. Applications will become more multi-threaded as the number of CPUs per die increase. I think AMD have already announced that eventually they will only sell multi-core CPUs (in this market).
Of course, this means that programmers will have to learn how to write parallel code
AMD's are released already in the Opteron range. Hopefully I have some heading my way right now (if I'm lucky!).
That is not necessarily true.
I know this article is talking about Intel dual core chips, but for well-designed CPUs with integrated memory controllers (Power5, Ultrasparc IV, Opteron), the difference between a single dual-core CPU and two single-core CPUs is significant.
On chips with built in memory controllers, as you increase the number of cores on a chip the memory bandwidth per core decreases, however as you increase the number of chips in a system, the memory bandwidth per core remains the same and the number of cores increases.
That can amount to a big performance difference when running memory-intensive jobs.
Intel seem to be really losing the plot here at the moment. In multichip configurations, Intel's memory bandwidth already sucks compared to Opteron. Multicore per chip is only going to make it FAR worse.
Keep your PIN (ie make sure no-one is watching you enter it) and card secure and nobody can clone your card (like they can with magstripe cards). They are far more secure than older cards.
Every TCO study I have seen into the cost benefits of Linux over Windows, and vice versa, seem to all be flawed. They are always paid for by someone with a vested interest in getting one "answer" or another. How can they be taken seriously... it's like going to Sun and IBM and saying "Whose hardware is better?"... I wonder what answer each company would give.
Are you sure that this book is still under the copyright of Ladybird?
On this page it claims...
"Copyright in a published edition expires 25 years from the end of the year in which the edition was first published."As 2004-1979 = 25 doesn't that make this book out of copyright now?
$2Bn cash now qualifies as "almost bankrupt"?
Damn, I wish I was almost bankrupt.
Surely version numbers are going up much faster now partly because a lot more work is being done on the kernels. There are far more people working on the kernel now than there were when 2.0 was about.
Speaking for myself, I've found 2.6 stable since I started using it when it hit 2.6pre. I've noticed no big changes in stability all the way from 2.0->2.6
Having said that, the idea of 2.6 progressing as a development tree certainly makes me feel uncomfortable.
There's a copy at http://www.kb.cert.org/vuls/id/323070. Right down at the bottom under "Use a different web browser".
As a University network administrator I thought I might answer this...
Universities already know that Bittorrent is not inherently bad. The problem is that there is a no way of distinguishing between a legitimate torrent (of say, a Linux distro) and a torrent of "unauthorised copyright material". If there were a way to easily differentiate between the two then I'm sure that many Universities would be quite willing to lift restrictions on bittorrents. Unfortunately, that's not the case.
I can guarantee, that if we altered our Bittorrent bandwidth restrictions to allow unfettered download/upload, our pipe would be saturated within a day.
I prefer being British! ;)
The TGV is still the worlds fastest commercial train. Whilst it normally cruises around 300km/h, it has travelled at 515.3km/h on a REAL piece of track (i.e. not a test track). That top speed was set in 1990.