Today the secret to good performance on games and 3d is to move the calculations to video cards which are specially designed for it. What is the reason not to do this?
As for AMD optimized code, well the Windows people can take care of themselves. Alan Cox has been given a couple of nice Athalons to try to optimize Linux on. They would be idiots not to try to do similar things with gcc. Which means that with open source you will be able to take advantage of it without having to depend on someone else for binaries.
Seriously, if you want a computer that ages gracefully it is far more important to buy extra RAM and hard drive than the fastest chip. A 30% difference in chip speed begins to feel very unimportant when you hit swap and you cannot install anything more on your computer...
If you want to know that you are writing standards-compliant html and you want to know exactly how it is supposed to look, then Amaya is a pretty good choice.
Unfortunately it stresses compliance over everything else. (Heck, by definition how it looks actually *IS* the standard...)
It is clear that right now AMD could bump performance significantly. As long as that remains true, every time Intel pushes the same old a little farther, AMD just stays ahead. The result? Intel is working well beyond what it can reliably produce, they have low chip yields, and AMD eats up more of their market.
Oh right, since AMD remains ahead on what is widely available, they keep on growing market share.
Intel needs to stop playing this losing game. Take the body blow, throttle back on the speed, play up reliability, point out the hidden costs of running the chips as hot as AMD runs them, then put the energy into a better architecture...
That the/. community doesn't have a single logical mind. It is a group of people with a signficant intersection of iterests. This is not the same as a borg.
A detail to pay attention to when dealing with groups of people. Suppose you have 3 choices, A, B, and C. Suppose 1/3 rank them A, B, and C. 1/3 rank them B, C, A. And the last third rank them C, A, B. Then the group can honestly be said by 2/3 to say that A is better than B, that B is better than C, and that C is better than A! Each of those statements have 2/3 support!
Think about this the next time you see someone wondering why the opinions of/. readers don't seem to make much logical sense...
*Sigh* Why is it called "common sense" when it appears to be so rare?
At least some satanists do not particularly believe in being honest. The very fact that true Wiccans are likely to be ticked off at the pretense is actually a motivation *to* pretend to be Wiccan. Think about it.
When I say, "Try to get away claiming to be" I am (among other things) implying that they are not.
I know very well what Wiccans are and are not. I am further aware of what Satanism is and is not. I know perfectly well the large divide between the two.
I am one. Always have been. The internet just gives me more people to talk with.
I remember when I found out in the early 90's that sci.math was read by tens of thousands of people, most of whom never posted. Wow. How could they not post?
By the time that dejanews began there was already a pretty good history on me had anyone bothered archiving it. Turns out that a few people did. Heck, by the time I found out that dejanews existed they already had a pretty good handle on me. That doesn't bother me. Most of what you will find publically is pretty innocuous. While there are a couple of items out there I would prefer to not have public, they are few and far between.
OTOH I cannot understand people who have online diaries. I don't mind chatting online, but I don't say anything that I would object to being announced in a large auditorium...
The first time I did a search on google for my name. I have wondered ever since who has linked to those ZDNet talkbacks. Every time I go back I always get surprises.
For instance Usenet articles from the early 90's have be immortalized by David Rusin. A very bad example of my early code has been rediscovered, a random letter in soc.culture.bahai has turned up, etc.
Sometimes I just have to wonder...
If you have been online for any time, try your own name. You may be surprised...
Is that you make parallelism a very small part of the code. That is good software design. You can do it and have it work. I really believe that. But you need to isolate out the parallelism and treat it with extreme prejudice.
As for why Java only has blocking IO, they very specifically wanted to keep people from using the standard C select() loop construct that inevitably has a small race in it. (Early versions of Apache were bitten by that one. It worked fine until you asked Apache to listen to multiple ports at once.)
BTW your repeated bull about my not understanding algorithms is getting mildly irritating.
As you say. If everything is designed properly, and implemented correctly, then threads give you better performance and frequently a simpler design as well. Which is pretty much what I said before you.
As you say. Sun uses blocking IO in Java to keep people from making design mistakes with threads. Which is pretty much what I said before you.
You don't say what happens if a mistake happens. I did. At length.
If you believe that mistakes never happen, then you are living in a fantasy world. Even if you were a perfect programmer, can you guarantee that every co-worker is? Heck, the Linux kernel folks regularly uncover race conditions, and it is not for nothing that people like Alan Cox and Linus Torvalds believe that very few programmers understand races well enough to properly avoid them.
In fact you may not have avoided them either. Do you know? Are you sure? Do you double-check everything that you are doing for correctness? Perhaps there are silent untested errors in your production code that just don't get hit very often?
Why does every competent programmer I know or have heard of from Knuth on down claim that they make mistakes and don't believe that there is anyone who doesn't?
Cisco is indeed burning up the back door with voice over IP. A lot of work has gone into that technology. It is far cheaper to send large amounts of data over IP than a phone. People don't want to have multiple wires. So on and so forth.
The needs of voice and network traffic are not exactly the same. (With data you care about perfect transmission of bursts of information, with voice you care about reliably holding a connection when you get it.) However it looks to me like the technology for data transmission is getting so much better than the technology for voice transmission that it is a question of time until the market says that we only need one of these...
This illusion lasts until the first race condiction that you need to debug.
For those who do not know, a race condition is a case where 2 things (threads, processes, whatever) can try to do the same thing at the same time when it is only safe for one to do it. Like incrementing a count of things in use, one figures out the current count, gets swapped out, the second figures out the current count, increments it and puts it back, the first comes back, increments what it had, and puts it back. So two attempts to increment the number happened, but it is marked that only one happened. This can lead to premature freeing of memory, and memory leaks.
While race conditions can happen with anything, the more you share, the more of a possibility you have for hitting them.
Now what are the characteristics of a race condition? Why they are quite simple:
1. They cannot be reliably reproduced.
2. The problem shows up somewhere different from the real mistake.
3. The probability rises sharply under load.
The last point is interesting. The possibilities for getting races vary with the number of pairs of things competing. So with 2 threads, you may have to wait many years to be hit by a race that nabs you every week when you have 50 threads going.
Anyways, look at those items for threads and think of what this looks like on the other side. Hmmm. Bugs that show up under load where nobody can figure out...does that sound familiar?
Yes, Microsoft products are heavily multi-threaded. Think carefully about that...
Cheers,
Ben
PS The design of Java very specifically forces people to program in a way that eliminates most race conditions. For instance that is why you have to spawn a thread and then use blocking IO...
Einstein did a *lot* besides the theory of relativity. (Hint, what did he win his Nobel for?) Furthermore his contributions continued for a very long time. For instance he was hardly a youngster when he and some friends came up with the EPR paradox!
Slashdot Mirror
Today the secret to good performance on games and 3d is to move the calculations to video cards which are specially designed for it. What is the reason not to do this?
As for AMD optimized code, well the Windows people can take care of themselves. Alan Cox has been given a couple of nice Athalons to try to optimize Linux on. They would be idiots not to try to do similar things with gcc. Which means that with open source you will be able to take advantage of it without having to depend on someone else for binaries.
:-)
Cheers,
Ben
If density is what matters, then the speed of sound through trolls should be much faster still..
Cheers,
Ben
The cost of the old ones will plummet. :-)
Seriously, if you want a computer that ages gracefully it is far more important to buy extra RAM and hard drive than the fastest chip. A 30% difference in chip speed begins to feel very unimportant when you hit swap and you cannot install anything more on your computer...
Cheers,
Ben
If you want to know that you are writing standards-compliant html and you want to know exactly how it is supposed to look, then Amaya is a pretty good choice.
Unfortunately it stresses compliance over everything else. (Heck, by definition how it looks actually *IS* the standard...)
Cheers,
Ben
It is clear that right now AMD could bump performance significantly. As long as that remains true, every time Intel pushes the same old a little farther, AMD just stays ahead. The result? Intel is working well beyond what it can reliably produce, they have low chip yields, and AMD eats up more of their market.
Oh right, since AMD remains ahead on what is widely available, they keep on growing market share.
Intel needs to stop playing this losing game. Take the body blow, throttle back on the speed, play up reliability, point out the hidden costs of running the chips as hot as AMD runs them, then put the energy into a better architecture...
Cheers,
Ben
That the /. community doesn't have a single logical mind. It is a group of people with a signficant intersection of iterests. This is not the same as a borg.
/. readers don't seem to make much logical sense...
A detail to pay attention to when dealing with groups of people. Suppose you have 3 choices, A, B, and C. Suppose 1/3 rank them A, B, and C. 1/3 rank them B, C, A. And the last third rank them C, A, B. Then the group can honestly be said by 2/3 to say that A is better than B, that B is better than C, and that C is better than A! Each of those statements have 2/3 support!
Think about this the next time you see someone wondering why the opinions of
*Sigh* Why is it called "common sense" when it appears to be so rare?
Cheers,
Ben
At least some satanists do not particularly believe in being honest. The very fact that true Wiccans are likely to be ticked off at the pretense is actually a motivation *to* pretend to be Wiccan. Think about it.
Ben
When I say, "Try to get away claiming to be" I am (among other things) implying that they are not.
I know very well what Wiccans are and are not. I am further aware of what Satanism is and is not. I know perfectly well the large divide between the two.
*sigh*
Ben
I am one. Always have been. The internet just gives me more people to talk with.
I remember when I found out in the early 90's that sci.math was read by tens of thousands of people, most of whom never posted. Wow. How could they not post?
By the time that dejanews began there was already a pretty good history on me had anyone bothered archiving it. Turns out that a few people did. Heck, by the time I found out that dejanews existed they already had a pretty good handle on me. That doesn't bother me. Most of what you will find publically is pretty innocuous. While there are a couple of items out there I would prefer to not have public, they are few and far between.
OTOH I cannot understand people who have online diaries. I don't mind chatting online, but I don't say anything that I would object to being announced in a large auditorium...
Cheers,
Ben
Let me make it easy.
The US Constitution. Complete with annotations.
Look for the word "Privacy", I dare you.
If you actually want to learn something about the legal state of privacy in the US you might want to pick up a good book on the topic...
Regards,
Ben
I have encountered Satanists who try to get away with claiming to be Wiccan instead of being honest about their beliefs.
Not many, but enough to make me think that the original post was neither mistaken nor ignorant.
Cheers,
Ben
Yeah, you are right. There seems to be a little interference with your signal.
/. for a while...
But anyone who cares can find stuff like a Debian email and that you have been on
Cheers,
Ben
As Mark Twain said, and Stephen J. Gould had opportunity to repeat, "The reports of my death are greatly exaggerated." :-)
Thanks for the laugh..
Cheers,
Ben
There are a couple of other Ben Tillys out there. I always wonder if there is any relation...
Cheers,
Ben
The first time I did a search on google for my name. I have wondered ever since who has linked to those ZDNet talkbacks. Every time I go back I always get surprises.
For instance Usenet articles from the early 90's have be immortalized by David Rusin. A very bad example of my early code has been rediscovered, a random letter in soc.culture.bahai has turned up, etc.
Sometimes I just have to wonder...
If you have been online for any time, try your own name. You may be surprised...
Cheers,
Ben
Designed by Microsoft!
.NET. Network enabling your OS from start to finish, without bothering to design any security into it.
Ba-dump!
For instance their latest initiative,
AAARRRGGGHHH!!!
Ben
Is that you make parallelism a very small part of the code. That is good software design. You can do it and have it work. I really believe that. But you need to isolate out the parallelism and treat it with extreme prejudice.
As for why Java only has blocking IO, they very specifically wanted to keep people from using the standard C select() loop construct that inevitably has a small race in it. (Early versions of Apache were bitten by that one. It worked fine until you asked Apache to listen to multiple ports at once.)
BTW your repeated bull about my not understanding algorithms is getting mildly irritating.
Regards,
Ben
As you say. If everything is designed properly, and implemented correctly, then threads give you better performance and frequently a simpler design as well. Which is pretty much what I said before you.
As you say. Sun uses blocking IO in Java to keep people from making design mistakes with threads. Which is pretty much what I said before you.
You don't say what happens if a mistake happens. I did. At length.
If you believe that mistakes never happen, then you are living in a fantasy world. Even if you were a perfect programmer, can you guarantee that every co-worker is? Heck, the Linux kernel folks regularly uncover race conditions, and it is not for nothing that people like Alan Cox and Linus Torvalds believe that very few programmers understand races well enough to properly avoid them.
In fact you may not have avoided them either. Do you know? Are you sure? Do you double-check everything that you are doing for correctness? Perhaps there are silent untested errors in your production code that just don't get hit very often?
Why does every competent programmer I know or have heard of from Knuth on down claim that they make mistakes and don't believe that there is anyone who doesn't?
Regards,
Ben
Cisco is indeed burning up the back door with voice over IP. A lot of work has gone into that technology. It is far cheaper to send large amounts of data over IP than a phone. People don't want to have multiple wires. So on and so forth.
The needs of voice and network traffic are not exactly the same. (With data you care about perfect transmission of bursts of information, with voice you care about reliably holding a connection when you get it.) However it looks to me like the technology for data transmission is getting so much better than the technology for voice transmission that it is a question of time until the market says that we only need one of these...
Cheers,
Ben
On paper. The designs look very sweet and simple.
This illusion lasts until the first race condiction that you need to debug.
For those who do not know, a race condition is a case where 2 things (threads, processes, whatever) can try to do the same thing at the same time when it is only safe for one to do it. Like incrementing a count of things in use, one figures out the current count, gets swapped out, the second figures out the current count, increments it and puts it back, the first comes back, increments what it had, and puts it back. So two attempts to increment the number happened, but it is marked that only one happened. This can lead to premature freeing of memory, and memory leaks.
While race conditions can happen with anything, the more you share, the more of a possibility you have for hitting them.
Now what are the characteristics of a race condition? Why they are quite simple:
1. They cannot be reliably reproduced.
2. The problem shows up somewhere different from the real mistake.
3. The probability rises sharply under load.
The last point is interesting. The possibilities for getting races vary with the number of pairs of things competing. So with 2 threads, you may have to wait many years to be hit by a race that nabs you every week when you have 50 threads going.
Anyways, look at those items for threads and think of what this looks like on the other side. Hmmm. Bugs that show up under load where nobody can figure out...does that sound familiar?
Yes, Microsoft products are heavily multi-threaded. Think carefully about that...
Cheers,
Ben
PS The design of Java very specifically forces people to program in a way that eliminates most race conditions. For instance that is why you have to spawn a thread and then use blocking IO...
Very, very busy.
I just do not have a site available that does such a good job dissecting it...
Cheers,
Ben
Trust me on this.
Wait for 5.6.1 (which will be out fairly shortly).
If you must use the new features, then figure out how to get the current development version of Perl and use that.
Otherwise you will get burned.
(Hint to the wise. Saying that *.0 releases are stable is a marketing label. This is definitely the case with 5.6.0.)
Ben
The interface.
Need I say more?
Cheers,
Ben
The earliest that I know of happened before you probably ever heard of the Internet. Go look up the Morris Worm.
And I am sure that was not the first, I heard of it because it was the last time that an individual accidentally took down the Internet.
Regards,
Ben
Don't abuse your ignorance by BSing about it.
Einstein did a *lot* besides the theory of relativity. (Hint, what did he win his Nobel for?) Furthermore his contributions continued for a very long time. For instance he was hardly a youngster when he and some friends came up with the EPR paradox!
Regards,
Ben