This probably depends on the compiler (and compiler flags). This is an optimization by the compiler to match memory allocation to cache line length for better performance. If you look, you'll find compiler options to allocate on whatever boundary you want.
Otherwise, it would just be a design of the memory allocation libraries (always allocate on 16-byte boundary) for similar reasons (and to lessen memory fragmentation).
What is being referred to here is the possibility of having different cores, not just two identical cores on the same silicon.... Another possibility is where the entire system is devoted to a single task (think HPC: fluid flow, weather simulations, etc) where you could have threads doing the intensive floating point calculations on one core, and the heavy integer arithmetic on the other, or maybe split up the cores based on memory accesses patterns, or cache use, or built-in ASICs!
The problem with this is that by designing resources to be used in "certain situations" you put resources in the mix that may not be usable most of the time. Using your example of having one core with 4 Integer units and 2 FPUs and another core with 2 Integer units and 4 FPUs, why not make a single core with 6 Integer units and 6 FPUs and let the core itself how to allocate those resources based on runtime requirements (as opposed to potentially having one of those cores idle until a "special" program is run)? In fact, why not just have two cores with the average of each (3 Integer and 3 FPUs) so that complex scheduling doesn't have to be written (and it would be complex... you'd have to know before the thread is scheduled which core would be more suited to it, either through analysis of the executable code or through hints given at compile time).
I guess that one of the things is that due to the widely varying usage patters on consumer PCs vs. complexity, you'd be better off making a compromise and designing based on average usages. If you are designing specifically for systems that will be heavy in CFD or the like, design a CPU that is an FPU beast. These two markets have pretty different requirements and by trying to blend them, you'll end up with something not really suited for either.
The other thing is that when you design specific silicon (different types of cores), you have to make sure that there's sufficient usage of each to justify the effort required. Having two cores where one is idle 99% of the time is a wasted effort.
Well... that's good information about you, I guess, but just because you think this way (or at least post that you think this way) doesn't mean that the rest of the world thinks the way you do.
The company has to at least release a Version01 of drivers to give basic functionality to start. If they release hardware with no drivers, then they will be gambling on how long it takes for drivers to appear. All the time they wait, they lose money.
It is expensive to make small runs of hardware, even if you have your on board manufacturing. This doesn't even take into account a chip foundry. Without economies of scale, the cards will be expensive, which pretty much is a death certificate for the things given the nature of the community for which it is targeted.
Folks tend to forget that one of the largest draws to the F/OSS world is the cost (free as in beer). Software and such takes little/no cost to copy and distribute. Hardware is an entirely different story because it is a physical device that takes materials and *must* be paid for because it will take significant time and materiel to produce.
If they can pull it off, more power to them, but I won't be investing any money in it because it is an extremely high risk endeavor.
Piracy is going to strengthen subscription model
on
GTA: San Andreas Leaked
·
· Score: 2, Insightful
Well, we've already seen it start happening, but I think that piracy will be one of the reasons most companies will move to a subscription model. Basically, the software that you get (whether legit or not) will be useless without a paid subscription. This will pretty much move all games towards a MMO model, even if they are small grouped like Diablo.
I could even see where a single player game may be subscription based where you have to be online and connected to a server and paying a subscription in order to play at all.
I guess as much as folks hate a software rental model, piracy may just be the thing to give the companies enough of a reason to switch to this model.
Folks will not like the model but they will have brought it on themselves.
There are so many things wrong with your post I won't even start. Some other folks have commented on them already. I'd guess that you probably are one of the people who should read the book. Many of your statements are just wrong out of ignorance.
Do you think because MyISAM isn't transactional that the whole database isn't?
Unfortunately, this is the default state of the system. Since most folks who jump on the MySQL bandwagon don't understand why you'd actually need data integrity or when to use it, they never both to learn more about databases and continue down their MyISAM path for everything. Just like most things, if someone doesn't understand it, they don't know what to use it for so they just keep going in the same way for all the problems they encounter. I guess ignorance truly is bliss.
I guess there are situations were people don't really know what they need, but then I would guess that most people do.
In my experience, I've found the opposite to be true. I find it much more common that people who "know" what they need and don't need in actuality don't understand their own problem or the technologies they can/should use to solve the problem. I've seen folks describing a problem that is clearly a transactional problem but then refuse to use transactions because of the "overhead" and they've had systems that "worked for years without using transactions so why do they need them now". I also have seen a number of folks who can't think of a single example of where one would use a transaction conclude that 'therefore, no one needs transactions'.
Most of the time, it is just ignorance combined with defensive pride. Folks who are "self taught" tend to arrive at a conclusion early in their programming career and doggedly stick to it because that's all they know, regardless of what is "correct".
If you need extreme performance in small commercial site, why you need an great and cpu-expensive Referential Integrity? In my opinion, in this case, MySQL is better. Comments?
Yeah, why would you want to link multiple purchases to a single account. Why would you want to want to make sure that the account you are associating with a purchase actually exists in your database and flag non-existance of a primary key as an error. Why would you want an error to be raised if you try to delete an account without dealing with the purchases associated with that account, thus leaving records of purchases that have no account associtated with them. Certainly you wouldn't want these things for accountability to your customers, tax purposes, or making sure the database isn't just garbage.
Unfortunately, a lot of folks who think they know a lot about databases really don't. They don't know when MySQL is sufficient or not because they don't understand what databases need to do and/or provide. Most wouldn't know the difference between MySQL, a big honkin text file, and a real RDBMS like Oracle. To them, they are jsut all places to fling a bunch of data to query back later. Most wouldn't know what transactions really are or when you should use them, although if you ask them, they'll assure you that they don't need that "overhead" in their problem.
an Athlon 64, it is definitely faster than an XP in the same MHz rating
Most of the time (like 95% of the time or more), this is true, but there are some things that rely on raw clock speed since the A64s in 32-bit mode are similar in architecture to the Athlon XPs.
I have three Athlon 64 3000+ (Newcastle) machines, an Athlon XP 2600+/333, and an Athlon XP 2400+/266. None of them are overclocked. Every once in a while, I can find something that my 2.083GHz Athlon XP 2600+/333 does a little faster than my 2.000GHz Athlon 64s can do it. These applications typically are doing heavy x87 code and the 83MHz clock speed difference is enough to show a difference, even if it is very small (usually, the difference is a linear increase based on clock speed = 2083/2000 as fast). Would I exchange my Athlon 64s for Athlon XP 2600+/333s? No way:) because they do most things better.
Myrinet is also interesting in that you can program the NIC yourself, at least you could the last time I fooled with it. It has a custom processor (LANai) with memory and a bunch of, basically, DMA channels on it. It doesn't do much of anything out of the box until you put an MCP (Myrinet Control Program) on it. My group made a few custom MCPs for Myrinet back in the day. Interesting programming since it is an embedded system. However, the complexity is/was pretty high and your MCP had to be debugged as well (try putting an IP stack on it in addition to "fast lanes" for high bandwidth, low latency MPI communication - that will surely keep you occupied:) I guess there are a few MCPs and MPI ports for it you can use out of the box, though.
A report on Newsforge (which is part of OSTG along with Slashdot) shows the flip side of the coin.
Actually, this report shows that the IT jobs in INDIA are increasing, not the IT jobs in the USA. Because the jobs are increasing in India does not mean that the jobs are (or are going to) increasing in the USA. In fact, it's quite probable the opposite will happen.
Ideal maybe in that they are the only broadly used format. I don't like PDFs and I don't like 32+MB acrobat applications hanging around doing nothing when you load a PDF into the browser. PDFs are big and slow. If there were another semi-standard cross-platform document format, I'd use it over PDF. I go to fair lengths to avoid PDFs.
I don't really use PriceWatch anyway. I'm not after the cheapest part sold by a fly-by-night company.
Re:I code C# for a living
on
Java 1.5 vs C#
·
· Score: 1
It was an honest question.
Re:All in it together
on
Java 1.5 vs C#
·
· Score: 2, Informative
Interfaces are your friends. Most of what people use multiple inheritance for can be achieved by using multiple interfaces.
Re:I code C# for a living
on
Java 1.5 vs C#
·
· Score: 1
Why do you think delegates are a mis-feature?
Re:I code C# for a living
on
Java 1.5 vs C#
·
· Score: 2, Insightful
How exactly does it pale in comparison? What features does it have that make it better? What features does VS.NET lack?
Re:With C#, stuck in windoze
on
Java 1.5 vs C#
·
· Score: 2, Informative
LoL... Write Once, Run Anywhere.... good one there... Java is Write Once, Debug Everywhere. Once you commit to one JVM, you are pretty much locked in because of the incompatibilities in other JVMs. A number of groups are using C# developed apps on both Windows and Linux without any issues at all (as long as the code uses GTK#).
There are C# development environments on other platforms... Linux for example (Mono) and one can even use Rotor on other platforms. And... at least C# is an open standard.
"If it's on the radio for free why do they care?".
Ahhh... but there's the rub... Radio play isn't free at all. Consumers can listen to it for "free" except they have to buy or build a device to listen to it. However, go tell a radio station operator that their business is "free". They'll show you how many people they have to pay and how many people are devoted to getting money from advertisers in order to pay the salaries, broadcast fees, and electrical bills that come with running a radio station.
We had a repeater of a good radio station here a while back that went belly up because they couldn't get enough money to keep broadcasting. People wanted to listen to the repeater for free. No money = no repeater and now we have to listen to other crap stations. If radio were free as you suggest, I'd be listening to that repeater right now. If, as many posters point out in these threads, there were people willing to pay for music, just not so much, why weren't there enough donations to the folks running that repeater.
The fact is, saying that folks would pay for music is just something to hide behind. They know that if somehow, some law was passed saying that all music was free and that the only money that could be earned was what folks donated, that the money that the artists and labels would make would be near 0. The "people will pay if the price is right and they don't want it to be zero cost" folks either know nothing about human nature or they are simply lying so they can get their agenda of free music on.
Slashdot Mirror
This probably depends on the compiler (and compiler flags). This is an optimization by the compiler to match memory allocation to cache line length for better performance. If you look, you'll find compiler options to allocate on whatever boundary you want.
Otherwise, it would just be a design of the memory allocation libraries (always allocate on 16-byte boundary) for similar reasons (and to lessen memory fragmentation).
Still good information to know about Mac OSX.
What is being referred to here is the possibility of having different cores, not just two identical cores on the same silicon. ...
Another possibility is where the entire system is devoted to a single task (think HPC: fluid flow, weather simulations, etc) where you could have threads doing the intensive floating point calculations on one core, and the heavy integer arithmetic on the other, or maybe split up the cores based on memory accesses patterns, or cache use, or built-in ASICs!
The problem with this is that by designing resources to be used in "certain situations" you put resources in the mix that may not be usable most of the time. Using your example of having one core with 4 Integer units and 2 FPUs and another core with 2 Integer units and 4 FPUs, why not make a single core with 6 Integer units and 6 FPUs and let the core itself how to allocate those resources based on runtime requirements (as opposed to potentially having one of those cores idle until a "special" program is run)? In fact, why not just have two cores with the average of each (3 Integer and 3 FPUs) so that complex scheduling doesn't have to be written (and it would be complex... you'd have to know before the thread is scheduled which core would be more suited to it, either through analysis of the executable code or through hints given at compile time).
I guess that one of the things is that due to the widely varying usage patters on consumer PCs vs. complexity, you'd be better off making a compromise and designing based on average usages. If you are designing specifically for systems that will be heavy in CFD or the like, design a CPU that is an FPU beast. These two markets have pretty different requirements and by trying to blend them, you'll end up with something not really suited for either.
The other thing is that when you design specific silicon (different types of cores), you have to make sure that there's sufficient usage of each to justify the effort required. Having two cores where one is idle 99% of the time is a wasted effort.
Well... that's good information about you, I guess, but just because you think this way (or at least post that you think this way) doesn't mean that the rest of the world thinks the way you do.
The company has to at least release a Version01 of drivers to give basic functionality to start. If they release hardware with no drivers, then they will be gambling on how long it takes for drivers to appear. All the time they wait, they lose money.
It is expensive to make small runs of hardware, even if you have your on board manufacturing. This doesn't even take into account a chip foundry. Without economies of scale, the cards will be expensive, which pretty much is a death certificate for the things given the nature of the community for which it is targeted.
Folks tend to forget that one of the largest draws to the F/OSS world is the cost (free as in beer). Software and such takes little/no cost to copy and distribute. Hardware is an entirely different story because it is a physical device that takes materials and *must* be paid for because it will take significant time and materiel to produce.
If they can pull it off, more power to them, but I won't be investing any money in it because it is an extremely high risk endeavor.
Well, we've already seen it start happening, but I think that piracy will be one of the reasons most companies will move to a subscription model. Basically, the software that you get (whether legit or not) will be useless without a paid subscription. This will pretty much move all games towards a MMO model, even if they are small grouped like Diablo.
I could even see where a single player game may be subscription based where you have to be online and connected to a server and paying a subscription in order to play at all.
I guess as much as folks hate a software rental model, piracy may just be the thing to give the companies enough of a reason to switch to this model.
Folks will not like the model but they will have brought it on themselves.
There are so many things wrong with your post I won't even start. Some other folks have commented on them already. I'd guess that you probably are one of the people who should read the book. Many of your statements are just wrong out of ignorance.
Do you think because MyISAM isn't transactional that the whole database isn't?
Unfortunately, this is the default state of the system. Since most folks who jump on the MySQL bandwagon don't understand why you'd actually need data integrity or when to use it, they never both to learn more about databases and continue down their MyISAM path for everything. Just like most things, if someone doesn't understand it, they don't know what to use it for so they just keep going in the same way for all the problems they encounter. I guess ignorance truly is bliss.
I guess there are situations were people don't really know what they need, but then I would guess that most people do.
In my experience, I've found the opposite to be true. I find it much more common that people who "know" what they need and don't need in actuality don't understand their own problem or the technologies they can/should use to solve the problem. I've seen folks describing a problem that is clearly a transactional problem but then refuse to use transactions because of the "overhead" and they've had systems that "worked for years without using transactions so why do they need them now". I also have seen a number of folks who can't think of a single example of where one would use a transaction conclude that 'therefore, no one needs transactions'.
Most of the time, it is just ignorance combined with defensive pride. Folks who are "self taught" tend to arrive at a conclusion early in their programming career and doggedly stick to it because that's all they know, regardless of what is "correct".
If you need extreme performance in small commercial site, why you need an great and cpu-expensive Referential Integrity? In my opinion, in this case, MySQL is better. Comments?
Yeah, why would you want to link multiple purchases to a single account. Why would you want to want to make sure that the account you are associating with a purchase actually exists in your database and flag non-existance of a primary key as an error. Why would you want an error to be raised if you try to delete an account without dealing with the purchases associated with that account, thus leaving records of purchases that have no account associtated with them. Certainly you wouldn't want these things for accountability to your customers, tax purposes, or making sure the database isn't just garbage.
Unfortunately, a lot of folks who think they know a lot about databases really don't. They don't know when MySQL is sufficient or not because they don't understand what databases need to do and/or provide. Most wouldn't know the difference between MySQL, a big honkin text file, and a real RDBMS like Oracle. To them, they are jsut all places to fling a bunch of data to query back later. Most wouldn't know what transactions really are or when you should use them, although if you ask them, they'll assure you that they don't need that "overhead" in their problem.
an Athlon 64, it is definitely faster than an XP in the same MHz rating
:) because they do most things better.
Most of the time (like 95% of the time or more), this is true, but there are some things that rely on raw clock speed since the A64s in 32-bit mode are similar in architecture to the Athlon XPs.
I have three Athlon 64 3000+ (Newcastle) machines, an Athlon XP 2600+/333, and an Athlon XP 2400+/266. None of them are overclocked. Every once in a while, I can find something that my 2.083GHz Athlon XP 2600+/333 does a little faster than my 2.000GHz Athlon 64s can do it. These applications typically are doing heavy x87 code and the 83MHz clock speed difference is enough to show a difference, even if it is very small (usually, the difference is a linear increase based on clock speed = 2083/2000 as fast). Would I exchange my Athlon 64s for Athlon XP 2600+/333s? No way
Myrinet is also interesting in that you can program the NIC yourself, at least you could the last time I fooled with it. It has a custom processor (LANai) with memory and a bunch of, basically, DMA channels on it. It doesn't do much of anything out of the box until you put an MCP (Myrinet Control Program) on it. My group made a few custom MCPs for Myrinet back in the day. Interesting programming since it is an embedded system. However, the complexity is/was pretty high and your MCP had to be debugged as well (try putting an IP stack on it in addition to "fast lanes" for high bandwidth, low latency MPI communication - that will surely keep you occupied :) I guess there are a few MCPs and MPI ports for it you can use out of the box, though.
The 'net was a lot better before 1993 in many regards. It's a shame they had to open the door and let so many idiots in.
A report on Newsforge (which is part of OSTG along with Slashdot) shows the flip side of the coin.
Actually, this report shows that the IT jobs in INDIA are increasing, not the IT jobs in the USA. Because the jobs are increasing in India does not mean that the jobs are (or are going to) increasing in the USA. In fact, it's quite probable the opposite will happen.
Heh... say you don't like something that the F/OSS people like, get modded as a Troll.
You have the freedom to like/use whatever you want, as long as it is blessed by the F/OSS community. Some freedom...
PDFs are ideal
Ideal maybe in that they are the only broadly used format. I don't like PDFs and I don't like 32+MB acrobat applications hanging around doing nothing when you load a PDF into the browser. PDFs are big and slow. If there were another semi-standard cross-platform document format, I'd use it over PDF. I go to fair lengths to avoid PDFs.
I don't really use PriceWatch anyway. I'm not after the cheapest part sold by a fly-by-night company.
It was an honest question.
Interfaces are your friends. Most of what people use multiple inheritance for can be achieved by using multiple interfaces.
Why do you think delegates are a mis-feature?
How exactly does it pale in comparison? What features does it have that make it better? What features does VS.NET lack?
LoL... Write Once, Run Anywhere.... good one there... Java is Write Once, Debug Everywhere. Once you commit to one JVM, you are pretty much locked in because of the incompatibilities in other JVMs. A number of groups are using C# developed apps on both Windows and Linux without any issues at all (as long as the code uses GTK#).
There are C# development environments on other platforms... Linux for example (Mono) and one can even use Rotor on other platforms. And... at least C# is an open standard.
Actually... whoever buys the remnants of Sun will "pick it up" and do whatever they want with it.
"If it's on the radio for free why do they care?".
Ahhh... but there's the rub... Radio play isn't free at all. Consumers can listen to it for "free" except they have to buy or build a device to listen to it. However, go tell a radio station operator that their business is "free". They'll show you how many people they have to pay and how many people are devoted to getting money from advertisers in order to pay the salaries, broadcast fees, and electrical bills that come with running a radio station.
We had a repeater of a good radio station here a while back that went belly up because they couldn't get enough money to keep broadcasting. People wanted to listen to the repeater for free. No money = no repeater and now we have to listen to other crap stations. If radio were free as you suggest, I'd be listening to that repeater right now. If, as many posters point out in these threads, there were people willing to pay for music, just not so much, why weren't there enough donations to the folks running that repeater.
The fact is, saying that folks would pay for music is just something to hide behind. They know that if somehow, some law was passed saying that all music was free and that the only money that could be earned was what folks donated, that the money that the artists and labels would make would be near 0. The "people will pay if the price is right and they don't want it to be zero cost" folks either know nothing about human nature or they are simply lying so they can get their agenda of free music on.
Hmmm.... Sudden rise in CO2 in the atmosphere... start of the USA Presidential election campaigns... I think I see something here...