A friend of mine, an overclocking expert (inventor of the "Goldfinger devices" if anyone remembers those) said that the new shrunk cores overclock to around 3GHz if you can get your FSB high enough (though this won't be an issue with the FX chips, which aren't clock-locked). To those paying attention, 2.2GHz in an Athlon64 can generally outperform a 3.4GHz PentiumIV, so this is a big deal.
I think it is unwise to try to use an unsupported (newsgroups do not count) product like Debian, especially in a business/enterprise environment. (Yes, I know you can buy 3rd party support for almost anything. It isn't IBM, and this is a conservative organization).
I know this isn't going to be popular (so read fast, this will be -1 in seconds), but risk management is very important in the corporate IT world, because mistakes or screw ups get people fired and can make the company lose millions. Not that Debian isn't up to the task--that isn't the point. If something does go wrong, it's your ass. If some of IBM's software dies, IBM gets to deal with management, and they aren't going to switch from IBM with such a large investment (especially when IBM makes many excellent products).
If your current stuff works, there is no great reason to change, as software licenses are a drop in the bucket compared to potential pitfalls.
If you have looked over the fact and believe Debian would be a good idea, I'd take it slowly. Mail server (or something) need more capacity? See what Qmail on a ReiserFS system can do, even on the same hardware. Need another file server? Set up a quickie and demo Debian/Samba3's performance in similar or weaker hardware. Nothing open-source can replace DB2, so show that it runs on Debian just fine, etc. Good luck!
Since my gaming system was "upgraded" to Windows XP from 2000, Starcraft decided that certain doors inside of buildings would cause a crash-to-desktop when opened. Seeing as how you must open some to get through the game, this was a bad thing. I tried various tricks (disable sound/change drivers/reinstall/different media/3 different systems) and even contact tech support. The *only* way I was able to get past these points was to run them in Cedega on my laptop.
Thus, in some situations, Cedega is more compatible with Windows games than Windows itself.
Agreed about Creative. Their drivers have consistantly been a steaming pile of dingo's kidneys (having been the one cause of system crashes on my Windows 2000 system for quite a while) and the software that comes with them is worse--not only is it crap, but it is horribly bloated and excessively flashy. The Audigy 2 is advertised as a "high-end" soundcard but the software lacks even the most basic functionality, such as directing bass to anything other than the subwoofer. (Some of us have real speakers, "Creative"!). While I do not know if Creative has ever submitted a driver for WHQL certification, even Microsoft's rather low standards (e.g. the WHQL ATI drivers that would cause data corruption if "Advanced performance" was turned on for the hard drives...eh?) wouldn't allow for Creative's garbage to pass the first level of QA. Either their programmers are insanely rushed or they are completely incompetent. (as a side note, I agree that VIA is crap, but would say that it is foolish to grant the same title to AMD, which has had no product recalls in many years and has had faster, better designed products for years [including processors and FLASH ram, but probably not chipsets]).
I think it is fairly apparent that the RIAA (and perhaps MPAA) aren't really concerned over P2P networks' effect on sales. It is a control issue. If they do not have control power over the distribution channel, they have less power.
P2P networks are decentralized and completely out of their sphere of influence. I am sure that the people running the RIAA are not morons--CD sales are up and there is greater and greater evidence that P2P networks slightly improve sales. However, this evidence is used to argue with a point that I think the RIAA is using to push anti-P2P legislation through, but not a point that the RIAA actually believes.
If it were purely a sales problem, the RIAA would be going after commercial CD pirates -- These are the real bad guys. They are commercially profiting from the work of others (not just sharing it), and have reportedly built an economy in and of themselves of 4.5 BILLION dollars.
Of course, this doesn't mean 4.5 billion in losses--that is a BSA/piracy argument which is quite absurd--but I am sure that they do lose some money.
I know that's the normal line, but it seems that a company that is so flagrantly linux-based that they would open a Linux-only computer shop (and a carwash with a Tux logo no less!) would make it a point to eat their own dog food and use a Linux server. It isn't as if it is difficult to find Linux hosting online. Many hosting providers specifically ask what operating system you would like to run, and there are far fewer Windows-only hosts than mixed OS hosts.
And ed doesn't waste space on my Timex Sinclair. Just look:
-rwxr-xr-x 1 root 24 Oct 29 1929/bin/ed -rwxr-xr-t 4 root 1310720 Jan 1 1970/usr/ucb/vi -rwxr-xr-x 1 root 5.89824e37 Oct 22 1990/usr/bin/emacs
Of course, on the system *I* administrate, vi is symlinked to ed. Emacs has been replaced by a shell script which 1) Generates a syslog message at level LOG_EMERG; 2) reduces the user's disk quota by 100K; and 3) RUNS ED!!!!!!
"Ed is the standard text editor."
Let's look at a typical novice's session with the mighty ed:
--- Note the consistent user interface and error reportage. Ed is generous enough to flag errors, yet prudent enough not to overwhelm the novice with verbosity.
How about a sequel Deus Ex (which is loosely based on System Shock ideas)? And no, "Invisible War: Console Style Edition for 12-year olds" doesn't count.:/
I read the Linux Journal (which, IMO, is entirely too openly biased) and Dr. Dobbs Journal. At least, I pretend to read the latter. It isn't really as interesting as it once was. Everything else I read online.
Interesting coincidence (to me at least)
on
Wired on McBride
·
· Score: 3, Interesting
I worked for IQorder.com at the time that it was apparently sued (Sorry, "asked to pay") by McBride's patent friends. This guy gets around. Suing random startups over BS patents and now suing Linux users over BS claims.
McBride of Frankenstein and Senator "Let's let the RIAA to hack file sharers legally" Hollings are some of the highest profile Mormons in their little church. I hope they get excommunicated (but then, they would lose all that tithing).
The scary thing is that McBride has seven kids. I hope that being a prick isn't genetic. I really hope that his kids see what kind of human being their dad is and don't follow in his footsteps.
According to Sandpile.org, the 3.4GHz Pentium IV Prescott can use up to 127W, and has a typical power usage of 103W (when browsing the web or reading email). In my opinion, it is rediculous for a single processor to single-handedly run up your power bill. That's like having two light bulbs on 24/7 (assuming you keep your computer on), not to mention the power needed to cool your PC, let alone your house's air conditioner.
I would take a VIA chip for low-performance stuff, and an Athlon64 for performance computing. support 64-bit software including 64-bit Linux distributions, are faster than Intel's best even running 32-bit software, and they have a maximum power usage of 89W. Because of Cool'n'Quiet mode, they spend most of the time running at 800MHz consuming about 30-35W and generally not requiring a loud and abnoxious cooling fan.
It is actually impressive what the chips can do at 800MHz. You can play a full screen DVD at 1400x1050, and the CPU usage tops out at about 5% (at 800MHz). If, of course, you run something that requires more power, like a video game or a compiler, the processor instantly switches to full speed. Handy, that.
Soon cellphone owners will be able to play games with realistic three-dimensional graphics rivaling those on PC's and game consoles.
It's great that cellphones are getting more advanced, but I'm afraid a mini-3D chip running off of a 2 ounce battery displaying on a 2" screen is not going to be rivaling PCs anytime soon.
Though who knows, maybe the new cell phones will have a DVI connector and a port of Doom 3.
StorageReview is planning to publish a review using a variety of RAID controllers testing their performance with TCQ (Tagged Command Queuing) for both user and server access patterns. It should be almost finished. The results so far are rather surprising to say the least..
In short, I would probably recommend RAID5 if you have 3+ drives. RAID5 gives you the most available space while still being redundant. It allows for exactly one hard drive failure. RAID5's write speed is usually terrible, especially with a small number of drives, but write speed isn't a big deal on my home file server. (Only you know about your needs).
RAID1+0 (NOT RAID 0+1, which is inferior) is great for performance. With 4 drives, you have potentially twice the STR of one drive (writing) and 4 times the STR of one drive reading. Of course, since STR is not important for most IO, this doesn't really effect your end performance much unless you are dealing with linearly reading/writing very large files. Writing performance will almost certainly be higher than with RAID5. You do lose quite a lot of space (especially when you use a large number of drives). If you used a 4-drive 1+0 array, you would have the space in two of those individual drives.
RAID1 is nice, and is very reliable, but is impractical with more than two drives unless you are incredibly paranoid. RAID1 simply makes all drives copies of the others, this, you always have as much free space as one drive would have, even if you have ten. If course, you could also handle 9 drive failures and not lose data. RAID1 is fine for 2-drive arrays though.
DO NOT FORGET that RAID is no substitute for regular backups. RAID will not help if your data loss is caused by FS corruption, a cracker, accidentally typing "rm -rf/", etc.
For lowest cost, I would use software RAID, such as Linux's LVM, FreeBSD's Vinum, or whatever Windows has. (RAID5 requires Windows server). (I would not use Windows as the file server myself). For slightly higher cost, try a Promise controller. I would avoid Highpoint and Silicon Image controllers. Highpoint, especially, is crap. (but it is very cheap, at least).
If you possibly can, I would recommend a nice 3Ware Escalade controller. Escalades are true hardware RAID cards, unlike Highpoint/SI and most of Promise's cards, and are OS independent and very stable (with certain exceptions for some unlikely configurations).
If you have any questions, you might try the StorageReview forums. There are a number of extremely knowledgeable people there, including engineers and executives-level researchers at hard drive companies. They can give far better advice than I can, I am sure.
By the way, all my comments assume that all drives are the same size. If not, treat all drives as if they are the same size as the smallest drive on the array (unless you are using JBOD, which is not redundant)
Re:Caught up with the speed, but still the ugliest
on
Java Faster Than C++?
·
· Score: 1
I don't know from where your impression about SWING performance comes.... neither on my MAC nor on my PC SWING applications (like CodeGuide, www.omnicore.com) seem in any way slower than native apps.
SWING tends to get slow when many controls are used. I am not sure what the specifications of your computer are, but try running a complex Java GUI app (one which is similar to a non-MFC Win32 app or a QT app for comparison) on a slower machine, or write up a quick test program which plays with 15 or so controls (hiding them, moving them, adding/removing data, creating them, etc.) a set number of times both using SWING and using the Win32 API or QT. SWING is quite a bit slower. Perhaps this has been improved in 1.5, but it is certainly the case in 1.4 and earlier. Regarding SWT performance, I've yet to see an example of a responsive, large SWT-based GUI app. Eclipse, while it is very nice in many ways, is slower than a dead slug stuck in frozen molasses compared to Microsoft Visual Studio or KDevelop, at least, in my experience. Perhaps I will have to try a server VM.
Java does RUNTIME optimization. C++ CAN'T do that. As long as C++ is not compiled to a VM it will fall more and more back behind Java the more the hot spot optimizations get improved.
I can see that runtime optimization has a few minor advantages. For example, a JVM can utilize processor-specific optimizations even if the CPU type is not known at "compile" time. That said, C++ and other compiled languages can do the same if the architecture is known (-march=). Because optimization is done at compile time, optimization time is not a very important factor, and the compiler can freely tune the hell out of the code. What optimizations can a JIT compiler do that a non-JIT compiler cannot? Let's say it sees that a function is being run quite a lot and the JIT decides to inline it. Well, C++ compilers are very good at guessing which functions will be run very often and inlining them, and the programmer can explicitly state that the function should be inlined if desired. Does the JIT try thousands of possible ways of executing the same code to see which is the fastest? That would be something that a C++ compiler cannot practically do, but compiled language optimizers are so finely tuned that I doubt there are many instances where a modern compiler, particularly an intelligent one such as Intel's or Compaq's, choose a path that is significantly slower than the best a JIT can come up with. That said, compiled languages would not have the overhead of optimizing the code at run time, not to mention the overhead of initializing the JVM. I realize that the latter is just at startup, but startup time is a problem in many modern applications more than run time. OpenOffice is a good (but extreme) example.
Further, Java lacks direct access to pointers, which allow for many snazzy optimization tricks, and cannot compare with compiled languages for bitwise operations. Perhaps I am not a sufficiently talented Java programmer, but I know of no way to do something like
if(ch >= 'a' && ch <= 'z') ch = ch&0xDF;
in Java, and I doubt that Java's JIT could figure out a similarly fast way to convert an ASCII character to uppercase, let alone a more sophisticated algorithm. Granted, in the RealWorld, developers do not often use ugly optimization hacks for readability and maintainability reasons alone. In those cases, pretty code in Java and C++ are likely comparable. Java probably has the advantage in many situations (as shown in the benchmarks in the article), but Java is a higher level language--It is much more removed from the machine, and as such, can never consistantly beat a sufficiently talented programmer until JITs begin including AI. Because you can have a strong measure of control over the code that is generated (going so far as to use inline asm if needed), there is no situation in which Ja
Re:Caught up with the speed, but still the ugliest
on
Java Faster Than C++?
·
· Score: 3, Interesting
Java might have finally caught up with the speed, but Swing is still the ugliest GUI out there.
SWING performance didn't catch up to anything. SWING and SWT are still FAR slower than QT or the Win32 API but a long, longshot. SWING especially is absurdly, stupidly slow.
Non-graphical Java code can indeed be very competitive with other languages, but it would help if the author bothered to implement the code for his tests intelligently. The Fibonacci code is recursive, which is about the slowest possible way to implement it, and much of the other code uses high-level features of C++ which are a convenience for the programmer, but are not used when worried about speed.
This fibo code, for example, should be faster:
const int max = 1000;
void fibonacci (unsigned long num) {
int fn = 1, fibo_array[max] = { 0 };
fibo_array[0] = 1;
cout << "1 ";
{
for(int i = 1 ; i < num ; i++)
{
cout << fn << ' ';
fibo_array[i] = fn;
fn += fibo_array[i-1];
}
}
return; }
This code was turned in by a student in a lab of mine. This was his first semester in CS, and this code outperforms the Java code quoted on the website considerably. (Try it!).
I am not saying that recursion and high-level C++ features should NOT be used, but I AM saying that if you are comparing the potential speed of languages, you should use tricks that each language provides to optimize speed.
Java will never be faster than properly optimized C++ compiled with an intelligent optimizing compiler except in bizarre corner cases, and tests like this are not terribly convincing demonstrations otherwise. Even the corner cases are removed by a sufficiently talented programmer. This is also not to say that Java is bad. I think Java is a great language (except for GUI programming with SWING), and definitely makes many programming tasks faster to code and easier to debug than one can do in C++.
They were never consistant in the first place! I put the model number vs. MHz on a scatterplot a while ago and it wasn't linear, though it was close.
Again, the numbers are not based on clockspeed. Even Intel's "Clockspeed is all that matters" platform does not scale linearly with clockspeed. Remember, during the life of the Athlon rating system, Intel's Pentium IV had minor overhauls that greatly boosted the performance-per-clock. The change to the Northwood core, the change to a two-channel DDR400 memory subsystem with a 200MHz (QDR) bus are two big examples. AMD had similar (but less significant) performance increases as well. If they would have stupidly stuck with Intel's "Clockspeed is performance" mantra, the model numbers would have eventually become extremely misleading. First generation Palomino Athlons do not perform as well as modern Thoroughbred Athlons anymore than Williamette Pentium IV's can compare to 800MHz FSB Northwoods.
If you plot your graph according to the average score of major benchmarks, you will find that up until about the AthlonXP 3200+ (possibly the 3000+), the rating system has been surprisignly accurate, and even a little conservative. The 3200+ rating is a bit overenthusiastic.
Athlon64's are now back to a conservative system of comparing performance.
AMD says its new model numbers are based on benchmarks developed by Synchromesh Computing. The scheme consists of the processor's family name (Geode NX or Geode GX) followed by its performance rating, followed by its power usage. Performance ratings reference performance relative to VIA's Centaur processors.
Thus, the models numbers are based on performance relative to a competitor's product, not on clockspeed. These are not, and have never been, the same thing. I suspect that the performance in this case does not scale linearly with the processor speed due to bottlenecks outside of the processor; perhaps the memory or chipset that the samples were provided with, or perhaps VIA's platform has significant performance tweaking in their higher-clockspeed cores. It does seem to be a fairly substantial difference within the same architecture.
It wasn't long ago that 1GHz was the magic number that both Intel and AMD were trying to hit. (AMD won). The performance of a 1GHz Athlon is plenty for a home server, and probably just fine for 90% of desktop PC users. My stepfather noticed zero difference moving from an Athlon 800 T-bird to an Athlon 1600+ Palomino, but it would be very noticeable for many people to not have the noise of a CPU cooling fan. Passively cooling a 6W processor would be a breeze (no pun intended). As an added bonus, the extremely low power usage and low heat output (thus lower air conditioning bills) would allow the chip to eventually pay for itself. I do hope that these chips are eventually made available through normal retail channels such as Newegg.com, since Transmeta products have certainly not been a choice outside of small laptops and diskless X terminals.
That would be cool, but they are really priced out of the workstation market. They are server chips all the way unless AMD makes them more affordable.
Er, the workstation market has traditionally been dominated by HP PA-RISC & Sun UltraSPARC-based systems, and other high-price hardware. The AMD Opteron 2xx series outperforms both and is several times cheaper. Twin Opteron 242's are around $415 USD for BOTH, and that is for consumers. Even if manufacturers like Iwill didn't get bulk discounts, that isn't even a drop in the bucket for a professional workstation.
Slashdot Mirror
A friend of mine, an overclocking expert (inventor of the "Goldfinger devices" if anyone remembers those) said that the new shrunk cores overclock to around 3GHz if you can get your FSB high enough (though this won't be an issue with the FX chips, which aren't clock-locked).
To those paying attention, 2.2GHz in an Athlon64 can generally outperform a 3.4GHz PentiumIV, so this is a big deal.
Yes, of course.
As you know, Intel is very eager to improve performance of processor architectures that are competing with their own.
I think it is unwise to try to use an unsupported (newsgroups do not count) product like Debian, especially in a business/enterprise environment. (Yes, I know you can buy 3rd party support for almost anything. It isn't IBM, and this is a conservative organization).
I know this isn't going to be popular (so read fast, this will be -1 in seconds), but risk management is very important in the corporate IT world, because mistakes or screw ups get people fired and can make the company lose millions.
Not that Debian isn't up to the task--that isn't the point. If something does go wrong, it's your ass. If some of IBM's software dies, IBM gets to deal with management, and they aren't going to switch from IBM with such a large investment (especially when IBM makes many excellent products).
If your current stuff works, there is no great reason to change, as software licenses are a drop in the bucket compared to potential pitfalls.
If you have looked over the fact and believe Debian would be a good idea, I'd take it slowly. Mail server (or something) need more capacity? See what Qmail on a ReiserFS system can do, even on the same hardware. Need another file server? Set up a quickie and demo Debian/Samba3's performance in similar or weaker hardware. Nothing open-source can replace DB2, so show that it runs on Debian just fine, etc.
Good luck!
Since my gaming system was "upgraded" to Windows XP from 2000, Starcraft decided that certain doors inside of buildings would cause a crash-to-desktop when opened. Seeing as how you must open some to get through the game, this was a bad thing.
I tried various tricks (disable sound/change drivers/reinstall/different media/3 different systems) and even contact tech support.
The *only* way I was able to get past these points was to run them in Cedega on my laptop.
Thus, in some situations, Cedega is more compatible with Windows games than Windows itself.
Agreed about Creative. Their drivers have consistantly been a steaming pile of dingo's kidneys (having been the one cause of system crashes on my Windows 2000 system for quite a while) and the software that comes with them is worse--not only is it crap, but it is horribly bloated and excessively flashy. The Audigy 2 is advertised as a "high-end" soundcard but the software lacks even the most basic functionality, such as directing bass to anything other than the subwoofer. (Some of us have real speakers, "Creative"!).
While I do not know if Creative has ever submitted a driver for WHQL certification, even Microsoft's rather low standards (e.g. the WHQL ATI drivers that would cause data corruption if "Advanced performance" was turned on for the hard drives...eh?) wouldn't allow for Creative's garbage to pass the first level of QA. Either their programmers are insanely rushed or they are completely incompetent.
(as a side note, I agree that VIA is crap, but would say that it is foolish to grant the same title to AMD, which has had no product recalls in many years and has had faster, better designed products for years [including processors and FLASH ram, but probably not chipsets]).
I think it is fairly apparent that the RIAA (and perhaps MPAA) aren't really concerned over P2P networks' effect on sales. It is a control issue. If they do not have control power over the distribution channel, they have less power.
.
P2P networks are decentralized and completely out of their sphere of influence. I am sure that the people running the RIAA are not morons--CD sales are up and there is greater and greater evidence that P2P networks slightly improve sales. However, this evidence is used to argue with a point that I think the RIAA is using to push anti-P2P legislation through, but not a point that the RIAA actually believes.
If it were purely a sales problem, the RIAA would be going after commercial CD pirates -- These are the real bad guys. They are commercially profiting from the work of others (not just sharing it), and have reportedly built an economy in and of themselves of 4.5 BILLION dollars
Of course, this doesn't mean 4.5 billion in losses--that is a BSA/piracy argument which is quite absurd--but I am sure that they do lose some money.
I know that's the normal line, but it seems that a company that is so flagrantly linux-based that they would open a Linux-only computer shop (and a carwash with a Tux logo no less!) would make it a point to eat their own dog food and use a Linux server. It isn't as if it is difficult to find Linux hosting online. Many hosting providers specifically ask what operating system you would like to run, and there are far fewer Windows-only hosts than mixed OS hosts.
Ironic that www.sub300.com and www.sub500.com, the Linux desktop resellers associated with this new store, Run Microsoft Windows 2000 webservers.
"Ed is the standard text editor."
/bin/ed /usr/ucb/vi /usr/bin/emacs
l o?
And ed doesn't waste space on my Timex Sinclair. Just look:
-rwxr-xr-x 1 root 24 Oct 29 1929
-rwxr-xr-t 4 root 1310720 Jan 1 1970
-rwxr-xr-x 1 root 5.89824e37 Oct 22 1990
Of course, on the system *I* administrate, vi is symlinked to ed.
Emacs has been replaced by a shell script which 1) Generates a syslog
message at level LOG_EMERG; 2) reduces the user's disk quota by 100K;
and 3) RUNS ED!!!!!!
"Ed is the standard text editor."
Let's look at a typical novice's session with the mighty ed:
golem$ ed
?
help
?
?
?
quit
?
exit
?
bye
?
hel
?
eat flaming death
?
^C
?
^C
?
^D
?
---
Note the consistent user interface and error reportage. Ed is
generous enough to flag errors, yet prudent enough not to overwhelm
the novice with verbosity.
"Ed is the standard text editor."
Ed, the greatest WYGIWYG editor of all.
How about a sequel Deus Ex (which is loosely based on System Shock ideas)? :/
And no, "Invisible War: Console Style Edition for 12-year olds" doesn't count.
I read the Linux Journal (which, IMO, is entirely too openly biased) and Dr. Dobbs Journal. At least, I pretend to read the latter. It isn't really as interesting as it once was.
Everything else I read online.
I worked for IQorder.com at the time that it was apparently sued (Sorry, "asked to pay") by McBride's patent friends. This guy gets around.
Suing random startups over BS patents and now suing Linux users over BS claims.
McBride of Frankenstein and Senator "Let's let the RIAA to hack file sharers legally" Hollings are some of the highest profile Mormons in their little church. I hope they get excommunicated (but then, they would lose all that tithing).
The scary thing is that McBride has seven kids. I hope that being a prick isn't genetic. I really hope that his kids see what kind of human being their dad is and don't follow in his footsteps.
It looks like I misinterpreted the information that I read. Thankyou for setting me straight.
According to Sandpile.org, the 3.4GHz Pentium IV Prescott can use up to 127W, and has a typical power usage of 103W (when browsing the web or reading email).
In my opinion, it is rediculous for a single processor to single-handedly run up your power bill. That's like having two light bulbs on 24/7 (assuming you keep your computer on), not to mention the power needed to cool your PC, let alone your house's air conditioner.
I would take a VIA chip for low-performance stuff, and an Athlon64 for performance computing. support 64-bit software including 64-bit Linux distributions, are faster than Intel's best even running 32-bit software, and they have a maximum power usage of 89W. Because of Cool'n'Quiet mode, they spend most of the time running at 800MHz consuming about 30-35W and generally not requiring a loud and abnoxious cooling fan.
It is actually impressive what the chips can do at 800MHz. You can play a full screen DVD at 1400x1050, and the CPU usage tops out at about 5% (at 800MHz). If, of course, you run something that requires more power, like a video game or a compiler, the processor instantly switches to full speed. Handy, that.
It's great that cellphones are getting more advanced, but I'm afraid a mini-3D chip running off of a 2 ounce battery displaying on a 2" screen is not going to be rivaling PCs anytime soon.
Though who knows, maybe the new cell phones will have a DVI connector and a port of Doom 3.
:blinks:
What brand of controllers failed the most often?
StorageReview is planning to publish a review using a variety of RAID controllers testing their performance with TCQ (Tagged Command Queuing) for both user and server access patterns. It should be almost finished. The results so far are rather surprising to say the least..
Why not read a few FAQ entries at StorageReview?
/", etc.
In short, I would probably recommend RAID5 if you have 3+ drives.
RAID5 gives you the most available space while still being redundant. It allows for exactly one hard drive failure.
RAID5's write speed is usually terrible, especially with a small number of drives, but write speed isn't a big deal on my home file server. (Only you know about your needs).
RAID1+0 (NOT RAID 0+1, which is inferior) is great for performance. With 4 drives, you have potentially twice the STR of one drive (writing) and 4 times the STR of one drive reading. Of course, since STR is not important for most IO, this doesn't really effect your end performance much unless you are dealing with linearly reading/writing very large files.
Writing performance will almost certainly be higher than with RAID5.
You do lose quite a lot of space (especially when you use a large number of drives). If you used a 4-drive 1+0 array, you would have the space in two of those individual drives.
RAID1 is nice, and is very reliable, but is impractical with more than two drives unless you are incredibly paranoid. RAID1 simply makes all drives copies of the others, this, you always have as much free space as one drive would have, even if you have ten. If course, you could also handle 9 drive failures and not lose data. RAID1 is fine for 2-drive arrays though.
DO NOT FORGET that RAID is no substitute for regular backups. RAID will not help if your data loss is caused by FS corruption, a cracker, accidentally typing "rm -rf
For lowest cost, I would use software RAID, such as Linux's LVM, FreeBSD's Vinum, or whatever Windows has. (RAID5 requires Windows server). (I would not use Windows as the file server myself).
For slightly higher cost, try a Promise controller.
I would avoid Highpoint and Silicon Image controllers. Highpoint, especially, is crap. (but it is very cheap, at least).
If you possibly can, I would recommend a nice 3Ware Escalade controller. Escalades are true hardware RAID cards, unlike Highpoint/SI and most of Promise's cards, and are OS independent and very stable (with certain exceptions for some unlikely configurations).
If you have any questions, you might try the StorageReview forums. There are a number of extremely knowledgeable people there, including engineers and executives-level researchers at hard drive companies. They can give far better advice than I can, I am sure.
By the way, all my comments assume that all drives are the same size. If not, treat all drives as if they are the same size as the smallest drive on the array (unless you are using JBOD, which is not redundant)
SWING tends to get slow when many controls are used. I am not sure what the specifications of your computer are, but try running a complex Java GUI app (one which is similar to a non-MFC Win32 app or a QT app for comparison) on a slower machine, or write up a quick test program which plays with 15 or so controls (hiding them, moving them, adding/removing data, creating them, etc.) a set number of times both using SWING and using the Win32 API or QT. SWING is quite a bit slower. Perhaps this has been improved in 1.5, but it is certainly the case in 1.4 and earlier.
Regarding SWT performance, I've yet to see an example of a responsive, large SWT-based GUI app. Eclipse, while it is very nice in many ways, is slower than a dead slug stuck in frozen molasses compared to Microsoft Visual Studio or KDevelop, at least, in my experience. Perhaps I will have to try a server VM.
I can see that runtime optimization has a few minor advantages. For example, a JVM can utilize processor-specific optimizations even if the CPU type is not known at "compile" time. That said, C++ and other compiled languages can do the same if the architecture is known (-march=). Because optimization is done at compile time, optimization time is not a very important factor, and the compiler can freely tune the hell out of the code. What optimizations can a JIT compiler do that a non-JIT compiler cannot? Let's say it sees that a function is being run quite a lot and the JIT decides to inline it. Well, C++ compilers are very good at guessing which functions will be run very often and inlining them, and the programmer can explicitly state that the function should be inlined if desired.
Does the JIT try thousands of possible ways of executing the same code to see which is the fastest? That would be something that a C++ compiler cannot practically do, but compiled language optimizers are so finely tuned that I doubt there are many instances where a modern compiler, particularly an intelligent one such as Intel's or Compaq's, choose a path that is significantly slower than the best a JIT can come up with. That said, compiled languages would not have the overhead of optimizing the code at run time, not to mention the overhead of initializing the JVM. I realize that the latter is just at startup, but startup time is a problem in many modern applications more than run time. OpenOffice is a good (but extreme) example.
Further, Java lacks direct access to pointers, which allow for many snazzy optimization tricks, and cannot compare with compiled languages for bitwise operations.
Perhaps I am not a sufficiently talented Java programmer, but I know of no way to do something like
in Java, and I doubt that Java's JIT could figure out a similarly fast way to convert an ASCII character to uppercase, let alone a more sophisticated algorithm.
Granted, in the RealWorld, developers do not often use ugly optimization hacks for readability and maintainability reasons alone. In those cases, pretty code in Java and C++ are likely comparable. Java probably has the advantage in many situations (as shown in the benchmarks in the article), but Java is a higher level language--It is much more removed from the machine, and as such, can never consistantly beat a sufficiently talented programmer until JITs begin including AI.
Because you can have a strong measure of control over the code that is generated (going so far as to use inline asm if needed), there is no situation in which Ja
Non-graphical Java code can indeed be very competitive with other languages, but it would help if the author bothered to implement the code for his tests intelligently.
The Fibonacci code is recursive, which is about the slowest possible way to implement it, and much of the other code uses high-level features of C++ which are a convenience for the programmer, but are not used when worried about speed.
This fibo code, for example, should be faster:This code was turned in by a student in a lab of mine. This was his first semester in CS, and this code outperforms the Java code quoted on the website considerably. (Try it!).
I am not saying that recursion and high-level C++ features should NOT be used, but I AM saying that if you are comparing the potential speed of languages, you should use tricks that each language provides to optimize speed.
Java will never be faster than properly optimized C++ compiled with an intelligent optimizing compiler except in bizarre corner cases, and tests like this are not terribly convincing demonstrations otherwise. Even the corner cases are removed by a sufficiently talented programmer.
This is also not to say that Java is bad. I think Java is a great language (except for GUI programming with SWING), and definitely makes many programming tasks faster to code and easier to debug than one can do in C++.
The change to the Northwood core, the change to a two-channel DDR400 memory subsystem with a 200MHz (QDR) bus are two big examples.
AMD had similar (but less significant) performance increases as well.
If they would have stupidly stuck with Intel's "Clockspeed is performance" mantra, the model numbers would have eventually become extremely misleading.
First generation Palomino Athlons do not perform as well as modern Thoroughbred Athlons anymore than Williamette Pentium IV's can compare to 800MHz FSB Northwoods.
If you plot your graph according to the average score of major benchmarks, you will find that up until about the AthlonXP 3200+ (possibly the 3000+), the rating system has been surprisignly accurate, and even a little conservative. The 3200+ rating is a bit overenthusiastic.
Athlon64's are now back to a conservative system of comparing performance.
It wasn't long ago that 1GHz was the magic number that both Intel and AMD were trying to hit. (AMD won).
The performance of a 1GHz Athlon is plenty for a home server, and probably just fine for 90% of desktop PC users. My stepfather noticed zero difference moving from an Athlon 800 T-bird to an Athlon 1600+ Palomino, but it would be very noticeable for many people to not have the noise of a CPU cooling fan. Passively cooling a 6W processor would be a breeze (no pun intended).
As an added bonus, the extremely low power usage and low heat output (thus lower air conditioning bills) would allow the chip to eventually pay for itself. I do hope that these chips are eventually made available through normal retail channels such as Newegg.com, since Transmeta products have certainly not been a choice outside of small laptops and diskless X terminals.
Ah, yes, that sucker is expensive, but it's a lot of CPU per watt, too (unless you consider Transmeta). :/
If only they were available for laptops.
Twin Opteron 242's are around $415 USD for BOTH, and that is for consumers. Even if manufacturers like Iwill didn't get bulk discounts, that isn't even a drop in the bucket for a professional workstation.