My 2GHz P5 is really feeling pokey, and I don't even play games! My engineering apps for school take a decent bit of horsepower, as does XSI EXP, which I tool around with occasionally. And I think I've finally found a compiler that eats CPUs even more ravenously than G++ --- the Gwydion Dylan d2c compiler:)
How are we going to explaing something this subtle
on
G5 vs Opteron, Finally
·
· Score: 3, Insightful
to Mac people??? The same people who thought that the "G3 was faster than the fastest Pentium II" for years!
Let me try to make this simple: neither Windows XP nor OS X are 64-bit OSs, and neither was running 64-bit programs. This is a much better situation for the G5 than the Opteron. 64-bit mode on the G5 really only allows for 64-bit instruction execution, and 64-bit pointers. On the Opteron, 64-bit mode enables a host of non-64-bit-related improvements, notably a doubling of the visible register set.
The bottom line is this:
The G5 will run 32-bit code just as fast (or faster, because of better cache utilization) than 64-bit code. The Opteron will run 32-bit bit code about 20% slower than 64-bit code, because of the architectural improvements in X86-64 long mode.
Note that none of the apps here would really benifet from 64-bit processing. Floating point is already 64-bit (actually, 80-bit) in both processors, and the only program that could concievably use 64-bit integer math would be Photoshop. Neither machine had more than 4GB of RAM, so 64-bit memory addressing was a non-factor.
That said, the G5 beat the Opteron by more than 20% in most of the benchmarks. I fully expect that with both CPUs running optimized 64-bit code, the G5 would still be faster, though the performance delta will be less.
That page doesn't explain anything. The real question is whether you can pass 64-bit pointers to system calls. Like, can I do a write() from a memory buffer that's above the 4GB limit? Otherwise, its more of a Windows NT PAE-type hack rather than actual support for 64-bit apps.
I too think that tax subsidies for OSS development is getting the government too involved, especially when there are already tax write-offs for donations. However, the original poster said that government brought *no* benefits to people, which was what I took exception with.
1. What is a mutated ccR5 protein? Mutated in which way? >>>>>>>>>> You can learn about ccR5 at this website.
Is this is the best you can do with "biological differences"? >>>>>>>>>>> I'm no biological expert, and I'm hazarding a guess that neither are you. I'm simply giving one example of a number of factors that could be involved.
Keep in mind that I do not believe that HIV exists. I'm entertaining your assumption because I think your claim will eventually fail. >>>>>>>>>>> That's not a very open-minded way to enter an argument. If you're convinced that you are right, and nothing can sway that convication, than this discussion is pointless.
I don't require a compelling argument. My position is skepticism, not believe. My position is, "I do not believe that HIV exists, nor do I believe that this alleged virus causes AIDS." >>>>>>>>>>> The skeptical position does not require any less proof than any other. AIDS researchers may not have proof that HIV causes AIDS, but they have strong evidence of corrolation. The onus is upon you to counter that evidence and offer your own on why AIDS should be different from other diseases in that its effects would not vary among people with different genetic makeups.
AZT was introduced in 1987, less than "a decade" before people were dying of AIDS. I see that you have not countered my claims that AZT does not prolong life and has side effects that match the symptoms of AIDS. Do you agree with these claims? >>>>>>>>>>> That was not your original claim. You asked how I know that all the American "victims" of AIDS really did not die of AZT-poisoning. My response was that people were dying of AIDS long before AZT was introduced. Thus, AZT-poisoning could not be responsible for at least most of the AIDS deaths throughout the 1980's (AZT was not in common use until the early 1990's). Also, you fail to note that the mortality rate as a result of AIDS drugs has been *decreasing* with *increases* in the use of those drugs. If the drugs were the problem, the mortality rate should be increasing as more people are medicated.
Furthermore, for those sufferers of AIDS before the introduction of AZT, how do you know that their symptoms of AIDS were not caused by AIDS cofactors? John Gallo originally denied the existence of AIDS cofactors but later revised his position. >>>>>>>>> I'm not denying (nor accepting) that claim. You said nothing about cofactors in the post to which I originally responded. You asked how I know that AIDS "victims" did not die of AZT-poisoning.
It matters not. You know as well as I that all of the millions of AIDS cases in Africa were not tested. If they were not tested, then how do we know that they have HIV? >>>>>>>>>> I don't see how it "matters not." AIDS cases in Africa *were* tested. Not all of them, because that would be stupid. A representative sample was tested and the results were extrapolated to the general population.
Do you then conclude that those samples apply to the entire population? >>>>>>>>>> Yes! That's generally how studies work. Unless you can give evidence that the sample size was insufficient (a surprisingly small sample can be surprisingly accurate)
Why is African AIDS so different from North American AIDS? Over there, it's a largely heterosexual problem. Over here, it has stayed almost completely confined to its original risk groups (promiscuous gay males and IV drug users). The HIV/AIDS Priests' explanation is something like this: "We can't expect those niggers over there in Africa not to fuck everything in sight." Will your explanation be less racist than theirs is? >>>>>>>>>>>> First, you're wrong about AIDS
Well that is your choice, but do note that modern economic theory considers government necessary for certain things that the free market is not adequate for. I'm probably more wary of government control than most people, but to believe that government brings no benifets to the people is a stupid and extremist notion.
Its the results from the Linux Text Project suite. 95% success rate, zero critical failures, means that 95% of the 2000 test cases completed successfully, and nothing crashed the kernel. To see what that means, just take a look at what test cases are in the LTP!
Which biological differences would make some people less susceptible to the alleged virus? >>>>>>>>>>>>>> For example, mutated ccR5 protiens on the surface of cells protects against AIDS because it prevents the virus from recognizing and infecting the cell. Other reasons will surely be found after research. Since the phenomena of some people being more resistant to disease is so common (indeed, a very close relative of HIV, SIV, can only infect primates even though their DNA is extremely similar to that of humans), the existence of long term non-progressors is very weak evidence. If you could prove why AIDS should be different from other viruses, in that it was not susceptible to genetic differences (more like a gun-shot to the head than the flu), then you'd have a more compelling argument.
I was talking about US AIDS cases. How do you know that all of the "AIDS victims" in the USA did not die of AZT poisioning? >>>>>>>>>>> Because people were dying of AIDS in the early 1980's, a decade before the introduction of AZT and other anti-AIDS drugs?
The "millions of people" dying of AIDS in Africa were *estimated* to have AIDS, not tested. How do you know that they are not dying of malnutrition? >>>>>>>>>>> Because studies do actually test people. I'll cite two studies from South Africa, which you can find here.In the first study, of 16,743 sexually active women at 396 locations, 24.5% were find to be infected. In the second study, of 8428 people in the general population, 11.4% were found to be infected. Moreover, my dad works in public health, and has done many projects in Africa. From what he's told me, AIDS is definately a real problem there, and many people are dying from it, without ever getting expensive drug treatments.
Really? Are you sure? How do you account for what the HIV/AIDS Priests call "long-term non-progressors"? --------- Because the virus doesn't necessarily work the same way in each person? There are significant biological differences between people, and a minority of the population may have differences that make them less susceptible to the effects of the virus.
Furthermore, how do you know that all of the "AIDS victims" did not actually die of AZT poisioning? ------------ Because there are millions of people in Africa that are dying of AIDS even though they don't have access to the AZT drug?
A multi-core processor appears as multiple processors to the software. Even SMT processors appear as multiple processors to the software. So the software scheduler is needed to juggle the available processors among the threads competing for them.
While Ars definately isn't targeted at the same audience as, say, KernelTrap, its nice to see there are a few technology websites/publications that aren't dumbed down. I remember when Byte magazine used to publish articles detailing the PowerPC architecture, down to the level of registers and the types of pipelines in the first set of implementations. Compare this to the ZD rags, which are a hair away from calling the CPU the "brain" of the computer!
the PowerPC instruction set was designed to be very easy for compilers to optimize for. >>>>>>>>>>>>>> Wrong. The PowerPC, like all RISC instruction sets, was designed to be more dependent on compiler optimizations for good performance. RISC processors are generally harder to optimize for, only because more aggressive optimizations are necessary to get the same level of performance.
You used them as examples when they weren't good examples. Its like saying that Ford introducing new models every year is a sign of their commitment to product innovation! Its something that everyone does, so its not an example of standing out.
What are you talking about? We were specifically talking about die shrinks and optimized compilers. IBM doesn't die shrink any more often than Intel, and its compiler isn't any more well tuned than Intel's.
And OS X isn't fast. Lmbench numbers show it to be a lot slower than Linux on the same machine. It *feels* faster, because the GUI back-buffers all windows, but in absolute terms its rather pokey.
That's a completely bogus post. Every chip manufacturer does die shrinks periodically. And Intel also makes a line of optimized compilers specialized for its chips. Apple and IBM are nothing special in this regard!
I enabled RenderAccel. Its a nice boost for Qt/KDE (where it makes text highlighting and scrolling much smoother), but using the nvidia driver at all (RenderAccel on or off), makes GTK+ resize even worse than it does already.
I don't get Indian movies either. I get a lot of exposure to them, but I can't say I enjoy most of them. Of course, I don't like classical music, musicals, or old movies, even though those genres were very popular at one time. And I don't understand daytime tv at all!
My point is that as (mostly) Westerners, we have a hard time understanding how anyone could like those movies, because our culture is so different. I'm sure they have a difficult time understanding why some of the most popular shows on TV now (Survivor, Fear Factor), have contestents routinely eating bugs and worms!
I generally can't stand Indian movies, but there are a couple of good ones. Somebody earlier mentioned Lagaan, and I heard Devdas was also pretty good. But you'll generally not see these movies on TV --- just as the only Hollywood movies I ever see on TV are ones like Billy Madison...
Slashdot Mirror
My 2GHz P5 is really feeling pokey, and I don't even play games! My engineering apps for school take a decent bit of horsepower, as does XSI EXP, which I tool around with occasionally. And I think I've finally found a compiler that eats CPUs even more ravenously than G++ --- the Gwydion Dylan d2c compiler :)
to Mac people??? The same people who thought that the "G3 was faster than the fastest Pentium II" for years!
Let me try to make this simple: neither Windows XP nor OS X are 64-bit OSs, and neither was running 64-bit programs. This is a much better situation for the G5 than the Opteron. 64-bit mode on the G5 really only allows for 64-bit instruction execution, and 64-bit pointers. On the Opteron, 64-bit mode enables a host of non-64-bit-related improvements, notably a doubling of the visible register set.
The bottom line is this:
The G5 will run 32-bit code just as fast (or faster, because of better cache utilization) than 64-bit code. The Opteron will run 32-bit bit code about 20% slower than 64-bit code, because of the architectural improvements in X86-64 long mode.
Note that none of the apps here would really benifet from 64-bit processing. Floating point is already 64-bit (actually, 80-bit) in both processors, and the only program that could concievably use 64-bit integer math would be Photoshop. Neither machine had more than 4GB of RAM, so 64-bit memory addressing was a non-factor.
That said, the G5 beat the Opteron by more than 20% in most of the benchmarks. I fully expect that with both CPUs running optimized 64-bit code, the G5 would still be faster, though the performance delta will be less.
That page doesn't explain anything. The real question is whether you can pass 64-bit pointers to system calls. Like, can I do a write() from a memory buffer that's above the 4GB limit? Otherwise, its more of a Windows NT PAE-type hack rather than actual support for 64-bit apps.
Yikes. Someone managed to get an eletro-fetish erotica story modded up on Slashdot.
I should be more surprised than I am...
I too think that tax subsidies for OSS development is getting the government too involved, especially when there are already tax write-offs for donations. However, the original poster said that government brought *no* benefits to people, which was what I took exception with.
1. What is a mutated ccR5 protein? Mutated in which way?
>>>>>>>>>>
You can learn about ccR5 at this website.
Is this is the best you can do with "biological differences"?
>>>>>>>>>>>
I'm no biological expert, and I'm hazarding a guess that neither are you. I'm simply giving one example of a number of factors that could be involved.
Keep in mind that I do not believe that HIV exists. I'm entertaining your assumption because I think your claim will eventually fail.
>>>>>>>>>>>
That's not a very open-minded way to enter an argument. If you're convinced that you are right, and nothing can sway that convication, than this discussion is pointless.
I don't require a compelling argument. My position is skepticism, not believe. My position is, "I do not believe that HIV exists, nor do I believe that this alleged virus causes AIDS."
>>>>>>>>>>>
The skeptical position does not require any less proof than any other. AIDS researchers may not have proof that HIV causes AIDS, but they have strong evidence of corrolation. The onus is upon you to counter that evidence and offer your own on why AIDS should be different from other diseases in that its effects would not vary among people with different genetic makeups.
AZT was introduced in 1987, less than "a decade" before people were dying of AIDS. I see that you have not countered my claims that AZT does not prolong life and has side effects that match the symptoms of AIDS. Do you agree with these claims?
>>>>>>>>>>>
That was not your original claim. You asked how I know that all the American "victims" of AIDS really did not die of AZT-poisoning. My response was that people were dying of AIDS long before AZT was introduced. Thus, AZT-poisoning could not be responsible for at least most of the AIDS deaths throughout the 1980's (AZT was not in common use until the early 1990's). Also, you fail to note that the mortality rate as a result of AIDS drugs has been *decreasing* with *increases* in the use of those drugs. If the drugs were the problem, the mortality rate should be increasing as more people are medicated.
Furthermore, for those sufferers of AIDS before the introduction of AZT, how do you know that their symptoms of AIDS were not caused by AIDS cofactors? John Gallo originally denied the existence of AIDS cofactors but later revised his position.
>>>>>>>>>
I'm not denying (nor accepting) that claim. You said nothing about cofactors in the post to which I originally responded. You asked how I know that AIDS "victims" did not die of AZT-poisoning.
It matters not. You know as well as I that all of the millions of AIDS cases in Africa were not tested. If they were not tested, then how do we know that they have HIV?
>>>>>>>>>>
I don't see how it "matters not." AIDS cases in Africa *were* tested. Not all of them, because that would be stupid. A representative sample was tested and the results were extrapolated to the general population.
Do you then conclude that those samples apply to the entire population?
>>>>>>>>>>
Yes! That's generally how studies work. Unless you can give evidence that the sample size was insufficient (a surprisingly small sample can be surprisingly accurate)
Why is African AIDS so different from North American AIDS? Over there, it's a largely heterosexual problem. Over here, it has stayed almost completely confined to its original risk groups (promiscuous gay males and IV drug users). The HIV/AIDS Priests' explanation is something like this: "We can't expect those niggers over there in Africa not to fuck everything in sight." Will your explanation be less racist than theirs is?
>>>>>>>>>>>>
First, you're wrong about AIDS
Well that is your choice, but do note that modern economic theory considers government necessary for certain things that the free market is not adequate for. I'm probably more wary of government control than most people, but to believe that government brings no benifets to the people is a stupid and extremist notion.
Its the results from the Linux Text Project suite. 95% success rate, zero critical failures, means that 95% of the 2000 test cases completed successfully, and nothing crashed the kernel. To see what that means, just take a look at what test cases are in the LTP!
Which biological differences would make some people less susceptible to the alleged virus?
>>>>>>>>>>>>>>
For example, mutated ccR5 protiens on the surface of cells protects against AIDS because it prevents the virus from recognizing and infecting the cell. Other reasons will surely be found after research. Since the phenomena of some people being more resistant to disease is so common (indeed, a very close relative of HIV, SIV, can only infect primates even though their DNA is extremely similar to that of humans), the existence of long term non-progressors is very weak evidence. If you could prove why AIDS should be different from other viruses, in that it was not susceptible to genetic differences (more like a gun-shot to the head than the flu), then you'd have a more compelling argument.
I was talking about US AIDS cases. How do you know that all of the "AIDS victims" in the USA did not die of AZT poisioning?
>>>>>>>>>>>
Because people were dying of AIDS in the early 1980's, a decade before the introduction of AZT and other anti-AIDS drugs?
The "millions of people" dying of AIDS in Africa were *estimated* to have AIDS, not tested. How do you know that they are not dying of malnutrition?
>>>>>>>>>>>
Because studies do actually test people. I'll cite two studies from South Africa, which you can find here.In the first study, of 16,743 sexually active women at 396 locations, 24.5% were find to be infected. In the second study, of 8428 people in the general population, 11.4% were found to be infected. Moreover, my dad works in public health, and has done many projects in Africa. From what he's told me, AIDS is definately a real problem there, and many people are dying from it, without ever getting expensive drug treatments.
Really? Are you sure? How do you account for what the HIV/AIDS Priests call "long-term non-progressors"?
---------
Because the virus doesn't necessarily work the same way in each person? There are significant biological differences between people, and a minority of the population may have differences that make them less susceptible to the effects of the virus.
Furthermore, how do you know that all of the "AIDS victims" did not actually die of AZT poisioning?
------------
Because there are millions of people in Africa that are dying of AIDS even though they don't have access to the AZT drug?
Eh? I think that's a large coffee at Starbucks. I think it goes "Java short," "Java tall," and "Java grande."
ewwww.
Men don't clip their nails?
A multi-core processor appears as multiple processors to the software. Even SMT processors appear as multiple processors to the software. So the software scheduler is needed to juggle the available processors among the threads competing for them.
While Ars definately isn't targeted at the same audience as, say, KernelTrap, its nice to see there are a few technology websites/publications that aren't dumbed down. I remember when Byte magazine used to publish articles detailing the PowerPC architecture, down to the level of registers and the types of pipelines in the first set of implementations. Compare this to the ZD rags, which are a hair away from calling the CPU the "brain" of the computer!
the PowerPC instruction set was designed to be very easy for compilers to optimize for.
>>>>>>>>>>>>>>
Wrong. The PowerPC, like all RISC instruction sets, was designed to be more dependent on compiler optimizations for good performance. RISC processors are generally harder to optimize for, only because more aggressive optimizations are necessary to get the same level of performance.
You used them as examples when they weren't good examples. Its like saying that Ford introducing new models every year is a sign of their commitment to product innovation! Its something that everyone does, so its not an example of standing out.
What are you talking about? We were specifically talking about die shrinks and optimized compilers. IBM doesn't die shrink any more often than Intel, and its compiler isn't any more well tuned than Intel's.
And OS X isn't fast. Lmbench numbers show it to be a lot slower than Linux on the same machine. It *feels* faster, because the GUI back-buffers all windows, but in absolute terms its rather pokey.
That's a completely bogus post. Every chip manufacturer does die shrinks periodically. And Intel also makes a line of optimized compilers specialized for its chips. Apple and IBM are nothing special in this regard!
Agh! The shame of it!
Aside from the dancing bits, the same is true for Hollywood movies.
For god's sake people! Fire in bursts! And MP5's don't have clips with 100 rounds!
What the hell is xpat?
I enabled RenderAccel. Its a nice boost for Qt/KDE (where it makes text highlighting and scrolling much smoother), but using the nvidia driver at all (RenderAccel on or off), makes GTK+ resize even worse than it does already.
I don't get Indian movies either. I get a lot of exposure to them, but I can't say I enjoy most of them. Of course, I don't like classical music, musicals, or old movies, even though those genres were very popular at one time. And I don't understand daytime tv at all!
My point is that as (mostly) Westerners, we have a hard time understanding how anyone could like those movies, because our culture is so different. I'm sure they have a difficult time understanding why some of the most popular shows on TV now (Survivor, Fear Factor), have contestents routinely eating bugs and worms!
I generally can't stand Indian movies, but there are a couple of good ones. Somebody earlier mentioned Lagaan, and I heard Devdas was also pretty good. But you'll generally not see these movies on TV --- just as the only Hollywood movies I ever see on TV are ones like Billy Madison...