How? Apple states a Pentium 4 gets 800 SPECInt by using a little used compiler called GCC, where Intel states it gets 1200 by using a more widely used compiler. It's clear that Apple is trying it's hardest to slow down the SPEC results for the Pentium 4 in ways no one else would think of. A normal person wouldn't even think to use GCC in a real-world app.
I'm having a very difficult time believing Apple's claims of it being the faster chip. I don't see why you do? It's simply untrue- You can run programs faster on an Intel Pentium 4 than you can on the PowerPC970. This statement comes from real-world application benchmarks from the SPEC Suite- It gzips faster than a PowerPC, it compiles a program faster than a PowerPC, it renders a 3-D image faster than a PowerPC. So, can you please retract any support for Apple's claims? No one else on slashdot believes them?
I disagree on not using SPEC as a real-world barometer of performance. SPEC was designed with real-world use in mind, and refined over the years. I've benchmarked several engineering apps on various CPU's, and they correlate directly with SPEC results.
Basically SPEC is a good way to find out how fast your computer is. You can use the numbers to make a purchasing decision. And from the SPEC website, the Pentium 4 is about 50% faster than a PowerPC 970.
I don't know if the PowerPC is faster, what gives you the impression that it could be? I haven't seen any better SPEC numbers than the numbers Apple put out. You're getting into theoretical "coulds/shoulds". And I dont know where you get the Intel compiler isn't used much in the real world, either? Please provide a reference. What's indisputable is the Pentium4 is faster than what Apple is stating- just head on over to the SPEC website at www.spec.org to see the 1200 SPECInt/FP numbers for the P4 yourself.
What makes you think SPEC results aren't real world tests? SPEC applications are stuff people use - one of the tests is to see how fast GCC can compile a program. GCC is used by Apple, for example... =^) Also, some of these programs can take days to run. I would use SPEC to buy a machine for my next engineering job. I don't care about internal specs, such as Memory Bandwidth to buy a machine. I only care about how fast my CPU intensive job gets finished, and SPEC measures this perfectly. This is because SPEC measure overall system performance - compilers, CPU, clock speeds, memory bandwidth, vector processing, etc. All these factors affect the runtime of the final results. If the memory bandwidth on the G5 is so great, it would make the results of the Apple machine better. Unfortunately for Apple, the Pentium 4 beats it in Real-world tests, as the individual components of the SPEC results state- I can Gzip a file faster on a Pentium 4 than a PowerPC 970, I can run my engineering simulations faster on a Pentium 4 than a PowerPC.
SPEC is very much a real-world application benchmark of CPU intensive tasks that people actually run and use, and not an arbitrary synthetic benchmark to measure CPU hardware performance only. The faster these benchmark programs run, the better scores you get, and the quicker you can go home after finishing your work at the office. SPEC's goal doesn't care how you get the job done, as long as it gets done. You could optimize the memory, compiler, ALU, or whatever. It seems Apple got themselves into a trap when they claimed fastest desktop based on SPEC results, as they clearly didn't understand the systems level benchmark objectives of SPEC CPU2000.
In the end, with a score of 1200, the Pentium 4 gets the job done much faster than PowerPC 970, with a score of 800. I'm not sure what Apple was thinking when they published the lower scores of the Pentium 4, as clearly the Pentium 4 could do better. The ultimate claim that Apple has the fastest desktop system is therefore incorrect.
From www.spec.org:
Q9: What source code is provided? What exactly makes up these suites?
A9: CINT2000 and CFP2000 are based on compute-intensive applications provided as source code. CINT2000 contains 11 applications written in C and one in C++ (252.eon) that are used as benchmarks:
Name Brief Description
164.gzip Data compression utility
175.vpr FPGA circuit placement and routing
176.gcc C compiler
181.mcf Minimum cost network flow solver
186.crafty Chess program
197.parser Natural language processing
252.eon Ray tracing
253.perlbmk Perl
254.gap Computational group theory
255.vortex Object-oriented database
256.bzip2 Data compression utility
300.twolf Place and route simulator
CFP2000 contains 14 applications (six FORTRAN77, four FORTRAN90 and four C) that are used as benchmarks:
Name Brief Description
168.wupwise Quantum chromodynamics
171.swim Shallow water modeling
172.mgrid Multi-grid solver in 3D potential field
173.applu Parabolic/elliptic partial differential equations
177.mesa 3D graphics library
178.galgel Fluid dynamics: analysis of oscillatory instability
179.art Neural network simulation: adaptive resonance theory
183.equake Finite element simulation: earthquake modeling
187.facerec Computer vision: recognizes faces
188.ammp Computational chemistry
189.lucas Number theory: primality testing
191.fma3d Finite-element crash simulation
200.sixtrack Particle accelerator model 301.apsi Solves problems regarding temperature, wind, distribution of pollutants
You don't design a processor without also creating a compiler architecture designed for that CPU- it is one and the same. Intel does not design a Pentium 4 CPU for it's own sake - it designs a system.
The hardware engineers know the kind of tradeoffs they will be making with the compilers when designing a CPU. The limitations of the compiler will influence the hardware architecture, an vice-versa. There's a lot of analysis on both the hardware and software side when designing a CPU. You are adding a variable to your scientific test when you use a suboptimal compiler. It throws off the design of the system completely.
Then they should include the results of that. The comparisons should be done on the fastest results available, and not be based on some arbitrary factoring out of the compiler capability.
In the end, SPEC is about measuring how fast something can be done in the real world - gzipping a file, large matrix multiplications, etc. Why attach an arbitrary compiler to achieve that end result? Get the fastest one, instead.
It's incorrect to normalize the compiler out when performing CPU benchmarks. Instead of measuring 2 different CPUs with the same compiler, they should be using the fastest compiler for each platform. The compiler is integral to CPU design- I could make a teraflops VLIW CPU that does 1000 floating point multiply-adds per instructions, but it would be useless if I gave it a compiler that wasn't designed for it.
So, the correct SPEC results for the 3GHz Intel CPU (from the www.spec.org website) should be 1200 SPECInt and 1229 SPECFp, vs. 800 SPECInt and 840 SPECFp for the PowerPC 970.
There needs to be a way for the IP addresses of known bad guys, like Bonzi and Gator, to be filtered out at the ISP level so they do not massively disrupt computers. Perhaps a central authority used to designate spyware domain names and IP addresses...
It would be nice if the Playstation 3 could be used for linear-algebra as well. I would think that 10 $200 Playstation 3s wouldbe better than 1 $2000 CPU/GPU system?
I guess it depends on how general purpose the Cell architecture is going to be vs. CPU/GPU combination. (64-bit precision, IO bandwith/latency, local memory size)
I'm just glad they brought up the number 23 prominently in the film.. Someone did their homework on that number! see: The Number 23
Did you know that... the letter W is the 23rd letter of the alphabet, which is used as the initials of the World Wide Web?? Or that the TCP/IP port for telnet is 23? or that generative cells have 23 chromosomes? =^)
Maybe it is a Human concoction as an experiment in AI, but I think using "Love" as a cop-out would be too Dark City.
Maybe each layer is meant to work on a specific human trait.
Overall this movie does think more than Part 1, as it introduces existence on a higher level than the usual and immediate "false reality" of Part 1. I do hope it doesn't go gratuitously cheesy on us by introducing several false realities as a needless plot twist, as that's been done before.
I think HD-DVD is going to be driven by the computer/PC market due to the massive data size increase over DVDs. I could only assume that all manufacturers will include HD-DVD's in their PCs within a few years after its release. The taiwanese/chinese manufacturers are going to create a cheap, possibly incompatible format no matter what. And once there are millions of these drives out there, the consumer electronics markets will follow.
Right now it takes DVD-Rs 2 discs to back up a movie (dual-layer vs. single layer). It would be nice to archive my movie collections using less media.
If it helps any Pioneer also sells the equivalent 50" Plasma for commercial use at a much lower price. The Pioneer 503CMX has the exact same picture quality and resolution as the Consumer model, but can be bought for around $6000, a $4000 savings.
Sucks that this thing doesn't do 24p. That could be used for easy film transfer for moviemaking.
Also, the long delay for something like this to come out is due to the need for an ASIC that can compress the data in real-time to MPEG2 format using very little power. This is the reason we didn't see an HD Camcorder before. The compression technology in the chips are still very new.
Note that there are also better versions of the video compression ASIC coming. Hopefully this will get us to 1080p resolutions soon. And maybe recordings onto solid-state mediums as well. I'd like to see real-time compression onto MPEG4 or WM9 formats. This will be about 3 years away.
>>Why is it NOT okay when Saddam defies the UN, but it's okay when President Bush does it? Could >> someone explain that to me, please?
>A fair question. The answer is very long, but the short version is this: Iraq is in defiance of the United Nations, but the United States and our parters are not.
It's hard to be in violations of UN laws when you can just veto any resolution that you don't agree with.
Having a seperate soundcard process all the data is expensive in terms of systems level design- you need a slot, it takes up rack space, needs a bigger box to package it for retail, adds weight to shipping costs, adds heat, consumes more power, etc.. Just more expensive in general.
A cheaper and better solution would be to some hardware sound processing integrated into the chipset like NVidia.
Soundcards or audio-only chips are just bad ideas these days. Full audio processing needs much less hardware resources than other processing tasks, such as 3-D graphics. It can be thrown in for cheap on any modern chipset.
They could have done what Pioneer has done in their top of the line Elite 49Txi receiver and 47Ai universal player- use an encrypted IEEE-1394 output instead of the optical/electrical digital link. They do this using approved standards for firewire audio transmission.
This is the only DVD-Audio or SACD player on the market capable of outputting a digitial audio signal (the article is wrong).
Of course the only thing that would be able to recieve this signal would be the 49Txi itself, but someone has to start somewhere to get high-end digital audio directly to the amplifier.
Actually, with a firewire audio output and a receiver that accepts firewire input, you wouldn't even need a soundcard... Anyone wanna try to write a device driver that can play audio through firewire???
The Movie trailer commercials were in High Definition. This is the first time I've seen commercials in HD. Wow! I love my plasma:)
First time commercials shown in High Definition?
on
Superbowl XXXVII
·
· Score: 1
This is the first time i've ever seen commercials in high definition. The Matrix trailer was so amazing that it might have looked better than watching it on film! They also had all the other movie trailers in high-def.
For all those that were fortunate enough to have seen this in high-def, is this a great thing or what??
I think the solution for radio's replacement would have to involve something way more than 100 channels offered by XM. Perhaps 10,000+ channels. This way no one company can monopolize the bandwidth.
Slashdot Mirror
How? Apple states a Pentium 4 gets 800 SPECInt by using a little used compiler called GCC, where Intel states it gets 1200 by using a more widely used compiler. It's clear that Apple is trying it's hardest to slow down the SPEC results for the Pentium 4 in ways no one else would think of. A normal person wouldn't even think to use GCC in a real-world app.
I'm having a very difficult time believing Apple's claims of it being the faster chip. I don't see why you do? It's simply untrue- You can run programs faster on an Intel Pentium 4 than you can on the PowerPC970. This statement comes from real-world application benchmarks from the SPEC Suite- It gzips faster than a PowerPC, it compiles a program faster than a PowerPC, it renders a 3-D image faster than a PowerPC. So, can you please retract any support for Apple's claims? No one else on slashdot believes them?
I disagree on not using SPEC as a real-world barometer of performance. SPEC was designed with real-world use in mind, and refined over the years. I've benchmarked several engineering apps on various CPU's, and they correlate directly with SPEC results.
Basically SPEC is a good way to find out how fast your computer is. You can use the numbers to make a purchasing decision. And from the SPEC website, the Pentium 4 is about 50% faster than a PowerPC 970.
I don't know if the PowerPC is faster, what gives you the impression that it could be? I haven't seen any better SPEC numbers than the numbers Apple put out. You're getting into theoretical "coulds/shoulds". And I dont know where you get the Intel compiler isn't used much in the real world, either? Please provide a reference. What's indisputable is the Pentium4 is faster than what Apple is stating- just head on over to the SPEC website at www.spec.org to see the 1200 SPECInt/FP numbers for the P4 yourself.
What makes you think SPEC results aren't real world tests? SPEC applications are stuff people use - one of the tests is to see how fast GCC can compile a program. GCC is used by Apple, for example... =^) Also, some of these programs can take days to run. I would use SPEC to buy a machine for my next engineering job. I don't care about internal specs, such as Memory Bandwidth to buy a machine. I only care about how fast my CPU intensive job gets finished, and SPEC measures this perfectly. This is because SPEC measure overall system performance - compilers, CPU, clock speeds, memory bandwidth, vector processing, etc. All these factors affect the runtime of the final results. If the memory bandwidth on the G5 is so great, it would make the results of the Apple machine better. Unfortunately for Apple, the Pentium 4 beats it in Real-world tests, as the individual components of the SPEC results state- I can Gzip a file faster on a Pentium 4 than a PowerPC 970, I can run my engineering simulations faster on a Pentium 4 than a PowerPC.
SPEC is very much a real-world application benchmark of CPU intensive tasks that people actually run and use, and not an arbitrary synthetic benchmark to measure CPU hardware performance only. The faster these benchmark programs run, the better scores you get, and the quicker you can go home after finishing your work at the office. SPEC's goal doesn't care how you get the job done, as long as it gets done. You could optimize the memory, compiler, ALU, or whatever. It seems Apple got themselves into a trap when they claimed fastest desktop based on SPEC results, as they clearly didn't understand the systems level benchmark objectives of SPEC CPU2000.
In the end, with a score of 1200, the Pentium 4 gets the job done much faster than PowerPC 970, with a score of 800. I'm not sure what Apple was thinking when they published the lower scores of the Pentium 4, as clearly the Pentium 4 could do better. The ultimate claim that Apple has the fastest desktop system is therefore incorrect.
From www.spec.org:
Q9: What source code is provided? What exactly makes up these suites?
A9: CINT2000 and CFP2000 are based on compute-intensive applications provided as source code. CINT2000 contains 11 applications written in C and one in C++ (252.eon) that are used as benchmarks:
Name Brief Description
164.gzip Data compression utility
175.vpr FPGA circuit placement and routing
176.gcc C compiler
181.mcf Minimum cost network flow solver
186.crafty Chess program
197.parser Natural language processing
252.eon Ray tracing
253.perlbmk Perl
254.gap Computational group theory
255.vortex Object-oriented database
256.bzip2 Data compression utility
300.twolf Place and route simulator
CFP2000 contains 14 applications (six FORTRAN77, four FORTRAN90 and four C) that are used as benchmarks:
Name Brief Description
168.wupwise Quantum chromodynamics
171.swim Shallow water modeling
172.mgrid Multi-grid solver in 3D potential field
173.applu Parabolic/elliptic partial differential equations
177.mesa 3D graphics library
178.galgel Fluid dynamics: analysis of oscillatory instability
179.art Neural network simulation: adaptive resonance theory
183.equake Finite element simulation: earthquake modeling
187.facerec Computer vision: recognizes faces
188.ammp Computational chemistry
189.lucas Number theory: primality testing
191.fma3d Finite-element crash simulation
200.sixtrack Particle accelerator model
301.apsi Solves problems regarding temperature, wind, distribution of pollutants
You don't design a processor without also creating a compiler architecture designed for that CPU- it is one and the same. Intel does not design a Pentium 4 CPU for it's own sake - it designs a system.
The hardware engineers know the kind of tradeoffs they will be making with the compilers when designing a CPU. The limitations of the compiler will influence the hardware architecture, an vice-versa. There's a lot of analysis on both the hardware and software side when designing a CPU. You are adding a variable to your scientific test when you use a suboptimal compiler. It throws off the design of the system completely.
Then they should include the results of that. The comparisons should be done on the fastest results available, and not be based on some arbitrary factoring out of the compiler capability.
In the end, SPEC is about measuring how fast something can be done in the real world - gzipping a file, large matrix multiplications, etc. Why attach an arbitrary compiler to achieve that end result? Get the fastest one, instead.
It's incorrect to normalize the compiler out when performing CPU benchmarks. Instead of measuring 2 different CPUs with the same compiler, they should be using the fastest compiler for each platform. The compiler is integral to CPU design- I could make a teraflops VLIW CPU that does 1000 floating point multiply-adds per instructions, but it would be useless if I gave it a compiler that wasn't designed for it.
So, the correct SPEC results for the 3GHz Intel CPU (from the www.spec.org website) should be 1200 SPECInt and 1229 SPECFp, vs. 800 SPECInt and 840 SPECFp for the PowerPC 970.
The Intel CPU wins (by a lot!)
I would particularly like to see the results at the SPEC org website: All published SPEC CPU2000 results
There needs to be a way for the IP addresses of known bad guys, like Bonzi and Gator, to be filtered out at the ISP level so they do not massively disrupt computers. Perhaps a central authority used to designate spyware domain names and IP addresses...
It would be nice if the Playstation 3 could be used for linear-algebra as well. I would think that 10 $200 Playstation 3s wouldbe better than 1 $2000 CPU/GPU system?
I guess it depends on how general purpose the Cell architecture is going to be vs. CPU/GPU combination. (64-bit precision, IO bandwith/latency, local memory size)
I'm just glad they brought up the number 23 prominently in the film.. Someone did their homework on that number! see: The Number 23
Did you know that... the letter W is the 23rd letter of the alphabet, which is used as the initials of the World Wide Web?? Or that the TCP/IP port for telnet is 23? or that generative cells have 23 chromosomes? =^)
Maybe it is a Human concoction as an experiment in AI, but I think using "Love" as a cop-out would be too Dark City.
Maybe each layer is meant to work on a specific human trait.
Overall this movie does think more than Part 1, as it introduces existence on a higher level than the usual and immediate "false reality" of Part 1. I do hope it doesn't go gratuitously cheesy on us by introducing several false realities as a needless plot twist, as that's been done before.
I think HD-DVD is going to be driven by the computer/PC market due to the massive data size increase over DVDs. I could only assume that all manufacturers will include HD-DVD's in their PCs within a few years after its release. The taiwanese/chinese manufacturers are going to create a cheap, possibly incompatible format no matter what. And once there are millions of these drives out there, the consumer electronics markets will follow.
Right now it takes DVD-Rs 2 discs to back up a movie (dual-layer vs. single layer). It would be nice to archive my movie collections using less media.
I got the Pioneer 503CMX from www.plasmaextreme.us, also a great source. $6000 vs. $10000.
If it helps any Pioneer also sells the equivalent 50" Plasma for commercial use at a much lower price. The Pioneer 503CMX has the exact same picture quality and resolution as the Consumer model, but can be bought for around $6000, a $4000 savings.
I bought one, and it's awesome!
Sucks that this thing doesn't do 24p. That could be used for easy film transfer for moviemaking.
Also, the long delay for something like this to come out is due to the need for an ASIC that can compress the data in real-time to MPEG2 format using very little power. This is the reason we didn't see an HD Camcorder before. The compression technology in the chips are still very new.
Note that there are also better versions of the video compression ASIC coming. Hopefully this will get us to 1080p resolutions soon. And maybe recordings onto solid-state mediums as well. I'd like to see real-time compression onto MPEG4 or WM9 formats. This will be about 3 years away.
Anyone have good pointers on how to retrofit an existing neighborhood similiar in design to this neighborhood to enable fiber-to-the-curb?
>>Why is it NOT okay when Saddam defies the UN, but it's okay when President Bush does it? Could
>> someone explain that to me, please?
>A fair question. The answer is very long, but the short version is this: Iraq is in defiance of the United Nations, but the United States and our parters are not.
It's hard to be in violations of UN laws when you can just veto any resolution that you don't agree with.
Having a seperate soundcard process all the data is expensive in terms of systems level design- you need a slot, it takes up rack space, needs a bigger box to package it for retail, adds weight to shipping costs, adds heat, consumes more power, etc.. Just more expensive in general.
A cheaper and better solution would be to some hardware sound processing integrated into the chipset like NVidia.
Soundcards or audio-only chips are just bad ideas these days. Full audio processing needs much less hardware resources than other processing tasks, such as 3-D graphics. It can be thrown in for cheap on any modern chipset.
They could have done what Pioneer has done in their top of the line Elite 49Txi receiver and 47Ai universal player- use an encrypted IEEE-1394 output instead of the optical/electrical digital link. They do this using approved standards for firewire audio transmission.
This is the only DVD-Audio or SACD player on the market capable of outputting a digitial audio signal (the article is wrong).
Of course the only thing that would be able to recieve this signal would be the 49Txi itself, but someone has to start somewhere to get high-end digital audio directly to the amplifier.
Actually, with a firewire audio output and a receiver that accepts firewire input, you wouldn't even need a soundcard... Anyone wanna try to write a device driver that can play audio through firewire???
Why stop at just laptops? Why not go straight to powering houses and cars?
The Movie trailer commercials were in High Definition. This is the first time I've seen commercials in HD. Wow! I love my plasma :)
This is the first time i've ever seen commercials in high definition. The Matrix trailer was so amazing that it might have looked better than watching it on film! They also had all the other movie trailers in high-def.
For all those that were fortunate enough to have seen this in high-def, is this a great thing or what??
I think the solution for radio's replacement would have to involve something way more than 100 channels offered by XM. Perhaps 10,000+ channels. This way no one company can monopolize the bandwidth.
And he would be cool if he named his kid Justin....