Efficient how? Power? usually. Space? hardly. Typical asynchronous designs can easily require 33% to 50% more gates than clocked designs. Think about your 25M gate CPU suddenly being 37.5M gates. Then you get into all the problems that large chips have. Clockless isn't free, you still have to pay for it, just in other ways. Of course, the arguments against clocked designs is that clock drivers can take up huge amounts of real estate on the chip too. I think I recall one of the Alpha CPUs having about 10% of its die dedicated to driving the clock signal.
OK, so you can... why would you? You run the risk of your ugly hacking being non-portable, which kind of defeats the purpose of Java. Not to mention that ugly hack code is a PITA to maintain a year down the road when the person who wrote the junk isn't on the project anymore.
I get far fewer religious group personel knocking on my door than unsolicited phone callers asking me to subscribe to some credit card and such. I also get fewer door callers than junk smail.
If I don't want to talk to the religious canvasers, I simply tell them I'm not interested and they leave me be and neither of us is worse for the wear.
I personally haven't witnessed discrimination that you describe, but to be honest I haven't watched for it either.
Yup, they understand it well. They have to make it fun enough for you to keep playing, but to make you play as long as possible. You paid $50 for the game and $70 for 7 months of playing. Looks like they did pretty good.
Even an non-twinked, non-powerleveled character can get to 50 in less than 20 days of playtime, btw.
That's the difference between "sponsoring" and "hiring" someone to do something for you. Sponsors give money to the NASCAR team. In exchange they get advertising space on the car. It's like paying for a commercial to be shown on TV. However, we hire software consultants to write code for us. When they are finished with the project, it is ours, not theirs.
...is I can't tell which group has the most and bigger assholes in it... the Atheists or the Cristians. In my experience, the Atheist group is the group that is more (negatively and abrasively) vocal than the Christians.
Regardless of what you believe, the tennants of the Bible are a good guide to live by. Simply by living according to the Ten Commandments promotes a better, less stressful, more peaceful life. I personally believe that people can do what they want as long as it doesn't negatively effect others. Believe what you want. You are welcome to express those beliefs to me but if I don't hold with them, go about your way and leave me be - there is no reason to then criticise me for what I believe. Unfortunately, it seems that few other people also believe this.
As to someone earlier who spoke of mistranslations, I had a Hebrew scholar tell me before that, just like any other language, Hebrew changes over the years. Idioms and vernacular interpretations did exist in ancient Hebrew that are different than would be interpreted even a few hundred years ago. One issue we discussed was the Commandment "Thou shalt not kill." He said that the word used for "kill" in the original texts could have actually meant "murder", which can put a bit of a different meaning on that Commandment.
Anyway, I just find it mildly amusing that I see far more posts from Evolutionists flaming Cristians than the other way around.
Re:Itanium at 1.6 GHz in 2003 ?
on
Intel's Big Chip
·
· Score: 1
Yes, to run with all the newest graphic models and such, you have to have a very high end system to have a good framerate.
I typically run it on an Athlon XP 1800+ with 768M DDR and a Geforce3 and in some areas it can choke.
This is correct. The way your software uses the cache(s) is what determines the performance. For instance, a naively written app may have horrible cache behavior even on large caches. Some study of the underlying algorithms and adjustment to the cache access patterns can speed up the app manyfold.
Binary compatibility is just a stinky way to be able to hide theire source.
I disagree strongly here. There are apps that were written by companies which are now defunct and also which are cornerstones of workflows of other companies who purchased that software. Having backwards compatibility is a big issue for them.
Just because there is newer/better/faster hardware is not sufficient reason to upgrade. I remember reading about a company that uses (and still uses I guess) AppleIIe computers for a control system for their company. They have custom controller boards and software that plug into those machines and still continue to do the job just fine. Unfortuantely, they have to lurk eBay a lot and other places to find replacement machines for them.
Anyway, remember the old saying: "If it ain't broke, don't fix it". There's no reason to upgrade a system that is working perfectly (and has been for years) unless you are a software writer and you need to continue your revenue flow:)
Users are already scheming a way to hack the hardware and already have ftp sites ready to dish out the software so they can pay nothing or as little as absolutely possible, even if they aren't paying the people who actually developed the port.
It is very hard to make money selling Linux software (been there, done that). No matter the price, the only question we were asked was if we had a free version, etc. Being that we were actually trying to make a living and support families with our products/skills, it just didn't add up. Couple this with the eagerness of the Linux group to pass around the software (pirate), it is just very hard to make money on Linux. Witness groups who try to create games for Linux and sell them.
Back in the 8-bit day, game piracy was so bad on some platforms that the big game makers stopped writing/porting games to them because it wasn't worth the time (it cost them money to try to support those platforms). Some of those platforms died out before their "time" (before 8-bit systems were replaced by 16-bit and 32-bit systems).
Until people can easily find good ways to make a profit on Linux, most, if not all, of the software will continue to be provided by hobbyists or other people who provide these services in their spare time. I know that I am greatful for the time/effort of these folks in providing the things I use, but many people just seem to want to know if they can get it for free. It's quite saddening.
Strange... I guess I'm so used to the "Linux is faster than Windows" posts that it surprises me to see so many posts saying that performance doesn't matter anymore when a compiler that costs money is released. Also, I haven't seen anyone mention that the benchmarks on that page show that the same codes when run on a WindowsXP box are faster than when run on a Linux box. I guess they are only relevant when Linux is in the lead.
Anyway, having used GCC for a long time, there are a couple of issues that I have gathered.
Postitive: It is wonderful to have a known compiler that works on just about everything. It really helps sysadmin install software for people to use as well as provide a common development environment. This was extremely beneficial in academia and somewhat beneficial in commercial activities.
Negative: GCC typically was ahead of the curve on support for features in the new standards. Unfortunately, this was the Turbo Pascal problem as well. Eventually you get people who have only used GCC who must use a different compiler and their code either won't compile or they have been trained to use constructs that aren't standard, basically requiring the continued use of GCC or some relearning when moving to a different platform/compiler. Yes, you can turn on strict ANSI checking but most people tend to not do that because it requires extra knowledge to know how to do that and the differences that you must watch out for and it requires more typing.
Older versions of GCC were horrid but it has improved greatly over the years (requiring a rewrite at some point iirc). For many people, just having software that runs is "good enough" (so my LaTeX document makes in 1:15 instead of 1:00, big deal... it works). However, for some people, any compiler improvements are a big deal (my simulation runs in 6 days instead of 7).
Unless your app was written with multi-processing in mind, any speedup you will see will be from the 2nd processor handling OS tasks and such (if this speedup is greater than about 10% at the extreme top end then your OS/configuration quite frankly has an incredible amount of overhead).
This is not "automagically" done. You must make specific thread creation calls and synchronization between processes/threads so that they work as expected. As another reader has posted, there has been tons of research work in automatic parallelizers and such but nothing that is ready for commercial release. It is far from trivial.
Yes, as I replied in another thread though. This machine is 2X the cost of a high-end PC with a Geforce3Ti500 card. Not only that but nVidia has used the MX designation to indicate the use of SDRAM or SGRAM on their video cards instead of DDR. So, the Geforce4MX is already running at 1/2 the memory bandwidth that it is capable. On the other hand, I'm eager to see the Geforce4 cards that come out for PCs (with DDR) =)
CUPERTINO, Calif., Jan. 28/PRNewswire-FirstCall/ -- Apple® (Nasdaq: AAPL - news) today announced the ultimate digital powerhouse for creative professionals -- the new Power Mac(TM) G4 featuring dual 1-GHz PowerPC G4 processors, the industry's first NVIDIA GeForce4 graphics card and a DVD/CD burning SuperDrive(TM), priced at just $2,999(US).
Yes. At almost 2X the cost of a high-end x86 machine with a Geforce3Ti500 card.
Spending $100 (maybe up to $1000 with all the other stuff) more to get an arguably very nice development suite isn't usually the make/break point for software (or game) developers. Even with buying a semi-high-end Windows box *and* spending the extra $$ on the development software you still weigh in less than the average Mac box.
Another thing (in response to another post) is that many Macs haven't been super game machines. Good, yes, but not super. An example here is the new iMac. It has a Geforce2MX in it. In the Windows world, this is an OK card for low-end games but it is not acceptable for many game players now, much less for newer game engines for which it is unacceptable (http://www.anandtech.com/video/showdoc.html?i=158 0). I'll be honest and don't know all the specs of the iMac (like what can be upgraded or not, if anything) but given that the iMac is introduced with "substandard" video card with respect to the game engines that are currently out and those coming soon, I wouldn't recommend the iMac as a gaming platform.
I'll be the first to admit that I don't know the reasoning behind why NeXT was used to develop those things. However, many times people use what they have available to develop their ideas at the start because, well, that's all they have.
Back in a previous incarnation, one of the projects that was going on at our place was called HECTOR. I didn't work on it but some of my friends did. It worked on a variety of UN*X flavors. Something similar to what you are talking about was that any processes that ran through it could be suspended, including file handles and sockets, and then be started on another machine (sockets only as long as all the processes connected via sockets were also running over it). It was used primarily with MPICH (but would work independently of MPICH). It was used to load balance a network of workstations and migrated processes around the cluster if node(s) became loaded beyond some threshold. To find links, search on http://www.google.com using "HECTOR RUSS" or "HECTOR ERC" (Dr. Sam Russ was the lead).
GFLOPS are useful if you are doing Photoshop or CFD codes. Even in those cases, you will not find a 100% instruction mix of only FPU instructions. The majority of instructions you run in any program will be integer based (don't forget that load/store instructions are not FPU instructions).
The only problem with the 1.6GHz G5 is that Apple will take about 2 years to get one of them into a Mac.
If you want to compare the CPUs, http://www.spec.org is the place to go. Unfortunately, I couldn't find these benchmarks run on a G4 on first glance or a quick search.
GFLOPS are useful if you are doing Photoshop or CFD codes. Even in those cases, you will not find a 100% instruction mix of only FPU instructions. The majority of instructions you run in any program will be integer based (don't forget that load/store instructions are not FPU instructions).
The only problem with the 1.6GHz G5 is that Apple will take about 2 years to get one of them into a Mac.
If you want to compare the CPUs, http://www.spec.org is the place to go. Unfortunately, I couldn't find these benchmarks run on a G4 on first glance or a quick search.
(The formatting sucks and I'm not gonna spend time fooling with it)
SPECint2000:
Company Name System Name #CPU Base Peak Full Disclosures
Advanced Micro Devic Epox 8KHA+ Motherboard, AMD Athlon (TM) XP 1900+ 1 677 701
Intel Corporation Intel D850GB motherboard (2.0 GHz, Pentium 4 pro 1 640 656 Text HTML PDF PS Config
SPECfp2000:
Company Name System Name #CPU Base Peak Full Disclosures
Advanced Micro Devic Epox 8KHA+ Motherboard, AMD Athlon (TM) XP 1900+ 1 588 634 Text HTML PDF PS Config
Intel Corporation Intel D850GB motherboard (2.0 GHz, Pentium 4 pro 1 704 714 Text HTML PDF PS Config
Re:You're going to be waiting a while...
on
New iMac Announced
·
· Score: 1
Except that the Power Architecture instruction set has more instructions in it than the x86 instruction set does...
The terms "RISC" and "CISC" are pretty much useless. Load/Store architectures are usually cleaner, but if you look at the code produced by most compilers nowdays, they stay clear of the more complex memory operand instructions. Also, the "CISC" processors are optimized for Load/Store type behavior as well, even though they have the circuitry to support the "nasty" instructions. At the heart of the Intel and AMD x86 processors is your "RISC" core.
Re:Some points to actually owning one?
on
New iMac Announced
·
· Score: 1
If you want to use the computer as an extension of the mind to visually express your thoughts and listen/record to music and view/create movies in a fun way, get the iMac.
Hmmmm... seems you have bitten deep into Apple's hype. I can do all of these things on my PC and I don't have to pretend that it is a new/revolutionary/religious experience. It's not the hardware that does all of that. It's the software and I like mine without fruit.
Ex-Apple owner and Mac fan
Slashdot Mirror
No, it was Transparent Aluminium. Star Trek IV.
Efficient how? Power? usually. Space? hardly. Typical asynchronous designs can easily require 33% to 50% more gates than clocked designs. Think about your 25M gate CPU suddenly being 37.5M gates. Then you get into all the problems that large chips have. Clockless isn't free, you still have to pay for it, just in other ways. Of course, the arguments against clocked designs is that clock drivers can take up huge amounts of real estate on the chip too. I think I recall one of the Alpha CPUs having about 10% of its die dedicated to driving the clock signal.
OK, so you can... why would you? You run the risk of your ugly hacking being non-portable, which kind of defeats the purpose of Java. Not to mention that ugly hack code is a PITA to maintain a year down the road when the person who wrote the junk isn't on the project anymore.
Maybe the reason that a project chooses a language is because there is little language neutrality?
I get far fewer religious group personel knocking on my door than unsolicited phone callers asking me to subscribe to some credit card and such. I also get fewer door callers than junk smail. If I don't want to talk to the religious canvasers, I simply tell them I'm not interested and they leave me be and neither of us is worse for the wear. I personally haven't witnessed discrimination that you describe, but to be honest I haven't watched for it either.
Yup, they understand it well. They have to make it fun enough for you to keep playing, but to make you play as long as possible. You paid $50 for the game and $70 for 7 months of playing. Looks like they did pretty good. Even an non-twinked, non-powerleveled character can get to 50 in less than 20 days of playtime, btw.
That's the difference between "sponsoring" and "hiring" someone to do something for you. Sponsors give money to the NASCAR team. In exchange they get advertising space on the car. It's like paying for a commercial to be shown on TV. However, we hire software consultants to write code for us. When they are finished with the project, it is ours, not theirs.
...is I can't tell which group has the most and bigger assholes in it... the Atheists or the Cristians. In my experience, the Atheist group is the group that is more (negatively and abrasively) vocal than the Christians.
Regardless of what you believe, the tennants of the Bible are a good guide to live by. Simply by living according to the Ten Commandments promotes a better, less stressful, more peaceful life. I personally believe that people can do what they want as long as it doesn't negatively effect others. Believe what you want. You are welcome to express those beliefs to me but if I don't hold with them, go about your way and leave me be - there is no reason to then criticise me for what I believe. Unfortunately, it seems that few other people also believe this.
As to someone earlier who spoke of mistranslations, I had a Hebrew scholar tell me before that, just like any other language, Hebrew changes over the years. Idioms and vernacular interpretations did exist in ancient Hebrew that are different than would be interpreted even a few hundred years ago. One issue we discussed was the Commandment "Thou shalt not kill." He said that the word used for "kill" in the original texts could have actually meant "murder", which can put a bit of a different meaning on that Commandment.
Anyway, I just find it mildly amusing that I see far more posts from Evolutionists flaming Cristians than the other way around.
Yes, to run with all the newest graphic models and such, you have to have a very high end system to have a good framerate. I typically run it on an Athlon XP 1800+ with 768M DDR and a Geforce3 and in some areas it can choke.
This is correct. The way your software uses the cache(s) is what determines the performance. For instance, a naively written app may have horrible cache behavior even on large caches. Some study of the underlying algorithms and adjustment to the cache access patterns can speed up the app manyfold.
Binary compatibility is just a stinky way to be able to hide theire source.
:)
I disagree strongly here. There are apps that were written by companies which are now defunct and also which are cornerstones of workflows of other companies who purchased that software. Having backwards compatibility is a big issue for them.
Just because there is newer/better/faster hardware is not sufficient reason to upgrade. I remember reading about a company that uses (and still uses I guess) AppleIIe computers for a control system for their company. They have custom controller boards and software that plug into those machines and still continue to do the job just fine. Unfortuantely, they have to lurk eBay a lot and other places to find replacement machines for them.
Anyway, remember the old saying: "If it ain't broke, don't fix it". There's no reason to upgrade a system that is working perfectly (and has been for years) unless you are a software writer and you need to continue your revenue flow
Users are already scheming a way to hack the hardware and already have ftp sites ready to dish out the software so they can pay nothing or as little as absolutely possible, even if they aren't paying the people who actually developed the port. It is very hard to make money selling Linux software (been there, done that). No matter the price, the only question we were asked was if we had a free version, etc. Being that we were actually trying to make a living and support families with our products/skills, it just didn't add up. Couple this with the eagerness of the Linux group to pass around the software (pirate), it is just very hard to make money on Linux. Witness groups who try to create games for Linux and sell them.
Back in the 8-bit day, game piracy was so bad on some platforms that the big game makers stopped writing/porting games to them because it wasn't worth the time (it cost them money to try to support those platforms). Some of those platforms died out before their "time" (before 8-bit systems were replaced by 16-bit and 32-bit systems).
Until people can easily find good ways to make a profit on Linux, most, if not all, of the software will continue to be provided by hobbyists or other people who provide these services in their spare time. I know that I am greatful for the time/effort of these folks in providing the things I use, but many people just seem to want to know if they can get it for free. It's quite saddening.
Strange... I guess I'm so used to the "Linux is faster than Windows" posts that it surprises me to see so many posts saying that performance doesn't matter anymore when a compiler that costs money is released. Also, I haven't seen anyone mention that the benchmarks on that page show that the same codes when run on a WindowsXP box are faster than when run on a Linux box. I guess they are only relevant when Linux is in the lead. Anyway, having used GCC for a long time, there are a couple of issues that I have gathered. Postitive: It is wonderful to have a known compiler that works on just about everything. It really helps sysadmin install software for people to use as well as provide a common development environment. This was extremely beneficial in academia and somewhat beneficial in commercial activities. Negative: GCC typically was ahead of the curve on support for features in the new standards. Unfortunately, this was the Turbo Pascal problem as well. Eventually you get people who have only used GCC who must use a different compiler and their code either won't compile or they have been trained to use constructs that aren't standard, basically requiring the continued use of GCC or some relearning when moving to a different platform/compiler. Yes, you can turn on strict ANSI checking but most people tend to not do that because it requires extra knowledge to know how to do that and the differences that you must watch out for and it requires more typing. Older versions of GCC were horrid but it has improved greatly over the years (requiring a rewrite at some point iirc). For many people, just having software that runs is "good enough" (so my LaTeX document makes in 1:15 instead of 1:00, big deal... it works). However, for some people, any compiler improvements are a big deal (my simulation runs in 6 days instead of 7).
Unless your app was written with multi-processing in mind, any speedup you will see will be from the 2nd processor handling OS tasks and such (if this speedup is greater than about 10% at the extreme top end then your OS/configuration quite frankly has an incredible amount of overhead). This is not "automagically" done. You must make specific thread creation calls and synchronization between processes/threads so that they work as expected. As another reader has posted, there has been tons of research work in automatic parallelizers and such but nothing that is ready for commercial release. It is far from trivial.
Yes, as I replied in another thread though. This machine is 2X the cost of a high-end PC with a Geforce3Ti500 card. Not only that but nVidia has used the MX designation to indicate the use of SDRAM or SGRAM on their video cards instead of DDR. So, the Geforce4MX is already running at 1/2 the memory bandwidth that it is capable. On the other hand, I'm eager to see the Geforce4 cards that come out for PCs (with DDR) =)
CUPERTINO, Calif., Jan. 28 /PRNewswire-FirstCall/ -- Apple® (Nasdaq: AAPL - news) today announced the ultimate digital powerhouse for creative professionals -- the new Power Mac(TM) G4 featuring dual 1-GHz PowerPC G4 processors, the industry's first NVIDIA GeForce4 graphics card and a DVD/CD burning SuperDrive(TM), priced at just $2,999(US).
Yes. At almost 2X the cost of a high-end x86 machine with a Geforce3Ti500 card.
Spending $100 (maybe up to $1000 with all the other stuff) more to get an arguably very nice development suite isn't usually the make/break point for software (or game) developers. Even with buying a semi-high-end Windows box *and* spending the extra $$ on the development software you still weigh in less than the average Mac box. Another thing (in response to another post) is that many Macs haven't been super game machines. Good, yes, but not super. An example here is the new iMac. It has a Geforce2MX in it. In the Windows world, this is an OK card for low-end games but it is not acceptable for many game players now, much less for newer game engines for which it is unacceptable (http://www.anandtech.com/video/showdoc.html?i=158 0). I'll be honest and don't know all the specs of the iMac (like what can be upgraded or not, if anything) but given that the iMac is introduced with "substandard" video card with respect to the game engines that are currently out and those coming soon, I wouldn't recommend the iMac as a gaming platform.
I'll be the first to admit that I don't know the reasoning behind why NeXT was used to develop those things. However, many times people use what they have available to develop their ideas at the start because, well, that's all they have.
Back in a previous incarnation, one of the projects that was going on at our place was called HECTOR. I didn't work on it but some of my friends did. It worked on a variety of UN*X flavors. Something similar to what you are talking about was that any processes that ran through it could be suspended, including file handles and sockets, and then be started on another machine (sockets only as long as all the processes connected via sockets were also running over it). It was used primarily with MPICH (but would work independently of MPICH). It was used to load balance a network of workstations and migrated processes around the cluster if node(s) became loaded beyond some threshold. To find links, search on http://www.google.com using "HECTOR RUSS" or "HECTOR ERC" (Dr. Sam Russ was the lead).
1. EverCrack.... er, EverQuest 2. Dark Ages of Camelot 3. Anarchy Online 4. Ultima Online 5. Asheron's Call
GFLOPS are useful if you are doing Photoshop or CFD codes. Even in those cases, you will not find a 100% instruction mix of only FPU instructions. The majority of instructions you run in any program will be integer based (don't forget that load/store instructions are not FPU instructions).
The only problem with the 1.6GHz G5 is that Apple will take about 2 years to get one of them into a Mac.
If you want to compare the CPUs, http://www.spec.org is the place to go. Unfortunately, I couldn't find these benchmarks run on a G4 on first glance or a quick search.
Summary:
Athlon XP 1900+ SPECint2000 677 SPECfp2000 588
Intel P4 2.0GHz SPECint2000 640 SPECfp2000 704
GFLOPS are useful if you are doing Photoshop or CFD codes. Even in those cases, you will not find a 100% instruction mix of only FPU instructions. The majority of instructions you run in any program will be integer based (don't forget that load/store instructions are not FPU instructions). The only problem with the 1.6GHz G5 is that Apple will take about 2 years to get one of them into a Mac. If you want to compare the CPUs, http://www.spec.org is the place to go. Unfortunately, I couldn't find these benchmarks run on a G4 on first glance or a quick search. (The formatting sucks and I'm not gonna spend time fooling with it) SPECint2000: Company Name System Name #CPU Base Peak Full Disclosures Advanced Micro Devic Epox 8KHA+ Motherboard, AMD Athlon (TM) XP 1900+ 1 677 701 Intel Corporation Intel D850GB motherboard (2.0 GHz, Pentium 4 pro 1 640 656 Text HTML PDF PS Config SPECfp2000: Company Name System Name #CPU Base Peak Full Disclosures Advanced Micro Devic Epox 8KHA+ Motherboard, AMD Athlon (TM) XP 1900+ 1 588 634 Text HTML PDF PS Config Intel Corporation Intel D850GB motherboard (2.0 GHz, Pentium 4 pro 1 704 714 Text HTML PDF PS Config
Except that the Power Architecture instruction set has more instructions in it than the x86 instruction set does...
The terms "RISC" and "CISC" are pretty much useless. Load/Store architectures are usually cleaner, but if you look at the code produced by most compilers nowdays, they stay clear of the more complex memory operand instructions. Also, the "CISC" processors are optimized for Load/Store type behavior as well, even though they have the circuitry to support the "nasty" instructions. At the heart of the Intel and AMD x86 processors is your "RISC" core.
If you want to use the computer as an extension of the mind to visually express your thoughts and listen/record to music and view/create movies in a fun way, get the iMac. Hmmmm... seems you have bitten deep into Apple's hype. I can do all of these things on my PC and I don't have to pretend that it is a new/revolutionary/religious experience. It's not the hardware that does all of that. It's the software and I like mine without fruit. Ex-Apple owner and Mac fan
Heh... What GUI do you use then? Sounds like you should be using a VT100 terminal emulator to me.