I don't know where TFA got the "globally addressable PB". I think someone was misquoted.
I can't find any mention of it in the NCSA webpages, and no shared memory system exists on this level, ccNUMA or otherwise (NASA Ames has a 4TB altix system, which is evidently the largest in the world that is publicly acknowledged).
Software distributed shared memory hasn't really gone anywhere either, so I think someone was fantasizing when they wrote the article... globally accessible filesystems, sure, but shared memory is something else altogether...
The GPU architecture has been progressively moving to a more "general" system with every generation. Originally the processing elements in the GPU could only write to one memory location, now the hardware supports scattered writes, for example.
As such I think the GPGPU method of casting algorithms into the GPU APIs (CUDA et. al) are going to die a quick death once Larabee comes out and people can simply run their threaded codes on these finely-grained co-processors.
Quantum computing is the new string-theory, ie. a theoretical physics quagmire. It's soaking up funding and diverting graduate student talent that could be better utilized in other areas.
We've been there, and picked up enough rocks to last a while. What else is there to do...?
Until we can build largely self-sustaining colonies and prove them on earth the fuel and resources would be better spent launching probes, satellites, telescopes, etc. - not sending people on moon vacations.
First thing is to upgrade the version of SA you're using, then configure it better (install good rule-sets), train bayes, in that order.
I have accounts on servers who have different policies/versions, and have experienced no (important!) false positives on one and had to whitelist on the other.
Convincing the people sending SPAM-ish looking mail to do otherwise could also help, rather than just accepting it:
Get thee to your local community radio station! Often this is at the university or community college. Also look for dj groups/clubs that may exist. These are usually the best places to find easy to approach people who will readily share their experiences.
Depending on what kind of media you are planning on investing in, your equipment costs will vary. If you are mixing tunes for your own consumption then you would do well with just software on your own pc. Being able to pitch-shift is the only necessity in this case, you can compose the beat-matched tracks in audacity.
If you want to play at the local bar/dancehall or whatever, cd/mp3 mixers will work just as well as vinyl. A lot of dj's went this route, especially in genres like trance and house. There are even hard-drive turntables now. If you want to start delving into turntablism, then you might want to buy a pair a pair of decent turntables and a mixer.. MP3 + turntable combos like final scratch (etc) are definitely cheaper than vinyl in the long run, most 12" singles are ~$15 CDN. You could also do live PA, adding your computer and/or other audio gear into the mix. The possibilities are endless..
If you are going to buy gear, either buy really cheap used gear or buy expensive brand new gear. If you are wary of dropping the cash, then try renting or meeting someone who will teach you first, and see if there is a place where you can get consistant access. Basic consumer instinct is to buy cheap and then upgrade, but if you drop $2000 CDN on a decent set of decks you will be able to resell it at 80-90% of cost. Anything less than this or packaged as a starter kit will be difficult to sell and will probably only be bought in used condition at 20-30% of cost.
I've found the ability to split channels in my headphones really helpful. Mathematically it's just making sure the waves have the same wavelength/amplitude and are in phase.
NAMD doesn't scale well, so unless you have a low latency network fabric you won't even come close to cashing in with more than 2^5 nodes on gigE. It's not really beowulf friendly.
Their Origin 2000 numbers show that the latest release had no speedup after 126 processors. That is an SMP...
I'm surprised no one has brought up the implications this has with respect to theories like Panspermia.
A virus could withstand extrasolar transit much better than even the most hardened extremophile bacteria.
Easy on the xenophopic slant (your last sentence doesn't help much).
I think climate change will soon eclipse the virus/bacteria/plague of the month/decade/century in terms of the negative effect they have on humanity. Regardless, how a government prioritizes has little to do with the actual desires of the populace in a nation state. Politics are so depressing.
The proper question to ask is:
Why are humans wasting so much effort launching actors into space when we could have accelerated automated programs that actually expand our understanding of the rest of the universe?
I'm surpised that there hasn't been more discussion about arXiv in these comments. At our institute (astrophysics) most people send their pre-prints to astro-ph before the journal version is published, and NASA ADS http://adswww.harvard.edu/ is the first place to go when looking for something. There are also frequent group meetings to discuss recent submissions to astro-ph, making it more talked about than any particular journal.
Personally I feel that research which is not made publicly available only helps re-inforce the white tower image of scientists as self-serving. The fact that funding is directly related to citations has firmly entrenched journals in the run of things. Of course they are going to make vague claims about how science will suffer without their editorial control, but maybe if money was not an issue things wouldn't be this way.
At the end of the day everything is only a collective hunch, and if you are trusting an editor to determine what you will read I don't think you will be less aware of the major issues.
Informed by which metric? Shouldn't the government radio address be all I need? Newspapers only print a particular set of wires, and they primarily exist to make advertising money. This is not concordant with my interests, so for me they are less informative.
I had an inane Toronto Star telemarketer yell at me once. I told her I only read online, she claims I'm half-informed and asks me if I'm proud of being ignorant. She must have been psychic.
I'm not sure. The draw for these shop owners is a certain crowd who aren't keen with having teenage h4XX0rs skipping class to take up seats in their poo-poo atmosphere. The majority of their market doesn't show up when a minority of others is there, in terms of profit.
Seems like they need to restrict or make it uncomfortable for campers. Get out in 20 minutes or ex-biker muffin-maker in the back comes out front to ask you to leave real nice-like. That sort of thing is standard at coffee shops in Northern Ontario.
Of course in todays sanitary and non-confrontational profit maximization society this can be implemented through some sort of MAC address based accounting system so that people can avoid any personal interaction and drain their wallets like mega while making a set piece for the passerby.
Which still doesn't solve the initial problem. They need to find out where the cash cows go when they aren't attending My_Little_Coffee_Shop(TM)... if it is to WIFI-less places they will just lose business unless they take it down, at least in the current market.
How much more mac does there need to be in the world? Back in the good ol' days we used to call this sort of thing archiving (! The wiki page doesn't even contain the word, so I must really be extinct by now). Pretty soon we'll just get pod-ears (tm) and then we will never have to listen to another moment of unscheduled audio... EVER!
Streamripper is a great way to archive your own content if your broadcaster of choice can't afford to serve what you want on demand.
Bonus points for the Napolean Dynamite reference!! You gonna eat those tots?
Peak is determined by having every floating point unit active during every clock. It is the theoretical maximum number of operations that you could ever possibly do. No program (other than making heat in a tight loop?) ever comes close to this.
Sustained is arbitrary. Basically you run a program that is optimized to avoid un-necessary cache misses and that produces some meaningful result (like the matrix multiply in linpack). Real world apps don't even come close to this level of effeciency, often having to wait for data to be sent from RAM or disk, which slows down the system even further (stalls the pipeline).
Fortran == Formula Translation, not general purpose bloated compiler. It's meant to convert mathematical operations to machine code as efficiently as possible. It just gives us a portable way to do more abacus/slide rule/calculator operations than assembly.
You can put a new marketing spin on anything. The point of the matter is that Fortran has evolved to suit coding mathematical systems, not to design applications. The mythical CS land where all languages are fully optimized and portable doesn't exist, at least yet anyway. If that were so people would just use symbolic math programs!
As far as syntax, I don't see how C is cleaner than Fortran. When you solve mathematical problems in physics, you always present it as a block structured set of equations. At that point, it's just a balance between how many keystrokes I have to make and how long it takes to compile/run. So why is something like exp(x,2) better than x**2?
Finally, I point to the following:
Programming languages should be designed not by piling feature on top of feature, but by removing the weaknesses and restrictions that make additional features appear necessary. (from Revised Report on Scheme, 1991)
We don't need fortran compilers that make for easier debugging, we need bug free fortran compilers. But that is a different story...
It should have never progressed past the national stucture. Now each echelon of the government and the associated financial community will balk at opening up the top level until it is economically viable, ie, in their best interest.
National Research Council of Canada: NRC is composed of over 20 institutes and national programs, spanning a wide variety of disciplines and offering a broad array of services. We are located in every province in Canada and play a major role in stimulating community-based innovation.
NRC institutes and programs are organized into three (3) key areas:
* Physical Sciences and Engineering
* Life Sciences and Information Technology
* Technology and Industry Support
Vectorization (SIMD) is built into the Intel compiler. There is no need to hack in assembly as the compiler will do it for you. This is the case with most vendor supplied compilers, as they want to fully exploit their hardware functionality.
The problem is bringing this functionality to OS compilers, which as far as I know, there is not even an OpenMP (threading) implementation, let alone internal vectorization.
Slashdot Mirror
I don't know where TFA got the "globally addressable PB". I think someone was misquoted.
I can't find any mention of it in the NCSA webpages, and no shared memory system exists on this level, ccNUMA or otherwise (NASA Ames has a 4TB altix system, which is evidently the largest in the world that is publicly acknowledged).
Software distributed shared memory hasn't really gone anywhere either, so I think someone was fantasizing when they wrote the article... globally accessible filesystems, sure, but shared memory is something else altogether...
The GPU architecture has been progressively moving to a more "general" system with every generation. Originally the processing elements in the GPU could only write to one memory location, now the hardware supports scattered writes, for example.
As such I think the GPGPU method of casting algorithms into the GPU APIs (CUDA et. al) are going to die a quick death once Larabee comes out and people can simply run their threaded codes on these finely-grained co-processors.
Good luck with that - it's all hype.
Quantum computing is the new string-theory, ie. a theoretical physics quagmire. It's soaking up funding and diverting graduate student talent that could be better utilized in other areas.
We've been there, and picked up enough rocks to last a while. What else is there to do...?
Until we can build largely self-sustaining colonies and prove them on earth the fuel and resources would be better spent launching probes, satellites, telescopes, etc. - not sending people on moon vacations.
Everyone loves Eric Raymond
Classic.
First thing is to upgrade the version of SA you're using, then configure it better (install good rule-sets), train bayes, in that order.
I have accounts on servers who have different policies/versions, and have experienced no (important!) false positives on one and had to whitelist on the other.
Convincing the people sending SPAM-ish looking mail to do otherwise could also help, rather than just accepting it:
http://wiki.apache.org/spamassassin/AvoidingFpsForSenders
It's pretty easy to get false positives depending on how you configure SpamAssasin.
Same basic idea.
This is basically what hoyle described as the next big step in nuclear energy production. in 1980.
Commonsense in Nuclear Energy: http://books.google.com/books?as_isbn=0435544322
Get thee to your local community radio station! Often this is at the university or community college. Also look for dj groups/clubs that may exist. These are usually the best places to find easy to approach people who will readily share their experiences.
Depending on what kind of media you are planning on investing in, your equipment costs will vary. If you are mixing tunes for your own consumption then you would do well with just software on your own pc. Being able to pitch-shift is the only necessity in this case, you can compose the beat-matched tracks in audacity.
If you want to play at the local bar/dancehall or whatever, cd/mp3 mixers will work just as well as vinyl. A lot of dj's went this route, especially in genres like trance and house. There are even hard-drive turntables now. If you want to start delving into turntablism, then you might want to buy a pair a pair of decent turntables and a mixer.. MP3 + turntable combos like final scratch (etc) are definitely cheaper than vinyl in the long run, most 12" singles are ~$15 CDN. You could also do live PA, adding your computer and/or other audio gear into the mix. The possibilities are endless..
If you are going to buy gear, either buy really cheap used gear or buy expensive brand new gear. If you are wary of dropping the cash, then try renting or meeting someone who will teach you first, and see if there is a place where you can get consistant access. Basic consumer instinct is to buy cheap and then upgrade, but if you drop $2000 CDN on a decent set of decks you will be able to resell it at 80-90% of cost. Anything less than this or packaged as a starter kit will be difficult to sell and will probably only be bought in used condition at 20-30% of cost.
I've found the ability to split channels in my headphones really helpful. Mathematically it's just making sure the waves have the same wavelength/amplitude and are in phase.
Don't get me wrong, I know very little about MD. Why doesn't NAMD scale near-linearly?
NAMD doesn't scale well, so unless you have a low latency network fabric you won't even come close to cashing in with more than 2^5 nodes on gigE. It's not really beowulf friendly.
Their Origin 2000 numbers show that the latest release had no speedup after 126 processors. That is an SMP...
I'm surprised no one has brought up the implications this has with respect to theories like Panspermia.
A virus could withstand extrasolar transit much better than even the most hardened extremophile bacteria.
In an academic setting (Canada) I've yet to encounter this sort of rationale.
I guess the best anyone can hope for is to enter a profession where merit is the prominent factor in determining whether or not you get the job.
Easy on the xenophopic slant (your last sentence doesn't help much).
I think climate change will soon eclipse the virus/bacteria/plague of the month/decade/century in terms of the negative effect they have on humanity. Regardless, how a government prioritizes has little to do with the actual desires of the populace in a nation state. Politics are so depressing.
The proper question to ask is:
Why are humans wasting so much effort launching actors into space when we could have accelerated automated programs that actually expand our understanding of the rest of the universe?
I'm surpised that there hasn't been more discussion about arXiv in these comments. At our institute (astrophysics) most people send their pre-prints to astro-ph before the journal version is published, and NASA ADS http://adswww.harvard.edu/ is the first place to go when looking for something. There are also frequent group meetings to discuss recent submissions to astro-ph, making it more talked about than any particular journal.
Personally I feel that research which is not made publicly available only helps re-inforce the white tower image of scientists as self-serving. The fact that funding is directly related to citations has firmly entrenched journals in the run of things. Of course they are going to make vague claims about how science will suffer without their editorial control, but maybe if money was not an issue things wouldn't be this way.
http://arxiv.org/blurb/pg02pr.html is a pretty insightful consideration of how peer review systems can be made more efficient.
At the end of the day everything is only a collective hunch, and if you are trusting an editor to determine what you will read I don't think you will be less aware of the major issues.
Informed by which metric? Shouldn't the government radio address be all I need?
Newspapers only print a particular set of wires, and they primarily exist to make advertising money. This is not concordant with my interests, so for me they are less informative.
I had an inane Toronto Star telemarketer yell at me once. I told her I only read online, she claims I'm half-informed and asks me if I'm proud of being ignorant. She must have been psychic.
I'm not sure. The draw for these shop owners is a certain crowd who aren't keen with having teenage h4XX0rs skipping class to take up seats in their poo-poo atmosphere. The majority of their market doesn't show up when a minority of others is there, in terms of profit.
Seems like they need to restrict or make it uncomfortable for campers. Get out in 20 minutes or ex-biker muffin-maker in the back comes out front to ask you to leave real nice-like. That sort of thing is standard at coffee shops in Northern Ontario.
Of course in todays sanitary and non-confrontational profit maximization society this can be implemented through some sort of MAC address based accounting system so that people can avoid any personal interaction and drain their wallets like mega while making a set piece for the passerby.
Which still doesn't solve the initial problem. They need to find out where the cash cows go when they aren't attending My_Little_Coffee_Shop(TM)... if it is to WIFI-less places they will just lose business unless they take it down, at least in the current market.
How much more mac does there need to be in the world? Back in the good ol' days we used to call this sort of thing archiving (! The wiki page doesn't even contain the word, so I must really be extinct by now). Pretty soon we'll just get pod-ears (tm) and then we will never have to listen to another moment of unscheduled audio... EVER!
Streamripper is a great way to archive your own content if your broadcaster of choice can't afford to serve what you want on demand.
Bonus points for the Napolean Dynamite reference!!
You gonna eat those tots?
Peak is determined by having every floating point unit active during every clock. It is the theoretical maximum number of operations that you could ever possibly do. No program (other than making heat in a tight loop?) ever comes close to this.
Sustained is arbitrary. Basically you run a program that is optimized to avoid un-necessary cache misses and that produces some meaningful result (like the matrix multiply in linpack). Real world apps don't even come close to this level of effeciency, often having to wait for data to be sent from RAM or disk, which slows down the system even further (stalls the pipeline).
What are Fortran's advantages?
Use the best tool for the job?
Fortran == Formula Translation, not general purpose bloated compiler. It's meant to convert mathematical operations to machine code as efficiently as possible. It just gives us a portable way to do more abacus/slide rule/calculator operations than assembly.
You can put a new marketing spin on anything. The point of the matter is that Fortran has evolved to suit coding mathematical systems, not to design applications. The mythical CS land where all languages are fully optimized and portable doesn't exist, at least yet anyway. If that were so people would just use symbolic math programs!
As far as syntax, I don't see how C is cleaner than Fortran. When you solve mathematical problems in physics, you always present it as a block structured set of equations. At that point, it's just a balance between how many keystrokes I have to make and how long it takes to compile/run. So why is something like exp(x,2) better than x**2?
Finally, I point to the following:
Programming languages should be designed not by piling feature on top of feature, but by removing the weaknesses and restrictions that make additional features appear necessary.
(from Revised Report on Scheme, 1991)
We don't need fortran compilers that make for easier debugging, we need bug free fortran compilers. But that is a different story...
It should have never progressed past the national stucture. Now each echelon of the government and the associated financial community will balk at opening up the top level until it is economically viable, ie, in their best interest.
It's probably because the majority of Canadians would prefer not to be fed figures by the CIA of the USA chez google.com
National Research Council of Canada:
NRC is composed of over 20 institutes and national programs, spanning a wide variety of disciplines and offering a broad array of services. We are located in every province in Canada and play a major role in stimulating community-based innovation.
NRC institutes and programs are organized into three (3) key areas:
* Physical Sciences and Engineering
* Life Sciences and Information Technology
* Technology and Industry Support
Vectorization (SIMD) is built into the Intel compiler. There is no need to hack in assembly as the compiler will do it for you. This is the case with most vendor supplied compilers, as they want to fully exploit their hardware functionality.
The problem is bringing this functionality to OS compilers, which as far as I know, there is not even an OpenMP (threading) implementation, let alone internal vectorization.