This Z-80 machine had a 2MHz/4MHz switch, a pair of 2K RAM boards, and most of the interaction was via the switch panel on the front. Later, we got a keyboard, a printer, and a paper tape reader/writer. Then we got 4K RAM boards, a TV Dazzler and Microsoft 4K basic. SPACEWAR and LIFE were great fun to play!
You mention you are interested in CFD. Intel Phi processors have been known to do well here: http://www.cfd-online.com/Foru... . In that linked story, a single Intel Phi processor beats a 1024 core cluster. Moreover, Thinkmate is literally giving away Intel Phi processors: http://www.thinkmate.com/syste... . But not all workloads fit the Phi, so you really need to do some benchmarking before you buy.
We know there are true but unprovable theorems in mathematics. Gödel showed this to be a rampant property of Peano arithmetic and anything more complex than that. Perhaps QM requires us to accept a similar fate: there are true but unprovable observations in reality.
When billionaires pay thousands of feeble-minded minions to act like millions of the American mainstream, democracy can be subverted: http://sunlightfoundation.com/...
In this case, can AI as an equalizer between moderately-funded NGOs like the Sunlight Foundation and plutocrats like the Koch brothers.
The question of whether AI kills, saves, or creates jobs thus can be reconsidered in the light of "who gets to choose what it is used for?" Capitalism's extremists will always prefer to maximize return on capital, despite whatever the short-term disruptions or long-term costs may be. AI in their hands is just as bad as any other technology. Those who are more socially, community, and humanity-minded will doubtless find ways to increase the agency of the individuals and groups they care about, just as they have with other technologies.
"Thousands of Internet tourists used their computers to tap into a central computer at Cygnus Support, a software company in Mountain View, Calif., to see the "xmastree." (The name itself is a joke to cyberspace insiders, who regularly use programs with names that start with "x," as in xterm or xwindows.)
"Two programmers at Cygnus had wired a real, 7-foot Christmas tree directly to the company's internal computer network, using simple controllers that enabled people on Cygnus Support's office network to turn the decorations, bells and lights on and off without leaving their computer terminals. The 6,000 or so outsiders who peered in from the Internet could view a simple computer rendering of the tree and check a status report to see which doodads were on and which were off, but only the people on Cygnus's local network could play with the switches."
The/. summary says "The computational complexity of this task is such that the time required to solve it increases in polynomial time with the number of images in the training set and the complexity of the "learned" feature." Moore's Law is such that any polynomial time problem will be trivially solved by the exponential advances of Moore's Law. If this problem were exponential in nature, not polynomial, then quantum computing might be our only hope. But polynomial-time problems are not the sweet spot for quantum computers.
I made my first open source contributions back in 1987, and I did so not by launching a new project, but by contributing to an existing project (GNU). Over time, those contributions took on a life of their own (GNU C++). It was quite some time (after starting Cygnus) that we had any need to launch new open source projects (such as automake, configure, Deja GNU, etc.)
My recommendation for corp OSS folks is (1) figure out how to make what you need out of existing projects and do that. If/when you reach those limits, explain the new problem you are trying to solve, see if there's interest (or even an existing solution), and then work from there. But never stop contributing to the ecosystem that likely surrounds the new code you're trying to launch.
If you only ever work on your own code, people will reciprocate by only working on their own code toward you. If you work on your own code and help improve the code that lives around it, you may well find many who want to join your project, too.
Slashdot Mirror
I wasn't talking to you. I was talking to the gripper.
Just sayin'
This Z-80 machine had a 2MHz/4MHz switch, a pair of 2K RAM boards, and most of the interaction was via the switch panel on the front. Later, we got a keyboard, a printer, and a paper tape reader/writer. Then we got 4K RAM boards, a TV Dazzler and Microsoft 4K basic. SPACEWAR and LIFE were great fun to play!
Umm...what part of transparency do they not understand?
'Nuff Said!
May be I am wrong, by I will try compare results. There is some data
http://www.hector.ac.uk/cse/di... and from topic starter
Xeon Phi for 50 time steps
grid size - 90^3 - 175^3
best time - 200s - 1500 s
Hectors 4 core of AMD 2.8GHz dual-core Opteron 5 time steps
grid size - 100^3 - 200^3
time - 795s - 8800 s
Hectors 1024 core of AMD 2.8GHz dual-core Opteron 40 time steps
grid size - 200^3
time - 1490 s
So, single Xeon Phi card for OpenFOAM is compatible with 1024 core cluster (for this benchmark)
You mention you are interested in CFD. Intel Phi processors have been known to do well here: http://www.cfd-online.com/Foru... . In that linked story, a single Intel Phi processor beats a 1024 core cluster. Moreover, Thinkmate is literally giving away Intel Phi processors: http://www.thinkmate.com/syste... . But not all workloads fit the Phi, so you really need to do some benchmarking before you buy.
We know there are true but unprovable theorems in mathematics. Gödel showed this to be a rampant property of Peano arithmetic and anything more complex than that. Perhaps QM requires us to accept a similar fate: there are true but unprovable observations in reality.
To paper over a deep problem with a shallow solution.
When billionaires pay thousands of feeble-minded minions to act like millions of the American mainstream, democracy can be subverted:
http://sunlightfoundation.com/...
In this case, can AI as an equalizer between moderately-funded NGOs like the Sunlight Foundation and plutocrats like the Koch brothers.
The question of whether AI kills, saves, or creates jobs thus can be reconsidered in the light of "who gets to choose what it is used for?" Capitalism's extremists will always prefer to maximize return on capital, despite whatever the short-term disruptions or long-term costs may be. AI in their hands is just as bad as any other technology. Those who are more socially, community, and humanity-minded will doubtless find ways to increase the agency of the individuals and groups they care about, just as they have with other technologies.
Anybody remember this: http://www.nytimes.com/1993/12... ?
"Thousands of Internet tourists used their computers to tap into a central computer at Cygnus Support, a software company in Mountain View, Calif., to see the "xmastree." (The name itself is a joke to cyberspace insiders, who regularly use programs with names that start with "x," as in xterm or xwindows.)
"Two programmers at Cygnus had wired a real, 7-foot Christmas tree directly to the company's internal computer network, using simple controllers that enabled people on Cygnus Support's office network to turn the decorations, bells and lights on and off without leaving their computer terminals. The 6,000 or so outsiders who peered in from the Internet could view a simple computer rendering of the tree and check a status report to see which doodads were on and which were off, but only the people on Cygnus's local network could play with the switches."
The /. summary says "The computational complexity of this task is such that the time required to solve it increases in polynomial time with the number of images in the training set and the complexity of the "learned" feature." Moore's Law is such that any polynomial time problem will be trivially solved by the exponential advances of Moore's Law. If this problem were exponential in nature, not polynomial, then quantum computing might be our only hope. But polynomial-time problems are not the sweet spot for quantum computers.
^ Good one! Is this better (when read with a thick Russian accent): In Soviet USA, history is subject of oligarchs.
I made my first open source contributions back in 1987, and I did so not by launching a new project, but by contributing to an existing project (GNU). Over time, those contributions took on a life of their own (GNU C++). It was quite some time (after starting Cygnus) that we had any need to launch new open source projects (such as automake, configure, Deja GNU, etc.) My recommendation for corp OSS folks is (1) figure out how to make what you need out of existing projects and do that. If/when you reach those limits, explain the new problem you are trying to solve, see if there's interest (or even an existing solution), and then work from there. But never stop contributing to the ecosystem that likely surrounds the new code you're trying to launch. If you only ever work on your own code, people will reciprocate by only working on their own code toward you. If you work on your own code and help improve the code that lives around it, you may well find many who want to join your project, too.