The site that provides more concrete data, particularly about their test methodology. After all, "In god we trust. All others must bring data."
ExtremeTech chose not to test at 2560x1600, despite using a beta driver that significantly alters the resolution scaling performance of the NVidia cards. This means their numbers and thus their conclusions are worthless for anybody who wants to us the higher resolution. On the other hand, they didn't test at 1680x1050 either, so their tests didn't reveal the significant advantage the ATI card has for people with smaller monitors.
HotHardware's review, linked above, did test at both 2560x1600 and 1920x1200, but also neglected 1680x1050. Overall, their testing seemed to be more thorough, but I stopped reading when I noticed that the labels on their first graphs didn't even come close to matching the explanation in the adjacent paragraph. I don't particularly trust numbers when I can't even tell what they're representing.
Anandtech's review seems to test the two new cards in more detail and offers more explanation of how the architectural differences will affect performance. However, they chose to only compare the two cards at that price point, eschewing the context of the more and less expensive products. This makes it harder for the consumer to figure out if it's worth it to spend $250 for a graphics card, but it seems clear to me that Anandtech provides the most reliable data about the two cards, and is certainly more concerned with informing the consumer than deciding for them by proclaiming an all-around winner when there isn't one.
nVidia is the graphics card vendor that hasn't abandoned their proprietary GPGPU framework in favor of the open standard yet. Sure, they are including support for OpenCL, but that strikes me as very similar to the way Microsoft makes it possible to use OpenGL, but does everything they can to get people to use DirectX.
nVidia is still trying get CUDA into a dominant position in the market, but they don't seem to realize that CUDA's position is much less tenable than Direct3D's position vs. OpenGL. nVidia isn't that far ahead of AMD when it comes to GPGPU performance, and OpenCL has the support of AMD, Apple, and Intel. CUDA is also soon to be facing two well funded competing APIs: OpenCL and DirectX11 Compute Shaders. To me, it looks like the only thing CUDA has going for it is an early-bird advantage that will be irrelevant in a matter of months.
Also, there's a difference between having myriad GPGPU apis and a bunch of physics engines implemented on top of them. It's good for the market to have only one or two low-level (but cross-platform) gpgpu apis, but there will always be a need to have a wide variety of middleware physics engines, because games and other simulations can have very different needs.
Most of the large computer labs at my university now have at least a few drop-in stations, some with large plasma screens to facilitate group work (or play).
Your university needs to stop being so cheap with it's MATLAB licenses. There's no reason why you should be running MATLAB only on servers when The MathWorks is quite willing to license MATLAB on terms that allow it to be installed on any student's laptop, with the only server necessary being the license server.
Mechanical engineers need to graduate with hands-on experience with a professional CAD package. Since these are far too expensive for students to buy, and there are no open-source alternatives, universities need to buy the software. When a university is buying CAD software, it makes sense to only buy one package, rather than waste money on several.
My university already has an established print quota system to charge students by the page for printing. A few years ago, they extended this system with a web interface that allows you to upload and PDF, PostScript, or plain text document to any printer on campus. This makes printing really easy, particularly for Mac users, who don't have to set up third-party software in order to create PDFs.
Don't blame Mozilla. Blame Apple for not retrofitting 10.3 with Universal Binary support. That way, you can be right, and also make the unreasonableness of your request apparent.
Once you pass the barrier where you're swapping, you fire up another instance of whatever code it is that you're running. (In my case, finite element method simulations that need at least 4GB for minimal precision.)
ext3 was merged to the mainline kernel in 2001. Git was created in 2005. I wouldn't trust any authorship evidence in a git repo for code predating the repo.
The journalling behavior of ext3 was probably decided by Stephen Tweedie
The response was a bit egotistical - I was homeschooled for several years by an atheist, but almost all the other homeschoolers I know do it for religious reasons. Sadly, many of them start going to public schools for high school when their parents realize they can't provide that kind of education (particularly with all the younger siblings to worry about). Those first few months in public school are usually pretty rough.
Homeschooling only fucks up kids when the parents are overprotective religious fundamentalists who want to protect their kids from the sinful real world. They're the most common kind of homeschoolers. But among those who are homeschooling for more rational reasons, the kids usually turn out to be at least as well-adjusted as those in public schools, particularly when it comes to dealing with adults as equals.
The above does nothing to reduce the need for good public schooling to be available to everybody.
Zero tolerance means more along the lines of "zero liability" and "zero responsibility". There will be zero common sense with or without "zero tolerance" policies.
By default, Tomato doesn't allow remote (from WAN port) administration. I don't know about the other WRT firmwares, but Tomato at least is secure from this exploit by default.
Good point. There's no excuse for application code breaking in a multiprocessor environment.
By the way, the piece of code I had in mind with my above comment is a simulation that's been developed, optimized, and debugged as a single threaded app over the past few years, with several million dollars poured into the project. The code may seem inefficient by being single threaded, but in this application, getting the right answer extremely important. With the economy the way it is, we can't afford to do the debugging that would be necessary to validate a multithreaded version. (That is, unless we get some of the stimulus money, which is still up in the air.) We're far from the only company in a situation like this.
It isn't. The whole point of TRIM is to erase a block before you're waiting to write something to it, ie. before your disk is full and you need to reuse the space. The garbage collection on a TI calculator is really defragmentation, to eliminate gaps between files. This is necessary because the calculators have the flash attached to the address bus, rather than behind a hard drive controller, and there's no MMU to give programmers a linear address space if there flash apps were to be discontiguous in memory. Thus, in order to load an app in to the calculator flash, there must be a contiguous region of available flash.
In a way, it is a physical problem. With hard drives, seek times are slow compared to read and write speeds, and reading and writing happen at the same speed. With flash, seek times are negligible, and read speeds are much faster than writes. It's that inequality of read and write speeds that messes up so many of the assumptions our current software makes, and leads to behavior that users find disturbing.
Apple can solve the storage issue by updating the Time Capsule software to make it a more generic NAS. Filesystem access is a non-issue, as the iPhone OS is already a full unix system. The only reasons access is currently restricted are security (sandboxing apps and all their data) and because it would overly complicate the user interface for such a small device.
If CPU vendors make good on their promises of low-power x86 chips, and if mobile graphics chips continue with their current trend, we could see a handheld device round out the low-end portions of the price range that the Mac Mini can't currently reach.
Slashdot Mirror
The site that provides more concrete data, particularly about their test methodology. After all, "In god we trust. All others must bring data."
ExtremeTech chose not to test at 2560x1600, despite using a beta driver that significantly alters the resolution scaling performance of the NVidia cards. This means their numbers and thus their conclusions are worthless for anybody who wants to us the higher resolution. On the other hand, they didn't test at 1680x1050 either, so their tests didn't reveal the significant advantage the ATI card has for people with smaller monitors.
HotHardware's review, linked above, did test at both 2560x1600 and 1920x1200, but also neglected 1680x1050. Overall, their testing seemed to be more thorough, but I stopped reading when I noticed that the labels on their first graphs didn't even come close to matching the explanation in the adjacent paragraph. I don't particularly trust numbers when I can't even tell what they're representing.
Anandtech's review seems to test the two new cards in more detail and offers more explanation of how the architectural differences will affect performance. However, they chose to only compare the two cards at that price point, eschewing the context of the more and less expensive products. This makes it harder for the consumer to figure out if it's worth it to spend $250 for a graphics card, but it seems clear to me that Anandtech provides the most reliable data about the two cards, and is certainly more concerned with informing the consumer than deciding for them by proclaiming an all-around winner when there isn't one.
Or he's at one of the schools (mine included) where some of the computer labs have xbox 360s.
That's because there are no OpenCL implementations, and the spec was only released last December.
Cool. I hadn't checked the status of JSR 292 in quite a while. It's good to see that it's actually being implemented, even if it is still a ways away.
The JVM architecture is still too restrictive to cleanly run truly dynamic languages, so the fact that it is fast isn't really useful.
You're quite right. The OpenGL ARB makes Microsoft's job far too easy.
nVidia is the graphics card vendor that hasn't abandoned their proprietary GPGPU framework in favor of the open standard yet. Sure, they are including support for OpenCL, but that strikes me as very similar to the way Microsoft makes it possible to use OpenGL, but does everything they can to get people to use DirectX.
nVidia is still trying get CUDA into a dominant position in the market, but they don't seem to realize that CUDA's position is much less tenable than Direct3D's position vs. OpenGL. nVidia isn't that far ahead of AMD when it comes to GPGPU performance, and OpenCL has the support of AMD, Apple, and Intel. CUDA is also soon to be facing two well funded competing APIs: OpenCL and DirectX11 Compute Shaders. To me, it looks like the only thing CUDA has going for it is an early-bird advantage that will be irrelevant in a matter of months.
Also, there's a difference between having myriad GPGPU apis and a bunch of physics engines implemented on top of them. It's good for the market to have only one or two low-level (but cross-platform) gpgpu apis, but there will always be a need to have a wide variety of middleware physics engines, because games and other simulations can have very different needs.
Most of the large computer labs at my university now have at least a few drop-in stations, some with large plasma screens to facilitate group work (or play).
Your university needs to stop being so cheap with it's MATLAB licenses. There's no reason why you should be running MATLAB only on servers when The MathWorks is quite willing to license MATLAB on terms that allow it to be installed on any student's laptop, with the only server necessary being the license server.
Mechanical engineers need to graduate with hands-on experience with a professional CAD package. Since these are far too expensive for students to buy, and there are no open-source alternatives, universities need to buy the software. When a university is buying CAD software, it makes sense to only buy one package, rather than waste money on several.
My university already has an established print quota system to charge students by the page for printing. A few years ago, they extended this system with a web interface that allows you to upload and PDF, PostScript, or plain text document to any printer on campus. This makes printing really easy, particularly for Mac users, who don't have to set up third-party software in order to create PDFs.
Don't blame Mozilla. Blame Apple for not retrofitting 10.3 with Universal Binary support. That way, you can be right, and also make the unreasonableness of your request apparent.
Once you pass the barrier where you're swapping, you fire up another instance of whatever code it is that you're running. (In my case, finite element method simulations that need at least 4GB for minimal precision.)
So, in other words, this won't dissuade the RIAA one whit, but it might make it hard for them to find lawyers willing to work for them.
ext3 was merged to the mainline kernel in 2001. Git was created in 2005. I wouldn't trust any authorship evidence in a git repo for code predating the repo.
The journalling behavior of ext3 was probably decided by Stephen Tweedie
The response was a bit egotistical - I was homeschooled for several years by an atheist, but almost all the other homeschoolers I know do it for religious reasons. Sadly, many of them start going to public schools for high school when their parents realize they can't provide that kind of education (particularly with all the younger siblings to worry about). Those first few months in public school are usually pretty rough.
Homeschooling only fucks up kids when the parents are overprotective religious fundamentalists who want to protect their kids from the sinful real world. They're the most common kind of homeschoolers. But among those who are homeschooling for more rational reasons, the kids usually turn out to be at least as well-adjusted as those in public schools, particularly when it comes to dealing with adults as equals.
The above does nothing to reduce the need for good public schooling to be available to everybody.
Zero tolerance means more along the lines of "zero liability" and "zero responsibility". There will be zero common sense with or without "zero tolerance" policies.
But as long as we have them, school teachers shouldn't get special treatment with respect to getting put on them.
By default, Tomato doesn't allow remote (from WAN port) administration. I don't know about the other WRT firmwares, but Tomato at least is secure from this exploit by default.
Good point. There's no excuse for application code breaking in a multiprocessor environment.
By the way, the piece of code I had in mind with my above comment is a simulation that's been developed, optimized, and debugged as a single threaded app over the past few years, with several million dollars poured into the project. The code may seem inefficient by being single threaded, but in this application, getting the right answer extremely important. With the economy the way it is, we can't afford to do the debugging that would be necessary to validate a multithreaded version. (That is, unless we get some of the stimulus money, which is still up in the air.) We're far from the only company in a situation like this.
Buying a faster chip is a lot cheaper and faster than rewriting something to be multithreaded.
It isn't. The whole point of TRIM is to erase a block before you're waiting to write something to it, ie. before your disk is full and you need to reuse the space. The garbage collection on a TI calculator is really defragmentation, to eliminate gaps between files. This is necessary because the calculators have the flash attached to the address bus, rather than behind a hard drive controller, and there's no MMU to give programmers a linear address space if there flash apps were to be discontiguous in memory. Thus, in order to load an app in to the calculator flash, there must be a contiguous region of available flash.
In a way, it is a physical problem. With hard drives, seek times are slow compared to read and write speeds, and reading and writing happen at the same speed. With flash, seek times are negligible, and read speeds are much faster than writes. It's that inequality of read and write speeds that messes up so many of the assumptions our current software makes, and leads to behavior that users find disturbing.
Apple can solve the storage issue by updating the Time Capsule software to make it a more generic NAS. Filesystem access is a non-issue, as the iPhone OS is already a full unix system. The only reasons access is currently restricted are security (sandboxing apps and all their data) and because it would overly complicate the user interface for such a small device.
If CPU vendors make good on their promises of low-power x86 chips, and if mobile graphics chips continue with their current trend, we could see a handheld device round out the low-end portions of the price range that the Mac Mini can't currently reach.