Well, I'm here at tech, and am still waiting for word on whether or not I've lost any friends today. One roommate of mine may have already lost one, and some friends have lost friends of theirs.
No guns are allowed on campus. We have a few full-time shrinks on campus. The engineering student was probably an american (most of our foreign students are grad students). So far the word is the kid was a senior with 3 engineering majors. We have lots of bible groups, etc on campus, but we're not really known for being very religious or very secular for that matter.
My guess would be stress. I've seen grown men cry over single assignments, several of them, over the years here @ VT. The engineering kids are pushed really hard, and many of them don't deal with it very well. 60 hours a week of real work are pretty normal, with classes that everage 27-50%, which are only curved at the end (and nobody knows the curve till then). Try that for 4 years while growing up... Many engineering students I know end up having fairly empty shells of personalities, as their entire lives so far have circled around work and thinly veiled attempts at having a life on the side.
3 engineering majors at once would break most people. Guaranteed.
So far, the big questions are: 1. Why didn't the students find out about the 7:15am shooting until 2 hours later 2. Why was only the dorm closed?
To be fair, we had two bomb threats (no bombs) over the last 2 weeks, the last one only 3 days ago. So maybe the administration was getting tired of interrupting school for non-issues.
Access Linux is the Cobalt APIs over a Linux Kernel. PalmSource did that for better hardware compatibility and easier driver writing for their licensees.
For me, seeing these things in person (I grew up in DC, so I ended up seeing lots of these things) made them more real for me. Reading about the Constitution from a textbook makes it seem like a bit of an abstract concept. But seeing the thing right in front of me reminds me of how it's a paper document that people wrote by hand.
I agree with the GP that a balance has to be kept. In this case, if it's so far gone as to be a fruit stand, it's probably best torn down and replaced with a new research lab.
Well, there's Just In Time compilation, which is part of the JRE we all use today. It runs a lightweight profiler and compiles the most-used section ("hotspots") to native machine code.
I think CW's buying the initial land in order to get the ball rolling on getting themselves involved in the transactional business of real estate on SL. Long after the current stuff is sold off, they want to be the agents you buy virtual real estate from later.
Atomic I/O generally (I'm from the posix side of things, so maybe there are different definitions on win32) means that the entire operation's done by the kernel at once. E.g. I write 16 bytes at offset 0 to this file, and another app's writing 16 bytes at offset 0 to this file. When we're both done writing with a single write() call, the file's got 16 bytes from one of us, not some from one and some from another. No amount of userspace I/O calls will guarantee this (outside some godawful locking, which essentially forces the same thing, only with more context switches and a lot more userland complexity).
In terms of filesystem changes, all the files are created (possibly overwriting other files) or, if one fails, they all undo. Same for registry changes. Them all running together to completion or not at all sounds great. I definitely want registry changes to be atomic, b/c there's a good chance other apps are mucking around with it too. Also, I want those changes to be consistent -- with an atomic change that includes both the registry and a filesystem change, I can make sure I don't change a file with the registry pointed to the old value, or visa versa. In userland I could roll back, but I couldn't stop that time period when it's inconsistent.
They also involve atomic I/O to multiple systems simultaneously. Userland can't do this. Databases work on one system, their own data files, and have full control over these files.
Userland apps don't have that kind of control over the registry. Hell they may not be sure to have that kind of control over the files they're manipulating.
Besides, I'd rather have this code once in a DLL than 10 times in 10 different apps. That's real bloat.
The Cell's pretty straightforward with the tools given. You use a pthread-like library to give them code to run, and all your global variables are valid I/O targets for DMA. See this for an introduction: http://gamasutra.com/features/20060721/chow_01.sht ml
Frankly, the programming's one of the reasons I bought a PS3. It just looks fun.
As for what to run on each one, some things come to mind: physics, collision, particle effects, AI, etc. Some of them can be split up across multiple SPEs. Run them on a work queue of objects, or assign each a calculable subset, (e.g. SPE1 gets all the objects with an even ID #, SPE2 gets all the odd ones).
Some people really hate getting multiple processors to work together on a problem. But, these people have a lot of problems coming their way, as more & more machines go that way. Hell the laptop I'm on right now is dual-core, and I doubt I'm alone in that fact.
No, if we do that, then we'd have a lot more Bushes elected to the white house.
Society needs an educated populace. The thing is people forget that 4 years isn't much time to learn enough for the next 50.
The current system lets people go to grad school, which is heavy thinking, when they want more. At 18-21, there's only so much thinking they're gonna do. It's also probably the only time they're intellectually green enough to have the patience for all that training (later on, people need to be sold on its necessity a lot more before investing in that kind of effort).
Economics. The most popular CPU architecture gets all $$ for expensive fabs and R&D. The cost of development & production's amortized quite a bit when you sell a whole lot of them. Intel's only competitor for speed is AMD, both essentially the same architecture. The top dogs get exponentially more money for that stuff than the kids after them, and it shows. No matter how much I'd like it the other way (I'm a mac/ppc fan myself, and loathe giving up codewarrior).
As for multi-CPU OS's, or whatever I've been reading on the comments here, why? That's a lot of debugging for all those CPU architectures (think of all the compiler backend bugs you can hit, endianness issues, binary & ABI management issues, etc), for what? The user doesn't see much of a difference between them, and "cleaner architecture" means jack squat to them.
Us computer scientists can wish for cleaner architectures (and just 4 more registers, please!) all we want, but that doesn't change the market. Besides, we shouldn't worry about it -- it's 2007, we've got much bigger fish to fry.
The two systems are not completely dissimilar: they both contain a PowerPC core running at 3.2 GHz, both have similarly-clocked GPUs, and both come with 512 MB of RAM
And then there's the matter of the 6 fully-programmable Vector DSPs (8-1 for production, -1 for sony's own uses) running on an on-die bus attached to the CPU on one, and then 2 other PPC procs on the other. If you don't know how much of a difference that makes, read a computer architecture book.
IMHO, I'd rather do the Cell any day of the week. Putting physics, collision, etc into the DSPs really screams 'absurd performance.' Get the algorithm implementations good once, and reuse them for the next 10 years of the PS3's lifetime. (plus, we're going to see similar algorithm implementations for the cell come out of academia, where it's used for other simulation uses). A lot of a game engine's contents can use DSPs pretty effectively. Save the PPC core for scripting.
1. They're ultimately looking for physical interaction. That's traceable. 2. They're not very good -- they've already been caught before. 3. They're not always very smart -- they may not think of multiple IDs.
And what would I run? What app do I need that doesn't have a better replacement on the mac? Remember, I don't use 10,000 apps, I use 5-10 apps, and they tend to mostly be the same ones others use. So it's not hard to find what I need in mac form. And the mac form is a lot nicer than X. The POSIX core of OS X runs everything I need, I'd only want X for the user-interface side... but why would I run something with an X interface?
Ok everyone, say it with me:
X is ugly and I won't take it anymore X is ugly and I won't take it anymore X is ugly and I won't take it anymore
And if "usability" just means "eye candy" to you, you've been drinking too much *NIX kool-aid.
Actually, chances are that a decent job out of high school will be earning the same amount you'd make (with 4 years of experience) as the starting job out of college.
Yeah. I'm really glad that my school skimmed over the theoretical parts of computer science, kept it light on things like politics and american history (GO BUSH!), and spent all its time on what was important to me at the time... COBOL and FORTRAN.
Seriously, the job you get right out of college is essentially entry-level, while the education you get in college is designed for your entire life. That education is also for a lot more than any job you ever get -- it's for the entire life: personal, political, social.
It's 2006, almost 2007. I seriously can't believe this argument's still being held. Of course, we did just watch an assload of jackasses trying to push creationism as truth...
This is really the best confirmation data we have. DPRK says they set off a nuke. Even if a nuke had fizzled, it would've been bigger than the 550T explosion the seismometers felt. From here, "A geology professor at Yale, Jeffrey Park, emails to tell me that the updated Richter magnitude for the North Korea event is 3.5, which he calls "mighty small for a crude nuke." And that's true: it suggests a very small yield. But the odd thing is that it's actually harder to build a 1 kiloton weapon than a 5 or 10 kiloton weapon, and it's unlikely North Korea has the expertise to do this."
So, nobody's really sure what to believe right now, and eventually it'll just fall to consensus on the data we already have.
The best place to hear about the debate's over at ArmsControlWonk. New radionucliotide data, insider info from some well-placed anonymous sources, and insights into the scientific cultures within dictatorships paints an interesting picture.
Slashdot Mirror
Well, I'm here at tech, and am still waiting for word on whether or not I've lost any friends today. One roommate of mine may have already lost one, and some friends have lost friends of theirs.
No guns are allowed on campus. We have a few full-time shrinks on campus. The engineering student was probably an american (most of our foreign students are grad students). So far the word is the kid was a senior with 3 engineering majors. We have lots of bible groups, etc on campus, but we're not really known for being very religious or very secular for that matter.
My guess would be stress. I've seen grown men cry over single assignments, several of them, over the years here @ VT. The engineering kids are pushed really hard, and many of them don't deal with it very well. 60 hours a week of real work are pretty normal, with classes that everage 27-50%, which are only curved at the end (and nobody knows the curve till then). Try that for 4 years while growing up... Many engineering students I know end up having fairly empty shells of personalities, as their entire lives so far have circled around work and thinly veiled attempts at having a life on the side.
3 engineering majors at once would break most people. Guaranteed.
So far, the big questions are:
1. Why didn't the students find out about the 7:15am shooting until 2 hours later
2. Why was only the dorm closed?
To be fair, we had two bomb threats (no bombs) over the last 2 weeks, the last one only 3 days ago. So maybe the administration was getting tired of interrupting school for non-issues.
Access Linux is the Cobalt APIs over a Linux Kernel. PalmSource did that for better hardware compatibility and easier driver writing for their licensees.
For me, seeing these things in person (I grew up in DC, so I ended up seeing lots of these things) made them more real for me. Reading about the Constitution from a textbook makes it seem like a bit of an abstract concept. But seeing the thing right in front of me reminds me of how it's a paper document that people wrote by hand.
I agree with the GP that a balance has to be kept. In this case, if it's so far gone as to be a fruit stand, it's probably best torn down and replaced with a new research lab.
Well, there's Just In Time compilation, which is part of the JRE we all use today. It runs a lightweight profiler and compiles the most-used section ("hotspots") to native machine code.
I think CW's buying the initial land in order to get the ball rolling on getting themselves involved in the transactional business of real estate on SL. Long after the current stuff is sold off, they want to be the agents you buy virtual real estate from later.
Atomic I/O generally (I'm from the posix side of things, so maybe there are different definitions on win32) means that the entire operation's done by the kernel at once. E.g. I write 16 bytes at offset 0 to this file, and another app's writing 16 bytes at offset 0 to this file. When we're both done writing with a single write() call, the file's got 16 bytes from one of us, not some from one and some from another. No amount of userspace I/O calls will guarantee this (outside some godawful locking, which essentially forces the same thing, only with more context switches and a lot more userland complexity).
In terms of filesystem changes, all the files are created (possibly overwriting other files) or, if one fails, they all undo. Same for registry changes. Them all running together to completion or not at all sounds great. I definitely want registry changes to be atomic, b/c there's a good chance other apps are mucking around with it too. Also, I want those changes to be consistent -- with an atomic change that includes both the registry and a filesystem change, I can make sure I don't change a file with the registry pointed to the old value, or visa versa. In userland I could roll back, but I couldn't stop that time period when it's inconsistent.
They also involve atomic I/O to multiple systems simultaneously. Userland can't do this. Databases work on one system, their own data files, and have full control over these files.
Userland apps don't have that kind of control over the registry. Hell they may not be sure to have that kind of control over the files they're manipulating.
Besides, I'd rather have this code once in a DLL than 10 times in 10 different apps. That's real bloat.
The Cell's pretty straightforward with the tools given. You use a pthread-like library to give them code to run, and all your global variables are valid I/O targets for DMA. See this for an introduction: http://gamasutra.com/features/20060721/chow_01.sht ml
Frankly, the programming's one of the reasons I bought a PS3. It just looks fun.
As for what to run on each one, some things come to mind: physics, collision, particle effects, AI, etc. Some of them can be split up across multiple SPEs. Run them on a work queue of objects, or assign each a calculable subset, (e.g. SPE1 gets all the objects with an even ID #, SPE2 gets all the odd ones).
Some people really hate getting multiple processors to work together on a problem. But, these people have a lot of problems coming their way, as more & more machines go that way. Hell the laptop I'm on right now is dual-core, and I doubt I'm alone in that fact.
You think it's a lack of intelligence that makes those conditions so rough? Try reading up on history & politics. Good God man.
No, if we do that, then we'd have a lot more Bushes elected to the white house.
Society needs an educated populace. The thing is people forget that 4 years isn't much time to learn enough for the next 50.
The current system lets people go to grad school, which is heavy thinking, when they want more. At 18-21, there's only so much thinking they're gonna do. It's also probably the only time they're intellectually green enough to have the patience for all that training (later on, people need to be sold on its necessity a lot more before investing in that kind of effort).
The APIs are exposed in Obj-C and in C.
Obj-C has some nice messaging, but it's a pain for anything big.
Most real Mac development's done in C++
Gotta give them props for that. Also, one VERY good reason to try your own OS on the PS3: the Cell CPU.
Economics. The most popular CPU architecture gets all $$ for expensive fabs and R&D. The cost of development & production's amortized quite a bit when you sell a whole lot of them. Intel's only competitor for speed is AMD, both essentially the same architecture. The top dogs get exponentially more money for that stuff than the kids after them, and it shows. No matter how much I'd like it the other way (I'm a mac/ppc fan myself, and loathe giving up codewarrior).
As for multi-CPU OS's, or whatever I've been reading on the comments here, why? That's a lot of debugging for all those CPU architectures (think of all the compiler backend bugs you can hit, endianness issues, binary & ABI management issues, etc), for what? The user doesn't see much of a difference between them, and "cleaner architecture" means jack squat to them.
Us computer scientists can wish for cleaner architectures (and just 4 more registers, please!) all we want, but that doesn't change the market. Besides, we shouldn't worry about it -- it's 2007, we've got much bigger fish to fry.
And then there's the matter of the 6 fully-programmable Vector DSPs (8-1 for production, -1 for sony's own uses) running on an on-die bus attached to the CPU on one, and then 2 other PPC procs on the other. If you don't know how much of a difference that makes, read a computer architecture book.
IMHO, I'd rather do the Cell any day of the week. Putting physics, collision, etc into the DSPs really screams 'absurd performance.' Get the algorithm implementations good once, and reuse them for the next 10 years of the PS3's lifetime. (plus, we're going to see similar algorithm implementations for the cell come out of academia, where it's used for other simulation uses). A lot of a game engine's contents can use DSPs pretty effectively. Save the PPC core for scripting.
Well, two things going on here:
1. They're ultimately looking for physical interaction. That's traceable.
2. They're not very good -- they've already been caught before.
3. They're not always very smart -- they may not think of multiple IDs.
Hopefully that helps.
And what would I run? What app do I need that doesn't have a better replacement on the mac? Remember, I don't use 10,000 apps, I use 5-10 apps, and they tend to mostly be the same ones others use. So it's not hard to find what I need in mac form. And the mac form is a lot nicer than X. The POSIX core of OS X runs everything I need, I'd only want X for the user-interface side... but why would I run something with an X interface?
Ok everyone, say it with me:
X is ugly and I won't take it anymore
X is ugly and I won't take it anymore
X is ugly and I won't take it anymore
And if "usability" just means "eye candy" to you, you've been drinking too much *NIX kool-aid.
It is Irix here. Hell it's usually 1 step away from making the shell suid root.
Actually, chances are that a decent job out of high school will be earning the same amount you'd make (with 4 years of experience) as the starting job out of college.
Yeah. I'm really glad that my school skimmed over the theoretical parts of computer science, kept it light on things like politics and american history (GO BUSH!), and spent all its time on what was important to me at the time... COBOL and FORTRAN.
Seriously, the job you get right out of college is essentially entry-level, while the education you get in college is designed for your entire life. That education is also for a lot more than any job you ever get -- it's for the entire life: personal, political, social.
It's 2006, almost 2007. I seriously can't believe this argument's still being held. Of course, we did just watch an assload of jackasses trying to push creationism as truth...
Yeah those evangelicals are a scary group... wait, who were you talking about?
I suspect that when the engines need the power, they'll take precedence. Sheesh, modern electronics do have a level of sophistication to them...
Is that a treo in your pocket or are you just happy to see me?
Yeah, it's a treo. Condoms are expensive.
Russia's likely got them, but they really don't want to piss off China by selling to DPRK.
(this is all my own speculation, of course).
This is really the best confirmation data we have. DPRK says they set off a nuke. Even if a nuke had fizzled, it would've been bigger than the 550T explosion the seismometers felt. From here, "A geology professor at Yale, Jeffrey Park, emails to tell me that the updated Richter magnitude for the North Korea event is 3.5, which he calls "mighty small for a crude nuke." And that's true: it suggests a very small yield. But the odd thing is that it's actually harder to build a 1 kiloton weapon than a 5 or 10 kiloton weapon, and it's unlikely North Korea has the expertise to do this."
So, nobody's really sure what to believe right now, and eventually it'll just fall to consensus on the data we already have.
The best place to hear about the debate's over at ArmsControlWonk. New radionucliotide data, insider info from some well-placed anonymous sources, and insights into the scientific cultures within dictatorships paints an interesting picture.