You seem to think an engineering degree should be awarded based on how knowledgeable you are, or how smart you are...
No engineering student would be expected to know anything close to what they need to know at their job. They might be expected to be able to learn it quickly, while the fuse is lit...
An engineering degree simply means that you are willing to work very hard and don't crack under inordinate deadlines, and you make the right calls in the midst of pressure. It's also meant to drive out the egotistical skew that "genius" high-schoolers graduate with. Colleges can't send kids back to high school to get another shot at turning their homework in, so.
It means either flunking them out or guiding them into different majors... but certainly not graduating them! Just because, what? They paid the money so they're entitled to a degree? Or they got a 12,000 SAT score in high school? No.
I know what my limits are now. I've worked many 90 hour weeks, and I've accomplished ludicrous deadlines, but I've never been pushed to the brink of madness since I graduated from Georgia Tech. Probably the success come from knowing exactly where that point of failure is. It was hard, but I've never wished it were easier.
So far George Bush reminds me of a person playing Bomberman for the first time-- "Hey Mr Terrorist, take that! No! Come back here. AHH! You have Kick. More bombs, I need more bombs..."
And now he's about to pick up SuperFlame.
Now, many Bomberman experts will tell you that despite the amazing offensive firepower of SuperFlame, your best chance of ultimate survival is to not use it...
debatable... but look at what is coming for the PSP in the next 2 months:
09/13 : Final Fantasy VII - Advent Children
09/13 : Burnout Legends
10/01 : Ghost in the Shell
10/01 : Legend of Heroes
10/24 : Grand Theft Auto - Liberty City Stories
11/15 : Metal Gear Acid 2
11/15 : Street Fighter
11/21 : Need For Speed - Most Wanted
Now, it sucks that it's taken this long to get that kind of line-up, but now there is. And what does the DS have again... Nintendogs?
You've just described real poker, where you're physically at a casino.
That misses the point entirely. There are 2 separate, distinct issues that the pokerbot addresses that are unique to online play:
1) Any player could be using this program to evaluate the current live hand in an off-line fashion. Attempting to weed them out by chatting is useless. As far as using a bot is cheating, this would be cheating (and many players "seeking to understand the game better" would deem it as excusable!)
2) Outright collusion. This can be done by two humans using the same on-line poker forum. No bots are neccessary. That bots also do it is irrelevant. The reason the bots can collude is because the program author thought that people need to be aware of the issue!
From a PS2 perspective there are 2 things that hit me as really cool improvements.
First the SPUs have the ability to initiate DMA. That means they can do stuff like calculate memory-mapped addresses and request more data, or select different destinations for a calculation. Or even load in a different program to do specialized execution. All independent of the main processor. BIG improvement.
2nd is the integer instructions. They really have everything... shifts, rotates, all SIMD. One of the big problems with PS2 VUs was that you had to resort to real sorcery to do simple things like shifts. But these seem to be real actual general-purpose CPUs. There's nothing that really strikes me as "OMG, I can't believe they didn't include X! Idiots!" Branch prediction, maybe.
No, the beliefs would just evolve to accomodate (or deny) the new discoveries like they always do.
You see, once there's a sudden change in the culture and the current belief system becomes unfit to propogate around the population, new amendments are inserted more-or-less randomly into the belief structure and whichever mutations are most fit to attract the greatest number of believers will become the basis for future generations of the religion.
This ability to adapt is really the cornerstone for modern day religion. It also provides us with a wide diversity and complexity of belief systems, yet which all have striking similarities.
It's always struck me as very odd how the teams line up in this format war. Matsushita, Toshiba and IBM are all in 6C, yet Matsushita swapped teams and lined up with Sony on Blu-Ray. IBM jumped on with Sony for the PS3. Toshiba, of course, backs HD-DVD but they are also partnered with Sony to do production for the Cell processors going into the PS3-- whose support of BluRay is arguably HD-DVD's greatest threat!
Really, I just don't get it. Is everyone (except Sony) playing both sides of the fence?
Who cares? There is the ever-present diminishing returns spectre when start talking about improving efficiency.
My round-trip commute costs $5.00 (40 miles at 8 miles per dollar). A car that is twice as efficient (which would be roughly 40mpg) would cost only $2.50. A car that is 250mpg (100 miles per dollar) will cost $0.40 per trip...
But you see, a simple doubling in efficiency yielded a greater benefit than an additional 6x improvement! Efficiency beyond 60mpg is simply irrelevant. People just don't travel far enough, or they're not poor enough to justify it.
"Of course the 360 Launch will flop." That's what Sony is thinking at least. They're saying "Our big competition is PGR3? Nobody's buying new hardware for that."
So what do you do when someone fires a spitwad at you? Fire one back... or wait a minute to compose yourself and then assault them with overwhelming force? Sony's not dumb. They saw what happened with the PSP... it's a niche market because 8 months after launch it'll finally be a launchable platform.
With the PS3, there isn't leeway. They'll get all their ducks in a row... launch titles, blu-ray, online, Blu-Movies, middleware... they'll get names in the pipeline to hit stores at the right times. Microsoft has shown their hand: "XBox 2 this Fall. Another Halo released the same day as PS3..." Sony has decided there is no point to engage that fight, so we'll see what they do and whether it pays off.
Maybe this is your point, but notice how vividly both Microsoft and Linux are moving away from x86? I mean, Microsoft must be sinking quite a lot of money into producing, then providing a development environment for an in-order heavily pipelined SMP PowerPC... a lot of cutting-edge technology that can't be leveraged into other projects, unless.......
IBM, meanwhile has already sunk a lot of money into making the Cell, and you can be sure that Linux is meant to be its baby. Sony has already announced you can buy it along with your PS3.
(P.S., I didn't mean 'Linux is moving away...' so much as 'IBM is morphing linux away...')
I downloaded the evaluation and tried to run a test on a program I'd written. Crash. Tried again, crash.
crash, crash, crash.
I checked the support forums and found other people having the same problems with the official response "VTune only supports Intel processors". Well fine, that's a bit assinine, but at least a damn dialog box would have saved me an hour!
But to just simply crash... that's for programs written by college kids. So I downloaded AMD's CodeAnalyst and (despite some quirks) it's been amazing... pipeline simulation-- yeah!. Just ditch Intel people. They don't really love you.
In theory, a good compiler could make this chip very, very fast, and in reality, as you see, this can be the case.
There is a lot of code a compiler just can't make fast. Code that will run fast has to be designed that way-- sequential memory access, predictable branches (or no branches), processing on multiple pieces of data at a time... Simply recompiling branch-heavy cache-agnostic pointer-aliased code with a smarter compiler will not give you huge gains.
And the reality, as we can see, is that the chip can run fast... but it takes forethought, like the kind chip makers use when writing benchmark code.
You're right that brackets aren't any better placed than in QWERTY.
However, I feel that's more than compensated for by the underscore (which lives where qwerty " is, on the home row). Of course, it depends on the naming standards you subscribe to, but many people use conventions where underscores heavily outweigh brackets.
Additionally, the comma and dot operators are on QWERTY 'w' and 'e' which is a pretty handy location.
And to answer the OP, switching back and forth is easy. Just learn dvorak to completion before trying to switch back for the first time.
I wonder if IBM will fall for the same trick in its suit with SCO?
The fork lift pulls up to the front steps of IBM headquarters and unloads several large crates marked "OpenServer 6". Crouched inside, Darl waits patiently, grinning, knife in hand. Hoping. Praying they take the bait...
That's completely inaccurate. The way you see collision equations written in textbooks, you're right, lits bits of code (trivial rejects) lots of branches. But you'd NEVER write real code like that for exactly the reason you mention! The recursion stuff, like descending a k-d tree can be done iteratively with predictable branches all the way down to the leaf.
Basically, you'd have to be a moron to process one thing at a time or branch at every trivial reject...
Here's a couple examples. For something simple like ray-triangle, you don't even bother with the trivial rejects... or your trivial reject is the ray triangle (leaving out the segment triangle, or calculating distance to triangle), and you can do this calculation on a whole batch of triangles at once. It's just 2 cross products and 3 dot products. You accumulate the succeses in a list (branchlessly again, using conditional writes), and do the distance/segment part of the test on groups of those at a time.
For something like box/triangle, it boils down to 4 dot products on each of 13 axes. After each axis, it may be possible to exit the test. But you don't have 13 branches! You use the axes of the box as the first three tests, do 12 dots, accumulate the results and branch once (which eliminates 90% of the non-intersecting cases). Then process each successive pair of potential axes in unison. Not only does the simultaneous execution hide the latencies in the calculation, but the branches are predictable.
Textbooks are not meant to be clever or optimal or tuned to any architecture, but they ARE accurate, and it's up to programmers to take it from there.
I haven't developed for either next-gen devkit, but I have done a lot with the PS2, and that is exactly what we do... except we do it with just 4kb on the VU0. And VU0 is only single-ported, so you can't DMA to it while it's running... the only way to feed it with data while running is for the CPU to write directly into its registers. I'm not kidding. And don't even start with complicated memory transfers, if you only knew what we had to do to sync textures and geometry for rendering.
On the PS3, we get 256 KB of memory with a vector processor running 3.2Ghz, and people are complaining? And we get a bunch of em. Whoever they talked to were not PS2 developers. The same people who made sweet PS2 games will be making sweet PS3 games, trust me. For everyone else it will be harder to get used to.
For one thing, you can kiss your virtual calls goodbye.
First of all, if you're going to rant you need some background. It's hugely important that you understand how this Hollywood wannabe mentality started.
A couple years ago Jason Rubin (of Crash Bandicoot/Jak&Daxter fame) gave a speech about Hollywood that seems to have been wildly misinterpreted. He likened the current state of the video game industry to the packaged goods business. People aren't buying the content of games, they're buying the box. They're buying the marketing, the [evil] publisher. The [evil] publisher wants it that way, they want to remove the public's association with talent from the purchasing of the game... they want consumers to think that all developers are the same and let hype take care of the rest (at this pointed he pointed out that Crash games are still being made, but not by Naughty Dog).
He then mentioned that if the top 300 game developers got in an airplane and it crashed, the industry would be set back a decade. If the top 300 marketing people fell into the same misfortune, the Industry wouldn't miss a beat. People hooted and cheered at this irony... laid out so eloquently, between where the publishers place the importance of moving products with where the real importance was.
He then confused a lot of people, talking about Hollywood is the future and getting invited to parties, and so that is what a lot of people walked away with... However, the real crux of the passionate speech was that Game Companies, not publishers, belong in big bold letters on the box. Game development is a talent industry, not a packaged good... Game Designers who consistently design good games deserve the same name recognition and the same selling power as the equivalent Hollywood celebrities, Robert Deniro, Kevin Spacey, etc. with their name Right There on the Box in the same way that Hollywood movies are marketed (And that there are more people making good games than just Will Wright and Miyamoto). Until developers make those demands, publishers will feel free to keep marketing and unloading the same crap on the unsuspecting public.
but floating point operations aren't exactly optimal for things like AI.
False.
FLOPs are not generally useful for things like scripted AI which are very branch heavy with a lot of indirection, and many possible branch targets and data requirements.
The techniques described in this game are highly mathematical in nature with a small memory foot-print, (adaptive neural networks and genetic programming via Kenneth Stanley's NEAT algorithm) and would benefit hugely from parallel vector proccessing.
Additionally, at the end of the day, the AI decision making is not nearly as expensive as the proximity-query and pathfinding routines that affect the decisions. These routines also benefit hugely from vector processors and high bus-bandwidth.
So fittingly, the AI will only suffer if the human intelligence can't adapt and make the fairly obvious decision to move toward more mathematical AI routines.
I feel like he should walk around with a mask and fluorescent yellow-and-orange spandex and wear a cape with a Red E emblazoned within a square... That guy is like the last real American hero.
Although, I doubt he could actually bring successful legal action against the 'household-name advertisers'... hopefully the threat is enough to choke off the money flow. Who cares about all the spigots when you can shut down the water main?
Slashdot Mirror
No engineering student would be expected to know anything close to what they need to know at their job. They might be expected to be able to learn it quickly, while the fuse is lit...
An engineering degree simply means that you are willing to work very hard and don't crack under inordinate deadlines, and you make the right calls in the midst of pressure. It's also meant to drive out the egotistical skew that "genius" high-schoolers graduate with. Colleges can't send kids back to high school to get another shot at turning their homework in, so.
It means either flunking them out or guiding them into different majors... but certainly not graduating them! Just because, what? They paid the money so they're entitled to a degree? Or they got a 12,000 SAT score in high school? No.
I know what my limits are now. I've worked many 90 hour weeks, and I've accomplished ludicrous deadlines, but I've never been pushed to the brink of madness since I graduated from Georgia Tech. Probably the success come from knowing exactly where that point of failure is. It was hard, but I've never wished it were easier.
Yeah, I suppose that's for "personal-use only"? Dude, they are so sued.
And now he's about to pick up SuperFlame.
Now, many Bomberman experts will tell you that despite the amazing offensive firepower of SuperFlame, your best chance of ultimate survival is to not use it...
debatable... but look at what is coming for the PSP in the next 2 months:
09/13 : Final Fantasy VII - Advent Children
09/13 : Burnout Legends
10/01 : Ghost in the Shell
10/01 : Legend of Heroes
10/24 : Grand Theft Auto - Liberty City Stories
11/15 : Metal Gear Acid 2
11/15 : Street Fighter
11/21 : Need For Speed - Most Wanted
Now, it sucks that it's taken this long to get that kind of line-up, but now there is.
And what does the DS have again... Nintendogs?
That misses the point entirely. There are 2 separate, distinct issues that the pokerbot addresses that are unique to online play:
1) Any player could be using this program to evaluate the current live hand in an off-line fashion. Attempting to weed them out by chatting is useless. As far as using a bot is cheating, this would be cheating (and many players "seeking to understand the game better" would deem it as excusable!)
2) Outright collusion. This can be done by two humans using the same on-line poker forum. No bots are neccessary. That bots also do it is irrelevant. The reason the bots can collude is because the program author thought that people need to be aware of the issue!
First the SPUs have the ability to initiate DMA. That means they can do stuff like calculate memory-mapped addresses and request more data, or select different destinations for a calculation. Or even load in a different program to do specialized execution. All independent of the main processor. BIG improvement.
2nd is the integer instructions. They really have everything... shifts, rotates, all SIMD. One of the big problems with PS2 VUs was that you had to resort to real sorcery to do simple things like shifts. But these seem to be real actual general-purpose CPUs. There's nothing that really strikes me as "OMG, I can't believe they didn't include X! Idiots!" Branch prediction, maybe.
No, the beliefs would just evolve to accomodate (or deny) the new discoveries like they always do.
You see, once there's a sudden change in the culture and the current belief system becomes unfit to propogate around the population, new amendments are inserted more-or-less randomly into the belief structure and whichever mutations are most fit to attract the greatest number of believers will become the basis for future generations of the religion.
This ability to adapt is really the cornerstone for modern day religion. It also provides us with a wide diversity and complexity of belief systems, yet which all have striking similarities.
Really, I just don't get it. Is everyone (except Sony) playing both sides of the fence?
My round-trip commute costs $5.00 (40 miles at 8 miles per dollar). A car that is twice as efficient (which would be roughly 40mpg) would cost only $2.50. A car that is 250mpg (100 miles per dollar) will cost $0.40 per trip...
But you see, a simple doubling in efficiency yielded a greater benefit than an additional 6x improvement! Efficiency beyond 60mpg is simply irrelevant. People just don't travel far enough, or they're not poor enough to justify it.
Just send me your PSP and a 6GB memory stick...
So what do you do when someone fires a spitwad at you? Fire one back... or wait a minute to compose yourself and then assault them with overwhelming force? Sony's not dumb. They saw what happened with the PSP... it's a niche market because 8 months after launch it'll finally be a launchable platform.
With the PS3, there isn't leeway. They'll get all their ducks in a row... launch titles, blu-ray, online, Blu-Movies, middleware... they'll get names in the pipeline to hit stores at the right times. Microsoft has shown their hand: "XBox 2 this Fall. Another Halo released the same day as PS3..." Sony has decided there is no point to engage that fight, so we'll see what they do and whether it pays off.
1) charge 1 coffee to connect.
2) Another per usage of the bathroom key.
3) Profit!
IBM, meanwhile has already sunk a lot of money into making the Cell, and you can be sure that Linux is meant to be its baby. Sony has already announced you can buy it along with your PS3.
(P.S., I didn't mean 'Linux is moving away...' so much as 'IBM is morphing linux away...')
Whoa! *IDEA*
Then you could have a never-ending supply of wind for your turbine! This is gonna be great, I'm getting goose bumps. Now where's my hammer?
I downloaded the evaluation and tried to run a test on a program I'd written. Crash. Tried again, crash.
crash, crash, crash.
I checked the support forums and found other people having the same problems with the official response "VTune only supports Intel processors". Well fine, that's a bit assinine, but at least a damn dialog box would have saved me an hour!
But to just simply crash... that's for programs written by college kids. So I downloaded AMD's CodeAnalyst and (despite some quirks) it's been amazing... pipeline simulation-- yeah!. Just ditch Intel people. They don't really love you.
There is a lot of code a compiler just can't make fast. Code that will run fast has to be designed that way-- sequential memory access, predictable branches (or no branches), processing on multiple pieces of data at a time... Simply recompiling branch-heavy cache-agnostic pointer-aliased code with a smarter compiler will not give you huge gains.
And the reality, as we can see, is that the chip can run fast... but it takes forethought, like the kind chip makers use when writing benchmark code.
However, I feel that's more than compensated for by the underscore (which lives where qwerty " is, on the home row). Of course, it depends on the naming standards you subscribe to, but many people use conventions where underscores heavily outweigh brackets.
Additionally, the comma and dot operators are on QWERTY 'w' and 'e' which is a pretty handy location.
And to answer the OP, switching back and forth is easy. Just learn dvorak to completion before trying to switch back for the first time.
The fork lift pulls up to the front steps of IBM headquarters and unloads several large crates marked "OpenServer 6". Crouched inside, Darl waits patiently, grinning, knife in hand. Hoping. Praying they take the bait...
Basically, you'd have to be a moron to process one thing at a time or branch at every trivial reject...
Here's a couple examples. For something simple like ray-triangle, you don't even bother with the trivial rejects... or your trivial reject is the ray triangle (leaving out the segment triangle, or calculating distance to triangle), and you can do this calculation on a whole batch of triangles at once. It's just 2 cross products and 3 dot products. You accumulate the succeses in a list (branchlessly again, using conditional writes), and do the distance/segment part of the test on groups of those at a time.
For something like box/triangle, it boils down to 4 dot products on each of 13 axes. After each axis, it may be possible to exit the test. But you don't have 13 branches! You use the axes of the box as the first three tests, do 12 dots, accumulate the results and branch once (which eliminates 90% of the non-intersecting cases). Then process each successive pair of potential axes in unison. Not only does the simultaneous execution hide the latencies in the calculation, but the branches are predictable.
Textbooks are not meant to be clever or optimal or tuned to any architecture, but they ARE accurate, and it's up to programmers to take it from there.
On the PS3, we get 256 KB of memory with a vector processor running 3.2Ghz, and people are complaining? And we get a bunch of em. Whoever they talked to were not PS2 developers. The same people who made sweet PS2 games will be making sweet PS3 games, trust me. For everyone else it will be harder to get used to.
For one thing, you can kiss your virtual calls goodbye.
A couple years ago Jason Rubin (of Crash Bandicoot/Jak&Daxter fame) gave a speech about Hollywood that seems to have been wildly misinterpreted. He likened the current state of the video game industry to the packaged goods business. People aren't buying the content of games, they're buying the box. They're buying the marketing, the [evil] publisher. The [evil] publisher wants it that way, they want to remove the public's association with talent from the purchasing of the game... they want consumers to think that all developers are the same and let hype take care of the rest (at this pointed he pointed out that Crash games are still being made, but not by Naughty Dog).
He then mentioned that if the top 300 game developers got in an airplane and it crashed, the industry would be set back a decade. If the top 300 marketing people fell into the same misfortune, the Industry wouldn't miss a beat. People hooted and cheered at this irony... laid out so eloquently, between where the publishers place the importance of moving products with where the real importance was.
He then confused a lot of people, talking about Hollywood is the future and getting invited to parties, and so that is what a lot of people walked away with... However, the real crux of the passionate speech was that Game Companies, not publishers, belong in big bold letters on the box. Game development is a talent industry, not a packaged good... Game Designers who consistently design good games deserve the same name recognition and the same selling power as the equivalent Hollywood celebrities, Robert Deniro, Kevin Spacey, etc. with their name Right There on the Box in the same way that Hollywood movies are marketed (And that there are more people making good games than just Will Wright and Miyamoto). Until developers make those demands, publishers will feel free to keep marketing and unloading the same crap on the unsuspecting public.
C'mon download download! Gooooooogle!
False.
FLOPs are not generally useful for things like scripted AI which are very branch heavy with a lot of indirection, and many possible branch targets and data requirements.
The techniques described in this game are highly mathematical in nature with a small memory foot-print, (adaptive neural networks and genetic programming via Kenneth Stanley's NEAT algorithm) and would benefit hugely from parallel vector proccessing.
Additionally, at the end of the day, the AI decision making is not nearly as expensive as the proximity-query and pathfinding routines that affect the decisions. These routines also benefit hugely from vector processors and high bus-bandwidth.
So fittingly, the AI will only suffer if the human intelligence can't adapt and make the fairly obvious decision to move toward more mathematical AI routines.
The hard part is getting the robots into outter space. Or... at least to the tops of very tall mountains.
Although, I doubt he could actually bring successful legal action against the 'household-name advertisers'... hopefully the threat is enough to choke off the money flow. Who cares about all the spigots when you can shut down the water main?