... in this era of high technology and global cooperation (note the recent launch of another Soyuz rocket to the international space station) why don't we use some of our new tech on things other than money making crap that is e-business? So I propose the Star Trek solution-- to seek out new life and new civilizations, to boldly go where no man has gone before. We have let our space program become so withered. If I see another dime spent on "global e-business solutions" (basically any IBM ad) I'm gonna puke.
You do not store polygon data in the framebuffer memory!!! The ps2 has 32MB of RDRAM main memory for storing polygon data with mad bandwidth to send it to the GPU which will write pixels into the frickin framebuffer!! This post was so ignorant. With 4MB of VRAM there is more than enough room for 1280 x 1024@24bpp. Jeesh...
the X-Box has Nvidia on board, Nintendo's cube has a specialized GPU from, who, ATI? The PS2
has something similar
The ps2 has more than just "something similar" to an nvidia or ati chip. They call it a "graphics synthesizer." Okay, so I'm sure they have a lot in common, but sixteen of these things were working in parallel to render in real time scenes from what was IIRC Toy Story at siggraph (data was coming from an sgi o2k). I forget the specs, but it does sound somewhat unique to the world of current GPUs. I could be and often am incorret and off-topic.
...or somthing. About the "Last Hacker" this guy may very well be him, although I know I can't be the only person like myself out there, us hackers have got to be flourishing our there in this era of high bandwidth. By hacker I'm not talking about script kiddies (although I know there are you legit crypto hackers out there, keep kickin ass, and keep stickin it to the man with DeCSS) but us mad hackers who think we can make games, etc, that can compete mainstream, we're all helping keep this movement alive. I'm gonna release my little widget set soon (okay, so it's like gtk, but I've got to contribute something) and it's openGL based and stuff... so you're all gonna laugh at me... it's not a desktop environment, but just some buttons and stuff that contributes to an app that'll let ya make some 3d worlds and stuff...
oh, and don't judge SGI on their stock price alone, they are still The Best Computer Company on the Planet (TM)
Mike
(some barley and hops from the Missouri area contributed to this post)
Yes, I agree. We're all living in a 13th floor type of Universe. On a recent trip out of town, as I neared Redmond, WA, the horizon disappeared in an unrendered wireframe...
The whole system has one contiguous view of memory. The NU means "non-uniform" as in the memory access time is non-linear. If a process's memory is located on the processor module it's running on, the memory access is fastest. If it has to jump one module away, the memory access time increases by 100 nanoseconds (roundtrip). Architecturally, it's completely different than anything Sun has to offer. Sun has been promising a NUMA mahcine for years and still hasn't delivered. The closest company to SGI is Compaq(DEC), and there top of the line offering can almost compete with Origin _2000_. All other companies high-end servers use symmetric multiprocessing, which becomes limited as more and more processors try to access the shared memory bus, ultimately bringing in negative returns as you add more processors. This NUMA architecture incurs very little (if any ) penalty by ading more processors, as long as the hardware and OS do a good job of placing processes and memory (keeping them physically near). Also, the machine is _not_ limited by 512 processors. To give an example of the power of this box, a company has certain calculations that they run day to day. On their top-of-the-line Sun hardware, it takes about seven hours. On O3k, it takes seven seconds! What does being a modular system have to do with being a cluster? By being "modular" it simply means that you can plug in more of whatever you want, whenever you want. I believe you can even mix faster cpu modules with existing ones as they become available, protecting your investment. This is not a cluster.
Creed, the #1 rock band in the US right now, held a contest where the largest number of online votes regiesterd from any city would win a free concert. The winning city... Lincoln, Nebraska!
There are plenty of rock shows in the midwest, especially in Omaha, NE; Ames, IA; and Kansas City, MO.
As for computing, as soon as a new product comes out, I can just as easily get it in Lincoln as anywhere else (okay, so we don't have "Fry's" in Lincoln)
The one exception, those cute little GM electric cars. I imagine they'd be a big hit anywhere (damn oil industry is keeping down the electric car, methinks, possibly the metric system, too).
As a footnote, I currently reside in Sunnyvale, CA (the heart of silicon valley) and work for a company that makes very attractive blue UNIX boxes.
While MS VS is nice, I find that I can be much more productive using the command line debugger, gdb, no pointing and clicking and waiting for the UI to reconfigure itself into "debug" mode. Once you know the commands, you can become extremely efficient at debugging your code (Of course, the best method would be to not write buggy code to begin with, but it happens). I should also mention at this point that my favorite editor is vim. To me, the 1-2 punch of gcc and gdb make development under linux ideal. Also, vim is a very elegant text editor that makes creating/modifying code very easy/quick to do.
Mike
Note: no vi/emacs wars, here, please, they are both excellent editors, I think preference is based on whichever one you learned first =)
I'd just as soon read a book celebrating what a wonderful thing Windows 3.0 is.
Who's to say that what was relevant then is not relevant now? Windows 3.0 is dead and gone, an imperfect system that was replaced with an imperfect system. X is not only still alive and kicking, it's off and running. I think the fact that this book is still relevant is a testament to the wonderful design of X.
You're are probably the same kid who complains that his math book is old. "I want the new stuff!" Hilarious.
I believe it was in the first qtest for windows, a message would come up: General Protection Fault in module foo: with a register dump followed by: You Are So Hosed!!!! Anyone else familiar with that one?
Me opening my big mouth w/o reading much, dammit, I haven't done that on/. for a long time.
uhhhm... you're talking about two different things. i said 'porting linux to a numa-q.' numa-q is a brand name of a type of (x86) server. so, there's really no porting of linux to be done.
But... just because it is x86 doesn't mean there is no coding to be done!
you said 'coding an OS for numa,
You're correct, I meant "coding numa for an OS". No, I've never used DYNIX. But you seem to be a pretty big fan, could you enlighten?
It's not that difficult.. "it" being porting linux to a NUMA-Q.
Really? So you've done this?
I wonder if IBM is actually gonna write a NUMA layer for linux?
SGI has the best in the business, and they're busy putting it into linux as we speak. However, IRIX is _the_ OS as far as NUMA goes, coupled with the sgi Origin. With sgi's massive linux movement, I'm sure linux will be up to par with IRIX in no time at all. Check em out at www.sgi.com/origin. Ever heard of ASCI Blue Mountain? Checkout the top 500 list.
Yes, your issue over price/value is right on the money. Case in point: people knocking sgi boxes because of low quake performance!!! (although the latest sgi linux/NT boxes run quake3 at mind boggling speeds) Also keep in mind these sgi boxes are designed for _professinal_ graphics, not games.
Oh, and wrt professonial graphics and ray-tracing renderers. The very best aren't done on pc-class machines, try sgi's origin boxes (not many examples to give, most work done on origins is classified, cept yer local weather forecast, oh, I guess most/. readers are familiar with Star Wars). Don't let the 400Mhz cpu fool you, it'll kick your pentium/athlon/g whatever to hell and back.
I could hack this myself in a couple of months. There is nothing ingenous about this UI. We've seen it before. You want a good desktop? Try indigo magic under irix, it is very nice, although I myself have found ways it could be improved. As a hacker, I think you'll all agree, nothing is better than plain old _Enlightenment_ written by hackers, for hackers. Hell, why waste time hacking to make an Aqua-like environment, you can configure E in a few hours to do the same (okay, efm is still in dev, but give raster a break) Aqua is nothing special.
If you have problem with your GUI toolkit, simply make your own. Here was my situation: I was working on a graphics app. I was using motif. I updated to newer compilers, etc. My motif library didn't work anymore. So I looked at other toolkits. I didn't like them. I read more about X than the knowledge I needed to blit graphics to the screen. In four weeks, I had push buttons, toggle buttons, scale sliding thingies, scrollbars, and text widgets. And it can be themed with nice png images. It's easily extendable, whatever I can dream up. What I can't believe is that it took so many years for someone to get sick of Motif and write their own toolkit (Gtk). Now I know I'll hear people say "you're reinventing the wheel, why not use something that's out there gtk! qt!" Well, I don't like them so kiss off. I like my own little tk, 2208 lines of code, 56k, and it does everything motif did in 75% less lines of gui code. So, if you're bitching, do it yourself!
Please tell me how motif is slow. Are you still running on a 386? Please, please, how is it slow? I don't like the look of it either, but come on, your post has no basis.
You can all quit complaining about nvidia's non-open source driver right now! Due to the way they integrate their linux and windows code base, we should be able to get drivers almost immediately for this amazing piece of hardware (I hope, cuz I can't wait to get my hands on one, I skipped the geforce generation and am still with my tnt and tnt2's.)
They now have their development environment setup so that their windows and linux drivers use nearly the same code base( methinks ), so that whenever they fix a bug/add a feature to the windows or linux driver the improvement shows up in both. I think this is very ingenius. It speeds up time to market. Give them some credit, they believe that releasing all the specs to their harware would give away trade secrets. I know how much you all love patents here on slashdot, and they have many. Come on, do you see Intel telling amd "here are the circuit layouts for our IA-64 architecure". I don't think linux is about open-source hardware. So yes, the driver is software and it's not open-source. If you want an open-source driver, buy some obsolete hardware that has an open-source driver, in the meantime I'll be enjoying my nvidia cards, closed-source drivers and all. If nvidia is willing to maintain linux drivers for you, what are you complaining about?
A line needs to be drawn somewhere. What should be free, what shouldn't be free? The guys at nvidia put a lot of hard work into their designs and can't afford to give them away. They have to make _a_living_ off of this and I don't think they'll jeopardize their livelihood in the name of open-source.
That rocket car story is just a hoax, meant to bring about FUD concerning the true rocket car which was recovered in Roswell, New Mexico in 1947 and is now housed in hangar 18.
These same customers don't want their work/applications to be wasted next year when they need to change their software to use a newer video card, or vice versa.
If these customers use OpenGL, the apps should work no matter what the underlying driver code is. The idea is a level of abstraction between your software and video hardware so you don't have to rewrite your software when you get new video hardware, you code for a standard interface.
Isn't porting 32-bit applications to 64-bit platforms merely an issue of size of data types and memory addressing?
I don't think 256-bit GPU's address 2^256 - 1 bits of memory, they just process information in 256-bit chunks.
Having a wider data path isn't always going to help that much. You can only do so much in parallel. For instance, I believe you need mutually exclusive access to ZBuffers, shadow buffers, alpha, etc for 3d graphics. That's gotta place a limit somewhere, or at least put a damper on all the things in 3d graphics that _can_ be done in parallel (transformations of vertices, culling, yada, yada)
Okay, I'm rambling, but please, don't always bring up the 'bit' thing. Saying that a chip is 64-bit can have wildly different meanings and may or may not give it an advantage over other chips.
Slashdot Mirror
... in this era of high technology and global cooperation (note the recent launch of another Soyuz rocket to the international space station) why don't we use some of our new tech on things other than money making crap that is e-business? So I propose the Star Trek solution-- to seek out new life and new civilizations, to boldly go where no man has gone before. We have let our space program become so withered. If I see another dime spent on "global e-business solutions" (basically any IBM ad) I'm gonna puke.
Mike
You do not store polygon data in the framebuffer memory!!! The ps2 has 32MB of RDRAM main memory for storing polygon data with mad bandwidth to send it to the GPU which will write pixels into the frickin framebuffer!! This post was so ignorant. With 4MB of VRAM there is more than enough room for 1280 x 1024@24bpp. Jeesh...
Mike
the X-Box has Nvidia on board, Nintendo's cube has a specialized GPU from, who, ATI? The PS2 has something similar
The ps2 has more than just "something similar" to an nvidia or ati chip. They call it a "graphics synthesizer." Okay, so I'm sure they have a lot in common, but sixteen of these things were working in parallel to render in real time scenes from what was IIRC Toy Story at siggraph (data was coming from an sgi o2k). I forget the specs, but it does sound somewhat unique to the world of current GPUs. I could be and often am incorret and off-topic.
Mike
...or somthing. About the "Last Hacker" this guy may very well be him, although I know I can't be the only person like myself out there, us hackers have got to be flourishing our there in this era of high bandwidth. By hacker I'm not talking about script kiddies (although I know there are you legit crypto hackers out there, keep kickin ass, and keep stickin it to the man with DeCSS) but us mad hackers who think we can make games, etc, that can compete mainstream, we're all helping keep this movement alive. I'm gonna release my little widget set soon (okay, so it's like gtk, but I've got to contribute something) and it's openGL based and stuff... so you're all gonna laugh at me... it's not a desktop environment, but just some buttons and stuff that contributes to an app that'll let ya make some 3d worlds and stuff...
oh, and don't judge SGI on their stock price alone, they are still The Best Computer Company on the Planet (TM)
Mike
(some barley and hops from the Missouri area contributed to this post)
Yes, I agree. We're all living in a 13th floor type of Universe. On a recent trip out of town, as I neared Redmond, WA, the horizon disappeared in an unrendered wireframe...
The telescope it replaces (designed to last 10 years) collapsed in 1988 after only 26 years
It lasted 2.6 times as long as expected and you're complaing?
The whole system has one contiguous view of memory. The NU means "non-uniform" as in the memory access time is non-linear. If a process's memory is located on the processor module it's running on, the memory access is fastest. If it has to jump one module away, the memory access time increases by 100 nanoseconds (roundtrip). Architecturally, it's completely different than anything Sun has to offer. Sun has been promising a NUMA mahcine for years and still hasn't delivered. The closest company to SGI is Compaq(DEC), and there top of the line offering can almost compete with Origin _2000_. All other companies high-end servers use symmetric multiprocessing, which becomes limited as more and more processors try to access the shared memory bus, ultimately bringing in negative returns as you add more processors. This NUMA architecture incurs very little (if any ) penalty by ading more processors, as long as the hardware and OS do a good job of placing processes and memory (keeping them physically near). Also, the machine is _not_ limited by 512 processors. To give an example of the power of this box, a company has certain calculations that they run day to day. On their top-of-the-line Sun hardware, it takes about seven hours. On O3k, it takes seven seconds! What does being a modular system have to do with being a cluster? By being "modular" it simply means that you can plug in more of whatever you want, whenever you want. I believe you can even mix faster cpu modules with existing ones as they become available, protecting your investment. This is not a cluster.
We sell our stuff on unparalleled performance _and_ cosmetics! =)
Mike
You mentioned only attracting old, dead bands.
Newsflash!
Creed, the #1 rock band in the US right now, held a contest where the largest number of online votes regiesterd from any city would win a free concert. The winning city... Lincoln, Nebraska!
There are plenty of rock shows in the midwest, especially in Omaha, NE; Ames, IA; and Kansas City, MO.
As for computing, as soon as a new product comes out, I can just as easily get it in Lincoln as anywhere else (okay, so we don't have "Fry's" in Lincoln)
The one exception, those cute little GM electric cars. I imagine they'd be a big hit anywhere (damn oil industry is keeping down the electric car, methinks, possibly the metric system, too).
As a footnote, I currently reside in Sunnyvale, CA (the heart of silicon valley) and work for a company that makes very attractive blue UNIX boxes.
Mike
While MS VS is nice, I find that I can be much more productive using the command line debugger, gdb, no pointing and clicking and waiting for the UI to reconfigure itself into "debug" mode. Once you know the commands, you can become extremely efficient at debugging your code (Of course, the best method would be to not write buggy code to begin with, but it happens). I should also mention at this point that my favorite editor is vim. To me, the 1-2 punch of gcc and gdb make development under linux ideal. Also, vim is a very elegant text editor that makes creating/modifying code very easy/quick to do.
Mike
Note: no vi/emacs wars, here, please, they are both excellent editors, I think preference is based on whichever one you learned first =)
I'd just as soon read a book celebrating what a wonderful thing Windows 3.0 is.
Who's to say that what was relevant then is not relevant now? Windows 3.0 is dead and gone, an imperfect system that was replaced with an imperfect system. X is not only still alive and kicking, it's off and running. I think the fact that this book is still relevant is a testament to the wonderful design of X.
You're are probably the same kid who complains that his math book is old. "I want the new stuff!" Hilarious.
I believe it was in the first qtest for windows, a message would come up: General Protection Fault in module foo: with a register dump followed by: You Are So Hosed!!!! Anyone else familiar with that one?
...MegaMaid, Sir. She's gone from suck to blow!"
uhhhm... you're talking about two different things. i said 'porting linux to a numa-q.' numa-q is a brand name of a type of (x86) server. so, there's really no porting of linux to be done.
But... just because it is x86 doesn't mean there is no coding to be done!
you said 'coding an OS for numa,
You're correct, I meant "coding numa for an OS". No, I've never used DYNIX. But you seem to be a pretty big fan, could you enlighten?
Really? So you've done this?
I wonder if IBM is actually gonna write a NUMA layer for linux?
SGI has the best in the business, and they're busy putting it into linux as we speak. However, IRIX is _the_ OS as far as NUMA goes, coupled with the sgi Origin. With sgi's massive linux movement, I'm sure linux will be up to par with IRIX in no time at all. Check em out at www.sgi.com/origin. Ever heard of ASCI Blue Mountain? Checkout the top 500 list.
Yes, your issue over price/value is right on the money. Case in point: people knocking sgi boxes because of low quake performance!!! (although the latest sgi linux/NT boxes run quake3 at mind boggling speeds) Also keep in mind these sgi boxes are designed for _professinal_ graphics, not games.
/. readers are familiar with Star Wars). Don't let the 400Mhz cpu fool you, it'll kick your pentium/athlon/g whatever to hell and back.
Oh, and wrt professonial graphics and ray-tracing renderers. The very best aren't done on pc-class machines, try sgi's origin boxes (not many examples to give, most work done on origins is classified, cept yer local weather forecast, oh, I guess most
Mike
I could hack this myself in a couple of months. There is nothing ingenous about this UI. We've seen it before. You want a good desktop? Try indigo magic under irix, it is very nice, although I myself have found ways it could be improved. As a hacker, I think you'll all agree, nothing is better than plain old _Enlightenment_ written by hackers, for hackers. Hell, why waste time hacking to make an Aqua-like environment, you can configure E in a few hours to do the same (okay, efm is still in dev, but give raster a break) Aqua is nothing special.
Mike
If you have problem with your GUI toolkit, simply make your own. Here was my situation: I was working on a graphics app. I was using motif. I updated to newer compilers, etc. My motif library didn't work anymore. So I looked at other toolkits. I didn't like them. I read more about X than the knowledge I needed to blit graphics to the screen. In four weeks, I had push buttons, toggle buttons, scale sliding thingies, scrollbars, and text widgets. And it can be themed with nice png images. It's easily extendable, whatever I can dream up. What I can't believe is that it took so many years for someone to get sick of Motif and write their own toolkit (Gtk). Now I know I'll hear people say "you're reinventing the wheel, why not use something that's out there gtk! qt!" Well, I don't like them so kiss off. I like my own little tk, 2208 lines of code, 56k, and it does everything motif did in 75% less lines of gui code. So, if you're bitching, do it yourself!
Mike
Please tell me how motif is slow. Are you still running on a 386? Please, please, how is it slow? I don't like the look of it either, but come on, your post has no basis.
Hasn't sgi always been MIPS?
You can all quit complaining about nvidia's non-open source driver right now! Due to the way they integrate their linux and windows code base, we should be able to get drivers almost immediately for this amazing piece of hardware (I hope, cuz I can't wait to get my hands on one, I skipped the geforce generation and am still with my tnt and tnt2's.)
Mike
Unanonymous Coward
They now have their development environment setup so that their windows and linux drivers use nearly the same code base( methinks ), so that whenever they fix a bug/add a feature to the windows or linux driver the improvement shows up in both. I think this is very ingenius. It speeds up time to market. Give them some credit, they believe that releasing all the specs to their harware would give away trade secrets. I know how much you all love patents here on slashdot, and they have many. Come on, do you see Intel telling amd "here are the circuit layouts for our IA-64 architecure". I don't think linux is about open-source hardware. So yes, the driver is software and it's not open-source. If you want an open-source driver, buy some obsolete hardware that has an open-source driver, in the meantime I'll be enjoying my nvidia cards, closed-source drivers and all. If nvidia is willing to maintain linux drivers for you, what are you complaining about?
A line needs to be drawn somewhere. What should be free, what shouldn't be free? The guys at nvidia put a lot of hard work into their designs and can't afford to give them away. They have to make _a_living_ off of this and I don't think they'll jeopardize their livelihood in the name of open-source.
Mike
Non-Anonymous Coward
That rocket car story is just a hoax, meant to bring about FUD concerning the true rocket car which was recovered in Roswell, New Mexico in 1947 and is now housed in hangar 18.
If these customers use OpenGL, the apps should work no matter what the underlying driver code is. The idea is a level of abstraction between your software and video hardware so you don't have to rewrite your software when you get new video hardware, you code for a standard interface.
Mike
Okay, can anyone out there help me out?
Isn't porting 32-bit applications to 64-bit
platforms merely an issue of size of data types
and memory addressing?
I don't think 256-bit GPU's address 2^256 - 1
bits of memory, they just process information
in 256-bit chunks.
Having a wider data path isn't always going to
help that much. You can only do so much in
parallel. For instance, I believe you need
mutually exclusive access to ZBuffers,
shadow buffers, alpha, etc for 3d graphics.
That's gotta place a limit somewhere, or at least
put a damper on all the things in 3d graphics
that _can_ be done in parallel (transformations
of vertices, culling, yada, yada)
Okay, I'm rambling, but please, don't always
bring up the 'bit' thing. Saying that a chip
is 64-bit can have wildly different meanings
and may or may not give it an advantage over
other chips.
Cheers,
Mike