I for one would like to thank Microsoft for giving colleges this opportunity, and giving away a high-quality office suite and reasonably good development tools to people that need them.
I don't know how this applies to other Universities, but no, Macrohard is NOT giving away anything. Here, we have a deal with them that makes me sick. The university pays M$ a lump sum payment and is then allowed to distribute any M$ title for $5. Does that include packages or install floppies? No, just the CDs. Guess what, that lump sum payment is seven digits before the decimal. Yes, seven digits. Care to guess how many CDs the university estimated it "gave away" to students? It was on the order of 5,000. Oh boy, Macrohard is getting over $200/CD that the university "gives away!" Isn't that a great deal for the university? Isn't that a great deal for the 30,000+ students and/or taxpayers that did not purchase any of these discs but have to foot the bill??? In discussing the "success" of the deal, university officials said they were working on paying a larger fee to Macrohard to continue this program. Yes, this program may have been beneficial to you, yes, it was definitely beneficial to Macrohard, but I can't see how you can claim it helped anyone else. Tell me the university officials pushing this deal aren't corrupt!
Let's talk about other fun things newbies can do with linux: edit the/etc/ppp/chap-secrets file to login to their isp. Use linuxconf (text-based) to get their connection online. Okay, I'm making this alittle hard - you could use kppp. =) After you login you find out you can't - it keeps disconnected.
If you ask me, the "Mom Myth" is true for all OSes, I don't care which one you talk about. How many times have I had to explain to my mom how to print a file in Windows? How many times have I had to explain to her how to simply browse the web? If someone is not inclined to using computers to begin with, no matter how "user friendly" you make them, if the UI is not organized the way that person thinks, they will struggle. I think all of us have at one point or another struggled to get dial-up connections working...wrong IPs, wrong submasks, something somewhere isn't done correctly. Those problems will exist no matter what you do, unless you want to try and force a standard deployment on everyone.
I often like to think of it like this: Computers are tools, just like say a hammer, perhaps one of the least complicated tools around. If you are skilled with the tool, of course it is easy to use, but if you haven't ever really used the tool, you are going to wind up with a bruised thumb.
Besides, methanol's only uses are as a fuel and as a raw material for making other compounds. When you burn or catalyze the Methanol as fuel, you get the CO2 back again, and when those other compounds are eventually consumed or destroyed, you get the CO2 back again (and other noxious compounds, by the way).
Methanol actually does have other uses. It makes a great solvent for semiconductor processing (we use it all the time), of course, that is electrical grade methanol, which may not be so easy to get from these little enzyme buggers.
It is implied in the article that yes, Macrohard did contest that claim:
A Microsoft spokeswoman, however, disputes these perspectives, claiming that Microsoft's closed-source software is more secure than ever.
These perspectives M$ is disputing would be that Linux is more secure than any M$ offering. Now, whether the author was putting a spin that Macrohard didn't mean, I can't say, but do you really think Macrohard didn't want everyone to think that was what they were saying???
I'm glad to see the free software/open source concept being recognized like this, but I think it would have been nice if these experts had taken the time to look at other alternatives.
Well, we have to remember, we are probably talking about beaurocrats here, and therefore quick to jump on the buzzwords (Linux). Sure they are the experts, but they are still the government experts, which probably means they are your standard beaurocrats becuase I would bet the private sector would be able to offer a lot more for their expertice...but of course, I could be wrong, wouldn't be the first time, definitely won't be the last time for that!
All I can say is my karma has gone up from moderation, not from posted comments since I am usually way late to threads (after the moderators have come through). I think the system is working, or at least a step in the proper direction.
I protect my privacy by linking my Netscape cookies files to/dev/null (the infamous bit- bucket). Netscape never knows the difference!
Well, I had never thought of doing that (being a relative newbie to Linux at home and having the powers to explore UNIX as a system), and yes, it certainly does seem to be a wonderful solution, with the small exception that here at least, each new pages requires me to log in again. A small price to pay I guess! Thanks for the suggestion.
It is mentioned that Linux is so stable, etc., and that people have incredible uptimes and such. I would love to have incredible uptimes and such, but being relatively new to Linux, I love to play with my kernel (doesn't that sound dirty?), which always requires a reboot. Does anyone have a way around this? This would be a really nice thing for me to be able to do, but from my limited system knowledge, it strikes me as an impossible dream....
Maybe add a feature to 2.3 or 2.5 to replace a kernel without rebooting ?
You got right to my question for people. I am under the understanding that reboots are necessary even in Linux, if you want to change kernels. Anyone know of a way around this? As a relatively new Linux user, I am enjoying playing around with my kernel and such, but it is such a pain to have to reboot all the time. It almost seems to me from my very limit system knowledge that I am asking the impossible? This would score as an incredibly cool feature in my book.
I hope someone starts another slashdot and keeps it less political.
And just how do you propose that they keep it less political? Seems to me that it is the readers that are bringing the heavy political slant, not the editors. If you don't allow the postings, then it isn't a/. is it? Just some thoughts as I get frustrated by people implying that these guys can really control what goes on here beyond shutting it down or making it not Slashdot.
Notice that they tie the web-server to the OS. So, if they call different Linux kernels different OSes (something the FTC would be hard pressed to deny, since the Linux community always talks about how fast Linux changes from version to version...), and each different version of UNIX, the numbers may actually work out.
I agree it is crap, but legally, I think it may be an easy statement to defend.
Have you forgotten that AMD beat Intel to market with enhanced 3D instructions by several months? The K6-2 was out WAY before the Pentium-III, which was simply a Pentium-II with Intel's version of AMD's 3D-Now! instructions. Happily, AMD didn't try and call this a whole new generation of processor like Intel did.
From what I have gathered reading the reviews on the web, it looks like AMD is having problems with the onboard cache at speeds over 450 MHz in the K6-3, so they are speeding up the K6-2 to keep in competition with the Celery chips from Intel. I hope that we will see the K6-3 chips speed up when the Athlon moves its L2 cache on die.
Athlons have steps of 50 MHz, there is no 733 MHz Athlon, it jumps from 700 MHz to 750 MHz. I saw an interesting review here (hope I can remember how to do links!):
Since I didn't see any mention that these parts were coming from Dresden, does anyone know if that means Austin has already switched over to the 0.18um process? I was under the impression that they were going to ramp up Dresden and then take Austin down for the switch.
OK, some thought streams I had while reading through the comments...nevermind, I am down to one thought since after moderating all morning I can't seem to put the others in words well:
It is very impressive to me how we always see people shouting and screaming that Intel has just as much of a monopoly over CPUs as Macrohard has over the OS market. I think this discussion pretty much dispells that idea. I haven't seen a single person crying about the EU suddenly being locked out of consumer computers until Intel changes its ways or the EU gives up this suggested embargo. Why not? Because AMD is there and with Dresden coming on line, should be able to supply them with chips (at one heck of a premium I would bet as supply and demand reign supreme). However, if we look at what would happen if they did ban Macrohard OSes, then you would have many people locked out of buying many other products (Winmodems come to mind immediately) that are tied to strickly Macrohard coded OSes. While many will say that people have the alternative of Linux, and that any company that ties themselves to Macrohard deserves to suffer, I can not agree with that. Linux is still a very difficult OS to operate effectively despite the fantastic progress that is being made (heck, my machine at home is still causing me some audio grief).
To me this shows that yes, AMD processors are substitutional products for Intel chips whereas Linux is not a substitutional product (in the economic sense) for Winblows products (which I think is a very good thing). I could go on with this for quite some time but will spare everyone the far reaching affects of AMD's capacity focusing on the EU and leaving us in the States high and dry....
*It does need to be noted, I have worked for both Intel and AMD, and I do tend to be biased on that front, but trying hard to not let that enter too heavily. However, I have no problems indicating my bias against Uncle Billy's company since I have not been directly affiliated with them*
Many thanks for the link to the pic. Yes, this is not at all what I had in mind, and they certainly deserve a new name for it (but I still don't get FinFET). The biggest problem I do see with this design is that it will only see large benefits (this is speculation, I haven't run any sims on a design like this) when you do have a very narrow width channel, and yes, that will kill your drive current. This will definitely limit the percentage of devices in a chip that would benefit from this design. However, I think it is still technologically important (not sure if my last post seemed to imply I felt otherwise). As for the concerns about needing more than one poly dep step, this should be a single dep, so I don't think this adds too much complexity to the process flow, only increases (dramatically I suspect) the alignment tolerances, which is already a pain at this scale.
OK, there sure seems to be a lot of confusion here. First off, this isn't extremely novel, dual/double channel devices (which is all this sounds to be) have been discussed for years in the literature, and have even been fabricated (but not with such a small gate LENGTH). As for wrapping the poly around the channel, it depends on their exact process, but it is very possible that they will do it with a single deposition, essentially growing the gate poly. You ask about going to "x-ray band" for masking, yes, they have to move away from optical lithography to get their 18 nm gate length, that is where the ebeam litho comes in. You seem confused about how the channel is "straddled" (an easy thing to be confused about by their wording). The "straddling" occurs on the top (where the gate usually lies) and the bottom, not the sides, so you can still have a WIDE MOSFET (when I say "gate," I mean gate vs. source vs. drain, not gate == transistor, they seem to use gate to mean both in the article, GRRR!) which will help your channel resistance. The resistance should actually be reduced because the whole channel will be inverted, limiting surface mobility effects (we have run some sims here to see a more than 2x current gain in dual vs. single gate designs). You ask about the increased gate capacitance (it seems you have seen a picture of their device, but I find none and would love to see images to make sure I have the correct mental image of what they have done), yes, you will increase you gate capacitance with the dual gate design (not due to thick poly, but due to increased area), but you should see more current gain as mentioned above. Finally, you ask whether this will be used for high-performance or circuit density, it should help with both, they tend to go hand in hand. I hope this helps you out a little bit, if you have more questions feel free to ask, and I will try and explain a bit better.
One more thing, how to they figure that 18 nm is about 100 atoms in length? Last I checked the atomic spacing in single crystal silicon (a wafer) was on the order of 5.43 Angstroms, so my handy dandy Rat Shack calculator says about 33 atoms would fit in a line 18 nm long....don't you just love PR press releases?!
One more thing (then I swear I will shut up), where did they come up with this stupid FinFET?? I would call this sucker a DGFET (Dual/Double Gate FET), but I guess they needed a new buzzword.
Just to join in the nit-picking, this isn't in a vacuum, but in a dielectric, so the speed of light won't be c necessarily. I don't think there is a huge difference (been a long time since I have done the calculations), but there is a difference, that's how the light is kept inside the cable!
How did they deliver the Microwave radiation? Any chance there were distracting harmonics only audible to the rats that might have polluted the study?
I don't think that would matter. My understanding was that the rats were exposed to the radiation after they had learned the location of the platform, not during the swimming lessons.
Where do you get the info that they are shipping 0.15 um? I am working in the long term R&D (grad school) and have not heard of anyone shipping below 0.18 um product. Please offer a source.
First off, check your core voltage. Going from 0.25 um to 0.18 um requires you to lower your voltage, possibly to a point your MB doesn't offer. Best of luck to you!
To the best of my knowledge, Dresden is not up yet, only dedicated. AMD isn't predicting product from Dresden until Q1 of 2000 according to their press releases. This means Intel has a few months of 0.18 um production while AMD ramps up their 0.18 um process. Sounds an awful lot like the move to 0.25 um when Intel had several months before AMD could ramp up Fab 25 (Austin). What should frighten Intel more is that AMD is pressing Intel's 0.18 um MHz hand with 0.25 um chips!
In terms of worrying about liquid oxygen forming and blowing things up, I would also worry about that liquid oxygen deciding to oxidize everything around it....say your connector pins, your board tracings, the steel case itself....
Oxygen is your best friend and your worst nightmare at the same time. Pretty neat if you ask me.
Slashdot Mirror
I for one would like to thank Microsoft for giving colleges this opportunity, and giving away a high-quality office suite and reasonably good development tools to people that need them.
I don't know how this applies to other Universities, but no, Macrohard is NOT giving away anything. Here, we have a deal with them that makes me sick. The university pays M$ a lump sum payment and is then allowed to distribute any M$ title for $5. Does that include packages or install floppies? No, just the CDs. Guess what, that lump sum payment is seven digits before the decimal. Yes, seven digits. Care to guess how many CDs the university estimated it "gave away" to students? It was on the order of 5,000. Oh boy, Macrohard is getting over $200/CD that the university "gives away!" Isn't that a great deal for the university? Isn't that a great deal for the 30,000+ students and/or taxpayers that did not purchase any of these discs but have to foot the bill??? In discussing the "success" of the deal, university officials said they were working on paying a larger fee to Macrohard to continue this program. Yes, this program may have been beneficial to you, yes, it was definitely beneficial to Macrohard, but I can't see how you can claim it helped anyone else. Tell me the university officials pushing this deal aren't corrupt!
Let's talk about other fun things newbies can do with linux: edit the /etc/ppp/chap-secrets file to login to their isp. Use
linuxconf (text-based) to get their connection online. Okay, I'm making this alittle hard - you could use kppp. =) After
you login you find out you can't - it keeps disconnected.
If you ask me, the "Mom Myth" is true for all OSes, I don't care which one you talk about. How many times have I had to explain to my mom how to print a file in Windows? How many times have I had to explain to her how to simply browse the web? If someone is not inclined to using computers to begin with, no matter how "user friendly" you make them, if the UI is not organized the way that person thinks, they will struggle. I think all of us have at one point or another struggled to get dial-up connections working...wrong IPs, wrong submasks, something somewhere isn't done correctly. Those problems will exist no matter what you do, unless you want to try and force a standard deployment on everyone.
I often like to think of it like this: Computers are tools, just like say a hammer, perhaps one of the least complicated tools around. If you are skilled with the tool, of course it is easy to use, but if you haven't ever really used the tool, you are going to wind up with a bruised thumb.
Besides, methanol's only uses are as a fuel and as a raw material for making other compounds. When you burn or
catalyze the Methanol as fuel, you get the CO2 back again, and when those other compounds are eventually consumed
or destroyed, you get the CO2 back again (and other noxious compounds, by the way).
Methanol actually does have other uses. It makes a great solvent for semiconductor processing (we use it all the time), of course, that is electrical grade methanol, which may not be so easy to get from these little enzyme buggers.
It is implied in the article that yes, Macrohard did contest that claim:
A Microsoft spokeswoman, however, disputes these perspectives, claiming that Microsoft's closed-source software is more secure than ever.
These perspectives M$ is disputing would be that Linux is more secure than any M$ offering. Now, whether the author was putting a spin that Macrohard didn't mean, I can't say, but do you really think Macrohard didn't want everyone to think that was what they were saying???
I'm glad to see the free software/open source concept being recognized like this, but I think it would have been nice if these experts had taken the time to look at other alternatives.
Well, we have to remember, we are probably talking about beaurocrats here, and therefore quick to jump on the buzzwords (Linux). Sure they are the experts, but they are still the government experts, which probably means they are your standard beaurocrats becuase I would bet the private sector would be able to offer a lot more for their expertice...but of course, I could be wrong, wouldn't be the first time, definitely won't be the last time for that!
All I can say is my karma has gone up from moderation, not from posted comments since I am usually way late to threads (after the moderators have come through). I think the system is working, or at least a step in the proper direction.
I protect my privacy by linking my Netscape cookies files to /dev/null (the infamous bit- bucket).
Netscape never knows the difference!
Well, I had never thought of doing that (being a relative newbie to Linux at home and having the powers to explore UNIX as a system), and yes, it certainly does seem to be a wonderful solution, with the small exception that here at least, each new pages requires me to log in again. A small price to pay I guess! Thanks for the suggestion.
It is mentioned that Linux is so stable, etc., and that people have incredible uptimes and such. I would love to have incredible uptimes and such, but being relatively new to Linux, I love to play with my kernel (doesn't that sound dirty?), which always requires a reboot. Does anyone have a way around this? This would be a really nice thing for me to be able to do, but from my limited system knowledge, it strikes me as an impossible dream....
Maybe add a feature to 2.3 or 2.5 to replace a kernel without rebooting ?
You got right to my question for people. I am under the understanding that reboots are necessary even in Linux, if you want to change kernels. Anyone know of a way around this? As a relatively new Linux user, I am enjoying playing around with my kernel and such, but it is such a pain to have to reboot all the time. It almost seems to me from my very limit system knowledge that I am asking the impossible? This would score as an incredibly cool feature in my book.
I hope someone starts another slashdot and keeps it less political.
/. is it? Just some thoughts as I get frustrated by people implying that these guys can really control what goes on here beyond shutting it down or making it not Slashdot.
And just how do you propose that they keep it less political? Seems to me that it is the readers that are bringing the heavy political slant, not the editors. If you don't allow the postings, then it isn't a
Notice that they tie the web-server to the OS. So, if they call different Linux kernels different OSes (something the FTC would be hard pressed to deny, since the Linux community always talks about how fast Linux changes from version to version...), and each different version of UNIX, the numbers may actually work out.
I agree it is crap, but legally, I think it may be an easy statement to defend.
Have you forgotten that AMD beat Intel to market with enhanced 3D instructions by several months? The K6-2 was out WAY before the Pentium-III, which was simply a Pentium-II with Intel's version of AMD's 3D-Now! instructions. Happily, AMD didn't try and call this a whole new generation of processor like Intel did.
From what I have gathered reading the reviews on the web, it looks like AMD is having problems with the onboard cache at speeds over 450 MHz in the K6-3, so they are speeding up the K6-2 to keep in competition with the Celery chips from Intel. I hope that we will see the K6-3 chips speed up when the Athlon moves its L2 cache on die.
Athlons have steps of 50 MHz, there is no 733 MHz Athlon, it jumps from 700 MHz to 750 MHz. I saw an interesting review here (hope I can remember how to do links!):
Anandtech Review
I do hope they get the L2 cache on die in a hurry, but of course, think of the transistor count after dumping in another 512 kB of memory!
Since I didn't see any mention that these parts were coming from Dresden, does anyone know if that means Austin has already switched over to the 0.18um process? I was under the impression that they were going to ramp up Dresden and then take Austin down for the switch.
OK, some thought streams I had while reading through the comments...nevermind, I am down to one thought since after moderating all morning I can't seem to put the others in words well:
It is very impressive to me how we always see people shouting and screaming that Intel has just as much of a monopoly over CPUs as Macrohard has over the OS market. I think this discussion pretty much dispells that idea. I haven't seen a single person crying about the EU suddenly being locked out of consumer computers until Intel changes its ways or the EU gives up this suggested embargo. Why not? Because AMD is there and with Dresden coming on line, should be able to supply them with chips (at one heck of a premium I would bet as supply and demand reign supreme). However, if we look at what would happen if they did ban Macrohard OSes, then you would have many people locked out of buying many other products (Winmodems come to mind immediately) that are tied to strickly Macrohard coded OSes. While many will say that people have the alternative of Linux, and that any company that ties themselves to Macrohard deserves to suffer, I can not agree with that. Linux is still a very difficult OS to operate effectively despite the fantastic progress that is being made (heck, my machine at home is still causing me some audio grief).
To me this shows that yes, AMD processors are substitutional products for Intel chips whereas Linux is not a substitutional product (in the economic sense) for Winblows products (which I think is a very good thing). I could go on with this for quite some time but will spare everyone the far reaching affects of AMD's capacity focusing on the EU and leaving us in the States high and dry....
*It does need to be noted, I have worked for both Intel and AMD, and I do tend to be biased on that front, but trying hard to not let that enter too heavily. However, I have no problems indicating my bias against Uncle Billy's company since I have not been directly affiliated with them*
Many thanks for the link to the pic. Yes, this is not at all what I had in mind, and they certainly deserve a new name for it (but I still don't get FinFET). The biggest problem I do see with this design is that it will only see large benefits (this is speculation, I haven't run any sims on a design like this) when you do have a very narrow width channel, and yes, that will kill your drive current. This will definitely limit the percentage of devices in a chip that would benefit from this design. However, I think it is still technologically important (not sure if my last post seemed to imply I felt otherwise). As for the concerns about needing more than one poly dep step, this should be a single dep, so I don't think this adds too much complexity to the process flow, only increases (dramatically I suspect) the alignment tolerances, which is already a pain at this scale.
OK, there sure seems to be a lot of confusion here. First off, this isn't extremely novel, dual/double channel devices (which is all this sounds to be) have been discussed for years in the literature, and have even been fabricated (but not with such a small gate LENGTH). As for wrapping the poly around the channel, it depends on their exact process, but it is very possible that they will do it with a single deposition, essentially growing the gate poly. You ask about going to "x-ray band" for masking, yes, they have to move away from optical lithography to get their 18 nm gate length, that is where the ebeam litho comes in. You seem confused about how the channel is "straddled" (an easy thing to be confused about by their wording). The "straddling" occurs on the top (where the gate usually lies) and the bottom, not the sides, so you can still have a WIDE MOSFET (when I say "gate," I mean gate vs. source vs. drain, not gate == transistor, they seem to use gate to mean both in the article, GRRR!) which will help your channel resistance. The resistance should actually be reduced because the whole channel will be inverted, limiting surface mobility effects (we have run some sims here to see a more than 2x current gain in dual vs. single gate designs). You ask about the increased gate capacitance (it seems you have seen a picture of their device, but I find none and would love to see images to make sure I have the correct mental image of what they have done), yes, you will increase you gate capacitance with the dual gate design (not due to thick poly, but due to increased area), but you should see more current gain as mentioned above. Finally, you ask whether this will be used for high-performance or circuit density, it should help with both, they tend to go hand in hand. I hope this helps you out a little bit, if you have more questions feel free to ask, and I will try and explain a bit better.
One more thing, how to they figure that 18 nm is about 100 atoms in length? Last I checked the atomic spacing in single crystal silicon (a wafer) was on the order of 5.43 Angstroms, so my handy dandy Rat Shack calculator says about 33 atoms would fit in a line 18 nm long....don't you just love PR press releases?!
One more thing (then I swear I will shut up), where did they come up with this stupid FinFET?? I would call this sucker a DGFET (Dual/Double Gate FET), but I guess they needed a new buzzword.
299,792,458m/s
Just to join in the nit-picking, this isn't in a vacuum, but in a dielectric, so the speed of light won't be c necessarily. I don't think there is a huge difference (been a long time since I have done the calculations), but there is a difference, that's how the light is kept inside the cable!
How did they deliver the Microwave radiation? Any chance there were distracting harmonics only audible to the rats that might have polluted the study?
I don't think that would matter. My understanding was that the rats were exposed to the radiation after they had learned the location of the platform, not during the swimming lessons.
Looks like lots of people are getting logged in. It did take a bit of time, but when doesn't /. take a bit of time these days?
Where do you get the info that they are shipping 0.15 um? I am working in the long term R&D (grad school) and have not heard of anyone shipping below 0.18 um product. Please offer a source.
First off, check your core voltage. Going from 0.25 um to 0.18 um requires you to lower your voltage, possibly to a point your MB doesn't offer. Best of luck to you!
To the best of my knowledge, Dresden is not up yet, only dedicated. AMD isn't predicting product from Dresden until Q1 of 2000 according to their press releases. This means Intel has a few months of 0.18 um production while AMD ramps up their 0.18 um process. Sounds an awful lot like the move to 0.25 um when Intel had several months before AMD could ramp up Fab 25 (Austin). What should frighten Intel more is that AMD is pressing Intel's 0.18 um MHz hand with 0.25 um chips!
In terms of worrying about liquid oxygen forming and blowing things up, I would also worry about that liquid oxygen deciding to oxidize everything around it....say your connector pins, your board tracings, the steel case itself....
Oxygen is your best friend and your worst nightmare at the same time. Pretty neat if you ask me.