From the second article to utterly crush Nr9's assessment of what Dr Myerson said last week
Quote:
Transistors will continue to shrink, Meyerson said.
IBM, Intel and other chip companies are steadfast in their determination to shrink process technology generations every two to three years. But recent chipmaking innovations such as strained silicon and silicon-on-insulator technology will grow more important with each successive process generation shrink and technologies such as virtualisation will become widespread, he said.
"There are trajectories forward that are enormously promising," Meyerson said.
As anyone who can see or who has followed any of this, Dr Myerson has been saying the same thing for near a year now, His most recent keynote address just put it in front of the largest audience to date. He has not said there are no more Hz anymore, he has not said scaling is dead; he has said a free lunch path described by Moore's law is no longer realistic or valid. Costs to follow that path are beginning to escalate and in the near future (he didn't say today or yesterday or last spring) it will make better business sense to look for performance gains in other areas.
IBM would not be attempting continued process shrinks if there were no performance gains to be expected. Period. Nr9's primary and possible only source has directly contradicted the main premise of Nr9's statements, that there are no more scaling gains to be made-ever.
Pretty simple really. That leaves us with the same set of realities we have known all along. It is becoming more difficult to scale and industry must begin preparations for the point scaling via feature size becomes prohibitively expensive. Nothing new there, and nothing at all like the errantly definitive statements we have been refuting all 7 pages so far.
Just stop Nr9. You're wrong.
We consistently have technical opinions and both primary (interview) and secondary (coverage of the keynote) sources that corroborate that. You SIGARCH bluff was called as well with nothing at all to help you case there. From here on out I think everyone can unconditionally and safely accept that scaling didn't die in March 2004 with the advent of the 970FX.
look, he said that cuz he doesnt want to scare away investors and possibly to scare competitors. the fact is, scaling is dead. on the article you linked is called "Chip industry needs a 'Plan B" process technology-wise, plan B does not exist, plan B is multicore. CMOS is at a dead end just like bipolar was a couple of years ago. problem is, there is no real alternative in sight.
anyone actually click on the article that airsluf links and look at it and see what its talking about?
"The days of relying on shrinking transistors to achieve performance gains were over, and the chip industry needed to enter a new era of innovation where system-level features were just as important as thinner transistor gates, vice-president and chief technologist at IBM's Systems and Technology Group, Bernie Meyerson, has told the Fall Processor Forum in a keynote address."
"Chips were now so small that atom-level defects on a silicon chip could cause power leakage up to 100 times the normal level, he said."
What I was trying to get at is that some people posting here seem to believe that their is no room for improvement at 90nm and that process scaling is dead.
Careful, you're over-generalizing the statements -- frequency scaling has hit a wall. That doesn't mean that smaller transistors aren't possible, nor does it mean that there isn't benefit to be gained from them. A 65nm transistor at 2 GHz is still more useful than a 130nm or 90nm transistor at 2 GHz.
frequency scaling has hit a wall. That doesn't mean that smaller transistors aren't possible, nor does it mean that there isn't benefit to be gained from them. A 65nm transistor at 2 GHz is still more useful than a 130nm or 90nm transistor at 2 GHz.
For the most part that's what I got from it as well.
Well the general is true as much as the specific. The idea that frequency scaling has hit the wall though simply doesn't hold water as there are examples, available now, that indicate that 90nm has been successful in allowing higher clock rates.
At 65nm I fully expect that at least a few manufactures will be successfull scaling clock rate at that node. Maybe that is not the direction that IBM is going but it is pretty clear that the rest of the world is still interested in improving single core performance.
I totally agree with you on the usefullness of the smaller transistors. I think we will start to see many more releases of SoC designs and much more in the way of intelligence in peripherals. The problem still reverts to the thought that dual cores, or even more for that matter, only offer so much in the way of performance increases. There is still a need to increase core performance and all those extra transistors can play a role in that performance increase. Clock rate increase though will not crop out ot the picture.
I'm not usually an optimistic person by any means. Maybe my opinion will change when I start to hear a larger crowd of scientist saying they don't see a solution in site. Right now that isn't the case.
Dave
Quote:
Originally posted by Programmer
Careful, you're over-generalizing the statements -- frequency scaling has hit a wall. That doesn't mean that smaller transistors aren't possible, nor does it mean that there isn't benefit to be gained from them. A 65nm transistor at 2 GHz is still more useful than a 130nm or 90nm transistor at 2 GHz.
Ahhh, finally a quote from our esteemed doomsayer! Huzzah! Yes, that was the first quasi-paraphrased paragraph of the second link's article. A rather nice run-in to the rest of the article that illustrates what could happen if a Plan-B wasn't contemplated.
Notice the tense in that paragraph Nr9, oh wait a second, you don't believe writing and communicating are relevant in technical engineers anymore. Well how about reading? 'Cause you just wrapped yourself around a past tense allegorical phrase and took it to be current day gospel. Maybe someone needs to brush up on their english skills a bit so they can make a proper analysis of what they read.
Careful, you're over-generalizing the statements -- frequency scaling has hit a wall. That doesn't mean that smaller transistors aren't possible, nor does it mean that there isn't benefit to be gained from them. A 65nm transistor at 2 GHz is still more useful than a 130nm or 90nm transistor at 2 GHz.
I disagree here. You are taking a relatively reasonable view, but it takes even more than reading between the lines to separate Nr9's statements from the absolute "scaling is dead" phrase, not "frequency scaling is dead". Heck he did it again in the post just above yours! I also give no quarter for "sloppy english", if he was in the field and spoke or wrote to his senior engineers in such a manner he would have a max seniority of days. There has been plenty of opportunity for his phraseology to catch up given the flak he has taken, but it doesn't. That leads me to surmise other things which don't need further mention here.
I think for the most part, most of us agree with you, but a small and immensely important difference in specific phraseology started this thread off, and has been repeated with great vigor. And that phraseology doesn't agree with yours, it's absolute, no more Hz, ever... We both know that's not true, the Hz increase rate is slowing but has not forever stopped.
Forgive me for not reading every page of this but unless Apple suddenly stops making Powermacs, me thinks they will get faster. I mean, powermacs have increased relatively steadily in speed since they were introduced in 1994.
I guess you mean that the 90nm 970 will not increase in hz. I disagree with this. While very hot, it runs under its rated temp...and there are process tweaks that haven't yet be implemented e.g. SOI (irc).
Even if they don't increase in hz, slapping another core on will speed things up no?
Ahhh, finally a quote from our esteemed doomsayer! Huzzah! Yes, that was the first quasi-paraphrased paragraph of the second link's article. A rather nice run-in to the rest of the article that illustrates what could happen if a Plan-B wasn't contemplated.
Notice the tense in that paragraph Nr9, oh wait a second, you don't believe writing and communicating are relevant in technical engineers anymore. Well how about reading? 'Cause you just wrapped yourself around a past tense allegorical phrase and took it to be current day gospel. Maybe someone needs to brush up on their english skills a bit so they can make a proper analysis of what they read.
Please, don't run with scissors.
plan b is multicore, not process improvment
i read the bestest, you obviously dont know how to read. when you read an article, the first thing you look is the title. what does the title of your quoted source. go back to grammar school. most of your sentences are poorly constructed and grammaticaly incorrect.
"Notice the tense in that paragraph Nr9, oh wait a second, you don't believe writing and communicating are relevant in technical engineers anymore. "
what the hell is "oh wait a second"
"Huzzah!"
nice english, i dont see it idctionary
" A rather nice run-in to the rest of the article that illustrates what could happen if a Plan-B wasn't contemplated. "
setence fragment
"Well how about reading?"
missing coma
"'Cause you just wrapped yourself around a past tense allegorical phrase and took it to be current day gospel. "
"Chip industry needs a 'Plan B" process technology-wise, plan B does not exist, plan B is multicore.
Apart from the fact what you just wrote is contradictory and a non-sequitur the articles specifically state the answer isn't multi-core alone. The problem is every company intends to lower frequencies on launching multi-core chips to get under the thermal limits. To some extent that's ok but again as you increase cores there's a point of diminishing returns unless you do something else. You either have to increase the frequency or you have to find smarter ways of doing the same thing.
On an aside IBM has been talking system level approaches ever since the POWER4 was released, since they believe themselves in a unique position to provide it, so this is nothing new. Plan B is smarter use of resources, new technologies and new materials.
On an aside that wasn't a sentence fragment, which I doubt you could identify yourself, and you'd find Huzzah is in most dictionaries, even pocket ones. It's an exclamation of delight. 'Cause is actually a word, it's the informal version of because. It could also be considered an abbreviated one but then so is "it's" so that wouldn't make it not a word either.
I would strongly suggest you don't attempt to build any further arguments around your supposed IBM employment or communicative skills as it has become quite clear that you lack both. If you want to make an argument, although you may find that difficult unless you use words to develop and outline your points, based around the technical side of microprocessor manufacture, from what you know, then by all means go crazy. I expect you'll have difficulty outlining much beyond your next insult though.
On a very general aside there have been transistors designed that can switch up to the terahertz rate. Obviously this doesn't directly translate to processor frequencies but scaling is ahead. It's just going to be slower and not a sole focus for performance enhancement anymore.
Are you serious about that statement? How can you attack other posts when yours have been horrid through out this whole thread. Like the above post claims, you have failed to prove any points related to the creation of this thread. Either your communication skills aren't up to par, or you are debating using emotional reasoning. Also quit using assumptions and biases; like you have throughout these posts. Define your terms and elaborate on each term. These suggestions I have made are very basic points using critical thinking and communication.
You have had many chances to prove your statement. So please quit trying to justify your topic post. At this point you have lost all credibility.
Only so many cores can be added to a chip. Clock rate will always be progressed on... even if it doesn't bring any major performance gains to the table. All aspects of cpu's will be revamped and or completely changed in the coming years. There isn't any aspect of a processor system that is perfect. Sure they are using some amazing technologies, but those technologies will always be progressed on by engineers.
Time for a new thread, I am done with this one. Have fun wasting the rest of your time trying to convince people... especially when you'll be proved wrong in about seven to ten months.
Are you serious about that statement? How can you attack other posts when yours have been horrid through out this whole thread. Like the above post claims, you have failed to prove any points related to the creation of this thread. Either your communication skills aren't up to par, or you are debating using emotional reasoning. Also quit using assumptions and biases; like you have throughout these posts. Define your terms and elaborate on each term. These suggestions I have made are very basic points using critical thinking and communication.
You have had many chances to prove your statement. So please quit trying to justify your topic post. At this point you have lost all credibility.
Only so many cores can be added to a chip. Clock rate will always be progressed on... even if it doesn't bring any major performance gains to the table. All aspects of cpu's will be revamped and or completely changed in the coming years. There isn't any aspect of a processor system that is perfect. Sure they are using some amazing technologies, but those technologies will always be progressed on by engineers.
Time for a new thread, I am done with this one. Have fun wasting the rest of your time trying to convince people... especially when you'll be proved wrong in about seven to ten months.
"Also quit using assumptions and biases; like you have throughout these posts. "
missing comma and use a comma instead of a semicolon
"These suggestions I have made are very basic points using critical thinking and communication. "
ambiguous modifier -> "using critical thinking and communcation" can refer to "points" or "have made"
"So please quit trying to justify your topic post. At this point you have lost all credibility. "
both sentences have missing commas
"Clock rate will always be progressed on... even if it doesn't bring any major performance gains to the table. "
use a comma, clock rate will progress, not will be progressed, active voice is good
"All aspects of cpu's will be revamped and or completely changed in the coming years."
capitalize your acronyms, what is "and or"
"Sure they are using some amazing technologies, but those technologies will always be progressed on by engineers. "
I disagree here. You are taking a relatively reasonable view, but it takes even more than reading between the lines to separate Nr9's statements from the absolute "scaling is dead" phrase, not "frequency scaling is dead". Heck he did it again in the post just above yours!
Extremist views are virtually always wrong, so I tend to ignore them or water them down automatically. Nr9, in particular, seems to be here to incite reactions rather than have level-headed discussions. Since that has clearly derailed the thread's original topic (again), we might as well just close this one now...
On an aside that wasn't a sentence fragment, which I doubt you could identify yourself, and you'd find Huzzah is in most dictionaries, even pocket ones. It's an exclamation of delight. 'Cause is actually a word, it's the informal version of because. It could also be considered an abbreviated one but then so is "it's" so that wouldn't make it not a word either.
that was definitely a sentence fragment
" A rather nice run-in to the rest of the article that illustrates what could happen if a Plan-B wasn't contemplated."
Before the Lock! - I was just wondering how they test chips to get the speed rating.
I'm imagining some sort of hydraulically operated heatsink or watercooler that comes down onto a chip, while on the bottom is a kind of socket to provide the power, and then they run it increasing the power until it becomes unstable. All automated of course.
Apart from the fact what you just wrote is contradictory and a non-sequitur the articles specifically state the answer isn't multi-core alone. The problem is every company intends to lower frequencies on launching multi-core chips to get under the thermal limits. To some extent that's ok but again as you increase cores there's a point of diminishing returns unless you do something else. You either have to increase the frequency or you have to find smarter ways of doing the same thing.
On an aside IBM has been talking system level approaches ever since the POWER4 was released, since they believe themselves in a unique position to provide it, so this is nothing new. Plan B is smarter use of resources, new technologies and new materials.
On an aside that wasn't a sentence fragment, which I doubt you could identify yourself, and you'd find Huzzah is in most dictionaries, even pocket ones. It's an exclamation of delight. 'Cause is actually a word, it's the informal version of because. It could also be considered an abbreviated one but then so is "it's" so that wouldn't make it not a word either.
I would strongly suggest you don't attempt to build any further arguments around your supposed IBM employment or communicative skills as it has become quite clear that you lack both. If you want to make an argument, although you may find that difficult unless you use words to develop and outline your points, based around the technical side of microprocessor manufacture, from what you know, then by all means go crazy. I expect you'll have difficulty outlining much beyond your next insult though.
On a very general aside there have been transistors designed that can switch up to the terahertz rate. Obviously this doesn't directly translate to processor frequencies but scaling is ahead. It's just going to be slower and not a sole focus for performance enhancement anymore.
"Apart from the fact what you just wrote is contradictory and a non-sequitur the articles specifically state the answer isn't multi-core alone. "
missing comma
"The problem is every company .."
insert a "that" between is and every
"To some extent that's ok but again as you increase cores there's a point of diminishing returns unless you do something else. "
missing coma, capitalize ok
"On an aside IBM has been talking system level approaches ever since the POWER4 was released, since they believe themselves in a unique position to provide it, so this is nothing new. "
missing comma after aside, you can't belive someone in a position, add a verb, "this is nothing new"<-- what does "this" refer to
"On an aside that wasn't a sentence fragment, which I doubt you could identify yourself, and you'd find Huzzah is in most dictionaries, even pocket ones. "
Comments
John Spooner, CNET News June 8, 04
From the first article to deflate Nr9's assessment of what Dr Myerson has been saying for quite some time
(Reporter)Does that mean performance stops?
(Myerson)It means that the rate of performance enhancement is becoming impacted.
Tom Krazit, IDG News Svc Oct 6, 04
From the second article to utterly crush Nr9's assessment of what Dr Myerson said last week
Transistors will continue to shrink, Meyerson said.
IBM, Intel and other chip companies are steadfast in their determination to shrink process technology generations every two to three years. But recent chipmaking innovations such as strained silicon and silicon-on-insulator technology will grow more important with each successive process generation shrink and technologies such as virtualisation will become widespread, he said.
"There are trajectories forward that are enormously promising," Meyerson said.
As anyone who can see or who has followed any of this, Dr Myerson has been saying the same thing for near a year now, His most recent keynote address just put it in front of the largest audience to date. He has not said there are no more Hz anymore, he has not said scaling is dead; he has said a free lunch path described by Moore's law is no longer realistic or valid. Costs to follow that path are beginning to escalate and in the near future (he didn't say today or yesterday or last spring) it will make better business sense to look for performance gains in other areas.
IBM would not be attempting continued process shrinks if there were no performance gains to be expected. Period. Nr9's primary and possible only source has directly contradicted the main premise of Nr9's statements, that there are no more scaling gains to be made-ever.
Pretty simple really. That leaves us with the same set of realities we have known all along. It is becoming more difficult to scale and industry must begin preparations for the point scaling via feature size becomes prohibitively expensive. Nothing new there, and nothing at all like the errantly definitive statements we have been refuting all 7 pages so far.
Just stop Nr9. You're wrong.
We consistently have technical opinions and both primary (interview) and secondary (coverage of the keynote) sources that corroborate that. You SIGARCH bluff was called as well with nothing at all to help you case there. From here on out I think everyone can unconditionally and safely accept that scaling didn't die in March 2004 with the advent of the 970FX.
anyone actually click on the article that airsluf links and look at it and see what its talking about?
"The days of relying on shrinking transistors to achieve performance gains were over, and the chip industry needed to enter a new era of innovation where system-level features were just as important as thinner transistor gates, vice-president and chief technologist at IBM's Systems and Technology Group, Bernie Meyerson, has told the Fall Processor Forum in a keynote address."
"Chips were now so small that atom-level defects on a silicon chip could cause power leakage up to 100 times the normal level, he said."
Originally posted by wizard69
What I was trying to get at is that some people posting here seem to believe that their is no room for improvement at 90nm and that process scaling is dead.
Careful, you're over-generalizing the statements -- frequency scaling has hit a wall. That doesn't mean that smaller transistors aren't possible, nor does it mean that there isn't benefit to be gained from them. A 65nm transistor at 2 GHz is still more useful than a 130nm or 90nm transistor at 2 GHz.
Originally posted by Programmer
frequency scaling has hit a wall. That doesn't mean that smaller transistors aren't possible, nor does it mean that there isn't benefit to be gained from them. A 65nm transistor at 2 GHz is still more useful than a 130nm or 90nm transistor at 2 GHz.
For the most part that's what I got from it as well.
themselves. A good place to start is the International
Technology Roadmap for Semiconductors.
http://public.itrs.net/
The executive summary for starters. Tables 4c and 4d
being particularly interesting.
http://public.itrs.net/Files/2003ITRS/ExecSum2003.pdf
Unless you buy the absolutist 'it's impossible' argument
it seems that most of the major players feel that
significant progress is possible. Perhaps the 2004
roadmap will say otherwise once published.
At 65nm I fully expect that at least a few manufactures will be successfull scaling clock rate at that node. Maybe that is not the direction that IBM is going but it is pretty clear that the rest of the world is still interested in improving single core performance.
I totally agree with you on the usefullness of the smaller transistors. I think we will start to see many more releases of SoC designs and much more in the way of intelligence in peripherals. The problem still reverts to the thought that dual cores, or even more for that matter, only offer so much in the way of performance increases. There is still a need to increase core performance and all those extra transistors can play a role in that performance increase. Clock rate increase though will not crop out ot the picture.
I'm not usually an optimistic person by any means. Maybe my opinion will change when I start to hear a larger crowd of scientist saying they don't see a solution in site. Right now that isn't the case.
Dave
Originally posted by Programmer
Careful, you're over-generalizing the statements -- frequency scaling has hit a wall. That doesn't mean that smaller transistors aren't possible, nor does it mean that there isn't benefit to be gained from them. A 65nm transistor at 2 GHz is still more useful than a 130nm or 90nm transistor at 2 GHz.
Notice the tense in that paragraph Nr9, oh wait a second, you don't believe writing and communicating are relevant in technical engineers anymore. Well how about reading? 'Cause you just wrapped yourself around a past tense allegorical phrase and took it to be current day gospel. Maybe someone needs to brush up on their english skills a bit so they can make a proper analysis of what they read.
Please, don't run with scissors.
Originally posted by Programmer
Careful, you're over-generalizing the statements -- frequency scaling has hit a wall. That doesn't mean that smaller transistors aren't possible, nor does it mean that there isn't benefit to be gained from them. A 65nm transistor at 2 GHz is still more useful than a 130nm or 90nm transistor at 2 GHz.
I disagree here. You are taking a relatively reasonable view, but it takes even more than reading between the lines to separate Nr9's statements from the absolute "scaling is dead" phrase, not "frequency scaling is dead". Heck he did it again in the post just above yours! I also give no quarter for "sloppy english", if he was in the field and spoke or wrote to his senior engineers in such a manner he would have a max seniority of days. There has been plenty of opportunity for his phraseology to catch up given the flak he has taken, but it doesn't. That leads me to surmise other things which don't need further mention here.
I think for the most part, most of us agree with you, but a small and immensely important difference in specific phraseology started this thread off, and has been repeated with great vigor. And that phraseology doesn't agree with yours, it's absolute, no more Hz, ever... We both know that's not true, the Hz increase rate is slowing but has not forever stopped.
I guess you mean that the 90nm 970 will not increase in hz. I disagree with this. While very hot, it runs under its rated temp...and there are process tweaks that haven't yet be implemented e.g. SOI (irc).
Even if they don't increase in hz, slapping another core on will speed things up no?
Originally posted by AirSluf
Ahhh, finally a quote from our esteemed doomsayer! Huzzah! Yes, that was the first quasi-paraphrased paragraph of the second link's article. A rather nice run-in to the rest of the article that illustrates what could happen if a Plan-B wasn't contemplated.
Notice the tense in that paragraph Nr9, oh wait a second, you don't believe writing and communicating are relevant in technical engineers anymore. Well how about reading? 'Cause you just wrapped yourself around a past tense allegorical phrase and took it to be current day gospel. Maybe someone needs to brush up on their english skills a bit so they can make a proper analysis of what they read.
Please, don't run with scissors.
plan b is multicore, not process improvment
i read the bestest, you obviously dont know how to read. when you read an article, the first thing you look is the title. what does the title of your quoted source. go back to grammar school. most of your sentences are poorly constructed and grammaticaly incorrect.
"Notice the tense in that paragraph Nr9, oh wait a second, you don't believe writing and communicating are relevant in technical engineers anymore. "
what the hell is "oh wait a second"
"Huzzah!"
nice english, i dont see it idctionary
" A rather nice run-in to the rest of the article that illustrates what could happen if a Plan-B wasn't contemplated. "
setence fragment
"Well how about reading?"
missing coma
"'Cause you just wrapped yourself around a past tense allegorical phrase and took it to be current day gospel. "
'cause is not a word
Originally posted by Nr9
"Chip industry needs a 'Plan B" process technology-wise, plan B does not exist, plan B is multicore.
Apart from the fact what you just wrote is contradictory and a non-sequitur the articles specifically state the answer isn't multi-core alone. The problem is every company intends to lower frequencies on launching multi-core chips to get under the thermal limits. To some extent that's ok but again as you increase cores there's a point of diminishing returns unless you do something else. You either have to increase the frequency or you have to find smarter ways of doing the same thing.
On an aside IBM has been talking system level approaches ever since the POWER4 was released, since they believe themselves in a unique position to provide it, so this is nothing new. Plan B is smarter use of resources, new technologies and new materials.
On an aside that wasn't a sentence fragment, which I doubt you could identify yourself, and you'd find Huzzah is in most dictionaries, even pocket ones. It's an exclamation of delight. 'Cause is actually a word, it's the informal version of because. It could also be considered an abbreviated one but then so is "it's" so that wouldn't make it not a word either.
I would strongly suggest you don't attempt to build any further arguments around your supposed IBM employment or communicative skills as it has become quite clear that you lack both. If you want to make an argument, although you may find that difficult unless you use words to develop and outline your points, based around the technical side of microprocessor manufacture, from what you know, then by all means go crazy. I expect you'll have difficulty outlining much beyond your next insult though.
On a very general aside there have been transistors designed that can switch up to the terahertz rate. Obviously this doesn't directly translate to processor frequencies but scaling is ahead. It's just going to be slower and not a sole focus for performance enhancement anymore.
Originally posted by Nr9
"Well how about reading?"
missing coma
Are you serious about that statement? How can you attack other posts when yours have been horrid through out this whole thread. Like the above post claims, you have failed to prove any points related to the creation of this thread. Either your communication skills aren't up to par, or you are debating using emotional reasoning. Also quit using assumptions and biases; like you have throughout these posts. Define your terms and elaborate on each term. These suggestions I have made are very basic points using critical thinking and communication.
You have had many chances to prove your statement. So please quit trying to justify your topic post. At this point you have lost all credibility.
Only so many cores can be added to a chip. Clock rate will always be progressed on... even if it doesn't bring any major performance gains to the table. All aspects of cpu's will be revamped and or completely changed in the coming years. There isn't any aspect of a processor system that is perfect. Sure they are using some amazing technologies, but those technologies will always be progressed on by engineers.
Time for a new thread, I am done with this one. Have fun wasting the rest of your time trying to convince people... especially when you'll be proved wrong in about seven to ten months.
Originally posted by Nr9
"Notice the tense in that paragraph Nr9, oh wait a second, you don't believe writing and communicating are relevant in technical engineers anymore. "
what the hell is "oh wait a second"
"Huzzah!"
nice english, i dont see it idctionary
" A rather nice run-in to the rest of the article that illustrates what could happen if a Plan-B wasn't contemplated. "
setence fragment
"Well how about reading?"
missing coma
"'Cause you just wrapped yourself around a past tense allegorical phrase and took it to be current day gospel. "
'cause is not a word
Every single one of your grammar "critiques" has a misspelling or punctuation error of its own.
Huzzah is a word.
You are beyond pathetic Nr9.
Originally posted by emig647
Are you serious about that statement? How can you attack other posts when yours have been horrid through out this whole thread. Like the above post claims, you have failed to prove any points related to the creation of this thread. Either your communication skills aren't up to par, or you are debating using emotional reasoning. Also quit using assumptions and biases; like you have throughout these posts. Define your terms and elaborate on each term. These suggestions I have made are very basic points using critical thinking and communication.
You have had many chances to prove your statement. So please quit trying to justify your topic post. At this point you have lost all credibility.
Only so many cores can be added to a chip. Clock rate will always be progressed on... even if it doesn't bring any major performance gains to the table. All aspects of cpu's will be revamped and or completely changed in the coming years. There isn't any aspect of a processor system that is perfect. Sure they are using some amazing technologies, but those technologies will always be progressed on by engineers.
Time for a new thread, I am done with this one. Have fun wasting the rest of your time trying to convince people... especially when you'll be proved wrong in about seven to ten months.
"Also quit using assumptions and biases; like you have throughout these posts. "
missing comma and use a comma instead of a semicolon
"These suggestions I have made are very basic points using critical thinking and communication. "
ambiguous modifier -> "using critical thinking and communcation" can refer to "points" or "have made"
"So please quit trying to justify your topic post. At this point you have lost all credibility. "
both sentences have missing commas
"Clock rate will always be progressed on... even if it doesn't bring any major performance gains to the table. "
use a comma, clock rate will progress, not will be progressed, active voice is good
"All aspects of cpu's will be revamped and or completely changed in the coming years."
capitalize your acronyms, what is "and or"
"Sure they are using some amazing technologies, but those technologies will always be progressed on by engineers. "
missia comma, dont use the passive
Originally posted by AirSluf
I disagree here. You are taking a relatively reasonable view, but it takes even more than reading between the lines to separate Nr9's statements from the absolute "scaling is dead" phrase, not "frequency scaling is dead". Heck he did it again in the post just above yours!
Extremist views are virtually always wrong, so I tend to ignore them or water them down automatically. Nr9, in particular, seems to be here to incite reactions rather than have level-headed discussions. Since that has clearly derailed the thread's original topic (again), we might as well just close this one now...
Originally posted by Telomar
On an aside that wasn't a sentence fragment, which I doubt you could identify yourself, and you'd find Huzzah is in most dictionaries, even pocket ones. It's an exclamation of delight. 'Cause is actually a word, it's the informal version of because. It could also be considered an abbreviated one but then so is "it's" so that wouldn't make it not a word either.
that was definitely a sentence fragment
" A rather nice run-in to the rest of the article that illustrates what could happen if a Plan-B wasn't contemplated."
that illustrates is a verbal describing article,
this is definitely fragment
I'm imagining some sort of hydraulically operated heatsink or watercooler that comes down onto a chip, while on the bottom is a kind of socket to provide the power, and then they run it increasing the power until it becomes unstable. All automated of course.
Originally posted by Telomar
Apart from the fact what you just wrote is contradictory and a non-sequitur the articles specifically state the answer isn't multi-core alone. The problem is every company intends to lower frequencies on launching multi-core chips to get under the thermal limits. To some extent that's ok but again as you increase cores there's a point of diminishing returns unless you do something else. You either have to increase the frequency or you have to find smarter ways of doing the same thing.
On an aside IBM has been talking system level approaches ever since the POWER4 was released, since they believe themselves in a unique position to provide it, so this is nothing new. Plan B is smarter use of resources, new technologies and new materials.
On an aside that wasn't a sentence fragment, which I doubt you could identify yourself, and you'd find Huzzah is in most dictionaries, even pocket ones. It's an exclamation of delight. 'Cause is actually a word, it's the informal version of because. It could also be considered an abbreviated one but then so is "it's" so that wouldn't make it not a word either.
I would strongly suggest you don't attempt to build any further arguments around your supposed IBM employment or communicative skills as it has become quite clear that you lack both. If you want to make an argument, although you may find that difficult unless you use words to develop and outline your points, based around the technical side of microprocessor manufacture, from what you know, then by all means go crazy. I expect you'll have difficulty outlining much beyond your next insult though.
On a very general aside there have been transistors designed that can switch up to the terahertz rate. Obviously this doesn't directly translate to processor frequencies but scaling is ahead. It's just going to be slower and not a sole focus for performance enhancement anymore.
"Apart from the fact what you just wrote is contradictory and a non-sequitur the articles specifically state the answer isn't multi-core alone. "
missing comma
"The problem is every company .."
insert a "that" between is and every
"To some extent that's ok but again as you increase cores there's a point of diminishing returns unless you do something else. "
missing coma, capitalize ok
"On an aside IBM has been talking system level approaches ever since the POWER4 was released, since they believe themselves in a unique position to provide it, so this is nothing new. "
missing comma after aside, you can't belive someone in a position, add a verb, "this is nothing new"<-- what does "this" refer to
"On an aside that wasn't a sentence fragment, which I doubt you could identify yourself, and you'd find Huzzah is in most dictionaries, even pocket ones. "
missing comma after aside
and on and on....
man your english sucks, go back o grammar school
Will someone PLEASE lock this monstrosity of a trainwreck?