Article is quite too enthuastic about x86 pushing to other domains.
Lets make it clear.Modern x86 decoder is more complex than entire simple single issue risc processor and it consumes more power. Yes, thats SINGLE unit in the front end of pipeline, that risc processors do not need.
About those "risc ops" which x86 instructions are translated as. First they are HUGE compared to risc instructions. ~4x as large since they need to map worst case size for each element for everyinstruction. Those are bits that need to be moved and processed in several pipeline stages before execution stage. And take there power.
Then extra local memory operations due to 2 operand instructions and register spills.
All the ugliness creates bad power efficiency in low power applications.
X86 has won in the desktop by sheer amount of engineering resources put in its development, and the engineering resources to get superior manufacturing process. Getting upwards to higher end is possible.
x86 is elephant, and you can make elephant to fly if you apply enough force, like Intel does. But there is serious doubt that no company can create enough force to push it through a key hole.
Actually the way they put the massive memory has nothing to do with the normal "desktop pc". The massive memory means, that spending several thousand USD you can get a system with 64GB of ram. And a company can spend hundreds of thousands to get a system with hundreds of GB:s of ram. And spending quite a lot you can get TB:s of ram, and some companies do that, and they do it for a reason. Some do it for huge calculations, others do it for running databases on ram. etc... Those things exists TODAY. It just opens new business opportunities when huge amounts of ram become available on cheaper systems too. [cheap as in under 100k$]
All it takes a good design. The hard part is debugging after loosing ability for reproducing bugs with equal input. As for 2 to many cores it really isn't that hard for many applications. And for many others it doesn't really matter because single threaded/process is fast enough. One thing that could use it is GAMES provided you actually are GUARANTEED to have that many processors available.
The syncronisations are easy if you design it properly. If your coding style is "think little, try something, debug until works" then you fail, the agile method for multithreaded is just asking trouble.
Design the application for parallerism instead of trying to parallerise each small block, you get far better result. Extreme waterfall for extreme numbers of processors. And using number of processors for scaling up such application is far easier than trying to speed up said application. Of course there are things that are hard to parallerise, luckily most of the time you don't need to parallerise them.
Of course I want PC:s with capability of executing 100's of threads at same time, and enough memory bandwith for that. Why? Because the easily parallerised problems include. game AI, optimizing compiler, graphics...
The hard problem is "Here's the code base, make it go faster by using more processors." , and thats lots of realworld situations, but it doesn't mean that parallerism is hard. Fitting parallerism for non-paraller design is hard. Its like trying to fit square through circle of equal area. For 2 to many cores, its hard for those who's entire code base is designed for single or two processors.
The palestinians actually choose their launching site so that retaliatory strike to that exact location wouldn't be a bad thing for them. They try to take two birds with one stone. In lebanon the trick was to launch as close to UN posts as possible. In west bank the trick is to go in christian village and launch from there. And no. Most certainly the inhabitants of said places cannot stop them, they don't have weapons and are minorities. Thats the way. You don't like some group, then launch your strike from their place. Or from place where israeli retaliatory action could become a diplomatic victory. Oh the christian arabs have been ethnicly cleansed by palestinian authority.
Its the Final Solution to the problem of having lots of geeks making inhouse software, just google for a FOSS alternative! We can get rid of them! We can increase our profits by this! And finally 90% of geeks get what they deserve , dumpster diving for you guys, we get the softwar for FREE;)
Think typical internet usage, its bursty. For more typical user case of relatively small files you get a nice boost in performance. If you download something bit then its another matter. The quick gives improves web surfer latency without taking much of the bandwith. They have limited bandwith and they probably cannot give that kind of speed up all the time.
Actually after reading the exact meaning of the word utilization it fits the context perfectly. But what he might have meant by such a choice of words, could be restated as "Not without increase in utilization". As we all know the queuing theory by heart, everything else becomes redundant in his post.
Please, bare my lousy English. English is my second language and its been nearly a decade from last time someone taught me it.
Actually what Tippet claims in using the same house analogy that you used. Is that people spend their money on bullet proof windows, and heavy steel doors with half a dosend different locks required to open it, and do not teach their children on NOT leaving their (ground floor) windows open when everyone leaves the house. Or leave the balcony door open...
What he claims is that we are not working on the best solution.
Actually there is reason to buy those cards. Its the more stable drivers optimized for the work. And the warranty service. For gaming cards the stability takes back seat compared to pure speed. In workstation its opposite. The problem still is that the manufacturer of gaming cards probably have cut corners somewhere to be a buck cheaper that others offering almost identical product. In professional arena the quality of capacitors should be higher, so that the card is less likely to fail at some point. I don't know about you but for me it seems like a business reason why to buy these cards. [not for hobbyist though].
Theoretical limits of human vision has been measured in 0.4 arc minute. Over 1 arc minute is considered below normal vision. The theoretical minimum visible object is normal computer monitor pixel (102ppi) from 14 feet away. Those with really perfect vision not just normal vision (which people speak as perfect vision).
(tan(pi/(360*60/0.4))*14feet) in inch = 0.00977384382 inch 1/0.00977384382 = 102.313892
Now the other way to consider the same limit is, 60/0.4=150 pixels per degrees of arch. 150*90= 10500 pixels horizontal gives you 90 degrees fov where you can concentrate your vision on any part of image and it is at limit of human perception. 60*120 = 7200 is a good target, 120 fov where each pixel is not smaller that worst of those who get 20/20 vision could see. 20/10 is still measurable vision and people get that also even people talk about 20/20 as "perfect" vision. But the market would be smaller and we are at the edge of what is possible.
>the maximum effective range of an AK-47 (the area at which you could expect to hit a large target firing horizontally, though I think a blimp is a bit above the large target in this standard) is generally estimated around 250m.
HAHAHAHA!
I had a 50% hit rate with human sized target on firing range at 300meters. And thats considered bad hit rate. Hitting a vehicle, I'd estimate that 500meters would be just fine for me. And for anyone who shoots regularly, hitting a human sized target at 500 meters would be just fine with it.
But hitting a small arms at balloon, the AK47 isn't the best choice, its range is still quite limited. I'd say that with some of the higher powered hunting rifles I could hit such balloon that fly 's upto 1 mile high and up to 1 miles distance. The target is huge, the cross section it has in the sights would be much bigger in 2km distance than what human has at 150 meters.
ps. Finns have a mandatory military service that I was in, we use local variant of AK-47.
1) dual channel chipsets require both channels operational. 2) More than 2 slots takes space. 3) There are no 1.5GB dimms.
Of course not everyone uses MS-OS And have a proper chipset for utilisation of 4GB.
Or this can be seen as manufacturers trying to create upgrade wave. If new machines have 4GB of ram the software will adapt to fill it within a year or two. Then OLD machines with 1GB or less require atleast new RAM. But while your at it why not buy entirely a new PC?
If you consider modern armoug against old RPG:s the result is obvious. Of course there is modern RPG:s available that can penetrate frontal armour of modern battle tank. However those things are heavy rpg:s not medium/light rpg:s (which is exacly what civilians think when talking about rpg:s).
Then there is modern tanks that are more concerned about mobility and firepower than armour, which medium rpg:s are still effective if you can hit it with one.
Originality does not signify novelty; a work may be original even though it closely resembles other works, so long as the similarity is fortuitous, not the result of copying. To illustrate, assume that two poets, each ignorant of the other, compose identical poems. Neither work is novel, yet both are original and, hence, copyrightable.
The tools that mathemacians produce need to be supplemented more with CS specific abstractions. And generic math classes become less important than more CS specific abstractions after certain point of mathematical prowess.
This is closer to real life situation. The point where more generic mathematics classes is not as useful any more is debatable.
Well Finland is good example of western democracy that has lower population density compared to USA, and far lower CO2 per capita figures. The answer to question how is 6 dollars per gallon. Most of fuel price is tax. That way we don't have to import as much oil, and it improves our trade balance, and because goverment spends the money they get from tax the overall economy don't get hurt by it.
America is the one who should take over 300 million muslim refugees. Thats right, there are over that many muslims living in countries where they are majority, which are going to get under seas. The neighbouring countries are NOT muslim countries. And after all americans have always talked about how large their country is they have enough space for them, and they are the ones that have always shot down any attempts of trying to solve problem before it becomes this kind of devastation.
In overall I don't expect man kind to go through the climate change without full scale nuclear war of somekind. There are just too many desparate people and too many nukes around to avoid that.
Firstly EU doesn't have federal law, and each and every country has its own laws based on these things. Secondly precedence of previous judgements isn't as important thing as it is in USA. And finally its isn't even Finnish Supreme court, in Finnish three level court system this happened in lowest level court. Don't make it what it isn't. What this means that it is unlikely no-one will be put to trial after this for breaking CSS, in Finland, since those interest groups they will probably loose with current wording of the law.
Now the obligatory IANAL, but a native Finnish citizen.
The shrinking transistors decrease transistor latencies, shrinking lines increase line latencies for given length. From the beginning of microprocessors to pentium 4 era we have been able to trust newer processes improving the frequency by good margin, but nowadays the transistor latencies are far less important in determining the frequency. At same time when frequencies increased there was also increase in power consumption. In very early days those where under 1 watt chips. We have hit wall there by capping the amount of power majority of people are willing to accept from their computer. Then there was also by reducing amount of transitors per pipeline stage. But now we are in part where shortening pipeline stages doesn't bring real world performance. As we can see in late P4 desings which were with overly long pipeline. We have hit LOTS of walls all the same time. Now adding more parallerism brings more performance than hunting the last bits of sequential performance.
Burning creates CO2. Total weight of atmosphere: 5.1480*10^18 kg , percentage of CO2 by weight 0.00035.->1.8*10^15 amount of CO2 in atmosphere. Molecular mass of C is 12 Molecular mass of O=16 So burning ton of PURE C creates 44/12 tons of CO2. For long hydro carbons its 44/14 tons of CO2 per ton of fuel. Last year the COAL burning alone, no agriculture, nor oil burning taken account created 10^13 kg of CO2. The coal burning between years 1980 and 2004 created 1.7*10^14 CO2. Thats about 10% of current CO2 levels. Now thats only coal and no Gasoline there, and only between years 80 to 2004. For the same period based on graph about world energy usage figures, the oil was twice the coal while natural gas was about equal to coal. So man kind has produced between years 1980 and year 2004 about as much CO2 as 40% of current amount of CO2 in atmosphere. And it isn't such a complex science to measure the amount of CO2 we produce by burning coal or burning oil, its just measuring the amount we burn. But just the HUGE amount we burn fossile fuels makes the difference. Sure nature have its own cycles and those probably have absorbed some of the CO2 we have spewn to atmosphere, I don't claim to understand those well, nor burning of those things, I calculated my own figures about CO2 production and got a lot bigger figure than mentioned here. But I took the book figure instead of using my own since I'd rather err in using lower estimate in here than using too high.
Slashdot Mirror
Article is quite too enthuastic about x86 pushing to other domains.
Lets make it clear.Modern x86 decoder is more complex than entire simple single issue risc processor and it consumes more power.
Yes, thats SINGLE unit in the front end of pipeline, that risc processors do not need.
About those "risc ops" which x86 instructions are translated as. First they are HUGE compared to risc instructions. ~4x as large since they need to map worst case size for each element for everyinstruction. Those are bits that need to be moved and processed in several pipeline stages before execution stage. And take there power.
Then extra local memory operations due to 2 operand instructions and register spills.
All the ugliness creates bad power efficiency in low power applications.
X86 has won in the desktop by sheer amount of engineering resources put in its development, and the engineering resources to get superior manufacturing process. Getting upwards to higher end is possible.
x86 is elephant, and you can make elephant to fly if you apply enough force, like Intel does. But there is serious doubt that no company can create enough force to push it through a key hole.
Actually the way they put the massive memory has nothing to do with the normal "desktop pc".
The massive memory means, that spending several thousand USD you can get a system with 64GB of ram.
And a company can spend hundreds of thousands to get a system with hundreds of GB:s of ram.
And spending quite a lot you can get TB:s of ram, and some companies do that, and they do it for a reason.
Some do it for huge calculations, others do it for running databases on ram. etc...
Those things exists TODAY.
It just opens new business opportunities when huge amounts of ram become available on cheaper systems too.
[cheap as in under 100k$]
All it takes a good design.
The hard part is debugging after loosing ability for reproducing bugs with equal input.
As for 2 to many cores it really isn't that hard for many applications. And for many others it doesn't really matter because single threaded/process is fast enough.
One thing that could use it is GAMES provided you actually are GUARANTEED to have that many processors available.
The syncronisations are easy if you design it properly. If your coding style is
"think little, try something, debug until works" then you fail, the agile method for multithreaded is just asking trouble.
Design the application for parallerism instead of trying to parallerise each small block, you get far better result. Extreme waterfall for extreme numbers of processors.
And using number of processors for scaling up such application is far easier than trying to speed up said application.
Of course there are things that are hard to parallerise, luckily most of the time you don't need to parallerise them.
Of course I want PC:s with capability of executing 100's of threads at same time, and enough memory bandwith for that.
Why?
Because the easily parallerised problems include.
game AI, optimizing compiler, graphics...
The hard problem is "Here's the code base, make it go faster by using more processors." , and thats lots of realworld situations, but it doesn't mean that parallerism is hard. Fitting parallerism for non-paraller design is hard.
Its like trying to fit square through circle of equal area.
For 2 to many cores, its hard for those who's entire code base is designed for single or two processors.
The palestinians actually choose their launching site so that retaliatory strike to that exact location wouldn't be a bad thing for them.
They try to take two birds with one stone. In lebanon the trick was to launch as close to UN posts as possible. In west bank the trick is to go in christian village and launch from there. And no. Most certainly the inhabitants of said places cannot stop them, they don't have weapons and are minorities. Thats the way. You don't like some group, then launch your strike from their place. Or from place where israeli retaliatory action could become a diplomatic victory. Oh the christian arabs have been ethnicly cleansed by palestinian authority.
unless they're also researching legs.
Or invent the wheel.
Its the Final Solution to the problem of having lots of geeks making inhouse software, just google for a FOSS alternative! We can get rid of them! We can increase our profits by this! And finally 90% of geeks get what they deserve , dumpster diving for you guys, we get the softwar for FREE ;)
ps. I'm CS student
I for one welcome our humorless overlords.
Imagine a beowulf cluster of THOSE.
Think typical internet usage, its bursty. For more typical user case of relatively small files you get a nice boost in performance. If you download something bit then its another matter. The quick gives improves web surfer latency without taking much of the bandwith. They have limited bandwith and they probably cannot give that kind of speed up all the time.
Actually after reading the exact meaning of the word utilization it fits the context perfectly.
But what he might have meant by such a choice of words, could be restated as "Not without increase in utilization".
As we all know the queuing theory by heart, everything else becomes redundant in his post.
Please, bare my lousy English. English is my second language and its been nearly a decade from last time someone taught me it.
Actually what Tippet claims in using the same house analogy that you used. ...
Is that people spend their money on bullet proof windows, and heavy steel doors with half a dosend different locks required to open it, and do not teach their children on NOT leaving their (ground floor) windows open when everyone leaves the house. Or leave the balcony door open
What he claims is that we are not working on the best solution.
Actually there is reason to buy those cards.
Its the more stable drivers optimized for the work. And the warranty service. For gaming cards the stability takes back seat compared to pure speed. In workstation its opposite. The problem still is that the manufacturer of gaming cards probably have cut corners somewhere to be a buck cheaper that others offering almost identical product. In professional arena the quality of capacitors should be higher, so that the card is less likely to fail at some point. I don't know about you but for me it seems like a business reason why to buy these cards. [not for hobbyist though].
Who is willing to relay their lifes on it?
I mean, that who non-HIV positive is willing to have sex with a HIV positive using the drug?
Oh wait, never mind this is slashdot.
Theoretical limits of human vision has been measured in 0.4 arc minute. Over 1 arc minute is considered below normal vision. The theoretical minimum visible object is normal computer monitor pixel (102ppi) from 14 feet away. Those with really perfect vision not just normal vision (which people speak as perfect vision).
(tan(pi/(360*60/0.4))*14feet) in inch = 0.00977384382 inch
1/0.00977384382 = 102.313892
Now the other way to consider the same limit is, 60/0.4=150 pixels per degrees of arch.
150*90= 10500 pixels horizontal gives you 90 degrees fov where you can concentrate your vision on any part of image and it is at limit of human perception.
60*120 = 7200 is a good target, 120 fov where each pixel is not smaller that worst of those who get 20/20 vision could see. 20/10 is still measurable vision and people get that also even people talk about 20/20 as "perfect" vision. But the market would be smaller and we are at the edge of what is possible.
>the maximum effective range of an AK-47 (the area at which you could expect to hit a large target firing horizontally, though I think a blimp is a bit above the large target in this standard) is generally estimated around 250m.
HAHAHAHA!
I had a 50% hit rate with human sized target on firing range at 300meters. And thats considered bad hit rate. Hitting a vehicle, I'd estimate that 500meters would be just fine for me. And for anyone who shoots regularly, hitting a human sized target at 500 meters would be just fine with it.
But hitting a small arms at balloon, the AK47 isn't the best choice, its range is still quite limited. I'd say that with some of the higher powered hunting rifles I could hit such balloon that fly 's upto 1 mile high and up to 1 miles distance. The target is huge, the cross section it has in the sights would be much bigger in 2km distance than what human has at 150 meters.
ps. Finns have a mandatory military service that I was in, we use local variant of AK-47.
1) dual channel chipsets require both channels operational.
2) More than 2 slots takes space.
3) There are no 1.5GB dimms.
Of course not everyone uses MS-OS And have a proper chipset for utilisation of 4GB.
Or this can be seen as manufacturers trying to create upgrade wave. If new machines have 4GB of ram the software will adapt to fill it within a year or two. Then OLD machines with 1GB or less require atleast new RAM. But while your at it why not buy entirely a new PC?
If you consider modern armoug against old RPG:s the result is obvious. Of course there is modern RPG:s available that can penetrate frontal armour of modern battle tank. However those things are heavy rpg:s not medium/light rpg:s (which is exacly what civilians think when talking about rpg:s).
Then there is modern tanks that are more concerned about mobility and firepower than armour, which medium rpg:s are still effective if you can hit it with one.
There ARE people who actually use it http://www.google.com/search?q=always+%22I'm+feeling+lucky%22+button&btnI=I'm+Feeling+Lucky
Sounds like Slashdot.
The tools that mathemacians produce need to be supplemented more with CS specific abstractions. And generic math classes become less important than more CS specific abstractions after certain point of mathematical prowess.
This is closer to real life situation. The point where more generic mathematics classes is not as useful any more is debatable.
Reply to parents signature.
I'm not interest in but:s of other slashdotters... posts.
Well Finland is good example of western democracy that has lower population density compared to USA, and far lower CO2 per capita figures. The answer to question how is 6 dollars per gallon. Most of fuel price is tax. That way we don't have to import as much oil, and it improves our trade balance, and because goverment spends the money they get from tax the overall economy don't get hurt by it.
America is the one who should take over 300 million muslim refugees. Thats right, there are over that many muslims living in countries where they are majority, which are going to get under seas. The neighbouring countries are NOT muslim countries. And after all americans have always talked about how large their country is they have enough space for them, and they are the ones that have always shot down any attempts of trying to solve problem before it becomes this kind of devastation.
In overall I don't expect man kind to go through the climate change without full scale nuclear war of somekind. There are just too many desparate people and too many nukes around to avoid that.
Firstly EU doesn't have federal law, and each and every country has its own laws based on these things.
Secondly precedence of previous judgements isn't as important thing as it is in USA.
And finally its isn't even Finnish Supreme court, in Finnish three level court system this happened in lowest level court. Don't make it what it isn't. What this means that it is unlikely no-one will be put to trial after this for breaking CSS, in Finland, since those interest groups they will probably loose with current wording of the law.
Now the obligatory IANAL, but a native Finnish citizen.
The shrinking transistors decrease transistor latencies, shrinking lines increase line latencies for given length.
From the beginning of microprocessors to pentium 4 era we have been able to trust newer processes improving the frequency by good margin, but nowadays the transistor latencies are far less important in determining the frequency. At same time when frequencies increased there was also increase in power consumption. In very early days those where under 1 watt chips.
We have hit wall there by capping the amount of power majority of people are willing to accept from their computer. Then there was also by reducing amount of transitors per pipeline stage. But now we are in part where shortening pipeline stages doesn't bring real world performance. As we can see in late P4 desings which were with overly long pipeline.
We have hit LOTS of walls all the same time. Now adding more parallerism brings more performance than hunting the last bits of sequential performance.
Burning creates CO2.
Total weight of atmosphere: 5.1480*10^18 kg , percentage of CO2 by weight 0.00035.->1.8*10^15 amount of CO2 in atmosphere.
Molecular mass of C is 12 Molecular mass of O=16 So burning ton of PURE C creates 44/12 tons of CO2.
For long hydro carbons its 44/14 tons of CO2 per ton of fuel.
Last year the COAL burning alone, no agriculture, nor oil burning taken account created 10^13 kg of CO2.
The coal burning between years 1980 and 2004 created 1.7*10^14 CO2.
Thats about 10% of current CO2 levels. Now thats only coal and no Gasoline there, and only between years 80 to 2004.
For the same period based on graph about world energy usage figures, the oil was twice the coal while natural gas was about equal to coal.
So man kind has produced between years 1980 and year 2004 about as much CO2 as 40% of current amount of CO2 in atmosphere.
And it isn't such a complex science to measure the amount of CO2 we produce by burning coal or burning oil, its just measuring the amount we burn. But just the HUGE amount we burn fossile fuels makes the difference. Sure nature have its own cycles and those probably have absorbed some of the CO2 we have spewn to atmosphere, I don't claim to understand those well, nor burning of those things, I calculated my own figures about CO2 production and got a lot bigger figure than mentioned here. But I took the book figure instead of using my own since I'd rather err in using lower estimate in here than using too high.