 |
|
| Linux Forum Index » General Linux Discussion » Where is 10Ghz?... |
|
Page 1 of 1 |
|
| Author |
Message |
| Jesse Dorland... |
 Posted: Mon Aug 17, 2009 2:59 am |
|
|
|
Guest
|
Guys
Remember the old days when Intel, and AMD were fighting over "single
core speed"? Well, then they change the rules of the games b/c they
couldn't deliver what they promise -- namely 10-50 ghz of cpu -- mind
you they were talking about single core cpu...
Now we have dual - quad core cpu... but none coming up with 10ghz
(each core around 10ghz)...
Have they hit they wall or something? |
|
|
| Back to top |
|
|
|
| Jesse Dorland... |
| Posted: Mon Aug 17, 2009 6:06 am |
|
|
|
Guest
|
On Aug 17, 9:13 am, Randy Yates <ya... at (no spam) ieee.org> wrote:
Quote: Jesse Dorland <jessedorl... at (no spam) gmail.com> writes:
Guys
Remember the old days when Intel, and AMD were fighting over "single
core speed"? Well, then they change the rules of the games b/c they
couldn't deliver what they promise -- namely 10-50 ghz of cpu -- mind
you they were talking about single core cpu...
Now we have dual - quad core cpu... but none coming up with 10ghz
(each core around 10ghz)...
Have they hit they wall or something?
Yes. They are hitting fundamental limitations imposed by semiconductor
device physics.
Most of the improvements thus far can be basically attributed to the
shinking of transistors. Have you heard the terms "120 nm (nanometer)
process" or "65 nm process"? They are referring to the width of the
"gate" of a complementary metal oxide semiconductor field effect
transistor (CMOS FET) in the process of manufacturing semiconductor ICs.
Basically, when you shrink the device size, you reduce the capacitance
of the device and increase its operating frequency.
However, this process can not go on indefinitely. At some point the
device characteristics are denigrated by unwanted effects that
practically prevent further reductions in size.
So in a strange way we don't have to worry about our present laptop
becoming obsolete. I mean my Pentium dual core laptop is good enough
for another five years or so? Safe for replacing lcd LCD backlight and
maybe hard-drive?
Ironically, windows xp over 11 years OS is refusing to die -- one
would think Microsoft would be happy rather then trying to replace it
with another vista's clone  |
|
|
| Back to top |
|
|
|
| Randy Yates... |
| Posted: Mon Aug 17, 2009 7:13 am |
|
|
|
Guest
|
Jesse Dorland <jessedorland at (no spam) gmail.com> writes:
Quote: Guys
Remember the old days when Intel, and AMD were fighting over "single
core speed"? Well, then they change the rules of the games b/c they
couldn't deliver what they promise -- namely 10-50 ghz of cpu -- mind
you they were talking about single core cpu...
Now we have dual - quad core cpu... but none coming up with 10ghz
(each core around 10ghz)...
Have they hit they wall or something?
Yes. They are hitting fundamental limitations imposed by semiconductor
device physics.
Most of the improvements thus far can be basically attributed to the
shinking of transistors. Have you heard the terms "120 nm (nanometer)
process" or "65 nm process"? They are referring to the width of the
"gate" of a complementary metal oxide semiconductor field effect
transistor (CMOS FET) in the process of manufacturing semiconductor ICs.
Basically, when you shrink the device size, you reduce the capacitance
of the device and increase its operating frequency.
However, this process can not go on indefinitely. At some point the
device characteristics are denigrated by unwanted effects that
practically prevent further reductions in size.
--
Randy Yates % "...the answer lies within your soul
Digital Signal Labs % 'cause no one knows which side
mailto://yates at (no spam) ieee.org % the coin will fall."
http://www.digitalsignallabs.com % 'Big Wheels', *Out of the Blue*, ELO |
|
|
| Back to top |
|
|
|
| J.O. Aho... |
| Posted: Mon Aug 17, 2009 10:23 am |
|
|
|
Guest
|
Jesse Dorland wrote:
Quote: On Aug 17, 9:13 am, Randy Yates <ya... at (no spam) ieee.org> wrote:
However, this process can not go on indefinitely. At some point the
device characteristics are denigrated by unwanted effects that
practically prevent further reductions in size.
So in a strange way we don't have to worry about our present laptop
becoming obsolete. I mean my Pentium dual core laptop is good enough
for another five years or so? Safe for replacing lcd LCD backlight and
maybe hard-drive?
It will become obsolete and it will not take too long, for there are other
optimizations to make than just making the CPU smaller, just compare a 65nm
CPU from today running with one core with a 65nm two years ago, both running
at 2GHz, you will see a big difference and when the day comes when you can't
improve the x86 architecture, you may see a new better designed CPU take over
(even microsoft do have experience of other architectures, even if they aren't
as good at it as the Linux community).
Quote: Ironically, windows xp over 11 years OS is refusing to die -- one
would think Microsoft would be happy rather then trying to replace it
with another vista's clone
Long lived products decreases the profit, if Bill & Co could, they most likely
would release a new version ever 12 months.
--
//Aho |
|
|
| Back to top |
|
|
|
| William Black... |
| Posted: Mon Aug 17, 2009 1:03 pm |
|
|
|
Guest
|
Jesse Dorland wrote:
Quote: Guys
Remember the old days when Intel, and AMD were fighting over "single
core speed"? Well, then they change the rules of the games b/c they
couldn't deliver what they promise -- namely 10-50 ghz of cpu -- mind
you they were talking about single core cpu...
Now we have dual - quad core cpu... but none coming up with 10ghz
(each core around 10ghz)...
Have they hit they wall or something?
Looks like two walls.
There is some sort of frequency wavelength problem when you get very
fast which means that eventually they'll have to use light rather than
electricity, but that needs some sort of fundamental breakthrough in
technology.
Also, they're running a bit warm these days.
In the end the old Cray supercomputers were more about pumping
refrigeration fluid than anything to do with clever processing systems.
If you need to pump two tonnes of freon around the bloody thing every
minute your major problem isn't processor speed, it's the plumbing...
--
William Black
"Any number under six"
The answer given by Englishman Richard Peeke when asked by the Duke of
Medina Sidonia how many Spanish sword and buckler men he could beat
single handed with a quarterstaff. |
|
|
| Back to top |
|
|
|
| William Black... |
| Posted: Mon Aug 17, 2009 1:11 pm |
|
|
|
Guest
|
Jesse Dorland wrote:
Quote: Guys
Remember the old days when Intel, and AMD were fighting over "single
core speed"? Well, then they change the rules of the games b/c they
couldn't deliver what they promise -- namely 10-50 ghz of cpu -- mind
you they were talking about single core cpu...
Now we have dual - quad core cpu... but none coming up with 10ghz
(each core around 10ghz)...
Have they hit they wall or something?
Looks like two walls.
There is some sort of frequency wavelength problem when you get very
fast which means that eventually they'll have to use light rather than
electricity, but that needs some sort of fundamental breakthrough in
technology.
Also, they're running a bit warm these days.
In the end the old Cray supercomputers were more about pumping
refrigeration fluid than anything to do with clever processing systems.
If you need to pump two tonnes of freon around the bloody thing every
minute your major problem isn't processor speed, it's the plumbing...
--
William Black
"Any number under six"
The answer given by Englishman Richard Peeke when asked by the Duke of
Medina Sidonia how many Spanish sword and buckler men he could beat
single handed with a quarterstaff. |
|
|
| Back to top |
|
|
|
| William Black... |
| Posted: Mon Aug 17, 2009 1:12 pm |
|
|
|
Guest
|
Jesse Dorland wrote:
Quote: Guys
Remember the old days when Intel, and AMD were fighting over "single
core speed"? Well, then they change the rules of the games b/c they
couldn't deliver what they promise -- namely 10-50 ghz of cpu -- mind
you they were talking about single core cpu...
Now we have dual - quad core cpu... but none coming up with 10ghz
(each core around 10ghz)...
Have they hit they wall or something?
Looks like two walls.
There is some sort of frequency wavelength problem when you get very
fast which means that eventually they'll have to use light rather than
electricity, but that needs some sort of fundamental breakthrough in
technology.
Also, they're running a bit warm these days.
In the end the old Cray supercomputers were more about pumping
refrigeration fluid than anything to do with clever processing systems.
If you need to pump two tonnes of freon around the bloody thing every
minute your major problem isn't processor speed, it's the plumbing...
--
William Black
"Any number under six"
The answer given by Englishman Richard Peeke when asked by the Duke of
Medina Sidonia how many Spanish sword and buckler men he could beat
single handed with a quarterstaff. |
|
|
| Back to top |
|
|
|
| Van Chocstraw... |
| Posted: Tue Aug 18, 2009 10:36 am |
|
|
|
Guest
|
Jesse Dorland wrote:
Quote: Guys
Remember the old days when Intel, and AMD were fighting over "single
core speed"? Well, then they change the rules of the games b/c they
couldn't deliver what they promise -- namely 10-50 ghz of cpu -- mind
you they were talking about single core cpu...
Now we have dual - quad core cpu... but none coming up with 10ghz
(each core around 10ghz)...
Have they hit they wall or something?
Yes, they would have to go back to vacuum tubes for faster speeds. |
|
|
| Back to top |
|
|
|
|
|
All times are GMT - 5 Hours
The time now is Tue Dec 08, 2009 5:25 am
|
|