And what is that extermely unhelpful thing suposed to mean?
Do you actually know something, or are you only adding your own one cent?
Quote:
Originally Posted by melgross
High input impedances will help protect the chips from high inrush voltages and currents, as well as capacitive input failure.
Well, I just think this doesn't make sense...
- but I think you're right to say that it's not a problem, in that manufacturers like Intel won't release products until they've got the leakage issues etc worked out - that's their job afterall
By the way, Intel won't be introducing 32nm until late-2009 (according to their press info)
- but I think you're right to say that it's not a problem, in that manufacturers like Intel won't release products until they've got the leakage issues etc worked out - that's their job afterall
By the way, Intel won't be introducing 32nm until late-2009 (according to their press info)
Why do you think it makes no sense? Have you developed electronic equipment? That was my business for some years, and I occasionally do some work even now.
High input impedance is a very important factor in protecting circuits.
In fact, some chips have impedances in the trillions of ohms. This keeps input current down, and also lowers the dangers of capacitance. With the very low voltages used in todays cpu's this is an important design element. high impedances isolate chips from the surrounding circuits.
I'm not saying that this is the only factor in chip life, but it does cause problems when things go wrong, or input voltages change with spiking, etc.
Below 32 nm, other factors become very important, including whiskering, which, though present as a problem for even 45 nm, has been solved at that size. At 32 nm it a worse problem, but has also been solved. At lower process sizes, it's still a problem, and solving it will be much more difficult.
By the way, whiskering is the main failure route for these thin lithium batteries we hear about that melt or burn.
32 nm will make an appearance in early 2009, or even late 2008, but you're right, I should have been more explicit. We will see the first sampling quantity of cpu's from 32 nm then, not actual shipping products.
High input impedance is a very important factor in protecting circuits.
In fact, some chips have impedances in the trillions of ohms. This keeps input current down, and also lowers the dangers of capacitance. With the very low voltages used in todays cpu's this is an important design element. high impedances isolate chips from the surrounding circuits.
I'm not saying that this is the only factor in chip life, but it does cause problems when things go wrong, or input voltages change with spiking, etc.
Are you talking about Intel's high-k dielectric process?
Here's a good overview of their 45nm process
- I think pages 15-25 are particularly entertaining
I had to leave for a meeting, and I just got back, so I haven had time toread the papers, though I think I've read the first one at some point. I'll get to them after dinner.
It is amazing how many people are waiting to upgrade when Nehalem systems are released. I know I will replace my G5 with one of these beasts! I have no doubt that Apple will continue to optimize OSX for muti-core and Intel will continue to optimize their compilers. What new technologies await us with Sandy Bridge?
I'm not so much waiting for Nehalem as I am waiting to move to a new flat and also to finish work on my top secret algorithm. It is intended to run on a PCI-E card.
I'm not so much waiting for Nehalem as I am waiting to move to a new flat and also to finish work on my top secret algorithm. It is intended to run on a PCI-E card.
Comments
No, Dunnington is a MP processor (7300 series), it cannot be used in the Mac Pro (DP chipset - 5400 series).
Good point...
And what is that extermely unhelpful thing suposed to mean?
Do you actually know something, or are you only adding your own one cent?
High input impedances will help protect the chips from high inrush voltages and currents, as well as capacitive input failure.
Well, I just think this doesn't make sense...
- but I think you're right to say that it's not a problem, in that manufacturers like Intel won't release products until they've got the leakage issues etc worked out - that's their job afterall
By the way, Intel won't be introducing 32nm until late-2009 (according to their press info)
Well, I just think this doesn't make sense...
- but I think you're right to say that it's not a problem, in that manufacturers like Intel won't release products until they've got the leakage issues etc worked out - that's their job afterall
By the way, Intel won't be introducing 32nm until late-2009 (according to their press info)
Why do you think it makes no sense? Have you developed electronic equipment? That was my business for some years, and I occasionally do some work even now.
High input impedance is a very important factor in protecting circuits.
In fact, some chips have impedances in the trillions of ohms. This keeps input current down, and also lowers the dangers of capacitance. With the very low voltages used in todays cpu's this is an important design element. high impedances isolate chips from the surrounding circuits.
I'm not saying that this is the only factor in chip life, but it does cause problems when things go wrong, or input voltages change with spiking, etc.
Below 32 nm, other factors become very important, including whiskering, which, though present as a problem for even 45 nm, has been solved at that size. At 32 nm it a worse problem, but has also been solved. At lower process sizes, it's still a problem, and solving it will be much more difficult.
By the way, whiskering is the main failure route for these thin lithium batteries we hear about that melt or burn.
32 nm will make an appearance in early 2009, or even late 2008, but you're right, I should have been more explicit. We will see the first sampling quantity of cpu's from 32 nm then, not actual shipping products.
- yes, Intel may have samples then
- but you & I both know Intel is on a 2 year cycle, Tick-Tock as they say
- introducing a new process, with a die-shrink CPU (like the Penryn in late-2007) every 2 years
- and introducing a new marchitecture on alternate years (like the Nahalem in late-2008).
- etc
http://www.intel.com/technology/arch..._45nm+rhc_32nm
High input impedance is a very important factor in protecting circuits.
In fact, some chips have impedances in the trillions of ohms. This keeps input current down, and also lowers the dangers of capacitance. With the very low voltages used in todays cpu's this is an important design element. high impedances isolate chips from the surrounding circuits.
I'm not saying that this is the only factor in chip life, but it does cause problems when things go wrong, or input voltages change with spiking, etc.
Are you talking about Intel's high-k dielectric process?
Here's a good overview of their 45nm process
- I think pages 15-25 are particularly entertaining
http://download.intel.com/technology...HiKMG_pres.pdf
Here's some waffle about the 32nm process, but it doesn't give much hard info
http://video.intel.com/index.jsp?fr_...5b6dba4da3ebe6
With regards to 32nm introduction, saying 'maybe late-2008' is misleading, and will get the less-well informed readers of this forum over-excited
- yes, Intel may have samples then
- but you & I both know Intel is on a 2 year cycle, Tick-Tock as they say
- introducing a new process, with a die-shrink CPU (like the Penryn in late-2007) every 2 years
- and introducing a new marchitecture on alternate years (like the Nahalem in late-2008).
- etc
http://www.intel.com/technology/arch..._45nm+rhc_32nm
Are you talking about Intel's high-k dielectric process?
Here's a good overview of their 45nm process
- I think pages 15-25 are particularly entertaining
http://download.intel.com/technology...HiKMG_pres.pdf
Here's some waffle about the 32nm process, but it doesn't give much hard info
http://video.intel.com/index.jsp?fr_...5b6dba4da3ebe6
[/quote]
I had to leave for a meeting, and I just got back, so I haven had time toread the papers, though I think I've read the first one at some point. I'll get to them after dinner.
The H-K process is something else, but it helps.
It is amazing how many people are waiting to upgrade when Nehalem systems are released. I know I will replace my G5 with one of these beasts! I have no doubt that Apple will continue to optimize OSX for muti-core and Intel will continue to optimize their compilers. What new technologies await us with Sandy Bridge?
I'm not so much waiting for Nehalem as I am waiting to move to a new flat and also to finish work on my top secret algorithm. It is intended to run on a PCI-E card.
I'm not so much waiting for Nehalem as I am waiting to move to a new flat and also to finish work on my top secret algorithm. It is intended to run on a PCI-E card.
Just one algorithm?
What kind of board. Oh, give us a hint.
The limit Intel is trying to reach is 0.1 nm!
Yes, under 1 nm.
In order to do this, they need to build a special new kind of factory to do this.
(this is all theoretical, and early design mode)
Actually, there is only one guy at Intel that thinks everything can be manufactured
at such a small scale, says he knows where to get the staff to work the plant.
Not many believe him, but his recent documentary has gotten a lot of publicity.
Guys name is Horton.
While you jest in an admirable fashion, people are actually working on single atom thick carbon connectors: Graphene sheets.