Searching \ for '[EE] Near future of 8086 Processors?' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/index.htm?key=near+future+8086
Search entire site for: 'Near future of 8086 Processors?'.

Exact match. Not showing close matches.
PICList Thread
'[EE] Near future of 8086 Processors?'
2008\01\31@175742 by Shawn Tan

flavicon
face
On Thursday 31 January 2008 20:24:42 Tobias Gogolin wrote:
> Im thinking as wide an as fast as possible (overclockable)
> But in the real wold there are 2 big manufacturers and so far dual and quad
> cores!
> Im also very power conscious and I want to be conservative, so Im guessing
> smallest feature size is best!

Hehe, people tend to forget VIA. They've pretty much cornered the low end
mini/nano/pico-itx business. And their latest Isaiah core is getting good
reviews. I've always been a fan of VIA x86 processors.

If you're extremely power conscious, look at VIA. I would be interested on how
they compare with Geode processors though. AMD seems to only do Geode
benchmarks againts other Geode processors.

> Its funny the other day I was giving a presentation on parallel computing
> and nobody guessed the actual transistor count that has been exponentially
> rising with Moores law!
> More than 300 Millions on a regular Graphics Processor, simply Amazing!

I'm actually hoping that we'll see an end to the madness. I'm hoping that the
transistor counts will stay constant or reduce. And I'm actually hoping to
see different architectures come into the mainstream.

Cheers.

--
with metta,
Shawn Tan

Aeste Works (M) Sdn Bhd - Engineering Elegance
http://www.aeste.net

2008\01\31@210335 by Tobias Gogolin

picon face
Yes There is transmeta, cyrix , via and maybe another few little ones, but
Im looking for the realistic affordable high end processor!
But what do you think 8 cores in 2008, 16 in 2009?


On Jan 31, 2008 2:57 PM, Shawn Tan <spam_OUTshawn.tanTakeThisOuTspamaeste.net> wrote:

{Quote hidden}

> -


'[EE] Near future of 8086 Processors?'
2008\02\01@003016 by William \Chops\ Westfield
face picon face

On Jan 31, 2008, at 6:03 PM, Tobias Gogolin wrote:

> But what do you think 8 cores in 2008, 16 in 2009?

Moore's law says 18 months for each doubling, and I think
it's slowed down somewhat.

But the state of the art of programming for multicore
processors had better start improving :-(

I find the prospect of new low-end x86 chips interesting
as well.  Apparently the initial set of patents on the
basic x86 architecture is expiring soon, so anyone will
be able to make x86 cpus...

BillW

2008\02\01@004929 by Tobias Gogolin

picon face
Yes but thats the point Multiprocessor ready programming will inevitably
come around!
And if there is not much of a penalty because of FSB contention then similar
clock frequencies on double the number of cores will be close to twice as
fast! Something that can admitably be seen so far with very few benchmarks!


On Jan 31, 2008 9:29 PM, William Chops Westfield <.....westfwKILLspamspam@spam@mac.com> wrote:

{Quote hidden}

> -

2008\02\01@011849 by Dr Skip

picon face


William "Chops" Westfield wrote:

> But the state of the art of programming for multicore
> processors had better start improving :-(

I don't think there will be a lot of impetus. Beyond the power user (and Vista
can allocate whole programs among cores) VMs will make use of it and allocate a
core to a VM. A whole different granularity, but that will probably keep people
busy chasing that for a good while... The market beyond that may be small.

I did see a vm appliance that claimed to provide a grid architecture among your
 cores and/or systems. The plan was to use grid computing at a cpu core
granularity. That might be the way it goes, with the average app never making
use of it.



2008\02\01@013421 by William \Chops\ Westfield

face picon face

On Jan 31, 2008, at 9:49 PM, Tobias Gogolin wrote:

> Multiprocessor ready programming will inevitably come around!

I'll believe it when I see it.  People have been working on it for 20
+ years now.  (of course, the software industry is quite ready to  
develop little bits of cpu-intensive eye candy, or security  
background processes and such that will happily eat up their cycles,  
and occasional apps and/or games will be carefully code to use more  
than one CPU, but in general multicore is still in the dark ages...

BillW

2008\02\01@015957 by Bob Blick

face picon face
William "Chops" Westfield wrote:
> On Jan 31, 2008, at 9:49 PM, Tobias Gogolin wrote:
>
>> Multiprocessor ready programming will inevitably come around!
>
> I'll believe it when I see it.  People have been working on it for 20
> + years now.  (of course, the software industry is quite ready to  
> develop little bits of cpu-intensive eye candy, or security  
> background processes and such that will happily eat up their cycles,  
> and occasional apps and/or games will be carefully code to use more  
> than one CPU, but in general multicore is still in the dark ages...

I agree. It's a "fix it in hardware", "fix it in software" thing. In
this case, multicores make it a little easier for the chipmakers, and
much, much harder to write compilers and rare to find ways to utilize
them at the programmer's level.

Cheers,

Bob

2008\02\01@021143 by Tobias Gogolin

picon face
Yes but look at things like Finite Elements Analysis, that are crying for as
many processors as you can imagine to have some day!
And I think its easy there are many search functions, loops compression and
decompression algorithms that can be unfolded...
Besides in the short term I don't want to have to feel bad that I usually
have my computer do x things at the same time, like playing music or videos
downloading and designing stuff...

On Jan 31, 2008 10:59 PM, Bob Blick <bbblickspamKILLspamsbcglobal.net> wrote:

{Quote hidden}

> -

2008\02\01@030226 by Richard Prosser

picon face
But isn't the compatibility requirement the hardest part? If the
program is only going to run on a quad core device, then in many cases
the tasks can be split, but if it has to also be able to run on
single, double and oct? (8) core devices, the problem gets a fair bit
harder.

RP

On 01/02/2008, Tobias Gogolin <.....usertogoKILLspamspam.....gmail.com> wrote:
{Quote hidden}

2008\02\01@040203 by Xiaofan Chen

face picon face
On 2/1/08, William Chops Westfield <westfwspamspam_OUTmac.com> wrote:
>
> I find the prospect of new low-end x86 chips interesting
> as well.  Apparently the initial set of patents on the
> basic x86 architecture is expiring soon, so anyone will
> be able to make x86 cpus...
>

Not so sure if any new vendors wants to go into the lower
end x86 business. It is a bit hard to beat ARM MCUs in the
lower end and ARM MPUs in the higher end for the non-PC
market. Even Intel and AMD have problems to go into that
market. Intel sold the ARM MPU segment to Marvell and
AMD also sold the MIPS based MPU product.


Xiaofan

2008\02\01@050501 by Jinx

face picon face
> I'm actually hoping that we'll see an end to the madness. I'm
> hoping that the transistor counts will stay constant or reduce

If you had a look at the post I made the other day,

[EE] Researchers make tiny radio from nanotubes

www.reuters.com/article/technologyNews/idUSN2848581520080129?pageNumb
er=1&virtualBrandChannel=0

one of the scientists said "Our goal is not to make tiny radios
per se, but really to develop nanotubes as a higher-performing
semiconductor"

I imagine with more breakthroughs the next computer era might
be that of carbon or organic materials, not silicon

2008\02\01@051625 by Shawn Tan

flavicon
face
On Friday 01 February 2008 10:04:10 Jinx wrote:
> > I'm actually hoping that we'll see an end to the madness. I'm
> > hoping that the transistor counts will stay constant or reduce
>
> one of the scientists said "Our goal is not to make tiny radios
> per se, but really to develop nanotubes as a higher-performing
> semiconductor"
>
> I imagine with more breakthroughs the next computer era might
> be that of carbon or organic materials, not silicon

Yes, that's what some of my friends, who're researching CNT, tell me. They're
talking about using CNTs for the gates of a transistor. But I believe that
it's still going to be mainly silicon.

But regardless of material used, I do think that computing architecture is
prime for change. There's only so far that multi-core can go as memory is
still shared.

Cheers.

--
with metta,
Shawn Tan

Aeste Works (M) Sdn Bhd - Engineering Elegance
http://www.aeste.net

2008\02\01@091223 by Martin Klingensmith

face
flavicon
face
All of these applications that can use multiple processors will work
with a single processor. They just have more threads or processes
running on the same processor. It's in the OS, not the program. The real
difficulty will always be splitting up the program into 'n' number of
parallel tasks.
-
Martin

Richard Prosser wrote:
> But isn't the compatibility requirement the hardest part? If the
> program is only going to run on a quad core device, then in many cases
> the tasks can be split, but if it has to also be able to run on
> single, double and oct? (8) core devices, the problem gets a fair bit
> harder.
>
> RP
>
>  

2008\02\02@131822 by Tobias Gogolin

picon face
I see the ultimate level of parallelization in good object oriented
programming!
I find that object oriented programming also can make positive use of the
vast memory amounts that are becoming affordable!
That is because objects exist in parallel and can pass messages and
information between each other continuously. Each object interaction
requiring a processor consciousness! The ideal program hence streams data
between object methods!
I admit a more visual way of developing programs may help to transition into
this world of faster computing...

On Feb 1, 2008 6:11 AM, Martin Klingensmith <@spam@martinKILLspamspamnnytech.net> wrote:

{Quote hidden}

> -

More... (looser matching)
- Last day of these posts
- In 2008 , 2009 only
- Today
- New search...