[Beowulf] A start in Parallel Programming?

Jim Lux James.P.Lux at jpl.nasa.gov
Wed Mar 14 12:57:20 PDT 2007


At 10:39 AM 3/14/2007, Joe Landman wrote:
>Robert G. Brown wrote:
>
>>So I'm not holding my breath on ML running out this week or next.  I'm
>>more interested in speculating on when the next massive superML jump on
>>TOP of ML will occur, when the next phase/paradigm shift is due that
>
>Ah....
>
>Ok, my concern is that silicon may be running out of room.  Now if 
>we all just put on our happy faces, and use GaAs, we can go much 
>faster. Yeah, it has some issues ...

Like
1) yield
2) wafer size
3) power consumption
4) maximum density

Note that what's available in GaAs tends to be things like RF 
amplifiers, 4 bit counters, etc.  I don't know that there's any LSI 
GaAs  (what they call LSI in the GaAs world is things like 
operational amplifiers with (gasp) dozens of transistors.. which is a 
very high level of integration compared to one or two transistors in 
the usual RF amplifier.

There have been some experimental GaAs switches (like crossbars) For 
instance, a google turns up the work of Yamada, et al., in 1999 with 
a 8x8 ATM switch using 0.5 micron gate GaAs clocking at 1.3 GHz. 
dissipating only a few tens of watts.

But... you CAN run it at a junction temperature of 200C.  GaN can run 
even hotter.

You're much better off pushing the CMOS processes (where they have 
fT's up in the 10s of GHz)

SiGe might be more attractive.. decent density, lower power 
consumption, ability to integrate CMOS structures, etc.

>  But think of how much faster we can be. (that and my thesis work 
> might suddenly have applicability, but that is another matter).
>
>More seriously, our computing industry can build silicon really 
>well. As it keeps shrinking problems arise.  Some of these problems 
>can be ameliorated by doing a material paradigm shift.  But our 
>computing industry does silicon really well ...


Really, really well...
Folks have been struggling with GaAs, GaN, InP, etc for a long time.


>Joe
>
>
>--
>
>Joseph Landman, Ph.D
>Founder and CEO
>Scalable Informatics LLC,
>email: landman at scalableinformatics.com
>web  : http://www.scalableinformatics.com
>phone: +1 734 786 8423
>fax  : +1 734 786 8452 or +1 866 888 3112
>cell : +1 734 612 4615

James Lux, P.E.
Spacecraft Radio Frequency Subsystems Group
Flight Communications Systems Section
Jet Propulsion Laboratory, Mail Stop 161-213
4800 Oak Grove Drive
Pasadena CA 91109
tel: (818)354-2075
fax: (818)393-6875 





More information about the Beowulf mailing list