[comp.lsi] Papers on low power digital circuits

luciano@canuck.Berkeley.EDU (Luciano Lavagno) (06/28/91)

Any pointer ? Design tricks, theoretical analyses, synthesis
methodologies:
everything is welcome ! I am especially interested in CMOS, by the
way...
Thanks !
Luciano

-- 
+----------------------------+------------------------------------+
|Luciano Lavagno             |  E-mail: luciano@ic.Berkeley.EDU   |
|Dept of EECS, Rm. 550B2-69  |                                    |
|UC Berkeley                 |  Phone: (415) 642-5012             |
|Berkeley, CA  94720 (USA)   |                                    |
+----------------------------+------------------------------------+

twolf@cs.utah.edu (Tom Wolf) (06/28/91)

In article <42494@ucbvax.BERKELEY.EDU> luciano@canuck.Berkeley.EDU (Luciano Lavagno) writes:
>Any pointer ? Design tricks, theoretical analyses, synthesis
>methodologies:
>everything is welcome ! I am especially interested in CMOS, by the
>way...

What about Asynchronous CMOS design? Martin at Cal Tech, Ginosar at Technion,
and I have all built CMOS circuits using Async design instead of clocked.
My results seem to agree with thiers, in that the circuits run over a wide
voltage range. Mine run between 0.70 volts and 5.5 volts (as high as my
tester goes). Of course, there is a greater than 10x performance difference.

Still, at less than 1.0 volts, they aren't using much power.

Tom

miyazaki@taichung (Takeshi Miyazaki) (06/29/91)

I don't know any papers, but I remember at least two semiconductor 
manufacturers in Japan (and probably some in the U.S.) sell microprocessors
which can operate at 1~2 V.

These chips must have been developed using an ordinary CMOS technology.
Difficult part may be I/O, and not use dynamic circuits, and change Vth.
Sorry, company's name escapes me.


Takehsi Miyazaki
miyazaki@ee.princeton.edu