( ESNUG 456 Item 1 ) -------------------------------------------- [07/17/06]

From: Karsten Matt <karsten.matt=user domain=amd spot calm>
Subject: A new unannounced "PrimeTime-PX" is to quietly replace PrimePower

Hi John,

For PrimePower users Synopsys has come out with a new extension to PrimeTime
called "PrimeTime-PX".  The PX stands for Power eXtension, and it is a new
power analysis option to PrimeTime.  According to Synopsys this new option
was officially released as a part of the 2006.06 PrimeTime release.

I was a beta tester of PrimeTime PX.  It seems like it's mostly PrimePower
integrated into PrimeTime with a few command name changes to match the
PrimeTime naming style.  I think this integration is not a bad thing.
PrimePower was good, but it is another standalone tool and it's based
on the Primetime engine.  Therefore, integrating PrimePower functionality
into PrimeTime is obvious.  I didn't get any official statement about the
future of PrimePower, but my AC indicated that they may quietly obsolete it.
 
PrimeTime-PX has many of the same commands and variables as PrimePower, but
Synopsys changed some names to match the PrimeTime naming style.  My AC gave
me a translation script (pp2ptpx) to convert PrimePower commands and
variable names into PrimeTime-PX names.

In our beta test of PrimeTime-PX, we ran it on several blocks from our last
chip and on the whole chip itself.  It completed with no surprizes, and we
did not find any major issues.  As a starting point for our vectorless power
analysis I used our existing PrimeTime script with some additions.  Synopsys
power analysis requires the timing windows and slew information from STA.
By putting power analysis into PrimeTime, Synopsys was able to eliminate
duplicate timing and slew calculations.  With update_power, PrimeTime-PX
re-uses the timing and slew data from update_timing in PrimeTime and does
not need to recalculate these numbers.

The power numbers PT-PX calculated matched the PrimePower numbers; so the
integration seems to have been done without impacting accuracy.  We did find
a slight increase in memory usage with the power analysis, though.  For a
60 K instance design, it increased by 8% compared to a regular PrimeTime run
without power (311 MB PT-PX versus 286 MB PrimeTime).  This increase seems
very reasonable to us, especially if you keep in mind, that you can save the
timing window and slew calculation for the power analysis by using these
values from the timing update.

Another aspect we liked was having PT-PX use RTL VCD to do an average power
analysis.  RTL VCD support means I don't have to run gate-level simulations
to generate switching activity.  Instead, I can just run a much faster RTL
simulation to generate a VCD file.

Overall, PrimeTime-PX works but the one little drawback is that you have to
translate your old PrimePower scripts into PT-PX terms.

    - Karsten Matt
      AMD                                        Dresden, Germany
Index    Next->Item






   
 Sign up for the DeepChip newsletter.
Email
 Read what EDA tool users really think.


Feedback About Wiretaps ESNUGs SIGN UP! Downloads Trip Reports Advertise

"Relax. This is a discussion. Anything said here is just one engineer's opinion. Email in your dissenting letter and it'll be published, too."
This Web Site Is Modified Every 2-3 Days
Copyright 1991-2024 John Cooley.  All Rights Reserved.
| Contact John Cooley | Webmaster | Legal | Feedback Form |

   !!!     "It's not a BUG,
  /o o\  /  it's a FEATURE!"
 (  >  )
  \ - / 
  _] [_     (jcooley 1991)