( ESNUG 477 Item 8 ) -------------------------------------------- [11/20/08]

Subject: ( ESNUG 476 #6 ) Hold on! Our Apache experience was quite bad

> I should mention that working with Apache AEs and R&D is unmatched with
> all the other EDA vendors.  Definitely there is no comparison with
> Cadence, Synopsys, Mentor, nor Magma.  I found dealing with Apache
> personnel at different levels (AEs, R&D, Marketing, and management) is
> as the difference between A and C grades.


From: [ King Tut ]

Hi, John,

Keep my name out of this.

I used Apache RedHawk 3+ years and Cadence VoltageStorm 3 years, too.  At
early beginning of using RedHawk, we felt it was powerful tool for power
analysis.  Few years late we tried VoltageStorm again, and found VSPE and
VSDG had lots improvement.  

After comparing their functions, we don't see VoltageStorm is behind Apache
RedHawk.  Now we use VoltageStorm for our 90nm and 65nm high speed
wireless network chips because:

  - VoltageStorm's library is easily shared for other Cadence tools.
    Consistent with SOCEncounter results, it reduces iterations.  When
    converting GDS data to RedHawk library, we had crashes many times.
    We wasted a lot of time getting the RedHawk library ready.

  - Cadence provides powerful on-line technical support, and on-site AE
    support.  Especially their AE is really expert not only at power
    analysis, but also on timing closure, floorplan, xtalk and etc. 
    They always gave us the timely support.

At very beginning, Apache sent a good AE worked with us.  After 6 months
when we purchased RedHawk, that AE quit the job.  Then they assigned another
AE in Texas supporting us.  It was inconvenience for us to explain Redhawk
problems to him.

When we wanted to eval Apache Dynamic Power Analysis for an urgent power
issue in our project, Apache asked us to promise to buy their tools if it
wasn't worse than VoltageStorm.  They didn't want to assign an AE to work
with us.  In contrast, Cadence sent us an eval license and an AE to help
to fix our problem.

    - [ King Tut ]

         ----    ----    ----    ----    ----    ----   ----

From: [ I Want My MTV ]

Hi John,

If you wish to post this, please make me anonymous.   I read ESNUG 476 #6
and was surprised at how much praise was being heaped on Redhawk.   The
user's post differs greatly from my own experience with both tools.

Through our 130 nm and 90 nm designs, we used Cadence's VoltageStorm for
static analysis.  Once we switched to 65 nm, we engaged with Apache.  For
our first 65 nm design we decided to use both Redhawk and VoltageStorm
for static and dynamic I/R analysis.

I found at the time Redhawk easy to use and it's runtimes slightly faster
than VoltageStorm.  Using the same input data, though, the problems
started when both tools reported vastly different I/R voltage drops in
both static and dynamic modes.  

The Redhawk results were roughly 4X more pessimistic than the VoltageStorm
results!

Of course, this caused great confusion, leaving us wondering which results
were correct.  We spent 2 weeks working with AEs from both companies trying
to determine the cause of the gap.  Both companies claimed that with the
inputs provided, their results were good.

Our chip works off of a battery, so obviously power was extremely important.
The Redhawk results threw us out of range with respect to our allowable I/R
drop, while with the VoltageStorm results we could proceed towards tapeout.
Not knowing which results were correct, we looked into ways we could add
metal to our power grids, add more vias, and adjust our routing to try and
free up space for more power routes.  All of this would've taken a
significant amount of effort and would have delayed our tapeout by weeks.

In the end, we decided not to add any additional power and taped out based
on the VoltageStorm results.  The thinking was -- we had used VoltageStorm
in the past with confidence, so we'd trust the tool now.   When we had
silicon, our decision was qualified.

Silicon measurements showed correlation to VoltageStorm, and that is the
true test of a tool.  Redhawk, as you can imagine, is no longer part of
our toolset.

Hope this helps someone.

    - [ I Want My MTV ]
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)