Subject: Re: [SI-LIST] update to "ack" program
From: Gregory R Edlund (gedlund@us.ibm.com)
Date: Mon Feb 25 2002 - 06:14:40 PST
Dear SI and IBIS compatriots:
Ray Anderson ran into a bus error when he ran ack on the Solaris platform.
Apparently there is still a bug lurking in my code that didn't show up when
I tested it on the aix and Windows platforms. Believe it or not, I used
the program on a couple dozen waveforms for a project I worked on over a
year ago, and I never ran into a snag. I'm going to check around and see
if I can come up with a Sun box that I can use for further testing. In the
meantime, if anyone wants to take a crack at debugging, be my guest! It's
probably another one of those memory conflicts that are so easy to fall
into when you're allocating memory in C.
Sorry for the trouble. You may still be able to run on Windows or aix
without trouble.
Greg Edlund
Electronic Packaging & Integration
IBM Server Technology Development
3605 Hwy. 52 N, Dept. HDC
Rochester, MN 55901
gedlund@us.ibm.com
----- Forwarded by Gregory R Edlund/Rochester/IBM on 02/25/02 08:06 AM
-----
Ray Anderson
<Raymond.Anderson To: Gregory R Edlund/Rochester/IBM@IBMUS
@sun.com> cc: raymond.anderson@sun.com
Subject: Re: [SI-LIST] update to "ack" program
02/22/02 03:35 PM
Please respond to
Ray Anderson
Greg-
I've compiled your ack.c code with both Gnu gcc and
also with Sun cc with no apparent errors.
When I try to run the program on your sample data using the
command line specified in your readme.txt file I get a "Bus Error"
crash (with both the gcc and cc compiled versions). Have you seen
this before? (I'm running on a Sun Ultra-1 with Solaris 8)
-Ray Anderson
Sun Microsystems
-------------------------------------------------------------------------------------
radium% ack -f1 hspice.dat -f2 lab.dat -e r -dx 0.05 -xu 1e-9 -yt 1.4 -yw
3.3 > out.dat
Accuracy Figure-of-Merit Calculator
Found 501 datapoints in hspice.dat.
Found 301 datapoints in lab.dat.
Waveform 1 crosses y = 1.4000e+00 at x = 1.6615e+00.
Waveform 2 crosses y = 1.4000e+00 at x = 1.1407e+00.
Shifting waveform 2 right by 5.2086e-01.
Bounding interval = ( 5.2086e-01, 6.5209e+00 )
Bus Error
-------------------------------------------------------------------------------------
>
>Hello Everyone,
>
>I recently made a long-overdue bug fix to "ack," a C program I wrote a
>while back that overlays two waveforms (in ASCII x-y format) and
>numerically compares the datapoints. This program is useful in comparing
>lab vs. simulation or behavioral vs. transistor-level simulations. The
bug
>was causing floating point overflows in certain cases.
>
>You can find the new code on the IBIS web site. It's a free-ware program
>that you have to compile yourself on your favorite system.
>
>http://www.vhdl.org/pub/ibis/accuracy/
This archive was generated by hypermail 2b28 : Mon Feb 25 2002 - 06:43:34 PST