3/25/2003
A. Coding for increasing efficiency.
Status: Using trimos-send.acc as a benchmark. Before starting effort, runs in 5.2 seconds on shang. After removing duplicate edges from stabilized list, 3.7 seconds. After removing visited field as termination point during propagation, 1.8 seconds. After removing visited field in timed and untimed stabilization routines, 1.25 seconds. Generated the following questions:
1. Are state assignments consistent for Atacs?
2. Is the state number usable for uniquely identifying states?
3. How to profile the verification run for comparison?
4. Propagating a 1 or 0 through a state with multiple incoming edges …
B. Code for reporting hazards of my specific
decomposition nodelists.
Status Coding done except for monotonicity check. Two ways I could go here …. Front burner.
C. Go through the monotonicity check code
carefully.
Status: Middle burner.
D. Getting gprof running under Cygwin and using
it for optimization.
Status: Not working and chances are slim.
E. Questions for Japan trip ….
F. Thesis writing and status …
D. Code to update the arrival times at the
inputs.
Status: Middle burner.
E. Code to do timed hazard checking with
level-based events.
Status: Back burner.
F. Code to output warning about timing issues.
Status: Back burner.
G. Code to read in and parse a library file in
SIS format.
Status. Back burner.
H. Code to free data structure memory after
use.
Status: Coding done but memory still leaking badly. Back burner.
I. Code to get the correct zone at the start of
timing analysis.
Status: Back burner.
A.
3. Other
issues.
A. Thesis.
1. Get style file, Peskin’s makefile up and
going, etc..
Status: Back burner.
B. ICCAD.
1. Get style file, etc.
Status: Middle burner.
C. Japan questions.
1.