COLLOQUIUM
Department of Computer Science and Engineering
University of South Carolina
 
Minimal Subset Evaluation:
Rapid Warm-up for Simulated Hardware State  

John Williams Haskins, Jr.
Department of Computer Science
University of Virginia 

Date: February 22, 2002 (Friday)
Time:  4:00-5:00PM
Place: Swearingen Faculty Lounge (SWGN 1A03)

Abstract

The Minimal Subset Evaluation (MSE) technique is presented as a way to
reduce time spent on large-structure warm-up during the
fast-forwarding portion of microprocessor simulations.  Warm up is
commonly used prior to full-detail simulation to avoid cold- start
bias in large on-chip structures like caches and branch predictors.
Unfortunately, warm up can be very time consuming, often representing
50% or more of total simulation time.  Previous techniques have used
the entire fast-forward interval to obtain accurate warm up, which may
be prohibitive for large parameter-space searches, or chosen a short
but ad-hoc warm-up length that reduces simulation time but may
sacrifice accuracy.  MSE uses combinatorics to determine---with
user-specified probability of accuracy---a minimally sufficient
fraction of contiguous fast-forward instructions that must be executed
for warm up to accurately produce state as it would have appeared had
the entire fast-forward interval been used for warm up.  For instance,
only the most recent memory references' cache transactions need to be
modeled to yield simulated cache state identical to the state that
would have resulted had all memory references been modeled.  By
modeling fewer memory references' cache transactions, simulation
running time is significantly reduced.My presentation motivates the
MSE technique by describing its mathematical underpinnings and
demonstrating its effectiveness for both single-large- sample and
multiple-sample simulation styles.  In preliminary experiments, MSE
yields errors of less than 1% in IPC measurements with cycle-accurate
simulation, while reducing simulation times by an average factor of
two or more.

John Haskins, Jr. is pursuing his doctorate at the University of
Virginia Department of Computer Science where he earned his MCS in
2000.  He earned his BS in computer science from Georgia Tech in 1997.
Between the conclusion of his undergraduate studies and the beginning
of graduate studies, he worked as an adjunct research staff member at
the IDA Center for Computing Sciences.  His technical interests
include computer architecture and operating systems.