Dear Collaborator,
I will like to share with you the letter from Satoshi, as well as the
more detailed report from Bruce Gibbard on the MDC-I result.
regards
Flemming
/enc
"
Our RHIC Mock Data Challenge 1 was successfully concluded on Monday,
October
19, 1998. Attached is a brief summary written by Bruce Gibbard on his
assessment of the challenge.
I take this opportunity to express my appreciation to all participants
for
their efforts in this challenge, and my sincere thanks to those who
supported it.
Satoshi Ozaki
Dr. Satoshi Ozaki Doris Rueger
RHIC Project Director Asst. to Dr. Ozaki
Brookhaven National Laboratory (tel) 516-344-5990/5663
Building 1005 UPTON, NY 11973 (fax) 516-344-2166
----------------------------------------------------------------------------
---------------------------
Brief Summary of RHIC Mock Data Challenge 1
The RHIC Mock Data Challenge 1 (MDC-1) which began on Tuesday, September
8,
was successfully concluded on Monday, October 19. This six-week
exercise
involved the RHIC Computing Facility, the DOE High Energy and the
Nuclear
Physics Computational Grand Challenge Initiative and the four RHIC
experimental collaborations, BRAHMS, PHENIX, PHOBOS, and STAR. The
stated
primary goals of the exercise were to demonstrate the performance of
three
basic functions, 1) Event Data Recording, 2) Event Reconstruction, and
3)
Data Mining, each for multiple experiments simultaneously. During the
exercise, aggregate event data recording rates into HPSS for the four
experiments as high as 18 MBytes/sec for an 8 hours period were
measured.
Event reconstruction by the four experiments on a farm consisting of up
to
104 Pentium II processors, ~1400 SPECint95 of CPU capacity, were
achieved
with CPU utilization efficiencies for a 16 hour period averaging 80
percent
across the experiments. During simultaneous event data mining by the
four
experiments, a variety of data access measurements were made. These
included the performance of a SUN server versus network connected
Pentium
farm machines, the use of Grand Challenge Project software to coordinate
queries, and the use of an ORNL developed system to batch files for
access
from HPSS tapes. The Grand Challenge Project and STAR were also able to
build and exercise a DST level Objectivity event data store. Secondary
objectives of MDC-1, including running multiple simultaneous functions
for a
subset of the experiments and running for extended periods for
individual
experiments, 7 days for PHENIX and STAR, were also achieved.
>From the perspective of the RCF, MDC-1 was extremely valuable in terms
of
verifying and detailing expected properties and limitation of the
current
facility and by revealing some unexpected problems. As anticipated the
Managed Data Server (MDS) and in particular HPSS were found to be the
single
most complex and critical components of the exercise. HPSS showed
itself to
be capable of high levels of performance and adequate to the goals
defined
for MDC-1. However, it was also clear that the less than two months
between
its initial installation at RCF and its large scale use in MDC-1 were
not
adequate to achieve desired levels of reliability in operation. The
limited
storage resources, in particular tape drives, available for this
exercise
also contributed to the stress on HPSS. Except for a delay in vendor
delivery, the performance of the Intel based Linux processor farms
during
MDC-1 were gratifyingly close to that anticipated. An unexpected issues
encountered was the performance of the RHIC Wide Area Network. While
the
need to do tuning of network parameters on RCF and collaborating remote
machines was anticipated, end-to-end problems including ESnet and/or
commercial components of connectivity were more serious and less
tractable
than anticipated.
>From the perspective of the RCF, the ability of all six parties to
participate effectively in a unified exercise was the most important
outcome. In preparation for and during MDC-1 clear synergy was evident
in
the efforts of all involved. To the extent that this synergy continues
and
convergent iteration based on MDC-1 can be achieved, it seems realistic
to
anticipate computing in support of the RHIC experimental program which
is
both timely and effective.
This archive was generated by hypermail 2b29 : Tue Feb 01 2000 - 20:35:19 EST