There is a 2% Scimark regression across all platforms.

http://alacrity.sfbay/nq/all/nes.jsp?base=3289,3290,3291,3292,3293,3294,3295,3296,3389&build=3452,3453,3454,3455,3456,3457,3458,3578,3580&pval=0.01

This is caused by the putback 20031023005430.ysr.cms since running the benchmark with the previous putback does not show any regression.

Although this might be related to cache alignment problems notified in bug 4914629, this regression is hitting all platforms and thus we want to make sure it is not caused by any gc changes from this putback.


benkim@pacecar ../scimark<61>% /export/benkim/1027/j2sdk1.5.0/bin/java -server -showversion -Xmx256m -Xms256m jnt.scimark2.commandline -large
java version "1.5.0-beta"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.5.0-beta-b26)
Java HotSpot(TM) Server VM (build 20031022101736.dcubed.service_hs_baseline, mixed mode)

Composite Score: 35.10459348805652
FFT (1048576): 8.468977309061597
SOR (1000x1000): 85.1249214437096
Monte Carlo : 7.456540444444443
Sparse matmult (N=100000, nz=1000000): 24.60592125981076
LU (1000x1000): 49.86660698325624

java.vendor: Sun Microsystems Inc.
java.version: 1.5.0-beta
os.arch: sparc
os.name: SunOS
os.version: 5.9


benkim@pacecar ../scimark<58>% /export/benkim/1027/j2sdk1.5.0/bin/java -server -showversion -Xmx256m -Xms256m jnt.scimark2.commandline -large
java version "1.5.0-beta"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.5.0-beta-b26)
Java HotSpot(TM) Server VM (build 20031023005430.ysr.cms, mixed mode)

Composite Score: 34.260402436887816
FFT (1048576): 6.931608189473579
SOR (1000x1000): 84.15291442311087
Monte Carlo : 7.62600710743802
Sparse matmult (N=100000, nz=1000000): 24.233244685436883
LU (1000x1000): 48.35823777897972

java.vendor: Sun Microsystems Inc.
java.version: 1.5.0-beta
os.arch: sparc
os.name: SunOS
os.version: 5.9