Hi,
seems there is a bug in LU.
I got a log like this:
LU Benchmark Completed
class_npb = C
Size = 162x 162x 162
Total threads = 16
Iterations = 250
Time in seconds = 48.91
Mop/s total = 41685.61
Operation type = floating point
Verification = SUCCESSFUL
Version = 4.1
Compile date = 27 Jan 2026
Compiler ver = 14.0.0
OpenMP version = 201811
Compile options:
CC = clang++ -std=c++14
CLINK = $(CC)
C_LIB = -lm
C_INC = -I../common
CFLAGS = -O3 -fopenmp -mcmodel=medium
CLINKFLAGS = -O3 -fopenmp -mcmodel=medium
RAND = (none)
----------------------------------------------------------------------
NPB-CPP is developed by:
Dalvan Griebler
Gabriell Araujo (Sequential Porting)
Júnior Löff (Parallel Implementation)
In case of questions or problems, please send an e-mail to us:
dalvan.griebler; gabriell.araujo; junior.loff@edu.pucrs.br
----------------------------------------------------------------------
SECTION Time (secs)
total : 48.914 (100.00%)
rhsx : 39.115 ( 79.97%)
rhsy : 69.802 (142.70%)
rhsz : 69.379 (141.84%)
rhs : 29.059 ( 59.41%)
--> sub-rhs: 178.296 (364.51%)
--> rest-rhs: -149.236 (-305.10%)
jacld : 3.317 ( 6.78%)
blts : 10.013 ( 20.47%)
jacu : 3.385 ( 6.92%)
buts : 10.522 ( 21.51%)
add : 2.211 ( 4.52%)
l2norm : 0.016 ( 0.03%)
You can see the "rest-rhs" is negative.
Hi,
seems there is a bug in LU.
I got a log like this:
You can see the "rest-rhs" is negative.