Benchmarking: CARMA Dev Kit - Linpack
Linkpack
file linpackc.new
for The LINPACK benchmark program done in C (improved version).
lang C
by Based on Bonnie Toy 5/88 but improved by Will Menninger.
wget http://www.netlib.org/benchmark/linpackc.new
Compile
mv linpackc.new linpack.c gcc -O -o linpack linpack.c -lm
CPU info
cat /proc/cpuinfo
Processor : ARMv7 Processor rev 9 (v7l) processor : 0 BogoMIPS : 1987.37 processor : 1 BogoMIPS : 1987.37 processor : 2 BogoMIPS : 1987.37 processor : 3 BogoMIPS : 1987.37 Features : swp half thumb fastmult vfp edsp neon vfpv3 tls CPU implementer : 0x41 CPU architecture: 7 CPU variant : 0x2 CPU part : 0xc09 CPU revision : 9 Hardware : cardhu Revision : 0000 Serial : 0000000000000000
Mem info
free -m
total used free shared buffers cached Mem: 1996 996 999 0 6 180 -/+ buffers/cache: 809 1186 Swap: 0 0 0
Results
Double Precision N=200
Enter array size (q to quit) [200]: 200
Memory required: 315K.
LINPACK benchmark, Double precision.
Machine precision: 15 digits.
Array size 200 X 200.
Average rolled and unrolled performance:
Reps Time(s) DGEFA DGESL OVERHEAD KFLOPS
----------------------------------------------------
64 0.52 84.62% 3.85% 11.54% 191072.464
128 1.04 85.58% 1.92% 12.50% 193172.161
256 2.08 90.87% 3.37% 5.77% 179374.150
512 4.15 89.16% 3.61% 7.23% 182635.498
1024 8.31 91.10% 2.89% 6.02% 180063.167
2048 16.59 89.87% 2.65% 7.47% 183230.402
Double Precision N=1000
Enter array size (q to quit) [200]: 1000
Memory required: 7824K.
LINPACK benchmark, Double precision.
Machine precision: 15 digits.
Array size 1000 X 1000.
Average rolled and unrolled performance:
Reps Time(s) DGEFA DGESL OVERHEAD KFLOPS
----------------------------------------------------
1 1.93 92.75% 0.52% 6.74% 93148.148
2 3.18 98.11% 0.63% 1.26% 106794.055
4 6.34 98.11% 0.63% 1.26% 107135.250
8 12.67 97.79% 0.95% 1.26% 107220.890
Double Precision N=5000
Enter array size (q to quit) [200]: 5000
Memory required: 195371K.
LINPACK benchmark, Double precision.
Machine precision: 15 digits.
Array size 5000 X 5000.
Average rolled and unrolled performance:
Reps Time(s) DGEFA DGESL OVERHEAD KFLOPS
----------------------------------------------------
1 315.83 99.57% 0.11% 0.32% 66256.896
Single Precision N=200
Enter array size (q to quit) [200]: 200
Memory required: 158K.
LINPACK benchmark, Single precision.
Machine precision: 6 digits.
Array size 200 X 200.
Average rolled and unrolled performance:
Reps Time(s) DGEFA DGESL OVERHEAD KFLOPS
----------------------------------------------------
128 0.84 89.29% 0.00% 10.71% 234382.266
256 1.69 88.76% 2.96% 8.28% 226821.516
512 3.38 88.76% 1.48% 9.76% 230539.859
1024 6.77 88.33% 3.25% 8.42% 226821.500
2048 13.52 88.17% 2.81% 9.02% 228665.875
Single Precision N=1000
Enter array size (q to quit) [200]: 1000
Memory required: 3914K.
LINPACK benchmark, Single precision.
Machine precision: 6 digits.
Array size 1000 X 1000.
Average rolled and unrolled performance:
Reps Time(s) DGEFA DGESL OVERHEAD KFLOPS
----------------------------------------------------
1 1.21 89.26% 0.83% 9.92% 153822.609
2 1.79 97.21% 0.56% 2.23% 191619.047
4 3.60 97.22% 0.83% 1.94% 189990.734
8 7.20 97.64% 0.56% 1.81% 189721.844
16 14.40 97.36% 0.83% 1.81% 189721.844
Single Precision N=5000
Enter array size (q to quit) [200]: 5000
Memory required: 97695K.
LINPACK benchmark, Single precision.
Machine precision: 6 digits.
Array size 5000 X 5000.
Average rolled and unrolled performance:
Reps Time(s) DGEFA DGESL OVERHEAD KFLOPS
----------------------------------------------------
1 167.10 99.49% 0.11% 0.39% 125320.430