x86 measurements
This commit is contained in:
committed by
Linus Baumberger
parent
c16a67a1f8
commit
3aa84d6b8f
@@ -1,33 +1,58 @@
|
||||
rm -f main_* sizes run_tests *.o *.s
|
||||
/usr/bin/g++ -O3 -march=native -std=c++23 -Wall -Wextra -flto -DFLOPCOUNT=0 -DRANDOM_SEED=0 -DROOFLINE=0 -DBENCHMARK=1 main.cpp ./variants/variant_inlining_7_symmetry_2_ilp_1.cpp -o main_inlining_7_symmetry_2_ilp_1
|
||||
=== N_P=50 N_ITER=1000 ===
|
||||
W(n) per iteration: 184500.00 flops
|
||||
rm -f main_* run_tests *.o *.s variants/*.s variants/*.o headers/*.s headers/*.o
|
||||
/usr/bin/g++ main-symmetry.cpp ./variants/variant_inlining_7_symmetry_2_ilp_1.cpp -o main_inlining_7_symmetry_2_ilp_1 -O3 -march=native -std=c++23 -Wall -Wextra -flto -DFLOPCOUNT=0 -DRANDOM_SEED=0 -DROOFLINE=0 -DBENCHMARK=1
|
||||
=== N_P=50 N_ITER=200 ===
|
||||
W(n) per iteration: 95500.00 flops
|
||||
Q(n) per iteration: 6512 bytes
|
||||
I(n): 28.33 flops/byte
|
||||
W(n): 184500000.00 flops
|
||||
T(n): 134117580.00 cycles
|
||||
P(n): 1.38 flops/cycle
|
||||
I(n): 14.67 flops/byte
|
||||
W(n): 19100000.00 flops
|
||||
T(n): 26906550.00 cycles
|
||||
P(n): 0.71 flops/cycle
|
||||
|
||||
=== N_P=100 N_ITER=1000 ===
|
||||
W(n) per iteration: 729000.00 flops
|
||||
=== N_P=100 N_ITER=200 ===
|
||||
W(n) per iteration: 371000.00 flops
|
||||
Q(n) per iteration: 12912 bytes
|
||||
I(n): 56.46 flops/byte
|
||||
W(n): 729000000.00 flops
|
||||
T(n): 529885740.00 cycles
|
||||
P(n): 1.38 flops/cycle
|
||||
I(n): 28.73 flops/byte
|
||||
W(n): 74200000.00 flops
|
||||
T(n): 103009840.00 cycles
|
||||
P(n): 0.72 flops/cycle
|
||||
|
||||
=== N_P=200 N_ITER=1000 ===
|
||||
W(n) per iteration: 2898000.00 flops
|
||||
=== N_P=200 N_ITER=200 ===
|
||||
W(n) per iteration: 1462000.00 flops
|
||||
Q(n) per iteration: 25712 bytes
|
||||
I(n): 112.71 flops/byte
|
||||
W(n): 2898000000.00 flops
|
||||
T(n): 2158215760.00 cycles
|
||||
P(n): 1.34 flops/cycle
|
||||
I(n): 56.86 flops/byte
|
||||
W(n): 292400000.00 flops
|
||||
T(n): 415634800.00 cycles
|
||||
P(n): 0.70 flops/cycle
|
||||
|
||||
=== N_P=500 N_ITER=1000 ===
|
||||
W(n) per iteration: 18045000.00 flops
|
||||
=== N_P=500 N_ITER=200 ===
|
||||
W(n) per iteration: 9055000.00 flops
|
||||
Q(n) per iteration: 64112 bytes
|
||||
I(n): 281.46 flops/byte
|
||||
W(n): 18045000000.00 flops
|
||||
T(n): 14658868260.00 cycles
|
||||
P(n): 1.23 flops/cycle
|
||||
I(n): 141.24 flops/byte
|
||||
W(n): 1811000000.00 flops
|
||||
T(n): 2788626920.00 cycles
|
||||
P(n): 0.65 flops/cycle
|
||||
|
||||
=== N_P=1000 N_ITER=200 ===
|
||||
W(n) per iteration: 36110000.00 flops
|
||||
Q(n) per iteration: 128112 bytes
|
||||
I(n): 281.86 flops/byte
|
||||
W(n): 7222000000.00 flops
|
||||
T(n): 11839944240.00 cycles
|
||||
P(n): 0.61 flops/cycle
|
||||
|
||||
=== N_P=2000 N_ITER=200 ===
|
||||
W(n) per iteration: 144220000.00 flops
|
||||
Q(n) per iteration: 256112 bytes
|
||||
I(n): 563.11 flops/byte
|
||||
W(n): 28844000000.00 flops
|
||||
T(n): 47471597160.00 cycles
|
||||
P(n): 0.61 flops/cycle
|
||||
|
||||
=== N_P=5000 N_ITER=200 ===
|
||||
W(n) per iteration: 900550000.00 flops
|
||||
Q(n) per iteration: 640112 bytes
|
||||
I(n): 1406.86 flops/byte
|
||||
W(n): 180110000000.00 flops
|
||||
T(n): 229239879220.00 cycles
|
||||
P(n): 0.79 flops/cycle
|
||||
|
||||
|
||||
@@ -1,33 +0,0 @@
|
||||
rm -f main_* sizes run_tests *.o *.s
|
||||
/usr/bin/g++ -O3 -march=native -std=c++23 -Wall -Wextra -flto -DFLOPCOUNT=0 -DRANDOM_SEED=0 -DROOFLINE=0 -DBENCHMARK=1 main.cpp ./variants/variant_inlining_7_symmetry_2_ilp_1.cpp -o main_inlining_7_symmetry_2_ilp_1
|
||||
=== N_P=50 N_ITER=1000 ===
|
||||
W(n) per iteration: 184500.00 flops
|
||||
Q(n) per iteration: 6512 bytes
|
||||
I(n): 28.33 flops/byte
|
||||
W(n): 184500000.00 flops
|
||||
T(n): 135847100.00 cycles
|
||||
P(n): 1.36 flops/cycle
|
||||
|
||||
=== N_P=100 N_ITER=1000 ===
|
||||
W(n) per iteration: 729000.00 flops
|
||||
Q(n) per iteration: 12912 bytes
|
||||
I(n): 56.46 flops/byte
|
||||
W(n): 729000000.00 flops
|
||||
T(n): 531086160.00 cycles
|
||||
P(n): 1.37 flops/cycle
|
||||
|
||||
=== N_P=200 N_ITER=1000 ===
|
||||
W(n) per iteration: 2898000.00 flops
|
||||
Q(n) per iteration: 25712 bytes
|
||||
I(n): 112.71 flops/byte
|
||||
W(n): 2898000000.00 flops
|
||||
T(n): 2149640340.00 cycles
|
||||
P(n): 1.35 flops/cycle
|
||||
|
||||
=== N_P=500 N_ITER=1000 ===
|
||||
W(n) per iteration: 18045000.00 flops
|
||||
Q(n) per iteration: 64112 bytes
|
||||
I(n): 281.46 flops/byte
|
||||
W(n): 18045000000.00 flops
|
||||
T(n): 14662439140.00 cycles
|
||||
P(n): 1.23 flops/cycle
|
||||
@@ -1,33 +0,0 @@
|
||||
rm -f main_* sizes run_tests *.o *.s
|
||||
/usr/bin/g++ -O3 -march=native -std=c++23 -Wall -Wextra -flto -DFLOPCOUNT=0 -DRANDOM_SEED=0 -DROOFLINE=0 -DBENCHMARK=1 main.cpp ./variants/variant_inlining_7_symmetry_2_ilp_1.cpp -o main_inlining_7_symmetry_2_ilp_1
|
||||
=== N_P=50 N_ITER=1000 ===
|
||||
W(n) per iteration: 184500.00 flops
|
||||
Q(n) per iteration: 6512 bytes
|
||||
I(n): 28.33 flops/byte
|
||||
W(n): 184500000.00 flops
|
||||
T(n): 132982240.00 cycles
|
||||
P(n): 1.39 flops/cycle
|
||||
|
||||
=== N_P=100 N_ITER=1000 ===
|
||||
W(n) per iteration: 729000.00 flops
|
||||
Q(n) per iteration: 12912 bytes
|
||||
I(n): 56.46 flops/byte
|
||||
W(n): 729000000.00 flops
|
||||
T(n): 530176140.00 cycles
|
||||
P(n): 1.38 flops/cycle
|
||||
|
||||
=== N_P=200 N_ITER=1000 ===
|
||||
W(n) per iteration: 2898000.00 flops
|
||||
Q(n) per iteration: 25712 bytes
|
||||
I(n): 112.71 flops/byte
|
||||
W(n): 2898000000.00 flops
|
||||
T(n): 2162889920.00 cycles
|
||||
P(n): 1.34 flops/cycle
|
||||
|
||||
=== N_P=500 N_ITER=1000 ===
|
||||
W(n) per iteration: 18045000.00 flops
|
||||
Q(n) per iteration: 64112 bytes
|
||||
I(n): 281.46 flops/byte
|
||||
W(n): 18045000000.00 flops
|
||||
T(n): 14697507480.00 cycles
|
||||
P(n): 1.23 flops/cycle
|
||||
Reference in New Issue
Block a user