Ahmed moreilp

This commit is contained in:
Ahmed Kallala
2026-06-17 03:01:19 +02:00
parent 42b3c7097e
commit 816626664c
167 changed files with 6415 additions and 269 deletions
@@ -5,15 +5,15 @@ W(n) per iteration: 95500.00 flops
Q(n) per iteration: 3256 bytes
I(n): 29.33 flops/byte
W(n): 19100000.00 flops
T(n): 9431654.00 cycles
P(n): 2.03 flops/cycle
T(n): 9376726.00 cycles
P(n): 2.04 flops/cycle
=== N_P=100 N_ITER=200 ===
W(n) per iteration: 371000.00 flops
Q(n) per iteration: 6456 bytes
I(n): 57.47 flops/byte
W(n): 74200000.00 flops
T(n): 33563136.00 cycles
T(n): 33509446.00 cycles
P(n): 2.21 flops/cycle
=== N_P=200 N_ITER=200 ===
@@ -21,23 +21,23 @@ W(n) per iteration: 1462000.00 flops
Q(n) per iteration: 12856 bytes
I(n): 113.72 flops/byte
W(n): 292400000.00 flops
T(n): 118019376.00 cycles
P(n): 2.48 flops/cycle
T(n): 118757464.00 cycles
P(n): 2.46 flops/cycle
=== N_P=500 N_ITER=200 ===
W(n) per iteration: 9055000.00 flops
Q(n) per iteration: 32056 bytes
I(n): 282.47 flops/byte
W(n): 1811000064.00 flops
T(n): 678422208.00 cycles
P(n): 2.67 flops/cycle
T(n): 676173376.00 cycles
P(n): 2.68 flops/cycle
=== N_P=1000 N_ITER=200 ===
W(n) per iteration: 36110000.00 flops
Q(n) per iteration: 64056 bytes
I(n): 563.73 flops/byte
W(n): 7222000128.00 flops
T(n): 2959266048.00 cycles
T(n): 2963177216.00 cycles
P(n): 2.44 flops/cycle
=== N_P=2000 N_ITER=200 ===
@@ -45,7 +45,7 @@ W(n) per iteration: 144220000.00 flops
Q(n) per iteration: 128056 bytes
I(n): 1126.23 flops/byte
W(n): 28843999232.00 flops
T(n): 12601305088.00 cycles
T(n): 12573572096.00 cycles
P(n): 2.29 flops/cycle
=== N_P=5000 N_ITER=200 ===
@@ -53,6 +53,6 @@ W(n) per iteration: 900550016.00 flops
Q(n) per iteration: 320056 bytes
I(n): 2813.73 flops/byte
W(n): 180110000128.00 flops
T(n): 144715956224.00 cycles
P(n): 1.24 flops/cycle
T(n): 143544270848.00 cycles
P(n): 1.25 flops/cycle
@@ -0,0 +1,58 @@
rm -f main_* run_tests *.o *.s variants/*.s variants/*.o headers/*.s headers/*.o
sudo /usr/bin/g++ main-float_symmetry.cpp ./variants/variant_float_inlining_SoA_symmetry_ILP_4.cpp -o main_float_inlining_SoA_symmetry_ILP_4 -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: 3256 bytes
I(n): 29.33 flops/byte
W(n): 19100000.00 flops
T(n): 9522074.00 cycles
P(n): 2.01 flops/cycle
=== N_P=100 N_ITER=200 ===
W(n) per iteration: 371000.00 flops
Q(n) per iteration: 6456 bytes
I(n): 57.47 flops/byte
W(n): 74200000.00 flops
T(n): 33704588.00 cycles
P(n): 2.20 flops/cycle
=== N_P=200 N_ITER=200 ===
W(n) per iteration: 1462000.00 flops
Q(n) per iteration: 12856 bytes
I(n): 113.72 flops/byte
W(n): 292400000.00 flops
T(n): 119636704.00 cycles
P(n): 2.44 flops/cycle
=== N_P=500 N_ITER=200 ===
W(n) per iteration: 9055000.00 flops
Q(n) per iteration: 32056 bytes
I(n): 282.47 flops/byte
W(n): 1811000064.00 flops
T(n): 733238656.00 cycles
P(n): 2.47 flops/cycle
=== N_P=1000 N_ITER=200 ===
W(n) per iteration: 36110000.00 flops
Q(n) per iteration: 64056 bytes
I(n): 563.73 flops/byte
W(n): 7222000128.00 flops
T(n): 3048842752.00 cycles
P(n): 2.37 flops/cycle
=== N_P=2000 N_ITER=200 ===
W(n) per iteration: 144220000.00 flops
Q(n) per iteration: 128056 bytes
I(n): 1126.23 flops/byte
W(n): 28843999232.00 flops
T(n): 13217501184.00 cycles
P(n): 2.18 flops/cycle
=== N_P=5000 N_ITER=200 ===
W(n) per iteration: 900550016.00 flops
Q(n) per iteration: 320056 bytes
I(n): 2813.73 flops/byte
W(n): 180110000128.00 flops
T(n): 145096245248.00 cycles
P(n): 1.24 flops/cycle
@@ -0,0 +1,58 @@
rm -f main_* run_tests *.o *.s variants/*.s variants/*.o headers/*.s headers/*.o
sudo /usr/bin/g++ main-float_symmetry.cpp ./variants/variant_float_inlining_SoA_symmetry_ILP_4.cpp -o main_float_inlining_SoA_symmetry_ILP_4 -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: 3256 bytes
I(n): 29.33 flops/byte
W(n): 19100000.00 flops
T(n): 9386540.00 cycles
P(n): 2.03 flops/cycle
=== N_P=100 N_ITER=200 ===
W(n) per iteration: 371000.00 flops
Q(n) per iteration: 6456 bytes
I(n): 57.47 flops/byte
W(n): 74200000.00 flops
T(n): 33523044.00 cycles
P(n): 2.21 flops/cycle
=== N_P=200 N_ITER=200 ===
W(n) per iteration: 1462000.00 flops
Q(n) per iteration: 12856 bytes
I(n): 113.72 flops/byte
W(n): 292400000.00 flops
T(n): 119032680.00 cycles
P(n): 2.46 flops/cycle
=== N_P=500 N_ITER=200 ===
W(n) per iteration: 9055000.00 flops
Q(n) per iteration: 32056 bytes
I(n): 282.47 flops/byte
W(n): 1811000064.00 flops
T(n): 676348160.00 cycles
P(n): 2.68 flops/cycle
=== N_P=1000 N_ITER=200 ===
W(n) per iteration: 36110000.00 flops
Q(n) per iteration: 64056 bytes
I(n): 563.73 flops/byte
W(n): 7222000128.00 flops
T(n): 2969299712.00 cycles
P(n): 2.43 flops/cycle
=== N_P=2000 N_ITER=200 ===
W(n) per iteration: 144220000.00 flops
Q(n) per iteration: 128056 bytes
I(n): 1126.23 flops/byte
W(n): 28843999232.00 flops
T(n): 12426423296.00 cycles
P(n): 2.32 flops/cycle
=== N_P=5000 N_ITER=200 ===
W(n) per iteration: 900550016.00 flops
Q(n) per iteration: 320056 bytes
I(n): 2813.73 flops/byte
W(n): 180110000128.00 flops
T(n): 140797083648.00 cycles
P(n): 1.28 flops/cycle