Merge branch 'elliot/mem-bench' into 'main'

Memory Read Bandwidth for x86

See merge request course-asl-fastcode26/team70!52

measurements/memory/x86/1.log

@@ -0,0 +1,37 @@
+sysbench 1.0.20 (using system LuaJIT 2.1.1761786044)
+
+Running the test with following options:
+Number of threads: 1
+Initializing random number generator from current time
+
+
+Running memory speed test with the following options:
+  block size: 1048576KiB
+  total size: 20480MiB
+  operation: read
+  scope: global
+
+Initializing worker threads...
+
+Threads started!
+
+Total operations: 20 (   37.08 per second)
+
+20480.00 MiB transferred (37967.45 MiB/sec)
+
+
+General statistics:
+    total time:                          0.5378s
+    total number of events:              20
+
+Latency (ms):
+         min:                                   25.99
+         avg:                                   26.87
+         max:                                   33.18
+         95th percentile:                       28.16
+         sum:                                  537.47
+
+Threads fairness:
+    events (avg/stddev):           20.0000/0.00
+    execution time (avg/stddev):   0.5375/0.00
+

measurements/memory/x86/2.log

@@ -0,0 +1,37 @@
+sysbench 1.0.20 (using system LuaJIT 2.1.1761786044)
+
+Running the test with following options:
+Number of threads: 1
+Initializing random number generator from current time
+
+
+Running memory speed test with the following options:
+  block size: 1048576KiB
+  total size: 20480MiB
+  operation: read
+  scope: global
+
+Initializing worker threads...
+
+Threads started!
+
+Total operations: 20 (   37.63 per second)
+
+20480.00 MiB transferred (38536.77 MiB/sec)
+
+
+General statistics:
+    total time:                          0.5299s
+    total number of events:              20
+
+Latency (ms):
+         min:                                   25.63
+         avg:                                   26.48
+         max:                                   36.38
+         95th percentile:                       27.66
+         sum:                                  529.61
+
+Threads fairness:
+    events (avg/stddev):           20.0000/0.00
+    execution time (avg/stddev):   0.5296/0.00
+

measurements/memory/x86/3.log

@@ -0,0 +1,37 @@
+sysbench 1.0.20 (using system LuaJIT 2.1.1761786044)
+
+Running the test with following options:
+Number of threads: 1
+Initializing random number generator from current time
+
+
+Running memory speed test with the following options:
+  block size: 1048576KiB
+  total size: 20480MiB
+  operation: read
+  scope: global
+
+Initializing worker threads...
+
+Threads started!
+
+Total operations: 20 (   37.35 per second)
+
+20480.00 MiB transferred (38249.60 MiB/sec)
+
+
+General statistics:
+    total time:                          0.5339s
+    total number of events:              20
+
+Latency (ms):
+         min:                                   26.01
+         avg:                                   26.68
+         max:                                   33.21
+         95th percentile:                       27.66
+         sum:                                  533.54
+
+Threads fairness:
+    events (avg/stddev):           20.0000/0.00
+    execution time (avg/stddev):   0.5335/0.00
+

measurements/variant_baseline/x86/2.log

@@ -0,0 +1,58 @@
+rm -f main_*  run_tests *.o *.s variants/*.s variants/*.o headers/*.s headers/*.o
+/usr/bin/g++ main.cpp ./variants/variant_baseline.cpp -o main_baseline -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: 183700.00 flops
+Q(n) per iteration: 6512 bytes
+I(n): 28.21 flops/byte
+W(n): 36740000.00 flops
+T(n): 27372765.00 cycles
+P(n): 1.34 flops/cycle
+
+=== N_P=100 N_ITER=200 ===
+W(n) per iteration: 727400.00 flops
+Q(n) per iteration: 12912 bytes
+I(n): 56.34 flops/byte
+W(n): 145480000.00 flops
+T(n): 107682440.00 cycles
+P(n): 1.35 flops/cycle
+
+=== N_P=200 N_ITER=200 ===
+W(n) per iteration: 2894800.00 flops
+Q(n) per iteration: 25712 bytes
+I(n): 112.59 flops/byte
+W(n): 578960000.00 flops
+T(n): 434952380.00 cycles
+P(n): 1.33 flops/cycle
+
+=== N_P=500 N_ITER=200 ===
+W(n) per iteration: 18037000.00 flops
+Q(n) per iteration: 64112 bytes
+I(n): 281.34 flops/byte
+W(n): 3607400000.00 flops
+T(n): 2803955380.00 cycles
+P(n): 1.29 flops/cycle
+
+=== N_P=1000 N_ITER=200 ===
+W(n) per iteration: 72074000.00 flops
+Q(n) per iteration: 128112 bytes
+I(n): 562.59 flops/byte
+W(n): 14414800000.00 flops
+T(n): 11379685260.00 cycles
+P(n): 1.27 flops/cycle
+
+=== N_P=2000 N_ITER=200 ===
+W(n) per iteration: 288148000.00 flops
+Q(n) per iteration: 256112 bytes
+I(n): 1125.09 flops/byte
+W(n): 57629600000.00 flops
+T(n): 46349779000.00 cycles
+P(n): 1.24 flops/cycle
+
+=== N_P=5000 N_ITER=200 ===
+W(n) per iteration: 1800370000.00 flops
+Q(n) per iteration: 640112 bytes
+I(n): 2812.59 flops/byte
+W(n): 360074000000.00 flops
+T(n): 312436846520.00 cycles
+P(n): 1.15 flops/cycle
+

scripts/mem-bench.sh

@@ -0,0 +1 @@
+sudo sysbench memory --memory-block-size=1G --memory-total-size=20G --memory-oper=read run