Fix measurements for Lightsaber benchmark

[wzhao18]

Essentially turn on the performance monitor built-in in lightsaber application and repeatedly run the benchmark.

Default benchmark parameters are changed to size 40000000 and runs 100 times. Some sizes will cause the program seg fault, such as 10000000.

Result now looks like:

root@7ec97438f7d7:~/lightsaber_bench/build# ./lightsaber 
error: unknown argument: '-internal-isystem'
error: -emit-llvm cannot be used when linking
[Post optimization module] dumped to /root/data/scabbard/GeneratedCode_0.ll
[DBG] set aggregate operator
[DBG] 1 threads
[DBG] CPU-only execution
Successfully set thread 140413114452032 to affinity to CPU 0
[DBG] bind worker 1 to core 1
Successfully set thread 140412842067712 to affinity to CPU 1
[MON] [MultiOperator] 1654481851000 S000
[MON] 1654481852015 q 0 t 0 w 0 b 0 p 0
[MON] 1654481853015 q 999 t 1026 w 1026 b 0 p 1
[MON] 1654481854015 S000 3122.330 MB/s (163700000 t/sec) (Average: 163700000 t/sec) output 3122.139 MB/s (163690000 t/sec)  q 977 t 1026 w 1026 b 0 p 1
[MON] 1654481855015 S000 3713.799 MB/s (194710000 t/sec) (Average: 179205000 t/sec) output 3713.417 MB/s (194690000 t/sec)  q 997 t 1026 w 1026 b 0 p 1
[MON] 1654481856016 S000 4098.799 MB/s (215110000 t/sec) (Average: 191173333 t/sec) output 4098.418 MB/s (215090000 t/sec)  q 0 t 1026 w 1026 b 0 p 1
[MON] 1654481857016 S000 3867.149 MB/s (202750000 t/sec) (Average: 194067500 t/sec) output 3866.768 MB/s (202730000 t/sec)  q 1003 t 1026 w 1026 b 0 p 1
[MON] 1654481858016 S000 3762.245 MB/s (197250000 t/sec) (Average: 194704000 t/sec) output 3761.864 MB/s (197230000 t/sec)  q 0 t 1026 w 1026 b 0 p 1
[MON] 1654481859016 S000 4197.693 MB/s (220080000 t/sec) (Average: 198933333 t/sec) output 4197.311 MB/s (220060000 t/sec)  q 998 t 1026 w 1026 b 0 p 1

If turn on remote sink:

root@9e3280b55ce9:~/lightsaber_bench/build# ./remotesink select 10000000
Successfully read 10000000 tuples.
Successfully read 10000000 tuples.
Successfully read 10000000 tuples.
Successfully read 10000000 tuples.
Successfully read 10000000 tuples.
Successfully read 10000000 tuples.
Successfully read 10000000 tuples.
Successfully read 10000000 tuples.
Successfully read 10000000 tuples.
Successfully read 10000000 tuples.

[wzhao18]

Making this a draft for now, want to look at whether it's possible for the main thread to wait for the work threads to finish.

[anandj91]

This looks like code copied from C++. Why is this part of a bash script?