chore(ilp): optimize parser code on hot path

[puzpuzpuz]

WIP

Intermediate TSBS benchmarks on local machine.

Before:

puzpuzpuz@apechkurov-laptop:~/projects/tsbs/bin$ GOMAXPROCS=2 ./tsbs_load_questdb --file /tmp/data --workers 1
time,per. metric/s,metric total,overall metric/s,per. row/s,row total,overall row/s
1655103689,2429550.36,2.430000E+07,2429550.36,242955.04,2.430000E+06,242955.04
1655103699,2499935.04,4.930000E+07,2464739.90,249993.50,4.930000E+06,246473.99
1655103709,2440140.53,7.370000E+07,2456541.00,244014.05,7.370000E+06,245654.10
1655103719,2389846.52,9.760000E+07,2439867.22,238984.65,9.760000E+06,243986.72
^C
puzpuzpuz@apechkurov-laptop:~/projects/tsbs/bin$ GOMAXPROCS=2 ./tsbs_load_questdb --file /tmp/data --workers 1
time,per. metric/s,metric total,overall metric/s,per. row/s,row total,overall row/s
1655103798,3679632.82,3.680000E+07,3679632.82,367963.28,3.680000E+06,367963.28
1655103808,3669999.70,7.350000E+07,3674816.50,366999.97,7.350000E+06,367481.65
1655103818,3610246.50,1.096000E+08,3653294.86,361024.65,1.096000E+07,365329.49
1655103828,3739876.95,1.470000E+08,3674940.75,373987.70,1.470000E+07,367494.07
^C
puzpuzpuz@apechkurov-laptop:~/projects/tsbs/bin$ GOMAXPROCS=2 ./tsbs_load_questdb --file /tmp/data --workers 1
time,per. metric/s,metric total,overall metric/s,per. row/s,row total,overall row/s
1655104001,3329710.60,3.330000E+07,3329710.60,332971.06,3.330000E+06,332971.06
1655104011,3180113.59,6.510000E+07,3254916.68,318011.36,6.510000E+06,325491.67
1655104021,2439851.58,8.950000E+07,2983221.93,243985.16,8.950000E+06,298322.19
1655104031,2470115.39,1.142000E+08,2854953.38,247011.54,1.142000E+07,285495.34
1655104041,2449927.56,1.387000E+08,2773947.36,244992.76,1.387000E+07,277394.74
1655104051,2429975.33,1.630000E+08,2716619.11,242997.53,1.630000E+07,271661.91
^C

After:

puzpuzpuz@apechkurov-laptop:~/projects/tsbs/bin$ GOMAXPROCS=2 ./tsbs_load_questdb --file /tmp/data --workers 1
time,per. metric/s,metric total,overall metric/s,per. row/s,row total,overall row/s
1655104218,2289775.51,2.290000E+07,2289775.51,228977.55,2.290000E+06,228977.55
1655104228,2340160.38,4.630000E+07,2314965.85,234016.04,4.630000E+06,231496.58
1655104238,2619890.98,7.250000E+07,2416609.38,261989.10,7.250000E+06,241660.94
1655104248,2680036.03,9.930000E+07,2482464.21,268003.60,9.930000E+06,248246.42
1655104258,2580051.56,1.251000E+08,2501981.14,258005.16,1.251000E+07,250198.11
1655104268,2319948.45,1.483000E+08,2471641.99,231994.85,1.483000E+07,247164.20
^C
puzpuzpuz@apechkurov-laptop:~/projects/tsbs/bin$ GOMAXPROCS=2 ./tsbs_load_questdb --file /tmp/data --workers 1
time,per. metric/s,metric total,overall metric/s,per. row/s,row total,overall row/s
1655104344,2959728.73,2.960000E+07,2959728.73,295972.87,2.960000E+06,295972.87
1655104354,2779998.84,5.740000E+07,2869867.88,277999.88,5.740000E+06,286986.79
1655104364,2790012.71,8.530000E+07,2843250.39,279001.27,8.530000E+06,284325.04
1655104374,3279948.63,1.181000E+08,2952423.84,327994.86,1.181000E+07,295242.38
1655104384,2990232.14,1.480000E+08,2959984.88,299023.21,1.480000E+07,295998.49
1655104394,2779797.71,1.758000E+08,2929951.99,277979.77,1.758000E+07,292995.20
^C
puzpuzpuz@apechkurov-laptop:~/projects/tsbs/bin$ GOMAXPROCS=2 ./tsbs_load_questdb --file /tmp/data --workers 1
time,per. metric/s,metric total,overall metric/s,per. row/s,row total,overall row/s
1655104453,2099837.06,2.100000E+07,2099837.06,209983.71,2.100000E+06,209983.71
1655104463,1990007.35,4.090000E+07,2044924.44,199000.74,4.090000E+06,204492.44
1655104473,2430062.96,6.520000E+07,2173298.56,243006.30,6.520000E+06,217329.86
1655104483,2359869.49,8.880000E+07,2219942.67,235986.95,8.880000E+06,221994.27
1655104493,1969981.45,1.085000E+08,2169951.08,196998.14,1.085000E+07,216995.11
^C

[ideoma]

[PR Coverage check]

:heart_eyes: pass : 22 / 23 (95.65%)

file detail

	path	covered line	new line	coverage
:large_blue_circle:	io/questdb/cutlass/line/tcp/LineTcpParser.java	22	23	95.65%

[puzpuzpuz]

I'll keep this PR pending for some time. The thing is that there is a source of jitter in the scenario under test and I'd like to understand and fix it first. Otherwise, it's very problematic to judge on any optimizations.

[puzpuzpuz]

Closing this one since it's been handing in the air for a while. Hope to return back to parser optimization one day.