CA: remove SIMDization and multithreading

• de-SIMDize and significantly cleanup the triplet construction

It is needed for further developments. We plan to add more flexibility to the triplet construction and implement approximated search windows. These windows will not require many calculations, therefore the SIMDization will not be needed anyhow.

• remove the (not working) multithreaded mode. Currently, the OpenMP code only complicates the tracker development. In the future, the code will be ported to GPU and will run in the OpenMP mode automatically via the XPU framework.

macro/L1/configs/ca_params_mcbm.yaml

@@ -26,7 +26,7 @@ ca:
       # 3) Turn the full TOF off
       #     inactive_stations: [TOF]
       #inactive_stations: ['MUCH']
-      inactive_stations: []
+      inactive_stations: ['MUCH']
 
       # Random seed
       random_seed: 1