Optimize `merge` algorithm for largest data sizes #208

Job	Run time
clang-format	10s
HOST,bknd=omp,cmplr=clang++,macos-latest,std=c++17,cfg=release	4m 8s
HOST,bknd=tbb,cmplr=icx-cl,windows-latest,std=c++17,cfg=release	17m 47s
HOST,bknd=tbb,cmplr=cl,windows-latest,std=c++17,cfg=release	16m 16s
CPU,bknd=dpcpp,cmplr=icpx,ubuntu-latest,std=с++17,cfg=release	15m 21s
CPU,bknd=dpcpp,cmplr=icx,windows-latest,std=c++17,cfg=release	12m 14s
FPGA_EMU,bknd=dpcpp,cmplr=icpx,ubuntu-latest,std=с++17,cfg=release	24m 46s
HOST,bknd=tbb,cmplr=icpx,ubuntu-latest,std=с++17,cfg=release	11m 47s
HOST,bknd=omp,cmplr=icpx,ubuntu-latest,std=с++17,cfg=release	9m 41s
HOST,bknd=omp,cmplr=g++,ubuntu-latest,std=с++17,cfg=release	6m 11s
HOST,bknd=tbb,cmplr=g++,ubuntu-latest,std=с++17,cfg=release	9m 2s
HOST,bknd=tbb,cmplr=g++,ubuntu-latest,std=с++17,cfg=release	8m 42s
HOST,bknd=tbb,cmplr=clang++,ubuntu-latest,std=с++17,cfg=release	11m 8s
HOST,bknd=omp,cmplr=clang++,ubuntu-latest,std=с++17,cfg=release	8m 34s
HOST,bknd=tbb,cmplr=clang++,ubuntu-latest,std=с++17,cfg=release	11m 2s
HOST,bknd=tbb,cmplr=icpx,ubuntu-latest,std=с++17,cfg=release	11m 34s
HOST,bknd=serial,cmplr=g++,ubuntu-latest,std=с++17,cfg=release	4m 53s
CPU,bknd=dpcpp,cmplr=icpx,ubuntu-latest,std=с++20,cfg=release	26m 33s
HOST,bknd=tbb,cmplr=icpx,ubuntu-latest,std=с++20,cfg=release	17m 3s
HOST,bknd=omp,cmplr=icpx,ubuntu-latest,std=с++20,cfg=release	17m 47s
	4h 4m 39s

Provide feedback