Optimize `merge` algorithm for largest data sizes #205

Job	Run time
clang-format	11s
HOST,bknd=omp,cmplr=clang++,macos-latest,std=c++17,cfg=release	3m 32s
HOST,bknd=tbb,cmplr=icx-cl,windows-latest,std=c++17,cfg=release	3m 24s
CPU,bknd=dpcpp,cmplr=icpx,ubuntu-latest,std=с++17,cfg=release	2m 45s
HOST,bknd=tbb,cmplr=cl,windows-latest,std=c++17,cfg=release	3m 26s
FPGA_EMU,bknd=dpcpp,cmplr=icpx,ubuntu-latest,std=с++17,cfg=release	3m 12s
CPU,bknd=dpcpp,cmplr=icx,windows-latest,std=c++17,cfg=release	3m 23s
HOST,bknd=tbb,cmplr=icpx,ubuntu-latest,std=с++17,cfg=release	3m 8s
HOST,bknd=omp,cmplr=icpx,ubuntu-latest,std=с++17,cfg=release	3m 7s
HOST,bknd=omp,cmplr=g++,ubuntu-latest,std=с++17,cfg=release	3m 21s
HOST,bknd=tbb,cmplr=g++,ubuntu-latest,std=с++17,cfg=release	3m 30s
HOST,bknd=tbb,cmplr=g++,ubuntu-latest,std=с++17,cfg=release	3m 30s
HOST,bknd=tbb,cmplr=clang++,ubuntu-latest,std=с++17,cfg=release	3m 29s
HOST,bknd=omp,cmplr=clang++,ubuntu-latest,std=с++17,cfg=release	3m 29s
HOST,bknd=tbb,cmplr=clang++,ubuntu-latest,std=с++17,cfg=release	3m 31s
HOST,bknd=tbb,cmplr=icpx,ubuntu-latest,std=с++17,cfg=release	3m 7s
HOST,bknd=serial,cmplr=g++,ubuntu-latest,std=с++17,cfg=release	3m 35s
CPU,bknd=dpcpp,cmplr=icpx,ubuntu-latest,std=с++20,cfg=release	1s
HOST,bknd=tbb,cmplr=icpx,ubuntu-latest,std=с++20,cfg=release	2m 34s
HOST,bknd=omp,cmplr=icpx,ubuntu-latest,std=с++20,cfg=release	1s
	56m 16s

Provide feedback