To harvest maximum power from a string operating under the partially shaded condition, it must work at its global power peak as the string characteristics follow changes in operating environment. The presence of multiple peaks in the power-voltage curve of a PV string increases computational complexity for the detection of the global power peak under rapidly changing climatic conditions. Traditional algorithms fail to operate satisfactorily under such dynamic changes and result in sub-optimal operation and higher power loss. In this paper, a novel algorithm is proposed for tracking of global maximum under rapidly changing solar radiation and temperature. The algorithm divides the search space into an optimal number of bands and then uses the golden search algorithm within each band to identify the local peaks and selects the global peak of the PV characteristics curve. The hardware in loop implementation of the proposed algorithm has been done using microcontroller by using Agilent E4360A for emulating the PV string characteristics and DC-DC boost converter resulting in various case studies. The comparison with the earlier proposed algorithm has been implemented on the same platform and the results obtained suggest that the proposed algorithm provides tracking of the global MPP with a fewer number of power perturbation steps. The hardware implementation with Arduino (Atmega 328) provides an economical and efficient solution. The practical results in conjunction with DC-DC boost converter show good agreement with the theoretical analysis, simulation and efficient tracking performance.