We have performed a detailed multi-wavelength analysis of the March 26, 1991 solar flare to study the physical processes responsible for the efficient acceleration of charged particles to high energies and the chromospheric response to the injection of the accelerated particles. This particular flare is of special interest because to date it is the only flare in which the gamma-ray emission with energies of 20-1000 MeV was registered throughout the entire development of the event. In this part of the study we used Ha , soft X-ray (SXR), radio (R), and gamma-ray observations to evaluate the energy deposition in the flare. Light curves of the SXR emission in two wavelength bands measured by the GOES-7 satellite with a temporal resolution of 3s were used to calculate the time. Using these data we calculated functions of energy deposition for the Ha , SXR and microwave emission by numerical solution of the governing differential equations. The energy deposition in the impulsive phase, which lasted for 200s, exhibited a complex time structure. The most pronounced pulses of the energy deposits IR and ISXR could be easily identified with corresponding peaks of the energy deposited into different areas of the Ha flare. The delayed phase began 250s after the beginning of the impulsive phase and showed itself in additional energy deposits into SXR and Ha emission in some of the areas. Comparison of the gamma-ray emission light curve with the energy deposited into other types of emission shows that the very beginning of the detection of the gamma-rays (when the count rate begins to exceed background) coincides with powerful energy deposits into the microwave, SXR and some areas of Ha emission. The observed hardening of the gamma-ray emission spectrum in the delayed phase indicates a more efficient acceleration of charged particles to high energies in this phase of the flare. We could not find any correlation between the intensity of the gamma-ray peaks and the power of the energy deposited into other types of emission. Thus we conclude that the intensity of the registered gamma-ray emission depends less on the number of accelerated high-energy particles and more on the position of the acceleration region and the anisotropy of the particle fluxes with respect to the observer.