Solving the Poisson-Boltzmann equation with the specialized computer chip MD-GRAPE-2

Area of Science:

• Computational Chemistry
• Biophysics
• Scientific Computing

Background:

• Accurate solvation effect calculations are crucial for computational chemistry.
• Implicit solvation models, using the Poisson-Boltzmann (PB) equation and Boundary Element Method (BEM), are common but computationally intensive.
• Existing methods require significant time on conventional computers.

Purpose of the Study:

• To present an efficient hardware-accelerated method for solving the PB equation.
• To leverage the MDGRAPE-2 special-purpose hardware for faster solvation calculations.
• To demonstrate the performance and accuracy of the new approach.

Main Methods:

• Recasting Boundary Element Method (BEM) equations for the PB solver.
• Implementing an iterative solution procedure on the MDGRAPE-2 hardware.
• Testing the hardware-accelerated PB solver on various molecular systems.

Main Results:

• Achieved significant speedups for PB equation solving, with acceleration factors ranging from 15-fold to 40-fold.
• Demonstrated the method's reliability and accuracy across small peptides and large proteins.
• Successfully implemented a hardware-accelerated implicit solvation solver.

Conclusions:

• The MDGRAPE-2 hardware enables a highly efficient and accurate solution of the PB equation for implicit solvation.
• This hardware acceleration significantly reduces computation time for critical molecular modeling tasks.
• The developed method offers a substantial performance improvement over conventional approaches without compromising accuracy.