摘要 |
<p>In carrying out concurrent analysis of fluid numerals in a concurrently processing computer having its own program in the arithmetic unit, an instruction queue in a fundamental calculation portion that is commonly processed among the arithmetic units is separated from an instruction queue that describe boundary setting portions which will be individually processed, and a common instruction is broadcast to the arithmetic units for each instruction. In this computer, the memory capacitance can be reduced irrespective of the number of loops, the number of repetitions and the amount of arithmetic operation.</p> |