摘要 |
<p>PROBLEM TO BE SOLVED: To enable efficient cache allocation by caching data structure after one avoidable cache error. SOLUTION: A computer system 100 is provided with plural subsystems 110, 120,...180 mutually linked through a global interconnect 190. Peculiar network node addresses are allocated to the respective subsystems. When a cache error is detected, it is determined whether that cache error can be avoided or not. Namely, when data are cached before and further remaining yet in a cache, it is determined whether the current cache error is to occur in spite of that or not. Only after the avoidable cache error occurs, data are to be cached. The data are cached only after one avoidable cache error at least in place of every access.</p> |