Design of the IBM Enterprise System/9000 high-end processor
暂无分享,去创建一个
The “high-end” water-cooled processors in the IBM Enterprise System/9000™ product family use a CPU organization and cache structure which depart significantly from previous designs. The CPU organization includes multiple execution elements which execute instructions out of sequence, and uses a new virtual register management algorithm to control them. It also contains a branch history table to remember recent branches and their target addresses so that instruction fetching and decoding can be directed more accurately. These models also use a two-level cache structure which provides a level 1 cache associated with each processor and a level 2 cache associated with central storage. The level 1 cache uses a store-through organization, and is split into two separate caches, one used for instruction fetching and the other for operand references. The level 2 cache uses a store-in method to handle stores.
[1] R. M. Tomasulo,et al. An efficient algorithm for exploiting multiple arithmetic units , 1995 .
[2] John S. Liptay,et al. Structural Aspects of the System/360 Model 85 II: The Cache , 1968, IBM Syst. J..
[3] S. G. Tucker,et al. The IBM 3090 System: An Overview , 1986, IBM Syst. J..