公开(公告)号：US20150127905A1

公开(公告)日：2015-05-07

申请号：US14072639

申请日：2013-11-05

Applicant: NETAPP, INC.

Inventor： Koling Chang ,  Ravikanth Dronamraju ,  Mark Smith ,  Naresh Patel

IPC: G06F12/08

CPC classification number: G06F12/0871 ,  G06F11/3447 ,  G06F11/3461 ,  G06F2201/885 ,  G06F2212/502 ,  G06F2212/601

Abstract: A system and method for determining an optimal cache size of a computing system is provided. In some embodiments, the method comprises selecting a portion of an address space of a memory structure of the computing system. A workload of data transactions is monitored to identify a transaction of the workload directed to the portion of the address space. An effect of the transaction on a cache of the computing system is determined, and, based on the determined effect of the transaction, an optimal cache size satisfying a performance target is determined. In one such embodiment the determining of the effect of the transaction on a cache of the computing system includes determining whether the effect would include a cache hit for a first cache size and determining whether the effect would include a cache hit for a second cache size different from the first cache size.

Abstract translation: 提供了一种用于确定计算系统的最佳高速缓存大小的系统和方法。 在一些实施例中，该方法包括选择计算系统的存储器结构的地址空间的一部分。 监视数据事务的工作负载以识别指向地址空间部分的工作负载的事务。 确定事务对计算系统的缓存的影响，并且基于所确定的事务的效果，确定满足性能目标的最佳高速缓存大小。 在一个这样的实施例中，事务对计算系统的高速缓存的影响的确定包括确定效果是否将包括第一高速缓存大小的高速缓存命中，并且确定该效果是否将包括不同于第二高速缓存大小的高速缓存命中 从第一个缓存大小。

公开(公告)号：US09405695B2

公开(公告)日：2016-08-02

申请号：US14072639

申请日：2013-11-05

Applicant: NETAPP, INC.

Inventor： Koling Chang ,  Ravikanth Dronamraju ,  Mark Smith ,  Naresh Patel

IPC: G06F12/08 ,  G06F11/34

CPC classification number: G06F12/0871 ,  G06F11/3447 ,  G06F11/3461 ,  G06F2201/885 ,  G06F2212/502 ,  G06F2212/601

Abstract: A system and method for determining an optimal cache size of a computing system is provided. In some embodiments, the method comprises selecting a portion of an address space of a memory structure of the computing system. A workload of data transactions is monitored to identify a transaction of the workload directed to the portion of the address space. An effect of the transaction on a cache of the computing system is determined, and, based on the determined effect of the transaction, an optimal cache size satisfying a performance target is determined. In one such embodiment the determining of the effect of the transaction on a cache of the computing system includes determining whether the effect would include a cache hit for a first cache size and determining whether the effect would include a cache hit for a second cache size different from the first cache size.

Abstract translation: 提供了一种用于确定计算系统的最佳高速缓存大小的系统和方法。 在一些实施例中，该方法包括选择计算系统的存储器结构的地址空间的一部分。 监视数据事务的工作负载以识别指向地址空间部分的工作负载的事务。 确定事务对计算系统的缓存的影响，并且基于所确定的事务的效果，确定满足性能目标的最佳高速缓存大小。 在一个这样的实施例中，事务对计算系统的高速缓存的影响的确定包括确定效果是否将包括第一高速缓存大小的高速缓存命中，并且确定该效果是否将包括不同于第二高速缓存大小的高速缓存命中 从第一个缓存大小。