Department of Pc Science And Engineering > 자유게시판

• 목록
• 답변
• 글쓰기
• 게시판 리스트 옵션
  • 수정
  • 삭제

In computing, interleaved memory is a design which compensates for the relatively gradual velocity of dynamic random-access memory (DRAM) or core memory, by spreading memory addresses evenly across memory banks. That means, contiguous memory reads and writes use every memory bank in turn, resulting in larger memory throughput as a result of diminished ready for memory banks to grow to be prepared for the operations. It's completely different from multi-channel memory architectures, primarily as interleaved memory doesn't add more channels between the main memory and the memory controller. Nonetheless, channel interleaving is also possible, for instance in freescale i.MX6 processors, which allow interleaving to be accomplished between two channels. With interleaved memory, memory addresses are allocated to every memory financial institution in flip. For instance, in an interleaved system with two memory banks (assuming word-addressable memory), if logical tackle 32 belongs to bank 0, then logical address 33 would belong to financial institution 1, logical tackle 34 would belong to financial institution 0, and so on. An interleaved memory is alleged to be n-approach interleaved when there are n banks and memory location i resides in bank i mod n.

Interleaved memory results in contiguous reads (which are common both in multimedia and execution of programs) and Memory Wave contiguous writes (which are used incessantly when filling storage or communication buffers) truly utilizing every memory financial institution in turn, as a substitute of using the identical one repeatedly. This leads to considerably higher memory throughput as every financial institution has a minimal waiting time between reads and writes. Main memory (random-entry memory, RAM) is normally composed of a collection of DRAM memory chips, where a variety of chips may be grouped together to kind a memory bank. It's then attainable, with a memory controller that supports interleaving, to put out these enhance memory retention banks so that the memory banks will probably be interleaved. Data in DRAM is saved in units of pages. Each DRAM bank has a row buffer that serves as a cache for Memory Wave accessing any page in the bank. Earlier than a web page in the DRAM bank is learn, it is first loaded into the row-buffer.

If the page is instantly learn from the row-buffer (or a row-buffer hit), it has the shortest memory entry latency in one memory cycle. If it is a row buffer miss, which can be known as a row-buffer battle, it is slower as a result of the new web page must be loaded into the row-buffer earlier than it is read. Row-buffer misses occur as access requests on totally different memory pages in the identical financial institution are serviced. A row-buffer conflict incurs a substantial delay for a memory access. In contrast, memory accesses to different banks can proceed in parallel with a excessive throughput. The problem of row-buffer conflicts has been effectively studied with an effective answer. The size of a row-buffer is generally the size of a memory page managed by the operating system. Row-buffer conflicts or misses come from a sequence of accesses to distinction pages in the identical memory bank. The permutation-primarily based interleaved memory methodology solved the issue with a trivial microarchitecture value.

Sun Microsystems adopted this the permutation interleaving technique rapidly of their merchandise. This patent-free methodology may be found in many industrial microprocessors, resembling AMD, Intel and enhance memory retention NVIDIA, for embedded systems, laptops, desktops, and enterprise servers. In conventional (flat) layouts, memory banks may be allotted a contiguous block of memory addresses, which is quite simple for the memory controller and provides equal performance in completely random access situations, when in comparison with performance ranges achieved by means of interleaving. However, in reality memory reads are hardly ever random due to locality of reference, and optimizing for close together entry gives far better performance in interleaved layouts. The way in which memory is addressed has no effect on the access time for memory areas which are already cached, having an affect solely on memory areas which have to be retrieved from DRAM. Zhao Zhang, Zhichun Zhu, and Xiaodong Zhang (2000). A Permutation-primarily based Web page Interleaving Scheme to cut back Row-buffer Conflicts and Exploit Data Locality. Division of Pc Science and Engineering, School of Engineering, Ohio State College. Mark Smotherman (July 2010). "IBM Stretch (7030) - Aggressive Uniprocessor Parallelism".