Write-back cache system using concurrent address transfers to setup requested address in main memory before dirty miss signal from cache

• US 4,858,111 A
• Filed: 10/20/1986
• Issued: 08/15/1989
• Est. Priority Date: 07/29/1983
• Status: Expired due to Term

• Current Assignee: Hewlett-Packard Company (HP Inc.)
• Original Assignee: Hewlett-Packard Company (HP Inc.)
• Inventors: Steps, Steven C.
• Primary Examiner(s): Williams, Jr., Archie E.
• Assistant Examiner(s): LEE, THOMAS C
• Application Number:
  US06/921,876
• Time in Patent Office: 1,030 Days
• Field of Search: 364/200 MS File, 364/900 MS File
• US Class Current: 711/143
• CPC Class Codes: G06F 12/0804 with main memory updating G06F12/0806 takes precedence
• CPC Class Codes: G06F 12/0884 Parallel mode, e.g. in parallel with main memory or CPU

Abstract

A computer system in which only the cache memory is permitted to communicate with main memory and the same address being used in the cache is also sent at the same time to the main memory. Thus, as soon as it is discovered that the desired main memory address is not presently in the cache, the main memory RAMs can be read to the cache without being delayed by the main memory address set up time. In addition, since the main memory is not accessable other than from the cache memory, there is also no main memory access delay caused by requests from other system modules such as the I/O controller. Likewise, since the contents of the cache memory is written into a temporary register before being sent to the main memory, a main memory read can be performed before doing a writeback of the cache to the main memory, so that data can be back to the cache in approximately the same amount of time required for a normal main memory access. The result is a significant reduction in the overhead time normally associated with cache memories.

Claims

1. In a computer system comprising a central processing unit (CPU), a write-back cache memory unit (cache) having a data store and a comparator for indicating whether information requested by the CPU from the cache is a hit, a clean miss, or a dirty miss, and a main memory unit (main memory), said units coupled so that the main memory can communicate only with the cache and so that the CPU can fetch information from and store information to the main memory only via the cache, and further coupled so that state machines representing the states of the cache and the main memory, respectively, are linked together, a method comprising the steps of:

(a) simultaneously sending an address of information desired to be read from the cache by the CPU from the CPU to both the cache and main memory at such a time that said address is set up in the main memory prior to receiving a dirty miss indication from the comparator;

(b) fetching the encached copy of said information from the data store and then temporarily buffering the fetched information in the computer system prior to receiving said dirty miss indication from said comparator; and

,

(c) storing a copy of the desired information fetched from the main memory into the cache if either a clean or dirty miss indication is provided by the comparator;

wherein step (c) takes place before dirty information buffered in the system is written back to the main memory if the comparator provides a dirty miss indication.

2. A method as in claim 1 wherein step (c) comprises signaling the main memory from the CPU to send read information to the cache as soon as it is known that a clean or dirty miss of the desired information in the cache has occurred;

and

step (b) comprises comparing an address of the cache to an address sent from the CPU with the comparator while buffering the encached information to a temporary register.

3. A method as in claim 2, if the CPU has requested fetch of information from the main memory to the cache, further comprising:

freezing the CPU as soon as a clean or dirty miss of the desired information in the cache is indicated by the comparator.

4. A method as in claim 3, further comprising:

fetching information from the main memory to the cache;

waiting for an indication from the comparator to the CPU that the information has been transferred from the main memory to the cache;

unfreezing the CPU; and

reading the information by the CPU from the cache substantially within one main memory access period after either a clean miss or a dirty miss.

5. A method as in claim 4 if a dirty miss of the cache is indicated by the comparator, further comprising after the step of fetching information from the main memory to the cache, the step of:

storing the information in the temporary register back to the main memory unit.

6. In a computer system comprising a central processing unit (CPU), a write-back cache memory unit (cache) having a data store and a comparator for indicating whether information requested by the CPU from the cache is a hit, a clean miss, or a dirty miss, and a main memory unit (main memory), said units coupled so that the main memory can communicate only with the cache and so that the CPU can fetch information from and store information to the main memory only via the cache, and further coupled so that state machines representing the states of the cache and the main memory, respectively, are linked together, a method comprising the steps of:

(a) comparing with said comparator an address of the cache to an address sent from the CPU while simultaneously fetching a copy of information from the data store encached at said address;

(b) storing the fetched information into a temporary register before a dirty miss indication is provided by said comparator;

(c) storing a copy of the information fetched from the main memory into the cache if either a clean or dirty miss indication is provided by the comparator, wherein step (c) takes place before dirty information stored in the temporary register is written back to the main memory if the comparator provides a dirty miss indication; and

,

(d) storing the information residing in the temporary register into the main memory if the comparator provides a dirty miss indication.

7. A method as in claim 6, if the CPU has requested a fetch of information from the main memory to the cache, further comprising:

freezing the CPU as soon as a clean or dirty miss of the desired information in the cache is indicated by the comparator.

8. A method as in claim 7, further comprising:

fetching information from the main memory to the cache;

waiting for an indication from the comparator to the CPU that the information has been transferred from the main memory to the cache;

unfreezing the CPU; and

reading the information by the CPU from the cache substantially within one main memory access period after either a clean miss or a dirty miss.