WO2005050455A1 - キャッシュメモリ及びその制御方法 - Google Patents
キャッシュメモリ及びその制御方法 Download PDFInfo
- Publication number
- WO2005050455A1 WO2005050455A1 PCT/JP2004/016272 JP2004016272W WO2005050455A1 WO 2005050455 A1 WO2005050455 A1 WO 2005050455A1 JP 2004016272 W JP2004016272 W JP 2004016272W WO 2005050455 A1 WO2005050455 A1 WO 2005050455A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- cache
- order
- data
- flag
- cache entry
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F12/00—Accessing, addressing or allocating within memory systems or architectures
- G06F12/02—Addressing or allocation; Relocation
- G06F12/08—Addressing or allocation; Relocation in hierarchically structured memory systems, e.g. virtual memory systems
- G06F12/12—Replacement control
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F12/00—Accessing, addressing or allocating within memory systems or architectures
- G06F12/02—Addressing or allocation; Relocation
- G06F12/08—Addressing or allocation; Relocation in hierarchically structured memory systems, e.g. virtual memory systems
- G06F12/12—Replacement control
- G06F12/121—Replacement control using replacement algorithms
- G06F12/126—Replacement control using replacement algorithms with special data handling, e.g. priority of data or instructions, handling errors or pinning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F12/00—Accessing, addressing or allocating within memory systems or architectures
- G06F12/02—Addressing or allocation; Relocation
- G06F12/08—Addressing or allocation; Relocation in hierarchically structured memory systems, e.g. virtual memory systems
- G06F12/12—Replacement control
- G06F12/121—Replacement control using replacement algorithms
Definitions
- the present invention relates to a cache memory for speeding up memory access of a processor and a control method therefor.
- a small-capacity, high-speed cache memory such as a static random access memory (SRAM) is arranged inside or near the microprocessor, and a part of data is stored in the cache memory.
- SRAM static random access memory
- a freeze function is provided in the cache memory.
- the freeze function is a function that copies a program that requires little access but needs to be processed at high speed once activated in the cache memory in advance, and prohibits the area from being rewritten. .
- the computer system can read out the program from the cache memory and execute it when necessary, thereby shortening the execution time.
- the purge function releases programs and data that are accessed frequently but whose execution speed is not so demanding, without storing them in the cache memory.
- the cache memory can be spared, and other programs and data with high priority can be loaded into the cache memory, thereby improving the use efficiency of the cache memory and improving the overall execution time. Is shortened.
- Patent Document 1 Japanese Patent Application Laid-Open No. 2003-200221
- freeze function and the purge function have been used, it is possible to keep frequently accessed data in the cache memory or to purge data from the cache memory when the access frequency is low. There is a problem that a complicated circuit is required for controlling the function.
- the cache memory of the present invention holds order data indicating an access order for each cache entry holding unit data of a cache, and replaces the cache entry indicating the oldest order.
- the cache memory includes a modifying unit for modifying the order data against an actual access order, and a selecting unit for selecting a cache entry to be replaced based on the modified order data.
- the altering means comprises: a specifying means for specifying a cache entry holding data belonging to a specified address range of the processor; and an ordering data of the specified cache entry, which is contrary to an actual access order. It may be configured to include the oldest means for old age.
- the cache entry is selected as the replacement target first. As a result, it is possible to reduce the frequency of access, and reduce the occurrence of cache misses due to data remaining in the cache memory.
- the specifying means converts the start address into a start line address indicating a start line included in the address range.
- the start line address may include a determination unit that determines whether there is a cache entry holding data corresponding to each line address up to the end line address.
- the processor can specify any address (or any size) from any address regardless of the line size of the cache memory and the address of the line boundary as the address range. it can. That is, it is not necessary for the processor to manage the line size and the address of the line boundary of the cache memory, so that the load for managing the cache memory can be eliminated.
- the oldest means may add an oldest flag indicating that the access order is the oldest to the order data.
- the access order is modified by indirectly adding the W flag without directly modifying the order data indicating the access order in the LRU scheme. This can be realized without adding.
- the selecting means selects the cache entry as a replacement target, If there is no cache entry to which the oldest flag is added, a cache entry to be replaced may be selected according to the order data.
- the cache entry has, as the order data, a 1-bit order flag indicating whether the access order is old or new, and the selecting means includes a cache to which an oldest flag is added. If the entry does not exist, a configuration may be adopted in which a cache entry indicating that the order flag is old is selected as a replacement target.
- the order data can be a 1-bit order flag! Since the data amount of the access order data is small and updating is easy, the hardware scale can be reduced. it can.
- the altering means alters the order data so that the access order of one cache entry is Nth, and N is a number indicating that (a) the access order is the oldest order , (B) a number indicating that the access order is the newest, (c) a number indicating that the access order is the oldest, and (d) a number indicating that the access order is also the Nth. It may be configured to be one of the numbers.
- the altering means includes an instruction detecting means for detecting that a memory access instruction with an access order alteration instruction has been executed, and writes order data to a cache entry accessed by the instruction. It may be configured to include a rewriting means for changing
- the altering means includes: holding means for holding an address range designated by a processor; searching means for searching for a cache entry holding data corresponding to the address range held by the holding means; Rewriting means for rewriting the order data so that the access order of the cache entry searched by the means is Nth.
- the cache memory control method of the present invention has the same means and operation as described above.
- data with a low access frequency can be replaced with data with higher priority than other data with a high access frequency. It is possible to prevent data that is frequently accessed from being replaced.
- the cache entry is selected as a replacement target first. As a result, it is possible to reduce the frequency of access and reduce the possibility of cache misses caused by data remaining in the cache memory.
- the load for managing the cache memory in the processor can be eliminated.
- FIG. 1 is a block diagram showing a schematic configuration of a system including a processor, a cache memory, and a memory according to a first embodiment of the present invention.
- FIG. 2 is a block diagram showing a configuration example of a cache memory.
- FIG. 3 shows a detailed bit configuration of a cache entry.
- FIG. 4 is a block diagram showing a configuration of a control unit.
- FIG. 5 is a block diagram showing a configuration example of a W flag setting unit.
- FIG. 6A shows an example of an instruction for writing a start address to a start address register.
- FIG. 6B shows an example of an instruction for writing a size to the size register.
- FIG. 6C shows an example of an instruction for writing a command to the command register.
- FIG. 6D shows an example of the command.
- FIG. 7 is an explanatory diagram of a start aligner and an end aligner.
- FIG. 8 is a flowchart showing a W flag setting process in a flag rewriting unit.
- FIG. 9 is a flowchart showing a replacement process in a replacement unit.
- FIG. 10 is a block diagram showing a configuration of a cache memory according to Embodiment 2 of the present invention.
- FIG. 11 shows a bit configuration of a cache entry.
- FIG. 12 is a block diagram showing a configuration of a control unit.
- FIG. 13 shows an example of updating a use flag by a replacement unit.
- FIG. 14B is an explanatory diagram showing a role of a weak flag W in a replacement process.
- FIG. 15 is a flowchart showing a U flag updating process in a flag updating unit.
- FIG. 16 is a flowchart showing a replacement process in a replacement unit. [17]
- FIG. 17 is a diagram illustrating another configuration example of the W flag setting unit.
- FIG. 18 is a diagram showing yet another configuration example of the W flag setting unit.
- FIG. 1 is a block diagram showing a schematic configuration of a system including a processor 1, a cache memory 3, and a memory 2 according to Embodiment 1 of the present invention. As shown in the figure, a cache memory 3 of the present invention is provided in a system having a processor 1 and a memory 2.
- the cache memory 3 is premised on replacement control for replacing a cache entry with an old access order by a so-called LRU method.
- the cache memory 3 according to the present embodiment replaces a cache entry holding data with a low access frequency by modifying order data indicating an access order for determining a replacement target against the access order. It is configured to be evicted from the cache memory.
- the order data is indirectly modified by adding a weak flag W to the cache entry indicating that the access order is considered to be the oldest. This eliminates the need for complicated circuits that directly modify the order data.
- FIG. 2 is a block diagram showing a configuration example of the cache memory 3.
- the cache memory 3 has an address register 20, a memory IZF21, a decoder 30, four ways 31a to 31d (hereinafter abbreviated as ways 0 to 3), four comparators 32a to 32d, and four AND circuits. 33a-33d, OR circuit 34, selectors 35 and 36, demultiplexer 37, and control unit 38.
- the address register 20 is a register that holds an access address to the memory 2. This access address shall be 32 bits. As shown in the figure, the access address includes a 21-bit tag address, a 4-bit set index (SI in the figure), and a 5-bit word index (WI in the figure) in order from the most significant bit.
- the tag address indicates the area in the memory that is mapped to the way (the size is the number of sets X blocks). The size of this area is determined by the address bits (A10—AO) that are lower than the tag address, that is, 2 kbytes, and is also the size of one way.
- the set index (SI) refers to one of multiple sets spanning ways 0-3. Since the number of sets is set index power bit, there are 16 sets.
- the cache entry specified by the tag address and the set index is a replacement unit, and when stored in the cache memory, is called line data or line.
- the size of the line data is the size determined by the address bits lower than the set index, that is, 128 bytes. If one word is 4 bytes, one line data is 32 words.
- the word index (WI) indicates one word among a plurality of words constituting the line data.
- the two least significant bits (Al, AO) in address register 20 are ignored during word access.
- the memory IZF 21 is an IZF for accessing the memory 2 from the cache memory 3 such as writing back data from the cache memory 3 to the memory 2 and loading data into the cache memory 3.
- the decoder 30 decodes 4 bits of the set index, and crosses over four ways 0-3.
- the four ways 0-3 are four ways having the same configuration and have a capacity of 4 x 2k bytes. Each way has 16 cache entries.
- FIG. 3 shows a detailed bit configuration in one cache entry. As shown in the figure, one cache entry 1 has a NORD flag V0-V3, a 21-bit tag, 128 Knots of line data, a weak flag W, and a dirty flag DO-D3.
- a tag is a copy of a 21-bit tag address.
- Line data is a copy of the 128-byte data in the block specified by the tag address and the set index, and consists of four 32-byte sub-lines.
- the NORD flag VO-V3 corresponds to four sub-lines and indicates whether the sub-line is valid.
- the dirty flag DO-D3 corresponds to the four sub-lines, and the processor power is written to those sub-lines. That is, the cached data exists in the sub-line but differs from the data in the memory due to the write. This indicates whether writing back to memory is necessary.
- the comparator 32a compares the tag address in the address register 20 with the tag of the way 0 among the four tags included in the set selected by the set index. The same applies to the comparators 32b-32c except that they correspond to the bays 31b-31d.
- the AND circuit 33a compares the NORD flag with the comparison result of the comparator 32a to determine whether or not the forces match. Let hO be the result of this comparison. If the comparison result hO is 1, it means that line data corresponding to the tag address and the set index in the address register 20 exists, that is, a hit has occurred in way 0. If the comparison result hO is 0, it means that a miss has occurred. The same applies to AND circuits 33b-33d except that they correspond to ways 31b-31d. The comparison result hi-h3 means the force that hit or missed way 1-13.
- the OR circuit 34 takes the OR of the comparison result hO—h3. Let the result of this OR be hit. hit is
- the selector 35 selects the line data of the hit way from the line data of the ways 0-3 in the selected set.
- the selector 36 selects one word indicated by the word index out of the 32-word line data selected by the selector 35.
- the demultiplexer 37 When writing data to the cache entry, the demultiplexer 37 outputs the write data to one of the ways 0-3. This write data may be in word units.
- the control unit 38 controls the entire cache memory 3. In particular, it sets the W flag and performs replacement control according to the W flag.
- FIG. 4 is a block diagram showing a configuration of the control unit 38. As shown in the figure, the control unit 38 includes a rib race unit 39 and a W flag setting unit 40.
- the W flag setting unit 40 sets the weak flag W according to a command from the processor 1.
- the processor 1 issues a command for setting the weak flag W to the cache memory 3 for a cache entry that is no longer written or read.
- FIG. 5 is a block diagram showing a configuration example of the W flag setting unit 40.
- the W flag setting unit 40 includes a command register 401, a start address register 402, and a size register 40.
- An adder 404, a start aligner 405, an end aligner 406, and a flag rewriting unit 407 are provided.
- the command register 401 is a register directly accessible from the processor 1, and holds the W flag setting command written by the processor 1.
- FIG. 6C shows an example of an instruction for writing a command to the command register 401. This instruction is a normal transfer instruction (mov instruction), in which a command is used as a source operand and a command is used as a destination operand. Command register (CR) is specified.
- FIG. 6D shows an example of the command. This command is a specific code indicating the W flag setting command. The W flag setting command sets the W flag for the cache entry that holds data corresponding to the address range of the size held in the size register 403 from the start address held in the start address register 402. It is a command to instruct.
- the start address register 402 is a register directly accessible from the processor 1, and holds the start address written by the processor 1. This start address indicates the start position of the address range in which the W flag should be set.
- FIG. 6A shows an example of an instruction for writing a start address to the start address register 402. This instruction is also a normal transfer instruction (mov instruction) as in FIG. 6 (c).
- the size register 403 is a register directly accessible from the processor 1, and holds the size written by the processor 1. This size indicates the address range from the start address.
- FIG. 6B shows an example of an instruction for writing the size in the size register 403. This instruction is also a normal transfer instruction (mov instruction) as in FIG. 6 (c).
- the unit of the size may be the number of bytes or the number of lines (the number of cache entries) as long as it is a predetermined unit! ,.
- the calorie calculator 404 adds the start address held in the start address register 402 and the size held in the size register 403.
- the addition result is an end address indicating the end position of the address range.
- the adder 404 may add the value as a byte address when the size is designated by the number of bytes, and add it as a line address when the size is designated by the number of lines.
- the start aligner 405 adjusts the start address to the position of the line boundary. With this adjustment, the processor 1 can specify an arbitrary address as the start address regardless of the line size and the line boundary.
- the end aligner 406 adjusts the end address to the position of the line boundary. By this adjustment, the processor 1 can specify any size as the above-mentioned size regardless of the line size and the line boundary.
- FIG. 7 is an explanatory diagram of the start aligner 405 and the end aligner 406. Smell
- the start address specified by processor 1 indicates an arbitrary position in the middle of line N.
- the start aligner 405 adjusts to point to the beginning of the next line (N + 1), and outputs the adjusted address as an align start address.
- the line indicated by the line start address is called the start line.
- the end address indicates an arbitrary position in the middle of the line M.
- the end aligner 406 adjusts to point to the head of the immediately preceding line (M-1), and outputs the adjusted address as an align end address.
- the line indicated by the line end address is called an end line.
- the W flag is set for each line (cache entry) from the start line (line (N + 1)) to the end line (line (M-1)).
- the start aligner 405 and the end aligner 406 are aligned inside the address range from the start address to the end address specified by the processor 1 only in the portions outside the line N and the line M. This is because writing or reading from the processor 1 may occur.
- the flag rewriting unit 407 controls the cache memory from the line card indicated by the line start address to the line indicated by the line end address (in the example of Fig. 7, the line (N + 1) force and the line (M-1)). If it is entered in 3! /, The W flag is set to 1.
- FIG. 8 is a flowchart illustrating an example of the W flag setting process in the flag rewriting unit 407.
- the flag rewriting unit 407 When the W flag setting command is held in the command register 401, the flag rewriting unit 407 outputs the line addresses from the start line card to the end line in order, and performs the processing of loop 1 (S82— S86). Since the flag rewriting unit 407 performs the same processing for each line, the processing for one line will be described here.
- the flag rewriting unit 407 outputs the line address to the address register 20 (S83), and outputs the tag address of the address register 20 and the tag of the cache entry. Are compared with the comparators 32a-32d, and it is determined whether or not there is a hit (S84). Further, the flag rewriting unit 407 sets the W flag to 1 for the hit cache entry in the case of a hit (S85), and None has been entered in the cache memory.
- the W flag is set to 1 when each line up to the start line force and the end line has been entered in the cache memory 3.
- FIG. 9 is a flowchart showing the replacement process in the replacement unit 39.
- the replacement unit 39 replaces one of the cache entries of the selected way in the set (step S96), and initializes the weak flag W of the cache entry to 0 after the replacement (step S97). At this time, the NORD flag V and the Dirty flag D are each initialized to 0.
- the order data indicating the access order in the conventional LRU method is directly modified. Since the access order is modified by indirectly adding the w flag, it can be realized without adding a complicated hardware circuit.
- cache memory of the present invention is not limited to the configuration of the above-described embodiment, and various modifications can be made. Hereinafter, some of the modified examples will be described.
- the order data is indirectly modified by the addition of the W flag.
- the order data may be directly modified.
- the weak flag W indicates that the access order is the oldest, but the weak flag W may indicate that the access order is not the latest or the oldest.
- the control unit 38 further includes an instruction detection unit for detecting a special store instruction.
- a flag setting unit for setting W l when writing by the store instruction.
- a 4-way 'set' associative cache memory has been described as an example, but the number of ways may be any number. Further, in the above embodiment, an example in which the number of sets is 16 has been described, but the number of sets may be shoes.
- each cache entry 1 may hold the same number of knowledge flags and dirty flags as the number of sublines.
- the first embodiment has described the configuration in which the access order is set to the oldest by the weak flag W on the premise of the normal LRU method.
- a pseudo LRU method in which order data indicating an access order is represented by a 1-bit flag, and a cache memory that performs the oldest dating using a weak flag are described below. Will be explained.
- FIG. 10 is a block diagram showing a configuration of a cache memory according to Embodiment 2 of the present invention.
- the cache memory shown in the figure is different from the configuration shown in FIG. 2 in that ways 131a-131d are provided instead of ways 31a-31d, and that a control unit 138 is provided instead of the control unit 38.
- ways 131a-131d are provided instead of ways 31a-31d
- a control unit 138 is provided instead of the control unit 38.
- description of the same points will be omitted, and different points will be mainly described.
- the way 131a is different from the way 31a in that the use flag U is added in each cache entry.
- Ways 131b-131d are similar.
- the use flag U is provided in place of the order data indicating the access order among the four ways, and is a flag that expresses the access order with 1 bit.
- FIG. 11 shows a bit configuration of a cache entry.
- the bit configuration in the figure is different from that in FIG. 3 in that a use flag U is added.
- the use flag U indicates whether or not the cache entry has been accessed, and is used instead of the access order data in the four-way cache entry at the time of replacement due to a mishit in the LRU method. More precisely, a use flag U of 1 means that an access has been made, and 0 does not. However, it is reset to 0 when the use flags of the four ways in one set all become 1. In other words, the usage flag U indicates the relative state of the access, old or new. In other words, a cache entry with a use flag U of 1 means that a cache entry with a use flag of 0 is more recently accessed.
- the control unit 138 is different from the control unit 38 in that instead of the access order information in the LRU scheme, In that replacement control is performed using the use flag u.
- FIG. 12 is a block diagram showing a configuration of the control unit 138.
- the control unit 138 shown in the drawing is different from the control unit 38 in that a replace unit 139 is provided instead of the replace unit 39, and that a flag update unit 41 is added.
- the replacement unit 139 performs a replacement process at the time of a cache miss by a pseudo LRU method using the use flag U as an access order. At that time, the replacement unit 139 treats the cache entry with the weak flag W as the oldest cache entry and selects the cache entry as the replacement target first.
- the flag updating unit 41 updates the use flag U when the cache memory is accessed.
- FIG. 13 shows an example of updating the use flag by the replacement unit 139.
- the upper, interrupted, and lower rows of the figure show the cache entries of the four ways that make up the set N that spans ways 0-3. 1 or 0 at the right end of each of the four cache entries is the value of the use flag.
- These four use flags U are described as UO-U3.
- the replacement unit 139 determines one of the ways 1 and 3 as the replacement target in the upper part of FIG. 5, determines the way 3 as the replacement target in the middle part of FIG. 5, and determines the way 0-2 in the lower part of FIG. One of them is determined as a replacement target.
- FIG. 14 (a) is a comparative example when it is assumed that a weak flag does not exist, and is a diagram showing a state where a cache entry is replaced.
- This figure also shows four cache entries that constitute a set N that spans layers 0 to 3 as in FIG. The 1 or 0 at the right end of each of the four cache entries is the value of the use flag.
- only data E is data with low access frequency
- data A, B, C, and D are data with high access frequency.
- data E that is not frequently used is not selected as a replacement target and remains in the cache memory.
- the access frequency (in the normal LRU method!), Low, the data E may cause a worst case of four cache misses in the case of 4-way.
- FIG. 14B is an explanatory diagram showing the role of the weak flag W in the replacement process.
- FIG. 15 is a flowchart showing the U flag updating process in the flag updating unit 41.
- the use flag U of the cache entry whose the NORD flag is 0 (invalid) has been initialized to 0.
- the flag update unit 41 sets the use flag U of the hit way in the set selected by set indices to 1 (step S62).
- the use flag U of the other way is read (step S63), and it is determined whether or not the read use flag U is all 1 (step S64). If it is, reset all use flags U of other ways to 0 (Step S65).
- the flag updating unit 41 updates the use flag U as in the update examples shown in Figs. 13 and 14 (a) and (b). [0100] ⁇ Replacement processing>
- FIG. 16 is a flowchart showing the replacement process in the replacement unit 139.
- the replacement unit 139 replaces the cache entry of the selected way in the set (step S96), and after the replacement, sets the use flag U of the cache entry to 1 and the weak flag W to 0. (Step S97). At this time, the normal flag V and the dirty flag D are initialized to 1 and 0, respectively.
- the pseudo LRU scheme using a 1-bit use flag instead of the data indicating the access order in the conventional LRU scheme is adopted. Since a 1-bit flag can be used as the access order data, the hardware size can be reduced because the data amount of the access order data is small and the update is easy.
- the processor 1 executes a load Z store instruction (hereinafter abbreviated as W-LZS instruction) for accessing data while setting the weak flag W to 1, and executes the control unit 38 or the control unit 138. May detect the execution of a W-LZS instruction and set the W flag to 1 immediately after the access by the W-LZS instruction.
- FIG. 17 shows a configuration example of the W flag setting unit 40a provided in the control unit 38 or 138 in that case.
- the W flag setting unit 40a includes an LDZST instruction detecting unit 410, a weak instruction detecting unit 411, an AND circuit 412, and a flag rewriting unit 413.
- LDZST instruction detecting section 410 detects that processor 1 has executed a load Z store instruction.
- the weak instruction detecting unit 411 detects whether a weak instruction is output from the processor 1 at the time of executing the load Z store instruction.
- the weak instruction can be detected, for example, by a signal line from the processor 1.
- the AND circuit 412 notifies the flag rewriting section 413 of the detection of the W-LZS instruction.
- the flag rewriting unit 413 sets the weak flag W to 1 for the cache entry holding the data accessed by the W-LZS instruction.
- the order data indicating the access order is directly transmitted instead of the weak flag W.
- the configuration may be modified.
- the weak instruction detecting unit 411 detects a number (Nth) indicating the access order to be set from the processor 1 executing the load Z store instruction with the access order to be set.
- the first may be any of the fourth (or 0 to 3).
- the flag rewriting unit 413 changes the order data of the cache entry holding the data accessed by the load Z store instruction with the access order designation to Nth.
- any order data can be directly
- the configuration may be changed to N.
- the W flag setting unit 40 shown in FIG. 5 may be replaced with a W flag setting unit 40b shown in FIG.
- the W flag setting unit 40b has a configuration in which a comparator 408 is added to the W flag setting unit 40, and a flag rewriting unit 407a is provided instead of the flag rewriting unit 407.
- the comparator 408 determines whether the line address of the end line output from the adder 404 matches the line address output from the flag rewriting unit 407a for setting a wick flag and held in the tag address register 20. ⁇ judge whether or not.
- This comparator 408 is used to determine the end address in loop 1 during the W flag setting process shown in FIG. That is, when the flag rewriting unit 407a determines that the comparator 408 matches, the flag rewriting unit 407a ends the setting of the quick flag.
- FIGS. 6 (a), (b) and (c) may be inserted into the program by the compiler.
- the compiler inserts the above-mentioned instructions at the program position without writing any more, such as writing array data or writing block data when decoding compressed moving image data. Just fine.
- the present invention is suitable for a cache memory for performing high-speed memory access, for example, an on-chip cache memory, an off-chip cache memory, a data cache memory, an instruction cache memory, and the like.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Memory System Of A Hierarchy Structure (AREA)
Abstract
Description
Claims
Priority Applications (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2005515574A JP4008946B2 (ja) | 2003-11-18 | 2004-11-02 | キャッシュメモリ及びその制御方法 |
US10/577,133 US7984243B2 (en) | 2003-11-18 | 2004-11-02 | Cache memory and method for cache entry replacement based on modified access order |
EP04799475A EP1686485A4 (en) | 2003-11-18 | 2004-11-02 | CACHE MEMORY AND ASSOCIATED CHECK |
US13/075,523 US20110179227A1 (en) | 2003-11-18 | 2011-03-30 | Cache memory and method for cache entry replacement based on modified access order |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2003387351 | 2003-11-18 | ||
JP2003-387351 | 2003-11-18 |
Related Child Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US13/075,523 Continuation US20110179227A1 (en) | 2003-11-18 | 2011-03-30 | Cache memory and method for cache entry replacement based on modified access order |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2005050455A1 true WO2005050455A1 (ja) | 2005-06-02 |
Family
ID=34616158
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2004/016272 WO2005050455A1 (ja) | 2003-11-18 | 2004-11-02 | キャッシュメモリ及びその制御方法 |
Country Status (7)
Country | Link |
---|---|
US (2) | US7984243B2 (ja) |
EP (1) | EP1686485A4 (ja) |
JP (1) | JP4008946B2 (ja) |
KR (1) | KR100801814B1 (ja) |
CN (1) | CN100424658C (ja) |
TW (1) | TW200530818A (ja) |
WO (1) | WO2005050455A1 (ja) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2008211681A (ja) * | 2007-02-27 | 2008-09-11 | Kyocera Corp | 情報機器および同機器における復号化処理方法 |
US7953935B2 (en) | 2005-04-08 | 2011-05-31 | Panasonic Corporation | Cache memory system, and control method therefor |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9003125B2 (en) * | 2012-06-14 | 2015-04-07 | International Business Machines Corporation | Cache coherency protocol for allowing parallel data fetches and eviction to the same addressable index |
US10248422B2 (en) * | 2016-07-02 | 2019-04-02 | Intel Corporation | Systems, apparatuses, and methods for snooping persistent memory store addresses |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPS5119453A (en) * | 1974-08-08 | 1976-02-16 | Fujitsu Ltd | Patsufua memoriseigyohoshiki |
JPS61290555A (ja) * | 1985-06-19 | 1986-12-20 | Toshiba Corp | キャッシュシステム |
JPH06348595A (ja) * | 1993-06-07 | 1994-12-22 | Hitachi Ltd | キャッシュ装置 |
US20020152361A1 (en) | 2001-02-05 | 2002-10-17 | International Business Machines Corporation | Directed least recently used cache replacement method |
Family Cites Families (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP0203601B1 (en) | 1985-05-29 | 1992-08-05 | Kabushiki Kaisha Toshiba | Cache system adopting an lru system, and magnetic disk controller incorporating it |
US5043885A (en) * | 1989-08-08 | 1991-08-27 | International Business Machines Corporation | Data cache using dynamic frequency based replacement and boundary criteria |
US5497477A (en) * | 1991-07-08 | 1996-03-05 | Trull; Jeffrey E. | System and method for replacing a data entry in a cache memory |
JP3146380B2 (ja) * | 1991-07-26 | 2001-03-12 | コニカ株式会社 | ハロゲン化銀カラー写真感光材料の処理方法 |
US5375216A (en) | 1992-02-28 | 1994-12-20 | Motorola, Inc. | Apparatus and method for optimizing performance of a cache memory in a data processing system |
JPH0869417A (ja) | 1994-08-29 | 1996-03-12 | Sanyo Electric Co Ltd | 計算機システム |
US6266742B1 (en) * | 1997-10-27 | 2001-07-24 | International Business Machines Corporation | Algorithm for cache replacement |
US6105115A (en) * | 1997-12-31 | 2000-08-15 | Intel Corporation | Method and apparatus for managing a memory array |
US6202129B1 (en) * | 1998-03-31 | 2001-03-13 | Intel Corporation | Shared cache structure for temporal and non-temporal information using indicative bits |
US6393525B1 (en) * | 1999-05-18 | 2002-05-21 | Intel Corporation | Least recently used replacement method with protection |
US6397298B1 (en) * | 1999-07-30 | 2002-05-28 | International Business Machines Corporation | Cache memory having a programmable cache replacement scheme |
US6516384B1 (en) * | 1999-12-30 | 2003-02-04 | Intel Corporation | Method and apparatus to perform a round robin and locking cache replacement scheme |
US6738865B1 (en) * | 2000-06-09 | 2004-05-18 | International Business Machines Corporation | Method, system, and program for demoting data from cache based on least recently accessed and least frequently accessed data |
EP1182566B1 (en) | 2000-08-21 | 2013-05-15 | Texas Instruments France | Cache operation based on range of addresses |
US6487638B2 (en) * | 2001-01-26 | 2002-11-26 | Dell Products, L.P. | System and method for time weighted access frequency based caching for memory controllers |
JP3946515B2 (ja) | 2001-12-27 | 2007-07-18 | 株式会社アマダエンジニアリングセンター | 折り曲げ加工装置 |
JP2003223360A (ja) | 2002-01-29 | 2003-08-08 | Hitachi Ltd | キャッシュメモリシステムおよびマイクロプロセッサ |
WO2005048112A1 (ja) * | 2003-11-12 | 2005-05-26 | Matsushita Electric Industrial Co., Ltd. | キャッシュメモリおよびその制御方法 |
-
2004
- 2004-11-02 EP EP04799475A patent/EP1686485A4/en not_active Withdrawn
- 2004-11-02 CN CNB2004800271169A patent/CN100424658C/zh not_active Expired - Fee Related
- 2004-11-02 US US10/577,133 patent/US7984243B2/en not_active Expired - Fee Related
- 2004-11-02 WO PCT/JP2004/016272 patent/WO2005050455A1/ja not_active Application Discontinuation
- 2004-11-02 KR KR1020067005718A patent/KR100801814B1/ko not_active IP Right Cessation
- 2004-11-02 JP JP2005515574A patent/JP4008946B2/ja not_active Expired - Fee Related
- 2004-11-17 TW TW093135224A patent/TW200530818A/zh unknown
-
2011
- 2011-03-30 US US13/075,523 patent/US20110179227A1/en not_active Abandoned
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPS5119453A (en) * | 1974-08-08 | 1976-02-16 | Fujitsu Ltd | Patsufua memoriseigyohoshiki |
JPS61290555A (ja) * | 1985-06-19 | 1986-12-20 | Toshiba Corp | キャッシュシステム |
JPH06348595A (ja) * | 1993-06-07 | 1994-12-22 | Hitachi Ltd | キャッシュ装置 |
US20020152361A1 (en) | 2001-02-05 | 2002-10-17 | International Business Machines Corporation | Directed least recently used cache replacement method |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7953935B2 (en) | 2005-04-08 | 2011-05-31 | Panasonic Corporation | Cache memory system, and control method therefor |
JP2008211681A (ja) * | 2007-02-27 | 2008-09-11 | Kyocera Corp | 情報機器および同機器における復号化処理方法 |
Also Published As
Publication number | Publication date |
---|---|
EP1686485A1 (en) | 2006-08-02 |
TW200530818A (en) | 2005-09-16 |
US20110179227A1 (en) | 2011-07-21 |
US7984243B2 (en) | 2011-07-19 |
US20080168232A1 (en) | 2008-07-10 |
CN100424658C (zh) | 2008-10-08 |
KR100801814B1 (ko) | 2008-02-11 |
JP4008946B2 (ja) | 2007-11-14 |
JPWO2005050455A1 (ja) | 2007-12-06 |
KR20060087579A (ko) | 2006-08-02 |
EP1686485A4 (en) | 2008-10-29 |
CN1853172A (zh) | 2006-10-25 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11416253B2 (en) | Multiple-table branch target buffer | |
US20060036811A1 (en) | Method for software controllable dynamically lockable cache line replacement system | |
JP5536655B2 (ja) | キャッシュメモリ、メモリシステム及びデータコピー方法 | |
KR100837479B1 (ko) | 캐시 메모리 및 그 제어 방법 | |
KR100826757B1 (ko) | 캐시 메모리 및 그 제어 방법 | |
US20080301372A1 (en) | Memory access control apparatus and memory access control method | |
US20020161976A1 (en) | Data processor | |
US5926841A (en) | Segment descriptor cache for a processor | |
JP3973129B2 (ja) | キャッシュメモリ装置及びそれを用いた中央演算処理装置 | |
US20110179227A1 (en) | Cache memory and method for cache entry replacement based on modified access order | |
WO1997034229A9 (en) | Segment descriptor cache for a processor | |
US8214601B2 (en) | Purging without write-back of cache lines containing spent data | |
US7376797B2 (en) | Cache memory system and method using reference bits | |
JPH0659977A (ja) | 明示的なライン置換え操作が可能なキャッシュメモリとその制御方法 | |
JPH0784886A (ja) | キャッシュメモリ制御方法およびキャッシュメモリ制御装置 | |
JP2004326187A (ja) | 情報処理装置およびその制御方法 | |
JPH06282487A (ja) | キャッシュ装置 | |
JPH06110787A (ja) | 半導体記憶装置およびその記憶制御方法 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
WWE | Wipo information: entry into national phase |
Ref document number: 200480027116.9 Country of ref document: CN |
|
AK | Designated states |
Kind code of ref document: A1 Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BW BY BZ CA CH CN CO CR CU CZ DE DK DM DZ EC EE EG ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NA NI NO NZ OM PG PH PL PT RO RU SC SD SE SG SK SL SY TJ TM TN TR TT TZ UA UG US UZ VC VN YU ZA ZM ZW |
|
AL | Designated countries for regional patents |
Kind code of ref document: A1 Designated state(s): BW GH GM KE LS MW MZ NA SD SL SZ TZ UG ZM ZW AM AZ BY KG KZ MD RU TJ TM AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IS IT LU MC NL PL PT RO SE SI SK TR BF BJ CF CG CI CM GA GN GQ GW ML MR NE SN TD TG |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
WWE | Wipo information: entry into national phase |
Ref document number: 2005515574 Country of ref document: JP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 1020067005718 Country of ref document: KR |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2004799475 Country of ref document: EP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 10577133 Country of ref document: US |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
WWW | Wipo information: withdrawn in national office |
Ref document number: DE |
|
WWP | Wipo information: published in national office |
Ref document number: 2004799475 Country of ref document: EP Ref document number: 1020067005718 Country of ref document: KR |