Apparatus and method for scratch pad memory management

Patent metadata

Field	Value
Publication number	US-9152547-B2
Application number	US-59194609-A
Country	US
Kind code	B2
Filing date	Dec 4, 2009
Priority date	Jul 16, 2009
Publication date	Oct 6, 2015
Grant date	Oct 6, 2015

Abstract

Disclosed is a scratch pad memory management device and a method thereof. The scratch pad memory management device divides a scratch pad memory into a plurality of unit blocks, maintains a memory allocation table corresponding to indices of the plurality of unit blocks in a main memory, and manages the scratch pad memory.

First claim

What is claimed is:

1. A scratch pad memory management device, the device comprising:
a scratch pad memory being divided into a plurality of unit blocks;
a main memory;
a memory allocation table; and
a memory managing unit being maintained in the main memory to manage the scratch pad memory,
wherein the memory managing unit generates a memory chunk comprising at least one of the plurality of unit blocks of the scratch pad memory, and classifies the memory chunk as one of a free memory, an allocated memory, and a fixed block pool to manage the memory chunk, and
wherein the memory managing unit stores, in the memory allocation table, metadata to manage the plurality of unit blocks, and
wherein a size of the fixed block pool is a non-integer multiple of the unit block.

2. The device of claim 1 , wherein indices of the plurality of unit blocks respectively correspond to indices of the metadata stored in the memory allocation table.

3. The device of claim 2 , wherein the scratch pad memory generates a free list where free memories are connected, the free list being generated for each size of the free memory, and stores, in a prior block of the free memory, an index of a previous free memory and an index of a subsequent free memory.

4. The device of claim 3 , wherein: the memory managing unit further comprises a memory allocating unit to manage memory allocation in response to a memory use request, the memory allocating unit allocates a memory of a fixed block pool when a size of a requested memory is a size managed by the fixed block pool, and allocates a free memory of a free list, the size of the free memory being greater than the size of the requested memory and closest to the size of the requested memory when the size of the requested memory is not a size managed by the fixed block pool.

5. The device of claim 4 , wherein, when a difference between the size of the free memory of the free list and the size of the requested memory is greater than or equal to the unit block, the memory allocating unit separates, from the free memory, a free memory corresponding to the difference and adds the separated free memory to a corresponding free list.

6. The device of claim 3 , wherein the memory managing unit further comprises a memory releasing unit to manage a memory release in response to a memory release request, wherein the memory releasing unit comprises:
a memory incorporating unit, when a memory prior or subsequent to a memory chunk to be released is a free memory, to incorporate the free memory and the memory chunk to be released; a memory allocation table updating unit to update the memory allocation table in response to the memory release; a free list updating unit to update the free list in response to the memory release; a location calculating unit to calculate a fixed block location of the memory chunk to be released; and a bitmap updating unit to update a bitmap that manages the allocation or release of the fixed block.

7. The device of claim 1 , wherein, when the classified memory chunk is classified as the free memory, the memory managing unit records a size of an empty memory in a prior one of the unit blocks and a subsequent block of the memory allocation table that manages the free memory.

8. The device of claim 1 , wherein, when the classified memory chunk is classified as the allocated memory, the memory managing unit records a size of a memory that is allocated in a prior one of the unit blocks and a subsequent block of the memory allocation table that manages the allocated memory.

9. The device of claim 1 , wherein, when the memory chunk is classified as the fixed block pool, the memory managing unit includes at least one of location information and identification information in the memory allocation table that manages the fixed block pool, the identification information identifying the fixed block pool, and the location information recognizing a location in the fixed block pool.

10. The device of claim 1 , wherein the fixed block pool of the scratch pad memory comprises:
a first storing area to store a size of a fixed block;
a second storing area to store an index of a previous fixed block pool;
a third storing area to store an index of a subsequent fixed block pool;
a fourth storing area to indicate an allocation condition and a release condition of each of the fixed blocks; and
a fifth storing area to store use data.

11. A scratch pad memory management method, the method comprising:
dividing a scratch pad memory into a plurality of unit blocks;
maintaining a memory allocation table in a main memory to manage the scratch pad memory; and
managing the scratch pad memory comprising using the memory allocation table having indices respectively corresponding to the plurality of unit blocks,
wherein the managing of the scratch pad memory comprises:
generating a memory chunk including at least one unit block among the plurality of unit blocks of the scratch pad memory; and
classifying the memory chunk as one of a free memory, an allocated memory, and a fixed block pool,
wherein a size of the fixed block pool is a non-integer multiple of the unit block.

12. The method of claim 11 , wherein the managing of the scratch pad memory comprises recording a size of an empty memory in a prior block and a subsequent block of a memory allocation table that manages the free memory, when the memory chunk is classified as the free memory.

13. The method of claim 11 , wherein the managing of the scratch pad memory comprises recording a size of a memory that is allocated, in a prior block and a subsequent block of a memory allocation table that manages the allocated memory, when the memory chunk is classified as the allocated memory.

14. The method of claim 11 , wherein the managing of the scratch pad memory comprises including at least one of location information and identification information in the memory allocation table that manages the fixed block pool, the identification information identifying the fixed block pool and the location information recognizing a location in the fixed block pool, when the memory chunk is classified as the fixed block pool.

15. The method of claim 11 , wherein the managing of the scratch pad memory comprises:
generating a free list where free memories are connected; and
storing an index of a previous free memory and an index of a subsequent free memory in the prior block of the free memory.

16. The method of claim 15 , wherein the managing of the scratch pad memory comprises:
allocating a memory in response to a memory use request; and
releasing the memory in response to a memory release request.

17. The method of claim 16 , wherein the allocating of the memory comprises allocating a memory of the fixed block pool when a size of the requested memory is a size managed by the fixed block pool, and allocating a free memory of a free list, the size of the free memory being greater than the size of the requested memory and closest to the size of the requested memory when the size of requested memory is not a size managed by the fixed block pool.

18. The method of claim 17 , wherein, when a difference between the size of the free memory of the free list and the size of the request memory is greater than or equal to the unit block, the allocating of the memory comprises:
separating, from the free memory, a free memory corresponding to the difference; and
adding the separated free memory to a corresponding free list.

19. A non-transitory computer readable recording media storing a program imp

Assignees

• Lee Jae Don
• Lee Shi Hwa
• Lee Seung Won
• Im Chae Seok
• Jeong Min Kyu
• Samsung Electronics Co Ltd

Inventors

• Lee Jae Don
• Lee Shi Hwa
• Lee Seung Won
• Im Chae Seok
• Jeong Min Kyu

Classifications

• G06F12/023 Primary

  Free address space management · CPC title

• G06F12/0292

  using tables or multilevel address translation means (G06F12/023 takes precedence; address translation in virtual memory systems G06F12/10) · CPC title

• G06F12/00

  Accessing, addressing or allocating within memory systems or architectures (digital input from, or digital output to record carriers, e.g. to disk storage units, G06F3/06) · CPC title

• G06F12/06 Primary

  Addressing a physical block of locations, e.g. base addressing, module addressing, memory dedication (G06F12/08 takes precedence) · CPC title