Single computer distributed memory computing environment and implementation thereof

• US 20040221122A1
• Filed: 04/30/2003
• Published: 11/04/2004
• Est. Priority Date: 04/30/2003
• Status: Active Grant

• Current Assignee: InteliTrac, Inc.
• Original Assignee: InteliTrac, Inc.
• Inventors: Chen, Tianlong, Wang, Yingbin, Vu, Jonathan, Wei, Yinong
• Granted Patent: US 7,051,158 B2
• Time in Patent Office: Days
• US Class Current: 711/202
• CPC Class Codes: G06F 12/0284 Multiple user address space allocation, e.g. using different base addresses interprocessor communication G06F15/163
• CPC Class Codes: G06F 12/0806 Multiuser, multiprocessor or multiprocessing cache systems

Abstract

A computing environment for a device having a processor, a device memory and an operating system. A first process running in the operating system has a memory and a memory cache. A second process running in the operating system has a second process memory accessible to the first process. A communication channel between the first process and second process makes the second process memory available for the first process memory to use. The device memory is sectioned into memory blocks, each having a starting address. A block ID identifies the memory block. Data is stored in various memory units of the memory block. An offset ID identifies the memory unit. A DMCE Virtual address contains the offset ID, the block Id and a device ID for identifying the memory unit. The DMCE virtual address of a memory unit used for the memory function is copied into the memory cache.

Claims

1. A computing environment comprising:

a device comprising a processor, a device memory and an operating system;

a first process running in said operating system, said first process having a first process memory and a memory cache within said first process memory;

a second process running in said operating system, said second process having a second process memory accessible to said first process; and

a communication channel between said first process and said second process;

wherein a total accessible memory of said first process comprises said first process memory and said second process memory.

2. A computing environment according to claim 1 wherein said communication channel between said first process and said second process comprises a memory allocation request from said first process to said second process.

3. A computing environment according to claim 1 wherein said communication channel between said first process and said second process comprises:

a memory address provided by said first process to said second process, the memory address identifying a set of information to be read corresponding to said first memory address in said second process memory.

4. A computing environment according to claim 3 wherein said set of information is stored in said memory cache of said first process.

5. A computing environment according to claim 1 wherein said communication channel between said first process and said second process comprises:

a memory address and a corresponding set of information provided by said first process to said second process to write to a memory location in said second process,

wherein said memory location corresponds to said memory address given by said first process.

6. A computing environment according to claim 1, said first process further comprises a stored content, wherein said stored content is written to said second process memory.

7. A computing environment according to claim 1 wherein said memory cache in said first process comprises at least one memory cache unit.

8. A computing environment according to claim 7 wherein said memory cache unit comprises at least:

a memory address; and

a content;

wherein said memory address is uniquely mapped to a memory location in said second process memory; and

wherein said content is a copy of content from said memory location in said second process memory.

9. A computing environment according to claim 7 wherein said memory cache unit comprises at least:

a memory address; and

a content;

wherein said memory address is uniquely mapped to a memory location in said second process memory; and

wherein said content may be written to said memory location in said second process memory.

10. A computing environment according to claim 8 wherein said memory address comprises a block ID, and an offset ID.

11. A computing environment according to claim 10 wherein said memory address further comprises a device ID.

12. A computing environment according to claim 8, said device memory comprising:

a memory block comprising at least one memory unit;

a device ID identifying said device;

a block ID identifying said memory block in said memory of said device;

an offset ID identifying a memory unit in said memory block; and

a virtual address identifying said memory unit in said memory block of said device;

wherein said virtual address contains the device ID, the block ID, and offset ID of a memory unit and said virtual address is copied into said memory cache.

13. A computing environment according to claim 8 wherein said memory cache in said first process comprises a memory device lookup table;

14. A computing environment according to claim 13 wherein said second process in said device comprises a device-process entry in said memory device lookup table.

15. A computing environment according to claim 14 wherein said device-process entry for said second process further comprises information on said communication channel between said first process and said second process.

16. A computing environment according to claim 1 wherein:

said communication channel further comprises a First-In-First-Out (FIFO) queue;

said first process further comprises a first thread, a corresponding messaging block for said first thread in shared memory, and a first thread ID for uniquely identifying said messaging block in said shared memory; and

said second process further comprises a second thread;

wherein said first thread in said first process puts requests and gets responses from said corresponding messaging block in said shared memory; and

wherein said second thread in said second process gets requests and put responses from said corresponding messaging block in said shared memory.

17. A computing environment according to claim 16, wherein said first thread in said first process puts request information in said messaging block memory and put its said ID of said first thread in said FIFO queue;

18. A computing environment according to claim 16, where said second thread in said second process checks said FIFO queue for available said ID from said first thread of said first process.

19. A computing environment according to claim 17, wherein said second thread in said second process gets a said ID of said first thread in said first process from said FIFO queue and removes the said ID from said FIFO queue, and read a request in said messaging block corresponding to said ID of said first thread in said first process.

20. A computing environment according to claim 19, wherein said second thread in said second process puts a response based on said request from said first thread in said first process in the corresponding said messaging block of said first thread in said first process.

21. A method according to claim 20, wherein said first thread in said first process gets a response from said second thread in said second process from the corresponding said messaging block in said shared memory.

23. The computing environment according to claim 16, wherein the FIFO queue has a structure selected from the group consisting of:

a circular array and a linked list.

24. A computing environment according to claim 11 wherein said memory address further comprises a process ID.

22. A method of expanding available memory in a computing environment having a device with a processor, a device memory and an operating system comprising:

running a first process in said operating system;

said first process having a first process memory and a memory cache within said first process memory;

running a second process in said operating system, said second process having a second process memory accessible to said first process; and

sending memory function requests between said first process and said second process through a communication channel;

generating a memory request result; and

caching said memory request result in said memory cache.