Neural processing device and method for transmitting data thereof

1. A processing device comprising processing circuitry comprising:

a plurality of processors;

a first memory shared by the plurality of processors; and

a cache comprising a second memory comprising a plurality of memory units, each of the plurality of memory units in the second memory being associated with a respective one of a plurality of request identifiers,

wherein the cache is configured to cause:

receiving, from at least one of the plurality of processors, a memory read request including a request identifier and a memory address,

identifying an allocated memory address identifier for the memory address,

accessing the first memory to read data of the memory address,

obtaining one or more request identifiers which requested data of the memory address from the second memory based on the allocated memory address identifier, and

transmitting the data of the memory address to one or more processors which requested data of the memory address based on the one or more request identifiers,

wherein the cache further comprises:

a third memory including a plurality of memory units, each of the plurality of memory units in the third memory being associated with a respective one of a plurality of memory address identifiers,

wherein the cache is further configured to cause:

reading a memory unit of the third memory based on the allocated memory address identifier to obtain a tail request identifier,

storing the request identifier of the memory read request to a memory unit of the second memory based on the tail request identifier, and

writing the request identifier of the memory read request as the tail request identifier to a memory unit of the third memory based on the allocated memory address identifier.

2. The processing device of claim 1 , wherein obtaining the one or more request identifiers comprises:

accessing a memory unit of the third memory based on the allocated memory address identifier to obtain a head request identifier and a tail request identifier,

accessing the second memory based on the head request identifier and the tail request identifier to obtain the one or more request identifiers, and wherein accessing the second memory based on the head request identifier and the tail request identifier comprises:

reading a memory unit of the second memory based on the head request identifier to obtain a next request identifier and reading a memory unit of the second memory based on the next request identifier unless the next request identifier is the tail request identifier to obtain the one or more request identifiers which requested data of the memory address.

3. The processing device of claim 1 , wherein the cache further comprises:

a fourth memory including a plurality of memory units, each of the plurality of memory units in the fourth memory being associated with a respective one of a plurality of memory address identifiers.

4. The processing device of claim 3 , wherein the cache is further configured to cause:

checking whether the memory address is stored in the fourth memory.

5. The processing device of claim 4 , wherein identifying the allocated memory address identifier for the memory address comprises:

when the memory address is not stored in the fourth memory, allocating the memory address an allocable memory address identifier among the plurality of memory address identifiers, and storing the memory address to a memory unit of the fourth memory based on the allocated memory address identifier.

6. The processing device of claim 4 , wherein identifying the allocated memory address identifier for the memory address comprises:

when the memory address is stored in the fourth memory, identifying a location of a memory unit storing the memory address in the fourth memory as the allocated memory address identifier.

7. The processing device of claim 1 , wherein the memory address is an address indicating a memory unit in the first memory, an address representing a memory array including a plurality of memory units in the first memory, or an address representing a memory pattern including a plurality of memory units in the first memory.

8. A method of operating a processing device comprising a plurality of processors, a shared memory shared by the plurality of processors, and a cache, wherein the cache comprises a first table comprising a plurality of elements, each of the plurality of elements in the first table being associated with a respective one of a plurality of request identifiers, the method comprising:

receiving, by the cache from at least one of the plurality of processors, a memory read request including a request identifier and a memory address;

identifying, by the cache, an allocated memory address identifier for the memory address,

accessing, by the cache, the shared memory to read data of the memory address;

obtaining, by the cache from the first table, one or more request identifiers which requested data of the memory address based on the allocated memory address identifier; and

transmitting, by the cache, the data of the memory address to one or more processors which requested data of the memory address based on the one or more request identifiers,

wherein the cache further comprises:

a second table including a plurality of elements, each of the plurality of elements in the second table being associated with a respective one of a plurality of memory address identifiers,

wherein the method is further comprising:

reading, by the cache, an element of the second table based on the allocated memory address identifier to obtain a tail request identifier;

storing, by the cache, the request identifier of the memory read request to an element of the first table based on the tail request identifier; and

writing, by the cache, the request identifier of the memory read request as the tail request identifier to an element of the second table based on the allocated memory address identifier.

9. The method of claim 8 , wherein obtaining the one or more request identifiers comprises:

accessing an element of the second table based on the allocated memory address identifier to obtain a head request identifier and a tail request identifier, and

accessing the first table based on the head request identifier and the tail request identifier to obtain the one or more request identifiers, and

wherein accessing the first table based on the head request identifier and the tail request identifier comprises:

reading an element of the first table based on the head request identifier to obtain a next request identifier and reading an element of the first table based on the next request identifier unless the next request identifier is the tail request identifier to obtain the one or more request identifiers which requested data of the memory address.

10. The method of claim 8 , wherein the cache further comprises:

a third table including a plurality of elements, each of the plurality of elements in the third table being associated with a respective one of a plurality of memory address identifiers.

11. The method of claim 10 , further comprising:

checking, by the cache, whether the memory address is stored in the third table.

12. The method of claim 11 , wherein identifying the allocated memory address identifier for the memory address comprises:

when the memory address is not stored in the third table, allocating the memory address an allocable memory address identifier among the plurality of memory address identifiers, and storing the memory address to an element of the third table based on the allocated memory address identifier.

13. The method of claim 11 , wherein identifying the allocated memory address identifier for the memory address comprises:

when the memory address is stored in the third table, identifying a location of an element storing the memory address in the third table as the allocated memory address identifier.

14. The method of claim 8 , wherein the memory address is an address indicating a memory unit in the shared memory, an address representing a memory array including a plurality of memory units in the shared memory, or an address representing a memory pattern including a plurality of memory units in the shared memory.