Data compression for direct memory access transfers

• US 9,026,568 B2
• Filed: 09/14/2012
• Issued: 05/05/2015
• Est. Priority Date: 03/30/2012
• Status: Active Grant

• Current Assignee: Altera Corporation (Intel Corporation)
• Original Assignee: Altera Corporation (Intel Corporation)
• Inventors: Wegener, Albert W.
• Primary Examiner(s): Mai, Tan V.
• Application Number:
  US13/617,205
• Publication Number:
  US 20130262538A1
• Time in Patent Office: 963 Days
• Field of Search: 708/203
• US Class Current: 708/203
• CPC Class Codes: G06F 13/28 using burst mode transfer, e.g. direct memory access DMA, cycle steal G06F13/32 takes precedence
• CPC Class Codes: H03M 7/30 Compression speech analysis-synthesis for redundancy reduction G10L19/00; for image communication H04N; Expansion; Suppression of unnecessary data, e.g. redundancy reduction

Abstract

Memory system operations are extended for a data processor by DMA, cache, or memory controller to include a DMA descriptor, including a set of operations and parameters for the operations, which provides for data compression and decompression during or in conjunction with processes for moving data between memory elements of the memory system. The set of operations can be configured to use the parameters and perform the operations of the DMA, cache, or memory controller. The DMA, cache, or memory controller can support moves between memory having a first access latency, such as memory integrated on the same chip as a processor core, and memory having a second access latency that is longer than the first access latency, such as memory on a different integrated circuit than the processor core.

Claims

1. A computer system, comprising:

a data processor and memory accessible by the data processor, the memory storing a data set; and

a direct memory access (DMA) controller in communication with the data processor, the DMA controller comprising circuitry to control movement of the data set from a first element of the memory to a second element of the memory, where the second element of the memory has an access latency for accesses by the DMA controller that is longer than that of the first element of the memory, the data processor provides parameters via a DMA descriptor for a data compression procedure by the DMA controller, the parameters including an identifier of the data set to be moved and parameters identifying characteristics of the data compression procedure, the DMA controller including logic to perform data compression according to the parameters identifying the characteristics of the data compression procedure on the data set in the first element of the memory to form a compressed data set, and to store the compressed data set and parameters identifying characteristics of the compressed data set in the second element of the memory,

wherein the data compression procedure produces one or more packets of compressed data of the compressed data set, wherein the parameters include indicators of a number of packets and a number of samples per packet in the compressed data set.

2. The system of claim 1, wherein the identifier includes indicators of a compressed data set location in the second element of the memory, an uncompressed data set location in the first element of the memory and a selected compression mode.

3. The system of claim 2, wherein operations of the DMA controller perform data move operations between the compressed data set location and the uncompressed data set location, in which data move operations one of compression and decompression is performed, as specified by the parameters.

4. The system of claim 1, wherein one of the parameters indicates one of a set of data compression modes, the set of data compression modes including at least one lossy mode.

5. The system of claim 1, wherein the identifier of the data set includes an identifier of a data type for the data set to be moved.

6. The system of claim 5, wherein the identifier of a data type identifies one of a supported set of data types, the set including one-dimensional and two-dimensional data types.

7. The system of claim 5, wherein the identifier of a data type identifies one of a supported set of data types, the set including integer data types and floating-point data types.

8. The system of claim 5, wherein the identifier of a data type identifies one of a supported set of data types, and the parameters of the DMA descriptor include a bit width indicator for at least one of the supported set of data types.

9. The system of claim 5, wherein the identifier of a data type identifies one of a supported set of data types, and the parameters of the DMA descriptor include a signed vs. unsigned indicator for at least one of the supported set of data types.

10. The system of claim 5, wherein the identifier of a data type identifies one of a supported set of data types, including a two-dimensional data type and the parameters of the DMA descriptor include indicators of a number of samples per row and a number of samples per column for the two-dimensional data type.

11. The system of claim 5, wherein the identifier of a data type identifies one of a supported set of data types, including a two-dimensional data type and the parameters of the DMA descriptor include indicators of a color encoding selection for the two-dimensional data type.

12. The system of claim 1, wherein the parameters identifying characteristics of a data compression procedure support a set of different algorithms selectable by the parameters, a selected one of the different algorithms specified by corresponding parameters of the DMA descriptor is applied for the data compression procedure.

13. The system of claim 12, wherein the set of different algorithms includes algorithms specialized for data types identified in the parameters of the DMA descriptor.

14. The system of claim 12, wherein the set of different algorithms includes an algorithm for compression of floating-point numbers and an algorithm for compression of integers.

15. The system of claim 12, wherein the set of different algorithms includes an algorithm for compression of image data.

16. The system of claim 12, wherein the set of different algorithms includes an algorithm specialized for lossless compression and an algorithm specialized for lossy compression.

17. The system of claim 12, wherein the set of different algorithms includes an algorithm specialized to achieve a compression ratio specified by one of the parameters of the DMA descriptor.

18. The system of claim 12, wherein the set of different algorithms includes an algorithm specialized to achieve a compression quality, specified by one of the parameters of the DMA descriptor.

19. A computer system, comprising:

a data processor and memory accessible by the data processor, the memory storing a data set; and

a direct memory access (DMA) controller in communication with the data processor, the DMA controller comprising circuitry to control movement of the data set from a first element of the memory to a second element of the memory, where the second element of the memory has an access latency for accesses by the DMA controller that is longer than that of the first element of the memory, the data processor provides parameters via a DMA descriptor for a data compression procedure by the DMA controller, the parameters including an identifier of the data set to be moved and parameters identifying characteristics of the data compression procedure, the DMA controller including logic to perform data compression according to the parameters identifying the characteristics of the data compression procedure on the data set in the first element of the memory to form a compressed data set, and to store the compressed data set and parameters identifying characteristics of the compressed data set in the second element of the memory,

wherein the data processor provides parameters via a decompression DMA descriptor for a data decompression procedure by the DMA controller, the decompression DMA descriptor parameters including an identifier of the data set to be moved, an indicator of a compressed data set location in the second element of memory, and an indicator of a decompressed data set location in the first element of memory, the DMA controller including logic, to perform data decompression according to identified characteristics of the data decompression procedure on the compressed data set in the second element of the memory to form a decompressed data set, and to store the decompressed data set in the first element of the memory.

20. The system of claim 19, wherein at least one parameter identifying at least one characteristic of the data decompression procedure is included within the compressed data set stored in the second element of the memory.

21. The system of claim 19, wherein at least one parameter identifying at least one characteristic of the data decompression procedure is stored in a separate file in the second element of memory.

22. The system of claim 19, wherein the data compression procedure produced one or more packets of compressed data of the compressed data set, the decompression DMA descriptor parameters including indicators of a number of packets and a number of samples per packet in the compressed data set, wherein at least one of the packets stores at least one identifier of at least one characteristic of the data decompression procedure.

23. The system of claim 19, wherein respective packets contain respective parameters identifying the characteristics of the data decompression procedure for the respective packet, the DMA controller to apply the data decompression procedure to the compressed data of the respective packet in accordance with the respective parameter.

24. A method for transferring a data set between a first element of a memory and a second element of the memory comprising:

initiating a data compression procedure in a direct memory access (DMA) controller in communication with a data processor and the memory, where the second element of the memory has an access latency for accesses by the DMA controller that is longer than that of the first element of the memory;

passing an identifier of an uncompressed data set in the first element of the memory from the data processor to the DMA controller;

passing parameters identifying characteristics of a data compression procedure to be applied to the uncompressed data set from the data processor to the DMA controller;

retrieving the uncompressed data set from the first element of the memory in accordance with the identifier;

compressing the uncompressed data set according to the identified characteristics of the data compression procedure to form a compressed data set; and

storing the compressed data set in the second element of the memory,

wherein said compressing produces one or more packets of compressed data of the compressed data set, wherein the parameters include indicators of a number of packets and a number of samples per packet in the compressed data set.

25. The method of claim 24, wherein the identifier of the uncompressed data set includes indicators of an uncompressed data set location in the first element of memory and compressed data set location in the second element of memory.

26. The method of claim 24, wherein the step of storing the compressed data set includes storing at least one parameter identifying a characteristic of a decompression procedure within the compressed data set stored in the second element of the memory.

27. The method of claim 24, wherein the step of storing the compressed data set includes storing at least one parameter identifying a characteristic of a decompression procedure for the compressed data set in a separate file in the second element of the memory.

28. The method of claim 24, wherein the step of storing the compressed data set in the second element of the memory includes storing parameters determined during said compressing within the compressed data set stored in the second element of the memory.

29. The method of claim 24, wherein the step of compressing the uncompressed data set applies a lossy compression mode in accordance with at least one of the parameters.

30. The method of claim 24, wherein the step of compressing the uncompressed data set applies a lossless compression mode in accordance with at least one of the parameters.

31. The method of claim 24, wherein the parameters identifying characteristics of a data compression procedure support a set of different algorithms selectable by the parameters, wherein said compressing applies a selected one of the different algorithms in accordance with corresponding parameters.

32. The method of claim 31, wherein the set of different algorithms includes algorithms specialized for data types identified for the uncompressed data set.

33. The method of claim 31, wherein the set of different algorithms includes an algorithm for compression of floating-point numbers and an algorithm for compression of integers.

34. The method of claim 31, wherein the set of different algorithms includes an algorithm for compression of image data.

35. The method of claim 31, wherein the set of different algorithms includes an algorithm specialized for lossless compression and an algorithm specialized for lossy compression.

36. The method of claim 31, wherein the set of different algorithms includes an algorithm specialized to achieve a compression ratio, specified by one of the parameters.

37. The method of claim 31, wherein the set of different algorithms includes an algorithm specialized to achieve a compression quality, specified by one of the parameters.

38. The method of claim 24, wherein the identifier of the uncompressed data set includes an identifier of a data type.

39. The method of claim 38, wherein the identifier of a data type identifies one of a supported set of data types, the set including one-dimensional and two-dimensional data types.

40. The method of claim 38, wherein the identifier of a data type identifies one of a supported set of data types, the set including integer data types and floating-point data types.

41. The method of claim 38, wherein the identifier of a data type identifies one of a supported set of data types, and the parameters include a bit width indicator for at least one of the supported set of data types.

42. The method of claim 38, wherein the identifier of a data type identifies one of a supported set of data types, and the parameters include a signed vs. unsigned indicator for at least one of the supported set of data types.

43. The method of claim 38, wherein the identifier of a data type identifies one of a supported set of data types, including a two-dimensional data type and the parameters include indicators of a number of samples per row and a number of samples per column for the two-dimensional data type.

44. The method of claim 38, wherein the identifier of a data type identifies one of a supported set of data types, including a two-dimensional data type and the parameters include indicators of a color encoding selection for the two-dimensional data type.

45. A method for transferring a data set between a first element of a memory and a second element of the memory comprising:

initiating a data compression procedure in a direct memory access (DMA) controller in communication with a data processor and the memory, where the second element of the memory has an access latency for accesses by the DMA controller that is longer than that of the first element of the memory;

passing an identifier of an uncompressed data set in the first element of the memory from the data processor to the DMA controller;

passing parameters identifying characteristics of a data compression procedure to be applied to the uncompressed data set from the data processor to the DMA controller;

retrieving the uncompressed data set from the first element of the memory in accordance with the identifier;

compressing the uncompressed data set according to the identified characteristics of the data compression procedure to form a compressed data set;

storing the compressed data set in the second element of the memory; and

transferring a compressed data set to be decompressed from the second element to the first element of the memory, comprising;

initiating a data decompression procedure in the DMA controller;

passing an identifier of the compressed data set to be decompressed from the data processor to the DMA controller;

retrieving the compressed data set from the second element of the memory in accordance with the identifier of the compressed data set to be decompressed;

retrieving a parameter identifying at least one characteristic of a data decompression procedure from the second element of the memory;

decompressing the compressed data set according to the identified characteristic of the data decompression procedure to produce a decompressed data set; and

storing the decompressed data set in the first element of the memory.

46. The method of claim 45, wherein the identifier of the compressed data set to be decompressed includes indicators of a compressed data set location in the second element of the memory and a decompressed data set location in the first element of the memory.

47. The method of claim 45, wherein the step of retrieving the parameter identifying at least one characteristic of the decompression procedure includes retrieving the parameter from within the compressed data set stored in the second element of the memory.

48. The method of claim 45, wherein the step of retrieving the parameter identifying at least one characteristic of the decompression procedure includes retrieving the parameter from a file in the second element of the memory.

49. The method of claim 45, wherein the data compression procedure produces one or more packets of compressed data of the compressed data set, the identifier of the compressed data set to be decompressed including indicators of a number of packets and a number of samples per packet in the compressed data set, wherein at least one of the packets stores the parameter identifying at least one characteristic of the data decompression procedure.

50. The method of claim 49, wherein respective packets contain respective parameters identifying the characteristics of the data decompression procedures to be applied to the respective packets, wherein said retrieving a parameter retrieves the respective parameters from the respective packets and said decompressing applies the data decompression procedure to the compressed data of the respective packet in accordance with the respective parameter.