Graphic processing unit and method for decompressing data

• US 9,013,473 B2
• Filed: 12/20/2011
• Issued: 04/21/2015
• Est. Priority Date: 02/09/2011
• Status: Active Grant

• Current Assignee: Samsung Electronics Co. Ltd.
• Original Assignee: Samsung Electronics Co. Ltd.
• Inventors: Ahn, Min Su, Kim, Do Kyoon, Rhee, Tae Hyun
• Primary Examiner(s): Ma, Tize
• Assistant Examiner(s): Chen, Yu
• Application Number:
  US13/331,943
• Publication Number:
  US 20120200559A1
• Time in Patent Office: 1,218 Days
• Field of Search: None
• US Class Current: 345/419
• CPC Class Codes: G06T 15/00 3D [Three Dimensional] image rendering
• CPC Class Codes: G06T 9/00 Image coding bandwidth or redundancy reduction for static pictures H04N1/41; coding or decoding of static colour picture signals H04N1/64; methods or arrangements for coding, decoding, compressing or decompressing digital video signals H04N19/00
• CPC Class Codes: H04N 19/20 using video object coding
• CPC Class Codes: H04N 19/436 using parallelised computational arrangements

Abstract

A graphic processing unit (GPU) and method for decompressing compressed 3-dimensional (3D) compressed data. The GPU may extract segment information by analyzing a compressed data header and decompress segments included in a bit stream based on the segment information.

Claims

1. A graphic processing unit (GPU) to decompress compressed 3-dimensional (3D) object data in association with a central processing unit (CPU), the GPU comprising:

a segment information extractor to extract segment information included in a compressed data header, by analyzing the compressed data header included in a bit stream of compressed data received from the CPU; and

a decompressor to

decompress segments of the compressed 3D object data included in the bit stream based on the segment information, the compressed 3D object data being 3D object data divided into the segments and compressed, and

store vertex position information and vertex connection information, which are a result of decompression, in a memory according to address information included in a segment header of each of the segments,

wherein the segment header of at least one of the segments comprises a write address information, and

wherein the write address information comprises address information indicating a start of the storing, in a second memory space, the vertex position information and the vertex connection information.

2. The GPU of claim 1, further comprising:

a memory allocator to allocate a first memory space adapted to store the bit stream of the compressed data and the second memory space adapted to store a result of the decompression.

3. The GPU of claim 2, wherein

the decompressor stores the vertex position information and the vertex connection information in the second memory space according to the write address information, and

each of the segments comprises segment data including the vertex position information, the vertex connection information, and a segment header.

4. The GPU of claim 1, wherein the segment header further comprises at least one of quantization information indicating bits of a quantized segment, information on a number of vertices of the segment, and information on a number of faces of the segment.

5. Te GPU of claim 1, further comprising:

a thread generator to generate threads adapted to decompress the segments based on the segment information.

6. The GPU of claim 5, wherein the segment information comprises at least one of a segment position information, a segment number information, a segment size information, and a bit stream information.

7. The GPU of claim 5, wherein

the thread generator generates the threads corresponding to the respective segments based on the segment position information, and

the decompressor decompresses the segments in parallel using the threads.

8. The GPU of claim 5, wherein, when a number of the segments is greater than a number of the threads generated based on the segment information, the decompressor decompresses a portion of the segments using the generated threads and then decompresses the remaining segments by reusing the generated threads.

9. The GPU of claim 1, further comprising:

a memory allocator to allocate the second memory space adapted to store a result of the decompression; and

a thread generator to generate threads respectively corresponding to the segments.

10. A non-transitory recording medium storing a bit stream of compressed 3-dimensional (3D) object data transmitted by a central processing unit (CPU) to a graphic processing unit (GPU),

wherein the compressed 3D object data is 3D object data divided into segments and compressed,

wherein the bit stream comprises a compressed data header and the segments of the compressed 3D object data,

wherein the compressed data header contains segment information of the segments,

wherein the segments are decompressed based on the segment information,

wherein vertex position information and vertex connection information, which are a result of decompression, are stored in a memory according to address information included in the segment header of each of the segments,

wherein the segment header of at least one of the segments comprises a write address information, and

wherein the write address information comprises address information indicating a start of the storing, in a second memory space, the vertex position information and the vertex connection information.

11. The non-transitory recording medium of claim 10, wherein the segment information comprises at least one of a segment number information, a segment size information, and a bit stream information.

12. A graphic processing method performed by a graphic processing unit (GPU) for decompressing compressed 3-dimensional (3D) object data in association with a central processing unit (CPU), the method comprising:

extracting segment information included in a compressed data header, by analyzing the compressed data header included in a bit stream of compressed data received from the CPU; and

decompressing segments of the compressed 3D object data included in the bit stream based on the segment information, the compressed 3D object data being 3D object data divided into the segments and compressed; and

storing vertex position information and vertex connection information, which are a result of decompression, in a memory according to address information included in a segment header of each of the segments,

wherein the segment header of at least one of the segments comprises a write address information, and

wherein the write address information comprises address information indicating a start of the storing, in a second memory space, the vertex position information and the vertex connection information.

13. The graphic processing method of claim 12, further comprising:

allocating a first memory space adapted to store the bit stream of the compressed data; and

allocating the second memory space adapted to store a result of decompressing the segments.

14. The graphic processing method of claim 13, wherein the decompressing comprises storing the vertex position information and the vertex connection information in the second memory space according to the write address information , and

each of the segments comprises segment data including the vertex position information, the vertex connection information, and a segment header.

15. The graphic processing method of claim 12, wherein the segment header further comprises at least one of quantization information indicating bits of a quantized segment, information on a number of vertices of the segment, and information on a number of faces of the segment.

16. The graphic processing method of claim 12, further comprising:

generating threads adapted to decompress the segments based on the segment information.

17. The graphic processing method of claim 16, wherein the segment information comprises a segment number information, a segment size information, and a bit stream information.

18. The graphic processing method of claim 16, further comprising:

the generating of the threads generates the threads respectively corresponding to the segments based on the segment position information included in the segment information, and

the decompressing decompresses the segments in parallel using the threads.

19. The graphic processing method of claim 16, wherein, when a number of the segments is greater than a number of the threads generated based on the segment information, the decompressing decompresses a portion of the segments using the generated threads and then decompresses the remaining segments by reusing the generated threads.

20. The graphic processing method of claim 12, further comprising:

allocating a first memory space adapted to store the bit stream of the compressed data;

allocating the second memory space adapted to store a result of decompressing the segments; and

generating threads respectively corresponding to the segments.

21. A non-transitory computer readable recording medium storing a program to cause a computer to implement the method of claim 12.

22. A computer system to decompress compressed 3-dimensional (3D) object data, the computer system comprising:

a segment information extractor to extract segment information from a compressed data header included in a bit stream of compressed data; and

a decompressor to

decompress segments of the compressed 3D object data included in the bit stream based on the segment information, the compressed 3D object data being 3D object data divided into the segments and compressed, and

store vertex position information and vertex connection information, which are a result of decompression, in a memory according to address information included in a segment header of each of the segments,

wherein the segment header of at least one of the segments comprises a write address information, and

wherein the write address information comprises address information indicating a start of the storing, in a second memory space, the vertex position information and the vertex connection information.

23. A graphic processing method performed by a graphic processing unit (GPU) for decompressing compressed 3-dimensional (3D) object, the method comprising:

extracting segment information from a compressed data header included in a bit stream of compressed data; and

decompressing segments of the compressed 3D object data included in the bit stream based on the segment information, the compressed 3D object data being 3D object data divided into the segments and compressed; and

storing vertex position information and vertex connection information, which are a result of decompression, in a memory according to address information included in a segment header of each of the segments,

wherein the segment header of at least one of the segments comprises a write address information, and

wherein the write address information comprises address information indicating a start of the storing, in a second memory space, the vertex position information and the vertex connection information.