Design and implementation of GPU-based SAR image processor
Download as pptx, pdf4 likes864 views
Najeeb Ahmad's master thesis discusses the design and implementation of a GPU-based synthetic aperture radar (SAR) image processor, focusing on accelerating SAR data processing through modern GPGPU technologies. The research involves implementing a simplified Range Doppler Algorithm on an NVIDIA Tesla GPU, comparing the performance of various implementations (CPU, MATLAB-GPU, and CUDA). The findings highlight significant speedups achieved with GPU processing and address future improvements in CUDA support and MATLAB functions.
Design and implementation of GPU-based SAR image processor
- 1. Najeeb Ahmad Master Thesis Presentation May, 2012 Supervisor: Dr. Sun Jinping Design and Implementation of GPU based SAR Image Processor School of Electronic Information Engineering Beihang University, Beijing China.
- 2. Contents 1. Introduction 2. GPU Computing 3. SAR Processing 4. Implementation 5. Conclusion & Future Work
- 3. 1.Introduction Problem Motivation Objective Methodology
- 4. PROBLEM Synthetic Aperture Radar data processing is a computationally intensive and time consuming task using conventional CPUs. Given the increasing popularity and use of GPU for scientific computing, it is required to accelerate simplified range Doppler SAR processing algorithm on GPU using modern GPGPU technology to achieve real/near real-time performance and to evaluate its suitability for SAR processing.
- 5. MOTIVATION Computationally intensive and time consuming nature of SAR processing algorithms. Inherent algorithm parallelism in most SAR processing algorithms. Advent of modern GPGPU technology and availability of commodity GPUs as general purpose computation engines. Architectural parallelism and availability of sufficient hardware resources in modern GPUs rendering them especially useful for handling large data quantities and parallel SAR algorithm implementation.
- 6. OBJECTIVE To implement and accelerate simplified range Doppler SAR processing algorithm on a modern NVIDIA TESLA GPU using CUDA and MATLAB- GPU capabilities. The resulting research will explore the areas like: Algorithm adaptation for parallel implementation. Suitability of MATLAB for algorithm implementation. Suitability of CUDA for algorithm implementation. Comparison of CPU/CUDA/MATLAB-GPU implementations. GPU as SAR processing platform.
- 7. METHODOLOGY Algorithm implementation and verification on Intel Xeon CPU using MATLAB. Identification of parallelizable portions of algorithm. Algorithm implementation on TESLA C1060 GPU using MATLAB’s native GPU capabilities. Algorithm implementation on TESLA C1060 GPU using CUDA. Analysis of CPU, MATLAB-GPU and CUDA implementations.
- 8. 2.GPU Computing Introduction to GPU Computing GPGPU: Brief History NVIDIA CUDA Writing efficient code
- 9. Introduction to GPU Computing Use of Graphics Processing Units (GPUs) for general purpose computing applications. CPU: Single, four or eight cores. Capable of handling few threads. Suitable for serial code. GPU: Hundreds of cores. Capable of handling hundreds of threads. Suitable for parallel code.
- 10. Introduction to GPU Computing GPU Computing Model: Heterogeneous computing model employing both CPU and GPU with serial computing on CPU, parallel computing on GPU.
- 11. GPGPU: Brief History First use of GPU as general purpose computing device, around 1999-2000 using graphics APIs. Huge performance boosts observed. Generally unpopular due to tedious programming. Introduction of NVIDIAs “CUDA” and AMDs “Stream Computing” in 2007. Beginning of modern GPGPU era. Other vendors introduced their own GPGPU systems. NVIDIAs CUDA gaining popularity due to its maturity and performance.
- 12. NVIDIA CUDA Compute Unified Device Architecture. Comprises of Instruction Set Architecture (ISA) and parallel compute engine in GPU programmable with high level languages extended for GPU computing. CUDA framework comprises of two parts; hardware and software. From software perspective, CUDA means extended C/C++, FORTRAN to support GPU computing. CUDA is “Single Instruction Multiple Thread” (SIMT) architecture.
- 13. CUDA Hardware Streaming multiprocessor (SM): Basic computing unit of the GPU. Comprises of eight streaming processors (SP) and memory. Different GPUs differ in number of SMs and SP clock frequency. SP SP SP SP SP SP SP SP SFU SFU MT IU Shared Memory
- 14. CUDA Memory Architecture Understanding of memory architecture critical for writing efficient CUDA programs. All CUDA-enabled hardware have following types of memory: Global memory Shared memory and registers. Texture memory and texture cache. Constant memory and constant cache. Local memory for register spilling. SP SP Shared memory SP SP SP Texture cache Constant cache SM n SP SP Shared memory SP SP SP Texture cache Constant cache SM 3 SP SP Shared memory SP SP SP Texture cache Constant cache SP SP Shared memory SP SP SP Texture cache Constant cache SM 1 SM 2 GPU Global memory (RAM) Local MemoryTexture memory Constant memory
- 15. NVIDIA TESLA C1060 GPU PCI Express 2.0 compliant computing processor board based on NVIDIA Tesla T10 graphics processing unit targeted for HPC applications. Feature highlights 30 SMs = 240 SPs. SP Clock = 1.296 GHz 4 GB DDR3 memory with 120 GB/s bandwidth. IEEE 754 single and double floating point compliant. 933 GFLOPS single and 78 GFLOPS double precision performance. Compute capability: 1.3 Supported by MATLAB for GPU computing
- 16. CUDA Programming Model At its core are thread groups, shared memory and barrier synchronization. Provides coarse-grained data and task parallelism and fine-grained data and thread parallelism providing expressivity and scalability. Thread hierarchy: Grid, blocks, threads. Kernels: Functions executed on device (GPU) in parallel threads. CUDA provides APIs to run and launch kernels in parallel threads and to synchronize them.
- 17. Processing Flow Copy input data from CPU to GPU memory. Load GPU program and execute, caching result on the device. Copy results from GPU to CPU. RAM CPU Host Global memory Constant Texture GPU Device
- 18. Writing Efficient Code High priority considerations Minimum CPU-GPU transfers. Use of coalesced data transfers. Use of shared memory instead of global memory whenever possible. Avoiding different execution paths within a warp. Medium priority considerations Access to shared memory should be planned to avoid serialization. Redundant data transfers from global memory should be avoided.
- 19. Writing Efficient Code Threads per block should be multiple of 32. Use of fast math library whenever possible. Low Priority Considerations Use of zero copy operations. For kernels with long argument list, some argument should be placed in constant memory. Expensive modulo, division operations should be avoided in favor of shift operations whenever possible. Automatic conversion of double to float should be avoided. Loop unrolling should be used whenever possible.
- 20. 3.SAR Processing What is Synthetic Aperture Radar SAR Processing Processing Algorithms Basic RDA Simplified RDA
- 21. What is Synthetic Aperture Radar An active microwave remote sensing imaging system. Employs long range propagation characteristics of radar and complex signal processing techniques to produce high resolution images. High resolution achieved by synthesizing long antenna aperture through signal processing techniques. Pros (in comparison with optical systems): All weather and day and night operation. No effects of constituents of atmosphere. Sensitivity to dielectric properties (can image ice, biomass etc.) Sensitivity to surface roughness (oceans, wind
- 22. What is Synthetic Aperture Radar Accurate measurement of distance. Sensitivity to man made objects. Sensitivity to target structure. Subsurface penetration. Cons Complex interactions (difficult to visualize and understand) Speckle effects (difficult in visual interpretation) Topographic effects
- 23. SAR Processing A set of procedures to obtain interpretable image from raw scattered in azimuth and range directions. In range, data is scattered by duration of transmitted FM pulse. In azimuth, data spread by duration point target is illuminated by the radar beam. SAR processing compresses this data taking into account range cell migration, earth curvature, earth rotation, air/spacecraft attitude noise to produce the final image. Given nature of SAR system and signals, signal processing rather than image processing provide appropriate tools for SAR processing.
- 24. SAR Processing Algorithms Mainstream SAR processing include: Range Doppler algorithm (RDA) High resolution images for low squint and for relatively smaller aperture sizes. Very popular. Chirp scaling algorithm (CSA) Two-dimensional operations with range independence followed by range corrections in range Doppler domain. Omega-K algorithm (ωKA) Efficient and accurate in two-dimensional frequency domain. SPECAN algorithm Good for medium to low resolution requirements.
- 25. Range Doppler Algorithm Versions of range Doppler: Basic RDA RDA with accurate SRC RDA with approximate SRC Simplified range Doppler
- 26. Basic RDA Raw data Range Compression Azimuth FFT RCMC Azimuth Compression Azimuth IFFT and lookup Summation Final Image Range FFT, matched filter multiply, range IFFT Data in range Doppler domain Interpolation operation in range Doppler domain Azimuth matched filter multiply To bring back signal into time domain.
- 27. Simplified RDA For narrower swath width and medium resolution requirements, RCM can be assumed independent of range.Raw data Pre-filtering Range Compression Azimuth FFTRCMCRange IFFT Azimuth Compression Azimuth IFFT and lookup Summation Final Image To remove Doppler centroid Range FFT, matched filter multiply (No range IFFT) Both range and azimuth in frequency domain RCM phase function multiply with each range line Data in range Doppler domain
- 28. 4.Implementation Hardware resources Software resources CPU Implementation MATLAB GPU Implementation CUDA Implementation Result Comparison
- 29. Hardware resources CPU GPU Name NVIDIA Tesla C1060 # of cores 240 SP Clock 1.296 GHz Memory 4 GB GDDR3 Maximum memory bandwidth 102 GB/s Memory interface 512 bit – PCI Express GFLOPS 933 single precision, 78 double precision Name Intel Xeon E5504 CPU Clock 2 GHz # of cores 4 System Memory 4 GB DDR3 Clock 800 MHz Maximum memory bandwidth 19.2 GB/s Memory type DDR3 PC3 PCI Slot PCI Express
- 30. Software resources CPU GPU Windows 7 Ultimate 64-bit MATLAB release 2010b Visual Studio 2008 SP1 CUDA Toolkit 4.1 MATLAB release 2010b NVIDIA Parallel Nsight Visual Profiler CUDA MEMCHECK CUFFT library
- 31. RADARSAT – I Data • CEOS Format • Raw data is required to be extracted from CEOS data before SAR processing algorithm can be applied. Parameter Value Units Sampling rate 32.317 MHz Range FM rate 0.7213 5 MHz/µs Pulse duration 41.74 µs Radar frequency 5.3 GHz Radar wavelength 0.0565 7 m Pulse repetition frequency 1256.9 8 Hz Effective radar velocity 7062 m/s Azimuth FM rate 1733 Hz/s Table RADARSAT – I data parameters CEOS data CEOS data extraction utility RAW SAR data
- 32. SAR Processing GUI Functions • CEOS data extraction. • MATLAB- CPU SAR processing. • MATLAB- GPU SAR processing • CUDA input/output manipulation. • CUDA program execution.
- 33. CPU Implementation Implemented using MATLAB FFT/IFFT using standard MATLAB functions
- 34. CPU Processed SAR image A 2048 x 4096 SAR image using CPU based implementation
- 35. MATLAB-GPU Implementation MATLAB started supporting GPU computing since MATLAB release 2010b. Implemented using native MATLAB-GPU functions only (no CUDA kernel calls). Vectorization strategy employed to implement vector- matrix multiplications on GPU. All FFT/IFFTs performed using MATLAB-GPU FFT/IFFT support functions. Column1 Column2 ………... Columnn Column1 Column2 ………... Columnn Column1 Column2 ………... Columnn
- 36. MATLAB-GPU Implementation Limit on maximum image size that can be calculated due to GPU memory constraints.
- 37. MATLAB-GPU Implementation Speedup as high as 21 achieved compared with CPU implementation
- 38. MATLAB-GPU Implementation A 2048 x 4096 SAR image using MATLAB-GPU based implementation
- 39. MATLAB-GPU Implementation Advantages Quick and easy to implement Sufficient speedups obtained with little effort Little knowledge of GPU hardware and no knowledge of optimization techniques required. Disadvantages Currently, limited number of MATLAB functions supported on GPU. Not all overloads of a function available for GPU. Lesser control of hardware resources and memory. Not many optimization options.
- 40. CUDA Implementation Strategy Signal data read as binary file Vectors, matched filters calculated on CPU Vectors/signal data transferred to GPU Following kernels executed in order on GPU Pre-filtering kernel Range compression kernel RCMC kernel Azimuth compression kernel Image pixel calculation kernel Data transferred from GPU to CPU and saved on disk as image.
- 41. Optimization considerations Chosen block size = 8 × 8 = 64. Conforms with memory coalescing requirements. Constant variables stored in constant memory Local variable and phase function calculation whenever possible to reduce global memory access. CPU-GPU data transfer kept to minimum by transferring data from CPUGPU at beginning and GPUCPU transfers at the end of algorithm. Using CUFFTs cufftPlanMany() plan for FFT/IFFTs along data columns.
- 42. CUDA Implementation Results A 2048 x 4096 SAR image using CUDA based implementation
- 47. MATLAB-GPU/CUDA speedup comparison Speedups as high as 53 times achieved in comparison with maximum speedup of 21 times in MATLAB.
- 48. 5. Conclusions & Future Work
- 49. Conclusions Feasibility of GPU for SAR processing Amount of data, computational effort and inherent algorithm parallelism makes SAR processing suitable on GPU. TESLA C1060 GPU offers enough memory to handle various common SAR image sizes. Cooling GPU may be a challenge in some environments. Scalability of CUDA will prove to be an advantage to port existing SAR code to newer GPUs. GPUs might not be suitable where customizable hardware is required or military hardware standards are to be adhered.
- 50. Conclusions MATLAB-GPU based SAR Processing Significant speedups compared with CPU. Quick and easy to implement. Has some limitations: Currently have lesser function support for GPU. Expected to improve with future MATLAB releases. Vectorization strategy needs more memory. Future release promise to take away need for vectorization (e.g. bsxfun in release 2012a). Lesser control over GPU resources (memory etc.). CUDA SAR Processing CUDA: Flexible and scalable with least learning curve. More control over GPU resources. Optimization strategies can be applied. Faster and more memory efficient than MATLAB implementation.
- 51. Conclusions Downsides of GPU Significant testing/verification effort might be required if GPU hardware have to be upgraded (due to old one becoming obsolete). Proprietary nature of CUDA might be problematic in case company discontinues CUDA or its support.
- 52. Future work CUDA kernels can be called in MATLAB code using MATLAB’s CUDA kernel calling support. MATLAB GPU implementation can be improved as newer and better functions become available. C/C++ based CPU implementation can be developed to better judge MATLAB-CPU/CUDA performance. Other SAR processing algorithms can be implemented using framework laid out in this project.
- 53. Q & A
- 54. Thank You