Fahad Zafar

Cross-Platform OpenCL Code and Performance Portability Investigated with a Climate and Weather Physics Model

Current generation of multicore computing platforms are vastly different. Sustenance of many core applications across heterogenous platforms is a daunting task, more so when dynamic nature of the application is factored in. Open Computing Language (OpenCL) was created to address this issue. Designed to run on CPUs, GPUs, FPGAs and other platforms. OpenCL is becoming a standard for cross-platform parallel programming. While current implementations of OpenCL compiler provide the capability to compile and run on the platforms mentioned above, most of the current literatures investigate the OpenCL performance on GPUs. In a previous work, Fahad et al [1] reported how low level implicit auto vectorization capability of OpenCL allows remarkable performance optimization on CPUs. In this paper we present our investigation results on OpenCL portability across CPU and GPU platforms in terms of code and performance via a representative climate and weather physics model, NASAs GEOS-5 solar radiation model (SOLAR). A single OpenCL implementation portable between CPUs and GPUs has been obtained. Through algorithm refactoring, OpenCLs vector-oriented programming paradigm and implicit vectorization led to significant performance gains.

Accelerating a Climate Physics Model with OpenCL

Open Computing Language (OpenCL) is fast becoming the standard for heterogeneous parallel computing. It is designed to run on CPUs, GPUs, and other accelerator architectures. By implementing a real world application, a solar radiation model component widely used in climate and weather models, we show that OpenCL multi-threaded programming and execution model can dramatically increase performance even on CPU architectures. Our preliminary investigation indicates that low-level vector instructions and code representations in OpenCL contribute to dramatic performance improvement over the serial version when compared with the execution of the serial code compiled across various compilers on multiple platforms with auto vectorization flags. However, the portability of OpenCL implementations needs to improve, even for CPU architectures.

Computational observers and visualization methods for stereoscopic medical imaging

As stereoscopic display devices become common, their image quality assessment evaluation becomes increasingly important. Most studies conducted on 3D displays are based on psychophysics experiments with humans rating their experience based on detection tasks. The physical measurements do not map to effects on signal detection performance. Additionally, human observer study results are often subjective and difficult to generalize. We designed a computational stereoscopic observer approach inspired by the mechanisms of stereopsis in human vision for task-based image assessment that makes binary decisions based on a set of image pairs. The stereo-observer is constrained to a left and a right image generated using a visualization operator to render voxel datasets. We analyze white noise and lumpy backgrounds using volume rendering techniques. Our simulation framework generalizes many different types of model observers including existing 2D and 3D observers as well as providing flexibility to formulate a stereo model observer approach following the principles of stereoscopic viewing. This methodology has the potential to replace human observer studies when exploring issues with stereo display devices to be used in medical imaging. We show results quantifying the changes in performance when varying stereo angle as measured by an ideal linear stereoscopic observer. Our findings indicate that there is an increase in performance of about 13-18% for white noise and 20-46% for lumpy backgrounds, where the stereo angle is varied from 0 to 30. The applicability of this observer extends to stereoscopic displays used for in the areas of medical and entertainment imaging applications.

Tiny encryption algorithm for parallel random numbers on the GPU

Random numbers have many uses in computer graphics, from Monte Carlo sampling for realistic image synthesis to noise generation for artistic shader construction. Perlin [1985] introduced the idea of using a repeatable band-limited noise function to add stochastic variation to procedural shaders. We show that the quality of random number generation directly affects the quality of the noise produced, however, good quality noise can still be produced with a lower quality random number generator. Further, we show that the Tiny Encryption Algorithm (TEA) [Reddy 2003] can serve as the basis of a fast and high quality random number generator, and by changing the number of encryption rounds we can trade speed for quality.

Characterization of crosstalk in stereoscopic display devices

Abstract Many different types of stereoscopic display devices are used for commercial and research applications. Stereoscopic displays offer the potential to improve performance in detection tasks for medical imaging diagnostic systems. Due to the variety of stereoscopic display technologies, it remains unclear how these compare with each other for detection and estimation tasks. Different stereo devices have different performance trade‐offs due to their display characteristics. Among them, crosstalk is known to affect observer perception of 3D content and might affect detection performance. We measured and report the detailed luminance output and crosstalk characteristics for three different types of stereoscopic display devices. We recorded the effect of other issues on recorded luminance profiles such as viewing angle, use of different eye wear, and screen location. Our results show that the crosstalk signature for viewing 3D content can vary considerably when using different types of 3D glasses for active stereo displays. We also show that significant differences are present in crosstalk signatures when varying the viewing angle from 0 degrees to 20 degrees for a stereo mirror 3D display device. Our detailed characterization can help emulate the effect of crosstalk in conducting computational observer image quality assessment evaluations that minimize costly and time‐consuming human reader studies.

GOTHIC: glare optimizer tool for high-dynamic-range images and content with implementation in video

Introduction: The luminance range of the sun to the night sky is approximately 14 orders of magnitude. Current display technology can present approximately three orders of magnitude, however this number is increasing as High-Dynamic-Range (HDR) technology develops to further emulate reality [Seetzen et al. 2004]. Another benefit to HDR technology is the increased bit-depth enabling the display of more information. However, a major limitation in the perception of added bit-depth is veiling glare. The increased luminance range in HDR displays have the ability to produce glare sources that can reduce the visible contrast in neighboring dark areas. This effect is especially undesirable in the visualization of scientific data and in medical images. The HDR presentation must be optimized so that the benefits of a wide luminance range are not diminished by glare in the human visual system. One important question is, what is the largest luminance range that avoids these veiling glare effects while presenting the most bit-depth? We have found that the answer is highly dependent on the spatial and luminance distribution in the image. Many models have been proposed to estimate the veiling glare in a given image. A well known model is High-Dynamic-Range Visual Difference Predictor 2 (HDR-VDP-2) [Mantiuk et al. 2011], a calibrated method able to determine the visibility of differences in HDR images. Building on a number of previous metrics of visible difference, this model operates in a broad range of viewing conditions, from scotopic to photopic vision. More importantly, HDR-VDP-2 can be used to represent the effects of visual glare in signal detection. The inputs of the HDR-VDP-2 are a luminance map of an image, a reference image, and an image with the target. The software outputs the probability of target detection accounting for various visual effects including veiling glare.

24.2: Estimating the Perceptual Limits of Mobile Displays

The quality of the displayed image in mobile devices is affected by the varying ambient illumination conditions. Determining appropriate viewing conditions for particular visual tasks requires appropriate instrumentation. However, visual tests can be more practical. We conducted experiments to explore the limitations of mobile displays in terms of the visibility of subtle targets. We explored the effect of background luminance and ambient lighting using two methodologies: a text detection task where the observer identifies characters embedded in noisy backgrounds and a threshold estimation staircase technique. We found that the text detection method holds promise as a surrogate for more complicated tests in the framework of a clinically practical implementation.

31.3: Experimental Methodology to Measure the Veiling Glare Limit for Detection Tasks in High-Dynamic-Range Displays

We describe an experimental methodology for quantifying the effect of veiling glare in high-dynamic-range displays for simple detection tasks using a sensitivity experiment. A Gaussian spot was located on white noise image backgrounds indicated with dark hairline markers and a ring was added to the image as the veiling glare source. A double random staircase technique with one-image-at-a-time paradigm was used to estimate intensity thresholds using published methods. Observer gaze position was recorded in real-time during the experiments and used to provide auditory feedback to ensure fixation on the region where the signal might be present and minimize significant changes in adaptation that would affect the thresholds.