Background: PLINK 1 is a widely used open-source C/C++ toolset for genome-wide association studies (GWAS) and research in population genetics. However, the steady accumulation of data from imputation and whole-genome sequencing studies has exposed a strong need for faster and scalable implementations of key functions, such as logistic regression, linkage disequilibrium estimation, and genomic distance evaluation. In addition, GWAS and population-genetic data now frequently contain genotype likelihoods, phase information, and/or multiallelic variants, none of which can be represented by PLINK 1’s primary data format. Findings: To address these issues, we are developing a second-generation codebase for PLINK. The first major release from this codebase, PLINK 1.9, introduces extensive use of bit-level parallelism, O √ n -time/constant-space Hardy-Weinberg equilibrium and Fisher’s exact tests, and many other algorithmic improvements. In combination, these changes accelerate most operations by 1-4 orders of magnitude, and allow the program to handle datasets too large to fit in RAM. We have also developed an extension to the data format which adds low-overhead support for genotype likelihoods, phase, multiallelic variants, and reference vs. alternate alleles, which is the basis of our planned second release (PLINK 2.0). Conclusions: The second-generation versions of PLINK will offer dramatic improvements in performance and compatibility. For the first time, users without access to high-end computing resources can perform several essential analyses of the feature-rich and very large genetic datasets coming into use.

/pdf/second-generation-plink-rising-to-the-challenge-of-larger-2i1coy4h8p.pdf

Second-generation PLINK: rising to the challenge of larger and richer datasets

This paper presents and characterizes the Princeton Application Repository for Shared-Memory Computers (PARSEC), a benchmark suite for studies of Chip-Multiprocessors (CMPs). Previous available benchmarks for multiprocessors have focused on high-performance computing applications and used a limited number of synchronization methods. PARSEC includes emerging applications in recognition, mining and synthesis (RMS) as well as systems applications which mimic large-scale multithreaded commercial programs. Our characterization shows that the benchmark suite covers a wide spectrum of working sets, locality, data sharing, synchronization and off-chip traffic. The benchmark suite has been made available to the public.

/pdf/the-parsec-benchmark-suite-characterization-and-jiajft2rem.pdf

The PARSEC benchmark suite: characterization and architectural implications

PLINK 1 is a widely used open-source C/C++ toolset for genome-wide association studies (GWAS) and research in population genetics. However, the steady accumulation of data from imputation and whole-genome sequencing studies has exposed a strong need for even faster and more scalable implementations of key functions. In addition, GWAS and population-genetic data now frequently contain probabilistic calls, phase information, and/or multiallelic variants, none of which can be represented by PLINK 1's primary data format. 
To address these issues, we are developing a second-generation codebase for PLINK. The first major release from this codebase, PLINK 1.9, introduces extensive use of bit-level parallelism, O(sqrt(n))-time/constant-space Hardy-Weinberg equilibrium and Fisher's exact tests, and many other algorithmic improvements. In combination, these changes accelerate most operations by 1-4 orders of magnitude, and allow the program to handle datasets too large to fit in RAM. This will be followed by PLINK 2.0, which will introduce (a) a new data format capable of efficiently representing probabilities, phase, and multiallelic variants, and (b) extensions of many functions to account for the new types of information. 
The second-generation versions of PLINK will offer dramatic improvements in performance and compatibility. For the first time, users without access to high-end computing resources can perform several essential analyses of the feature-rich and very large genetic datasets coming into use.

We explore the effect of dimensionality on the nearest neighbor problem. We show that under a broad set of conditions (much broader than independent and identically distributed dimensions), as dimensionality increases, the distance to the nearest data point approaches the distance to the farthest data point. To provide a practical perspective, we present empirical results on both real and synthetic data sets that demonstrate that this effect can occur for as few as 10-15 dimensions. These results should not be interpreted to mean that high-dimensional indexing is never meaningful; we illustrate this point by identifying some high-dimensional workloads for which this effect does not occur. However, our results do emphasize that the methodology used almost universally in the database literature to evaluate high-dimensional indexing techniques is flawed, and should be modified. In particular, most such techniques proposed in the literature are not evaluated versus simple linear scan, and are evaluated over workloads for which nearest neighbor is not meaningful. Often, even the reported experiments, when analyzed carefully, show that linear scan would outperform the techniques being proposed on the workloads studied in high (10-15) dimensionality!.

When is nearest neighbor meaningful

Similarity search finds application in specialized database systems handling complex data such as images or videos, which are typically represented by high-dimensional features and require specific indexing structures. This paper tackles the problem of better utilizing GPUs for this task. While GPUs excel at data-parallel tasks, prior approaches are bottlenecked by algorithms that expose less parallelism, such as k-min selection, or make poor use of the memory hierarchy. 
We propose a design for k-selection that operates at up to 55% of theoretical peak performance, enabling a nearest neighbor implementation that is 8.5x faster than prior GPU state of the art. We apply it in different similarity search scenarios, by proposing optimized design for brute-force, approximate and compressed-domain search based on product quantization. In all these setups, we outperform the state of the art by large margins. Our implementation enables the construction of a high accuracy k-NN graph on 95 million images from the Yfcc100M dataset in 35 minutes, and of a graph connecting 1 billion vectors in less than 12 hours on 4 Maxwell Titan X GPUs. We have open-sourced our approach for the sake of comparison and reproducibility.

Billion-scale similarity search with GPUs

Techniques for generated and presenting images of items within user selected context images are presented herein. In an example embodiment, an access module can be configured to receive a first environment image. A simulation module coupled to the access module may process the environment image to identify placement areas within the image, and an imaging module may merge an item image with the environment image and filter the merged image in an erosion area. In various embodiments, the items and environments may be selected by a user and presented to a user in real-time or near-real time as part of an online shopping experience. In further embodiments, the environments may be processed from images taken by a device of the user.

Systems and methods for generating virtual contexts

Techniques for three-dimensional garment simulation using parallel computing are presented herein. An access module can be configured to access a three-dimensional garment model of a garment. The garment model can include garment points that represent a surface of the garment. A processor, having a plurality of cores, can be configured by a garment simulation module to calculate one or more exerted forces on a subset of garment points. Additionally, the garment simulation module can generate cross pairs and apportion the generated cross pairs among the plurality of cores. Moreover, the garment simulation module can determine, using the plurality of vector execution units in parallel based on an organized data layout, whether boundaries of the first subgroup of cross pairs are overlapping based on the one or more exerted forces. Subsequently, the garment simulation module can calculate one or more simulated forces acting on the garment points based on the determination.

Garment simulation using thread and data level parallelism

Techniques for generated and presenting images of items within user selected context images are presented herein. In an example embodiment, an access module can be configured to receive a first environment model and a first wearable item model. A simulation module coupled to the access module may process the environment model to identify placement volumes within the environment model and to place a clothed body model within the placement volume to generate a context model. A rendering module may then generate a context image from the context model. In various embodiments, the environment model used for the context, the wearable item positioned within the environment model, and rendering values used to generate context images may be changed in response to user inputs to generate new context images that are displayed to a user.

Systems and methods for generating virtual contexts from three dimensional models

A texture unit may be used to perform general purpose mathematical computations such as dot products. This enables some general purpose computations and operations to be offloaded from a central processing unit to the texture unit. The texture unit may use linear interpolators in order to perform the dot product calculations.

Texture Unit for General Purpose Computing

Embodiments of techniques and systems for parallel processing of B+ trees are described. A parallel B+ tree processing module with partitioning and redistribution may include a set of threads executing a batch of B+ tree operations on a B+ tree in parallel. The batch of operations may be partitioned amongst the threads. Next, a search may be performed to determine which leaf nodes in the B+ tree are to be affected by which operations. Then, the threads may redistribute operations between each other such that multiple threads will not operate on the same leaf node. The threads may then perform B+ tree operations on the leaf nodes of the B+ tree in parallel. Subsequent modifications to nodes in the B+ may similarly be redistributed and performed in parallel as the threads work up the tree.

Jatin Chhugani

Papers

Systems and methods for generating virtual contexts

Garment simulation using thread and data level parallelism

Systems and methods for generating virtual contexts from three dimensional models

Texture Unit for General Purpose Computing

Parallel operation on b+ trees