LOGAN: High-Performance GPU-Based X-Drop Long-Read Alignment

Zeni, Alberto; Guidi, Giulia; Ellis, Marquita; Ding, Nan; Santambrogio, Marco D.; Hofmeyr, Steven; Buluç, Aydın; Oliker, Leonid; Yelick, Katherine

doi:10.1109/ipdps47924.2020.00055

Cited by 38 publications

(23 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The cost calculation takes a long time if we increase the number of nodes, which means the size of the network has become larger. For this reason, we propose a hybridization between GA and parallel architectures to end the search process quickly, especially for large map network, our results showed that the use of the parallel method, is powerful in this case of our problem [18].…”

Section: Problem Studied "Ad Hoc Network Routing Optimization Problem"mentioning

confidence: 97%

Parallel genetic approach for routing optimization in large ad hoc networks

Khankhour

Abdoun

Abouchabaka

2022

IJECE

View full text Add to dashboard Cite

<span>This article presents a new approach of integrating parallelism into the genetic algorithm (GA), to solve the problem of routing in a large ad hoc network, the goal is to find the shortest path routing. Firstly, we fix the source and destination, and we use the variable-length chromosomes (routes) and their genes (nodes), in our work we have answered the following question: what is the better solution to find the shortest path: the sequential or parallel method?. All modern systems support simultaneous processes and threads, processes are instances of programs that generally run independently, for example, if you start a program, the operating system spawns a new process that runs parallel elements to other programs, within these processes, we can use threads to execute code simultaneously. Therefore, we can make the most of the available central processing unit (CPU) cores. Furthermore, the obtained results showed that our algorithm gives a much better quality of solutions. Thereafter, we propose an example of a network with 40 nodes, to study the difference between the sequential and parallel methods, then we increased the number of sensors to 100 nodes, to solve the problem of the shortest path in a large ad hoc network.</span>

show abstract

Section: Problem Studied "Ad Hoc Network Routing Optimization Problem"mentioning

confidence: 97%

Parallel genetic approach for routing optimization in large ad hoc networks

Khankhour

Abdoun

Abouchabaka

2022

IJECE

View full text Add to dashboard Cite

show abstract

“…There has only been a few prior works [23,22,25,26] which try to improve the performance of (accelerate) mm2. Zeni et al [25] and Feng et al [26] accelerate the base-level alignment step which is no longer the dominant bottleneck as reads have grown longer in length. Guo et al [22] and Kalikar et al [23] remove the MAX SKIP heuristic for speed in mm2 in order to extract intra-range parallelism and parallelizes chain score generation for each anchor (MAX SKIP is set to INF).…”

Section: Prior Workmentioning

confidence: 99%

Accelerating Minimap2 for accurate long read alignment on GPUs

Sadasivan

Maric

Dawson

et al. 2022

Preprint

View full text Add to dashboard Cite

BACKGROUND: Long read sequencing technology is becoming increasingly popular for Precision Medicine applications like variant calling from Whole Genome Sequencing (WGS) and for metagenomics applications like microbial abundance estimation. Minimap2 is the state-of-the-art aligner and mapper used by the leading long read sequencing technologies, today. However, Minimap2 is very slow for long noisy reads. 60-70% of the run-time on a CPU comes from the highly sequential chaining step in Minimap2. Most Point-of-Care computational workflows in long read sequencing use Graphics Processing Units (GPUs). We present minimap2-accelerated (mm2-ax), a heterogeneous design for sequence mapping and alignment where the compute intensive chaining step of minimap2 is sped up on the GPU and demonstrate its time and cost benefits. RESULTS: We extract better intra-read parallelism from chaining without loosing mapping accuracy by forward transforming Minimap2's chaining algorithm . Further, we utilize the high memory available on modern cloud instances for better performance on the GPU by converting a sparse vector which defines the chaining workload to a dense one in order to optimize for better arithmetic intensity (more operations per byte of data fetched from high-latency global memory) on the GPU. We also optimize for better workload balancing, data locality and minimal branch divergence on the GPU. We show mm2-ax on an NVIDIA A100 GPU improves the chaining step with 12.6-5X Speedup and 9.44-3.77X Speedup:Costup over the fastest version of Minimap2, mm2-fast, benchmarked on a single Google Cloud Platform instance of 30 SIMD cores (Intel Cascade Lake with AVX-512).

show abstract

“…A second thrust of ExaBiome involves protein clustering and annotation. ExaBiome's HipMCL [49] and PASTIS [50] code-the latter developed jointly with ExaGraph-provide a scalable protein clustering pipeline, whereas a new prototype deep learning framework [51] shows promising results for functional annotation. HipMCL runs on thousands of nodes and effectively uses GPUs.…”

Section: Exabiomementioning

confidence: 99%