اشترك بالحزمة الذهبية واحصل على وصول غير محدود شمرا أكاديميا
تسجيل مستخدم جديدRevisiting Huffman Coding: Toward Extreme Performance on Modern GPU Architectures
141
0
0.0
(
0
)
اسأل ChatGPT حول البحث
ﻻ يوجد ملخص باللغة العربية
Todays high-performance computing (HPC) applications are producing vast volumes of data, which are challenging to store and transfer efficiently during the execution, such that data compression is becoming a critical technique to mitigate the storage burden and data movement cost. Huffman coding is arguably the most efficient Entropy coding algorithm in information theory, such that it could be found as a fundamental step in many modern compression algorithms such as DEFLATE. On the other hand, todays HPC applications are more and more relying on the accelerators such as GPU on supercomputers, while Huffman encoding suffers from low throughput on GPUs, resulting in a significant bottleneck in the entire data processing. In this paper, we propose and implement an efficient Huffman encoding approach based on modern GPU architectures, which addresses two key challenges: (1) how to parallelize the entire Huffman encoding algorithm, including codebook construction, and (2) how to fully utilize the high memory-bandwidth feature of modern GPU architectures. The detailed contribution is four-fold. (1) We develop an efficient parallel codebook construction on GPUs that scales effectively with the number of input symbols. (2) We propose a novel reduction based encoding scheme that can efficiently merge the codewords on GPUs. (3) We optimize the overall GPU performance by leveraging the state-of-the-art CUDA APIs such as Cooperative Groups. (4) We evaluate our Huffman encoder thoroughly using six real-world application datasets on two advanced GPUs and compare with our implemented multi-threaded Huffman encoder. Experiments show that our solution can improve the encoding throughput by up to 5.0X and 6.8X on NVIDIA RTX 5000 and V100, respectively, over the state-of-the-art GPU Huffman encoder, and by up to 3.3X over the multi-thread encoder on two 28-core Xeon Platinum 8280 CPUs.
قيم البحث
اقرأ أيضاً
The trend towards highly parallel multi-processing is ubiquitous in all modern computer architectures, ranging from handheld devices to large-scale HPC systems; yet many applications are struggling to fully utilise the multiple levels of parallelism
exposed in modern high-performance platforms. In order to realise the full potential of recent hardware advances, a mixed-mode between shared-memory programming techniques and inter-node message passing can be adopted which provides high-levels of parallelism with minimal overheads. For scientific applications this entails that not only the simulation code itself, but the whole software stack needs to evolve. In this paper, we evaluate the mixed-mode performance of PETSc, a widely used scientific library for the scalable solution of partial differential equations. We describe the addition of OpenMP threaded functionality to the library, focusing on sparse matrix-vector multiplication. We highlight key challenges in achieving good parallel performance, such as explicit communication overlap using task-based parallelism, and show how to further improve performance by explicitly load balancing threads within MPI processes. Using a set of matrices extracted from Fluidity, a CFD application code which uses the library as its linear solver engine, we then benchmark the parallel performance of mixed-mode PETSc across multiple nodes on several modern HPC architectures. We evaluate the parallel scalability on Uniform Memory Access (UMA) systems, such as the Fujitsu PRIMEHPC FX10 and IBM BlueGene/Q, as well as a Non-Uniform Memory Access (NUMA) Cray XE6 platform. A detailed comparison is performed which highlights the characteristics of each particular architecture, before demonstrating efficient strong scalability of sparse matrix-vector multiplication with significant speedups over the pure-MPI mode.
Convolution layers are prevalent in many classes of deep neural networks, including Convolutional Neural Networks (CNNs) which provide state-of-the-art results for tasks like image recognition, neural machine translation and speech recognition. The c
omputationally expensive nature of a convolution operation has led to the proliferation of implementations including matrix-matrix multiplication formulation, and direct convolution primarily targeting GPUs. In this paper, we introduce direct convolution kernels for x86 architectures, in particular for Xeon and XeonPhi systems, which are implemented via a dynamic compilation approach. Our JIT-based implementation shows close to theoretical peak performance, depending on the setting and the CPU architecture at hand. We additionally demonstrate how these JIT-optimized kernels can be integrated into a lightweight multi-node graph execution model. This illustrates that single- and multi-node runs yield high efficiencies and high image-throughputs when executing state-of-the-art image recognition tasks on CPUs.
Efficient optimal prefix coding has long been accomplished via the Huffman algorithm. However, there is still room for improvement and exploration regarding variants of the Huffman problem. Length-limited Huffman coding, useful for many practical app
lications, is one such variant, in which codes are restricted to the set of codes in which none of the $n$ codewords is longer than a given length, $l_{max}$. Binary length-limited coding can be done in $O(n l_{max})$ time and O(n) space via the widely used Package-Merge algorithm. In this paper the Package-Merge approach is generalized without increasing complexity in order to introduce a minimum codeword length, $l_{min}$, to allow for objective functions other than the minimization of expected codeword length, and to be applicable to both binary and nonbinary codes; nonbinary codes were previously addressed using a slower dynamic programming approach. These extensions have various applications -- including faster decompression -- and can be used to solve the problem of finding an optimal code with limited fringe, that is, finding the best code among codes with a maximum difference between the longest and shortest codewords. The previously proposed method for solving this problem was nonpolynomial time, whereas solving this using the novel algorithm requires only $O(n (l_{max}- l_{min})^2)$ time and O(n) space.
Designing efficient and scalable sparse linear algebra kernels on modern multi-GPU based HPC systems is a daunting task due to significant irregular memory references and workload imbalance across the GPUs. This is particularly the case for Sparse Tr
iangular Solver (SpTRSV) which introduces additional two-dimensional computation dependencies among subsequent computation steps. Dependency information is exchanged and shared among GPUs, thus warrant for efficient memory allocation, data partitioning, and workload distribution as well as fine-grained communication and synchronization support. In this work, we demonstrate that directly adopting unified memory can adversely affect the performance of SpTRSV on multi-GPU architectures, despite linking via fast interconnect like NVLinks and NVSwitches. Alternatively, we employ the latest NVSHMEM technology based on Partitioned Global Address Space programming model to enable efficient fine-grained communication and drastic synchronization overhead reduction. Furthermore, to handle workload imbalance, we propose a malleable task-pool execution model which can further enhance the utilization of GPUs. By applying these techniques, our experiments on the NVIDIA multi-GPU supernode V100-DGX-1 and DGX-2 systems demonstrate that our design can achieve on average 3.53x (up to 9.86x) speedup on a DGX-1 system and 3.66x (up to 9.64x) speedup on a DGX-2 system with 4-GPUs over the Unified-Memory design. The comprehensive sensitivity and scalability studies also show that the proposed zero-copy SpTRSV is able to fully utilize the computing and communication resources of the multi-GPU system.
This paper investigates the multi-GPU performance of a 3D buoyancy driven cavity solver using MPI and OpenACC directives on different platforms. The paper shows that decomposing the total problem in different dimensions affects the strong scaling per
formance significantly for the GPU. Without proper performance optimizations, it is shown that 1D domain decomposition scales poorly on multiple GPUs due to the noncontiguous memory access. The performance using whatever decompositions can be benefited from a series of performance optimizations in the paper. Since the buoyancy driven cavity code is latency-bounded on the clusters examined, a series of optimizations both agnostic and tailored to the platforms are designed to reduce the latency cost and improve memory throughput between hosts and devices efficiently. First, the parallel message packing/unpacking strategy developed for noncontiguous data movement between hosts and devices improves the overall performance by about a factor of 2. Second, transferring different data based on the stencil sizes for different variables further reduces the communication overhead. These two optimizations are general enough to be beneficial to stencil computations having ghost changes on all of the clusters tested. Third, GPUDirect is used to improve the communication on clusters which have the hardware and software support for direct communication between GPUs without staging CPUs memory. Finally, overlapping the communication and computations is shown to be not efficient on multi-GPUs if only using MPI or MPI+OpenACC. Although we believe our implementation has revealed enough overlap, the actual running does not utilize the overlap well due to a lack of asynchronous progression.
سجل دخول لتتمكن من نشر تعليقات
التعليقات
جاري جلب التعليقات
سجل دخول لتتمكن من متابعة معايير البحث التي قمت باختيارها
