SpChar: Characterizing the sparse puzzle via decision trees

The booming growth in AI and machine learning is drastically reshaping the landscape of high performance computing. Traditional HPC addresses scientific problems that are driven by simulation, modeling and analysis inscience domains. Algorithms like linear algebra and methods like differential equations are at the core of solutions to such problems. However, emerging machine learning tasks rest on a new set of algorithms and models. The computation and data movement patterns inherent in learning algorithms cannot be directly mapped to the computation motifs in the physics, chemistry, or biology simulation problems. As a result, the high performance libraries that originate in the traditional scientific domains thus cannot be straightforwardly applied to emerging ML tasks to deliver required performance. This thesis focuses on performance optimizations of computation kernelsin emerging machine learning applications, spanning across a diverse range from dense, regular to sparse and irregular kernels. In this work, we demonstrate how code specialization and generation together with expert-built performance models and learned dispatch strategies can together enable ML motifs to achieve better performance on modern processors. First, we investigate the performance optimization of dense kernels with a focus on the convolutional neural networks (CNN). The computation of convolution layers in deep neural networks typically relies on high performance matrix-multiplication routines to improve performance. However, these routines are not optimized for performing convolution. Extra memory overhead is incurred due to data transformation, and the performance obtained is also less than conventionally expected. We demonstrate that direct convolution, when implemented correctly, eliminates all memory overhead, and yields performancethat is between 10% to 4x better than existing high performance implementations on conventional and embedded CPU architectures. We present a model-guided optimization approach which utilizes the characteristics of system architectures to guide the optimization choices of loop ordering, blocking, and memory layout transformation. We show that a high performance direct convolution exhibits better performance scaling than expert-tuned matrix implementation, i.e. suffers less performance drop, when increasing the number of threads. Sparse kernel is an equally important computation kernel appearing in many machine learning applications such as graph analytics and genetic sequencing. One factor that prevents sparse kernels from achieving high performance on modern processors results from the prohibitively large number of different implementations and data structures for sparse problems. We start with the observation that the complicated sparse computations can be distilled into primitive set of operators such as join, merge, and difference. To acceleratethose operators on modern processors with data parallelism, we propose a vectorization and code specialization approach which can eliminate the control divergences of these operators. Next, we explore the design space for vectorization on CPUs with various vector width, based on which we present the code generation algorithm that takes the data width and operations as input and generates various implementations. We then demonstrate the acceleration of the General Sparse Matrix-Matrix Multiplication (SpGEMM) on GPUs. We show how the SpGEMM implementation can leverage join/merge operatorsto compose a variety of implementations. Another challenge when optimizing sparse kernels is that their performance behavior is data dependent, while the input characteristics may change online during iterative updates. To leveragethe different implementations offered by the code generator, we propose a low-overhead mechanism that collects the data characteristic information to learn online dispatch decisions over iterations. Overall, in this thesis, we demonstrate the interplay of code specialization and generation, together with performance modeling, learned dispatch, can enable high performance kernels for the emerging machine learning applications.

We examine the problem of evaluating performance of supercomputer architectures on sparse (matrix) computations and lay out the details of a benchmark package for this problem. Whereas there already exists a number of benchmark packages for scientific computations, such as the Livermore Loops, the Linpack benchmark and the Los Alamos benchmark, none of these deals with the specific nature of sparse computations. Sparse matrix techniques are characterized by the relatively small number of operations per data element and the irregularity of the computation. Both facts may significantly increase the overhead time due to memory traffic. For this reason, the performance evaluation of sparse computations should not only take into account the CPU performance but also the degradation of performance caused by high memory traffic. Furthermore, sparse matrix techniques comprise a variety of different types of basic computations. Taking these considerations into account we propose a benchmark package that consists of several independent modules, each of which has a distinct role.

AMG is one of the most efficient and widely used methods for solving sparse linear systems. The computational process of AMG mainly consists of a series of iterative calculations of generalized sparse matrix-matrix multiplication (SpGEMM) and sparse matrix-vector multiplication (SpMV). Optimizing these sparse matrix calculations is crucial for accelerating solving linear systems. In this paper, we first focus on optimizing the SpGEMM algorithm in AmgX, a popular AMG library for GPUs. We propose a new algorithm called SpGEMM-upper, which achieves an average speedup of 2.02× on Tesla V100 and 1.96× on RTX 3090 against the original algorithm. Next, through experimental investigation, we conclude that no single SpGEMM library or algorithm performs optimally for most sparse matrices, and the same holds true for SpMV. Therefore, we build machine learning-based models to predict the optimal SpGEMM and SpMV used in the AMG calculation process. Finally, we integrate the prediction models, SpGEMM-upper, and other selected algorithms into a framework for adaptive sparse matrix computation in AMG. Our experimental results prove that the framework achieves promising performance improvements on the test set.

In many scientific applications, the majority of the execution time is spent within a few basic sparse kernels such as sparse matrix vector multiplication (SMV). Such sparse kernels can utilize only a fraction of the available processing speed because of their relatively large number of data accesses per floating point operation, and limited data locality and data re-use. Algorithmic changes and tuning of codes through blocking and loop unrolling schemes can improve performance but such tuned versions are typically not available in benchmark suites such as the SPEC CFP 2000. In this paper, we consider sparse SMV kernels with different levels of tuning that are representative of this application space. We emulate certain memory subsystem optimizations using SimpleScalar and Wattch to evaluate improvements in performance and energy metrics. We also characterize how such an evaluation can be affected by the interplay between code tuning and memory subsystem optimizations. Our results indicate that the optimizations reduce execution time by over 40%, and the energy by over 85%, when used with power control modes of CPUs and caches. Furthermore, the relative impact of the same set of memory subsystem optimizations can vary significantly depending on the level of code tuning. Consequently, it may be appropriate to augment traditional benchmarks by tuned kernels typical of high performance sparse scientific codes to enable comprehensive evaluations of future systems.

Sparse matrix-vector semiring computation is a key operation in sparse matrix computations, with performance strongly dependent on both program design and the features of the sparse matrices. Given the diversity of sparse matrices, designing a tailored program for each matrix is challenging. To address this, we propose SRSparse 1 an program generator that creates tailored programs by automatically combining program designing methods to fit specific input matrices. It provides two components: the problem definition configuration , which declares the computation, and the scheduling language , which can be leveraged by an auto-tuner to specify the program designs. The two are lowered to the intermediate representations of SRSparse, the Format IR and Kernel IR , which respectively generates format conversion routine and kernel code. We evaluate SRSparse on four representative sparse kernels and three format conversion routines. For sparse kernels, SRSparse achieves median speedups over handwritten programs: COO (3.50 ×), CSR-Adaptive (5.36 ×), CSR5 (2.06 ×), ELL (1.63 ×), Gunrock (1.57 ×), and GraphBLAST (1.96 ×); over an auto-tuner: AlphaSparse (1.16 ×); and over a compiler: TACO (1.71 ×). For format conversion routines, SRSparse achieves median speedups over handwritten implementations: Intel MKL (7.60 ×), SPARSKIT (2.61 ×), CUSP (2.77 ×), and Ginkgo (1.74 ×); and over a compiler: TACO (4.04 ×).

Machine learning (ML) and high-performance computing (HPC) applications increasingly combine dense and sparse memory access computations to maximize storage efficiency. However, existing central processing units (CPUs) and graphics processing units (GPUs) struggle to flexibly handle these heterogeneous workloads with consistently high compute efficiency. We present Occamy, a 432-core, 768-DP-GFLOP/s, dual-HBM2E, dual-chiplet RISC-V system with a latency-tolerant hierarchical interconnect and in-core streaming units (SUs) designed to accelerate dense and sparse FP8-to-FP64 ML and HPC workloads. We implement Occamy’s compute chiplets in 12-nm FinFET and its passive interposer, Hedwig, in a 65-nm node. On dense linear algebra (LA), Occamy achieves a competitive floating-point unit (FPU) utilization of 89%. On stencil codes, Occamy reaches an FPU utilization of 83% and a technology-node-normalized compute density of 11.1 DP-GFLOP/s/mm2, leading state-of-the-art (SoA) processors by <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$1.7\times $ </tex-math></inline-formula> and <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$1.2\times $ </tex-math></inline-formula>, respectively. On sparse-dense LA, it achieves 42% FPU utilization and a normalized compute density of 5.95 DP-GFLOP/s/mm2, surpassing the SoA by <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$5.2\times $ </tex-math></inline-formula> and <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$11\times $ </tex-math></inline-formula>, respectively. On sparse–sparse LA, Occamy reaches a throughput of up to 187 GCOMP/s at 17.4 GCOMP/s/W and a compute density of 3.63 GCOMP/s/mm2. Finally, we reach up to 75% and 54% FPU utilization on and dense (large language model) and graph-sparse (graph convolutional network) ML inference workloads. Occamy’s register transfer level (RTL) description is freely available under a permissive open-source license.

This paper introduces three new compiler transformations for representing and transforming sparse matrix computations and their data representations. In cooperation with run-time inspection, our compiler derives transformed matrix representations and associated transformed code to implement a variety of representations targeting different architecture platforms. This systematic approach to combining code and data transformations on sparse computations, which extends a polyhedral transformation and code generation framework, permits the compiler to compose these transformations with other transformations to generate code that is on average within 5% and often exceeds manually-tuned, high-performance sparse matrix libraries CUSP and OSKI. Additionally, the compiler-generated inspector codes are on average 1.5 faster than OSKI and perform comparably to CUSP, respectively.

Recommendation systems have become an important field of research in computer science and physics. In recent years, breakthroughs have been achieved in social, biological, and research cooperation networks. With the popularization of big data and deep learning technology development, graph structures are increasingly being used to represent large-scale and complex data in the real world. In this paper, we reviewed the progress made in recommendation systems research in the past 20 years and comprehensively classified recommender systems based on the heterogeneous input features. We introduced layering in the classification of recommendation systems. Furthermore, we proposed a new hierarchical classification model of recommendation systems divided into three layers: feature input, feature learning, and output layers. In the feature learning layer, existing recommendation systems were divided into graph-based, text-based, behavior-based, spatiotemporal-based, and hybrid recommendation systems. Additionally, we provided evaluation index, open-source implementation, experimental comparison and the relative merits for each recommendation method. Subsequently, future development directions of recommendation systems are discussed.

At present recommender systems (RS) are incorporating contextual and social data of the client, delivering context aware RS where it take request response approach in which the recommendations are given to the client upon his solicitation. Later on they will utilize verifiable data of the client anywhere from the Internet of Things (IoT). Currently, a proactive RS that pushes proposals to the client when the present circumstance appears proper, without unequivocal client demand has been presented in the exploration research field of RS. In this paper, an outline of a context aware RS that prescribes recommendations under the IoT worldview is proposed. We also did the empirical analysis of the Machine Learning (ML) techniques like Genetic Algorithm optimized Artificial Neural Network (GA-ANN), decision tree, ANN, bagging, boosting which will perform the reasoning of the context. All inputs are virtually derived from the IoT and its output scores are calculated based on ML techniques which will decide if to push a recommendation to the user or not. Benchmark data of the Chicago restaurant is analyzed and it was observed that for the 98% of the contexts, GA-ANN produced correct recommendations with higher accuracy and efficiency in the correct times and context.

Sparse matrix-vector multiplication is often employed in many data-analytic workloads in which low latency and high throughput are more valuable than exact numerical convergence. FPGAs provide quick execution times while offering precise control over the accuracy of the results thanks to reduced-precision fixed-point arithmetic. In this work, we propose a novel streaming implementation of Coordinate Format (COO) sparse matrix-vector multiplication, and study its effectiveness when applied to the Personalized PageR-ank algorithm, a common building block of recommender systems in e-commerce websites and social networks. Our implementation achieves speedups up to 6x over a reference floating-point FPGA architecture and a state-of-the-art multi-threaded CPU implementation on 8 different data-sets, while preserving the numerical fidelity of the results and reaching up to 42x higher energy efficiency compared to the CPU implementation.

Recommendations are useful suggestions used by people from all walks of life. However, the usage of recommender systems plays a vital role in modern applications. They are found in different domains such as E-commerce. Concerning the health care industry, recommendations play a very crucial role. This industry has significance as it is linked to the lives of people and their well-being. Human health depends on the diet followed. Keeping this fact in mind, in this paper, we investigated healthy diet recommendations. The recommender systems that are existing in healthcare focused a little in this area. From the literature, it is understood that most of the frameworks on health recommendations are theoretical in nature. As food decides health, it is to be given paramount importance. In this paper, we proposed a hybrid mechanism based on Artificial Intelligence (AI) for big data analytics. Particularly we used Machine Learning (ML) for generating healthy diet recommendations. The proposed system is known as Hybrid Recommender System (HRS). It involves a hybrid approach with Natural Language Processing (NLP) and machine learning. An algorithm named Intelligent Recommender for Healthy Diet (IR-HD) is proposed to analyze data and provide healthy diet recommendations. IR-HD could generate recommendations on a healthy diet and outperform existing models. Python data science platform is used to implement the recommender system. The results of experiments showed that the system is capable of providing quality recommendations and it has performance improvement over the state of the art.

Introduction: In the current educational landscape, universities provide a vast array of diverse and obscure courses. However, the primary challenge students face is decision overload, as the abundance of options can make it difficult to choose courses that align with their interests, career aspirations, and strengths. The development and application of proper course suggestion systems might assist students in overcoming the challenges associated with choosing the right courses. However, these recommendation systems still require improvement to address explainability issues, security issues, and cold start. Our analysis reveals that there is limited research addressing how course recommendation systems can assist 12th-grade students in selecting courses for higher education. Therefore, this study presents a novel personalized recommendation system, namely "Interpretability-Driven Course Recommendation: A Random Forest Approach with Explainable AI", that chooses the top-3 courses for students based on their intermediate class grades/- marks. This research provides accurate and interpretable recommendations using the Random Forest algorithm with Explainable AI that ensures transparency by explaining why a particular course is recommended. Method: The system uses a Random Forest classifier with SHAP (SHapley Additive exPlanations) to forecast the top-3 courses with the greatest expected scores for appropriateness. The system uses SHAP (SHapley Additive exPlanations) values to integrate Explainable AI (XAI) to guarantee openness and trust. The proposed model solves the cold start problem and provides data security. Results: In terms of precision (0.90), recall (0.92), and F1-score (0.92), Random Forest fared better than any other classifier. By combining predictions from several decision trees, its ensemble learning technique increases stability, decreases overfitting, and improves generalization across a broad range of input data patterns. Discussions: A major drawback of conventional black-box machine learning models is their lack of transparency, which is addressed by the incorporation of SHAP into the course recommendation system. In this model, students get suggestions in addition to information on which topic scores had the biggest influence on their choice. Conclusion: This study presents a practical and interpretable course recommendation system designed for 12th-grade students transitioning into higher education. Through the use of Explainable AI-enhanced Random Forest, the model provides precise, transparent, and customized recommendations. It addresses the main drawbacks of conventional systems, such as their lack of explainability and their cold start problems. The model';s excellent precision, recall, and F1-score indicate its superior performance, which makes it a useful tool for student guidance and academic advice. For even more thorough recommendations, future research may consider combining long-term academic objectives, aptitude test results, and student interests.

E-commerce has been overgrown due to the ongoing advancements in Internet Technology and the notable enhancement of network fundamentals. Nowadays, the most popular type of electronic commerce is online shopping. These days, user profile-based online recommender systems are popular in engineering and research fields. Getting users' profiles right is essential for making recommendations. Many studies on extracting user profiles, including content-based recommendations, have been initiated recently. We have introduced an intelligent recommendation system for e-commerce, aiming to streamline online shopping by addressing information overload. This system leverages customer behaviour to tailor personalized suggestions and enhance user experiences. It combines a genetic algorithm and a decision tree model for e-commerce platforms. The goal is to smooth the shopping journey, ensuring users find what they seek amidst the vast online options. This approach prioritizes user preferences, reducing the risk of overlooking crucial details when navigating online stores. Through experiments and analysis, we assess its effectiveness compared to conventional recommendation systems. We tested the model using publicly available datasets, comparing its performance against existing recommendation systems. The suggested approach is targeted to decrease recommendation sparsity and increase recommendation accuracy.

This research presents an in-depth hardware compatibility evaluation of Meta's Llama 3.2 model on ten varied tasks—from long-context summarization and code generation to multi-turn dialogue and structured output generation—to discern performance trends and computational bottlenecks. Outputs show that hardware requirements differ widely by task type: long-context processing (1,915 input tokens) maximizes memory bandwidth (34% utilization), whereas code generation (659 tokens) and creative storytelling (2,321 tokens) maximize Central Processing Unit (CPU) parallelism (355% utilization on 4 cores). Multi-turn dialog shows growing latency (45 → 58 seconds) and token output increase (115→846 tokens) due to the compounding overhead of context retention, highlighting inefficiency in recurrent attention mechanisms. Repetitive work (1,000 "A"s) shows excellent token throughput (373 tokens/47 seconds), in contrast with mathematical calculations (square root of 123,456,789 × π), which overwhelm CPUs out of proportion (382% utilization) with trivial outputs, indicating Large Language Models (LLM) weakness in numeric accuracy. Ordered outputs (JSON) and translation work further reveal formatting-centric CPU overhead (355–382%), with tax policy questions (1,110 input tokens) uncovering context-parsing latency (53 seconds to process 148 tokens), highlighting semantic extraction inefficiencies from high-density texts. These results develop LLM work in three fronts: (1) Hardware Optimization, promoting task-specific settings (code/math with multi-core CPUs, plenty of RAM for context-intensive processes); (2) Model Architecture, calling for enhanced context handling and arithmetic blocks; and (3) Deployment Strategy, emphasizing resource allocation in keeping with use case operationalization (e.g., memory over legal/text analysis versus clock speed over creative applications). By correlating task types with hardware profiles, this research offers a framework for LLM scalability optimization, inference cost reduction, and informing future research into energy-efficient designs. The research highlights the need for interdisciplinary collaboration between Artificial Intelligence (AI) researchers and systems engineers to close the gap between algorithmic innovation and hardware capabilities, enabling sustainable large-scale LLM adoption.

With the promotion of social information construction and the rapid update and replacement of large capacity storage equipment, the amount of data from every field grows exponentially. Reportedly, the amount of the data accumulated by Shandong Hi-speed Group is very large. These data can satisfy us some daily usefulness, such as query, retrieval, statistics, statements etc. But what is more important is that how can we discover some useful information from the information ocean. This information can be used in real life such as auxiliary decision. This paper is proposed in this historical background. Data mining is a powerful tool for acquiring knowledge from massive data. Some methods of data mining, such as decision tree, support vector machine, Bayesian decision theory, artificial neural network, k-nearest neighbor, association rule mining etc, are commonly used. In this paper, we design a recommendation system of highway ETC card by using the theory of decision tree. The recommendation system can predict whether a car owner is a potential ETC customer or not through the analysis of the vehicle information. Experiments proved that the accuracy rate of the recommendation system is larger than 90%, so it can provide effective information for the extension of China's ETC card.