John Keiser, Daniel Lemire
The majority of text is stored in UTF-8, which must be validated on ingestion. We present the lookup algorithm, which outperforms UTF-8 validation routines used in many libraries and languages by more than 10 times using commonly available SIMD instructions. To ensure reproducibility, our work is freely available as open source software.
Wojciech MuÅa, Daniel Lemire
Web developers use base64 formats to include images, fonts, sounds and other resources directly inside HTML, JavaScript, JSON and XML files. We estimate that billions of base64 messages are decoded every day. We are motivated to improve the efficiency of base64 encoding and decoding. Compared to state-of-the-art implementations, we multiply the speeds of both the encoding (~10x) and the decoding (~7x). We achieve these good results by using the single-instruction-multiple-data (SIMD) instructions available on recent Intel processors (AVX2). Our accelerated software abides by the specification and reports errors when encountering characters outside of the base64 set. It is available online as free software under a liberal license.
Jaël Champagne Gareau, Daniel Lemire
Converting binary integers to variable-length decimal strings is a fundamental operation in computing. Conventional fast approaches rely on recursive division and small lookup tables. We propose a SIMD-based algorithm that leverages integer multiply-add instructions available on recent AMD and Intel processors. Our method eliminates lookup tables entirely and computes multiple quotients and remainders in parallel. Additionally, we introduce a dual-variant design with dynamic selection that adapts to input characteristics: a branch-heavy variant optimized for homogeneous digit-length distributions and a branch-light variant for heterogeneous datasets. Our single-core algorithm consistently outperforms all competing methods across the full range of integer sizes, running 1.4-2x faster than the closest competitor and 2-4x faster than the C++ standard library function std::to_chars across tested workloads.
Daniel Lemire, Nathan Kurz, Christoph Rupp
Arrays of integers are often compressed in search engines. Though there are many ways to compress integers, we are interested in the popular byte-oriented integer compression techniques (e.g., VByte or Google's Varint-GB). They are appealing due to their simplicity and engineering convenience. Amazon's varint-G8IU is one of the fastest byte-oriented compression technique published so far. It makes judicious use of the powerful single-instruction-multiple-data (SIMD) instructions available in commodity processors. To surpass varint-G8IU, we present Stream VByte, a novel byte-oriented compression technique that separates the control stream from the encoded data. Like varint-G8IU, Stream VByte is well suited for SIMD instructions. We show that Stream VByte decoding can be up to twice as fast as varint-G8IU decoding over real data sets. In this sense, Stream VByte establishes new speed records for byte-oriented integer compression, at times exceeding the speed of the memcpy function. On a 3.4GHz Haswell processor, it decodes more than 4 billion differentially-coded integers per second from RAM to L1 cache.
Dmytro Ivanchykhin, Sergey Ignatchenko, Daniel Lemire
Random hashing can provide guarantees regarding the performance of data structures such as hash tables---even in an adversarial setting. Many existing families of hash functions are universal: given two data objects, the probability that they have the same hash value is low given that we pick hash functions at random. However, universality fails to ensure that all hash functions are well behaved. We further require regularity: when picking data objects at random they should have a low probability of having the same hash value, for any fixed hash function. We present the efficient implementation of a family of non-cryptographic hash functions (PM+) offering good running times, good memory usage as well as distinguishing theoretical guarantees: almost universality and component-wise regularity. On a variety of platforms, our implementations are comparable to the state of the art in performance. On recent Intel processors, PM+ achieves a speed of 4.7 bytes per cycle for 32-bit outputs and 3.3 bytes per cycle for 64-bit outputs. We review vectorization through SIMD instructions (e.g., AVX2) and optimizations for superscalar execution.
Jeff Plaisance, Nathan Kurz, Daniel Lemire
We consider the ubiquitous technique of VByte compression, which represents each integer as a variable length sequence of bytes. The low 7 bits of each byte encode a portion of the integer, and the high bit of each byte is reserved as a continuation flag. This flag is set to 1 for all bytes except the last, and the decoding of each integer is complete when a byte with a high bit of 0 is encountered. VByte decoding can be a performance bottleneck especially when the unpredictable lengths of the encoded integers cause frequent branch mispredictions. Previous attempts to accelerate VByte decoding using SIMD vector instructions have been disappointing, prodding search engines such as Google to use more complicated but faster-to-decode formats for performance-critical code. Our decoder (Masked VByte) is 2 to 4 times faster than a conventional scalar VByte decoder, making the format once again competitive with regard to speed.
Wayne Xin Zhao, Xudong Zhang, Daniel Lemire et al.
Compression algorithms are important for data oriented tasks, especially in the era of Big Data. Modern processors equipped with powerful SIMD instruction sets, provide us an opportunity for achieving better compression performance. Previous research has shown that SIMD-based optimizations can multiply decoding speeds. Following these pioneering studies, we propose a general approach to accelerate compression algorithms. By instantiating the approach, we have developed several novel integer compression algorithms, called Group-Simple, Group-Scheme, Group-AFOR, and Group-PFD, and implemented their corresponding vectorized versions. We evaluate the proposed algorithms on two public TREC datasets, a Wikipedia dataset and a Twitter dataset. With competitive compression ratios and encoding speeds, our SIMD-based algorithms outperform state-of-the-art non-vectorized algorithms with respect to decoding speeds.
Xiaodan Zhu, Peter Turney, Daniel Lemire et al.
The importance of a research article is routinely measured by counting how many times it has been cited. However, treating all citations with equal weight ignores the wide variety of functions that citations perform. We want to automatically identify the subset of references in a bibliography that have a central academic influence on the citing paper. For this purpose, we examine the effectiveness of a variety of features for determining the academic influence of a citation. By asking authors to identify the key references in their own work, we created a data set in which citations were labeled according to their academic influence. Using automatic feature selection with supervised machine learning, we found a model for predicting academic influence that achieves good performance on this data set using only four features. The best features, among those we evaluated, were those based on the number of times a reference is mentioned in the body of a citing paper. The performance of these features inspired us to design an influence-primed h-index (the hip-index). Unlike the conventional h-index, it weights citations by how many times a reference is mentioned. According to our experiments, the hip-index is a better indicator of researcher performance than the conventional h-index.
Daniel Lemire, Leonid Boytsov, Nathan Kurz
Sorted lists of integers are commonly used in inverted indexes and database systems. They are often compressed in memory. We can use the SIMD instructions available in common processors to boost the speed of integer compression schemes. Our S4-BP128-D4 scheme uses as little as 0.7 CPU cycles per decoded integer while still providing state-of-the-art compression. However, if the subsequent processing of the integers is slow, the effort spent on optimizing decoding speed can be wasted. To show that it does not have to be so, we (1) vectorize and optimize the intersection of posting lists; (2) introduce the SIMD Galloping algorithm. We exploit the fact that one SIMD instruction can compare 4 pairs of integers at once. We experiment with two TREC text collections, GOV2 and ClueWeb09 (Category B), using logs from the TREC million-query track. We show that using only the SIMD instructions ubiquitous in all modern CPUs, our techniques for conjunctive queries can double the speed of a state-of-the-art approach.
Daniel Lemire, Leonid Boytsov
In many important applications -- such as search engines and relational database systems -- data is stored in the form of arrays of integers. Encoding and, most importantly, decoding of these arrays consumes considerable CPU time. Therefore, substantial effort has been made to reduce costs associated with compression and decompression. In particular, researchers have exploited the superscalar nature of modern processors and SIMD instructions. Nevertheless, we introduce a novel vectorized scheme called SIMD-BP128 that improves over previously proposed vectorized approaches. It is nearly twice as fast as the previously fastest schemes on desktop processors (varint-G8IU and PFOR). At the same time, SIMD-BP128 saves up to 2 bits per integer. For even better compression, we propose another new vectorized scheme (SIMD-FastPFOR) that has a compression ratio within 10% of a state-of-the-art scheme (Simple-8b) while being two times faster during decoding.