Simple and Fast BlockQuicksort using Lomuto's Partitioning Scheme

Aumüller, Martin; Hass, Nikolaj

doi:10.1137/1.9781611975499.2

Cited by 1 publication

(6 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, I1S 2 Ra outperforms IppRadix by a factor of 1.76 to 1.88 for the largest input size on A1x64, A1x16, and I4x20. On I2x16, I1S 2 Ra is even a factor of 3.00 faster. We want to note that IppRadix is surprisingly fast for (mostly small) Uniform distributed inputs with data type uint32 (see Fig.…”

Section: Running Times For Uniform Inputmentioning

confidence: 95%

“…After the decision tree of S 4 o has assigned element e to bucket b i , StringPS 4 o updates the bucket to introduce additional equality buckets for elements corresponding to a splitter: Element e goes to bucket b 2i+1 e =s i if i < k − 1, otherwise e goes to bucket b 2i . 2 The case distinction i < k − 1 is necessary as a k −1 is undefined.…”

Section: Definitions and Preliminariesmentioning

confidence: 99%

“…Overall, I1S 2 Ra is significantly faster than our fastest radix sort competitor SkaSort. For example, I1S 2 Ra is for all instances at least a factor of 1.10 faster than SkaSort and for even 63 % of the instances more than a factor of 1.40. The radix sorter IppRadix is faster than I1S 2 Ra in some special cases.…”

Section: Comparison Of Average Slowdownsmentioning

confidence: 99%

“…Our algorithm IPS 4 o can be called "winner" of this complicated comparison in the following sense: (1) It outperforms all competing implementations for most cases. (2) In many of these cases, the performance difference is large. (3) When IPS 4 o is slower than a competitor this is only by a small percentage in the overwhelming majority of cases -the only exceptions are for easy instances that are handled very quickly by some contenders that however perform poorly for more difficult instances.…”

Section: Introductionmentioning

confidence: 99%

“…(3) When IPS 4 o is slower than a competitor this is only by a small percentage in the overwhelming majority of cases -the only exceptions are for easy instances that are handled very quickly by some contenders that however perform poorly for more difficult instances. Our radix sorter IPS 2 Ra complements this by being even faster in some cases involving few processor cores and small, uniformly distributed keys. IPS 2 Ra also outperforms other radix sorters in many cases.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Engineering In-place (Shared-memory) Sorting Algorithms

Axtmann¹,

Witt²,

Ferizovic³

et al. 2020

Preprint

View full text Add to dashboard Cite

We present new sequential and parallel sorting algorithms that now represent the fastest known techniques for a wide range of input sizes, input distributions, data types, and machines. Somewhat surprisingly, part of the speed advantage is due to the additional feature of the algorithms to work in-place, i.e., they do not need a significant amount of space beyond the input array. Previously, the in-place feature often implied performance penalties. Our main algorithmic contribution is a blockwise approach to in-place data distribution that is provably cache-efficient. We also parallelize this approach taking dynamic load balancing and memory locality into account.Our new comparison-based algorithm In-place Superscalar Samplesort (IPS 4 o), combines this technique with branchless decision trees. By taking cases with many equal elements into account and by adapting the distribution degree dynamically, we obtain a highly robust algorithm that outperforms the best previous inplace parallel comparison-based sorting algorithms by almost a factor of three. That algorithm also outperforms the best comparison-based competitors regardless of whether we consider in-place or not in-place, parallel or sequential settings.Another surprising result is that IPS 4 o even outperforms the best (in-place or not in-place) integer sorting algorithms in a wide range of situations. In many of the remaining cases (often involving near-uniform input distributions, small keys, or a sequential setting), our new In-place Parallel Super Scalar Radix Sort (IPS 2 Ra) turns out to be the best algorithm.Claims to have the -in some sense -"best" sorting algorithm can be found in many papers which cannot all be true. Therefore, we base our conclusions on an extensive experimental study involving a large part of the cross product of 21 state-of-the-art sorting codes, 6 data types, 10 input distributions, 4 machines, 4 memory allocation strategies, and input sizes varying over 7 orders of magnitude. This confirms the claims made about the robust performance of our algorithms while revealing major performance problems in many competitors outside the concrete set of measurements reported in the associated publications. This is particularly true for integer sorting algorithms giving one reason to prefer comparison-based algorithms for robust general-purpose sorting.

show abstract

Section: Running Times For Uniform Inputmentioning

confidence: 95%

Section: Definitions and Preliminariesmentioning

confidence: 99%

Section: Comparison Of Average Slowdownsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations