A factoring method is presented which, heuristically, splits composite n in O(n1/4+ ) steps. There are two ideas: an integer approximation to √ (q/p) provides an O(n1/2+ ) algorithm in which n is represented as the difference of two rational squares; observing that if a prime m divides a square, then m2 divides that square, a heuristic speed-up to O(n1/4+ ) steps is achieved. The method is well...

Scalable Flat Cache Only Memory Architectures (Flat COMA) are designed for reduced memory access latencies while minimizing programmer and operating system involvement. Indeed, to keep memory access latencies low, neither the programmer needs to perform clever data placement nor the operating system needs to perform page migration. The hardware automatically replicates the data and migrates it ...

External mergesort is normally implemented so that each run is stored contiguously on disk and blocks of data are read exactly in the order they are needed during merging. We investigate two ideas for improving the performance of external mergesort: interleaved layout and a new reading strategy. Interleaved layout places blocks from diierent runs in consecutive disk addresses. This is done in t...

68 • IATSS RESEARCH Vol.27 No.1, 2003 PERSONALITY AND SPEEDING – Some Policy Implications – Richard TAY Philip CHAMPNESS Barry WATSON Associate Professor in Road Safety Centre for Accident Research & Road Safety Centre for Accident Research & Road Safety Centre for Accident Research & Road Safety Queensland University of Technology Queensland University of Technology Queensland University of Te...

This paper describes several speedups and simplifications for XTR. The most important results are new XTR double and single exponentiation methods where the latter requires a cheap precomputation. Both methods are on average more than 60% faster than the old methods, thus more than doubling the speed of the already fast XTR signature applications. An additional advantage of the new double expon...

Time required for model training is an important limiting factor for faster pace of progress in the field of deep learning. The faster the model training, the more options researchers are able to try in the same amount of time, and the higher the quality of their results. In this work we stacked a set of techniques to optimize training time of the ResNet model with 20 layers and achieved a subs...

In this note, we describe how to achieve a simple yet substantial speed up of Miller’s algorithm, when not using denominator elimination, and working over quadratic extension fields.

This paper presents a model and implementation techniques for speeding up constraint propagation. Two fundamental approaches to improving constraint propagation are explored: keeping track of which propagators are at fixpoint, and choosing which propagator to apply next. We show how idempotence reasoning and events help track fixpoints more accurately. We improve these methods by using them dyn...

Speeding up inferences made from large knowledge bases is a key to scaling up AI systems. The query-tree is a powerful tool for analyzing KBs containing Horn rules which takes into consideration the semantics of interpreted predicates that appear in the rules (e.g., order and sort predicates). It is a nite structure that encodes all derivations of a given set of queries and tells us which rules...

For over thirty years until 2004-5 the speed of processors followed the Moore’s Law and was exponentially increasing. Approximately every 18 months processors speed doubled. This rapid and steady progress was very economically satisfying. It was enough to wait several months for new processor generation without the need for redeveloping software. However shortly after the electronic technology ...