Bull Sas

Simon Derr

Philippe Couvee

Benoit Welterlen

Philippe Garrigues

Grégoire Pichon

Gaël Goret

Sylvain Jeaugey

Benoît Welterlen

Antoine Percher

The invention concerns a storage method for storing, on data servers (), data file (to) slices (to) from the execution of a plurality of processes (to) of one or more applications (), comprising: distributing the stored data file (to) slices (to) over different data servers (), characterized in that: this distribution is carried out in such a way that the data file (to) slices (to) likely to be subsequently accessed simultaneously by different application () processes (to) are stored on different data servers () so as to reduce the subsequent access, to each of all or part of these data servers () by too many application () processes (to) simultaneously, and in that: the determination of the data file (to) slices (to) likely to be subsequently accessed simultaneously by different application () processes (to) has been carried out, during a prior phase of executing these application () processes (to), by observing the behavior of these application () processes (to) in order to access these stored data file (to) slices (to) over time.

Data storage method and method for executing an application with reduced access time to the stored data

The method for preloading data of a file containing the following steps of defining a plurality of bins of predetermined sizes in a file, for each input and/or output operation executed on the file, determining the bin involved in the operation, counting the number of input and/or output operations executed in each bin of the file by taking into account only a predetermined number of last operations on the whole file, and when the sum of the operations counted in a bin is greater than a predetermined threshold, loading, in a memory medium, at least one area of the file determined on the basis of this bin.

Method for preloading data

A computer system for managing the execution of threads including at least one central processing unit which performs interleaved execution of a plurality of threads throughout a plurality of virtual processors from said same central processing unit, and a handler for distributing the execution of the threads throughout the virtual processors. The computer system further includes means for classifying threads to be executed according to several predetermined types, and the handler for distributing the execution of threads directs each thread to be executed to a virtual processor according to the type thereof.

System and method for managing the interleaved execution of threads

The invention relates in particular to the optimization of the execution of a software application in a system having multiprocessor architecture including a plurality of input/output controllers and secondary processing units. After determining () the system topology, a call to a function to be executed by a secondary processing unit is intercepted (). The main processor that generated said call is identified (). A secondary processing unit is then identified () according to the main processor identified and according to the topology of said system. Advantageously, the secondary processing unit thus identified is the secondary processing unit that is the closest to the identified main processor. The call is then modified () in order to force the execution of at least one part of the function called in the identified secondary processing unit.

Method and device for optimizing execution of software applications in a multiprocessor architecture comprising several input/output controllers and secondary computing units

Growing community of inventors

Fontaine, France

Simon Derr