The use of specialized hardware, such as accelerators or coprocessors offers an interesting approach to overcoming the physical limits encountered by processor architects. As a result, many machines are now equipped with one or several accelerators (e.g. a GPU), in addition to the usual processor(s). While significant efforts have been devoted to offloading computation onto such accelerators, very little attention has been paid to portability concerns on the one hand, and to the possibility of having heterogeneous accelerators and processors interact on the other hand.
│ │ │StarPU is a runtime system that provides support for heterogeneous multicore architectures. It not only offers a unified view of the computational resources (i.e. CPUs and accelerators simultaneously) but also takes care of efficiently mapping and executing tasks onto an heterogeneous machine while transparently handling low-level issues such as data transfers in a portable manner.
│ │ │StarPU is a software tool designed to enable programmers to harness the computational capabilities of both CPUs and GPUs, all while sparing them the need to meticulously adapt their programs for specific target machines and processing units.
│ │ │ ├── html2text {} │ │ │ │ @@ -1,15 +1,15 @@ │ │ │ │ StarPU Handbook │ │ │ │ Loading... │ │ │ │ Searching... │ │ │ │ No Matches │ │ │ │ 1. Introduction │ │ │ │ FFoorreewwoorrdd │ │ │ │ This manual documents the version 1.4.10 of StarPU. Its contents was last │ │ │ │ -updated on 2025-12-10. │ │ │ │ +updated on 2025-12-11. │ │ │ │ MMoottiivvaattiioonn │ │ │ │ The use of specialized hardware, such as accelerators or coprocessors offers an │ │ │ │ interesting approach to overcoming the physical limits encountered by processor │ │ │ │ architects. As a result, many machines are now equipped with one or several │ │ │ │ accelerators (e.g. a GPU), in addition to the usual processor(s). While │ │ │ │ significant efforts have been devoted to offloading computation onto such │ │ │ │ accelerators, very little attention has been paid to portability concerns on │ │ ├── ./usr/share/doc/libstarpu-dev/manual/html_dev/index.html │ │ │ @@ -90,15 +90,15 @@ │ │ │ │ │ │
This part presents how to write a StarPU application from an existing application.
│ │ │Some of the applications presented in the following chapters and some others are available in the git repository https://gitlab.inria.fr/starpu/starpu-applications
│ │ │A full StarPU tutorial which can be run with Docker is available at https://starpu.gitlabpages.inria.fr/tutorials/docker/
│ │ │
This part presents the basic knowledge of StarPU. It should be read to understand how StarPU works and how to execute a basic StarPU application.
│ │ │
This part explains the advanced concepts of StarPU. It is intended for users whose applications need more than basic task submission.
│ │ │You can learn more knowledge about some important and core concepts in StarPU:
This part explains how to better tune your application to achieve good performance, and also how to fix some difficulties you may encounter while implementing your applications.
│ │ │
This parts shows a basic usage of StarPU and how to execute the provided examples or your own applications.
│ │ │
The use of specialized hardware, such as accelerators or coprocessors offers an interesting approach to overcoming the physical limits encountered by processor architects. As a result, many machines are now equipped with one or several accelerators (e.g. a GPU), in addition to the usual processor(s). While significant efforts have been devoted to offloading computation onto such accelerators, very little attention has been paid to portability concerns on the one hand, and to the possibility of having heterogeneous accelerators and processors interact on the other hand.
│ │ │StarPU is a runtime system that provides support for heterogeneous multicore architectures. It not only offers a unified view of the computational resources (i.e. CPUs and accelerators simultaneously) but also takes care of efficiently mapping and executing tasks onto an heterogeneous machine while transparently handling low-level issues such as data transfers in a portable manner.
│ │ │
This part shows how StarPU which is natively written in C, has been extended to allow applications written in other languages to use it.
│ │ │
This part shows how to measure application performances.
│ │ │