{"id":105781,"date":"2022-12-02T03:36:38","date_gmt":"2022-12-02T01:36:38","guid":{"rendered":"https:\/\/prohoster.info\/blog\/novosti-interneta\/v-yadro-linux-6-2-vojdyot-podsistema-dlya-uskoritelej-vychislenij"},"modified":"2022-12-02T03:36:38","modified_gmt":"2022-12-02T01:36:38","slug":"v-yadro-linux-6-2-vojdyot-podsistema-dlya-uskoritelej-vychislenij","status":"publish","type":"post","link":"https:\/\/prohoster.info\/en\/blog\/news\/v-yadro-linux-6-2-vojdyot-podsistema-dlya-uskoritelej-vychislenij","title":{"rendered":"The Linux 6.2 kernel will include a subsystem for acceleration.","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"
In the DRM-Next branch, which is planned for inclusion in the Linux 6.2 kernel, code for a new subsystem called \"accel\" has been accepted, implementing a framework for computation accelerators. This subsystem is built on the DRM\/KMS foundation, as developers have already split the GPU representation into components that include sufficiently independent aspects of \"graphics output\" and \"computations\", allowing the subsystem to work with display controllers that do not have a computation block, as well as with computation blocks that do not have their own display controller, such as the ARM Mali GPU, which is essentially an accelerator. <\/p>\n
These abstractions turned out to be quite close to what was needed for a more general implementation of support for computation accelerators, so it was decided to expand the computation subsystem and rename it to \"accel\", as some of the supported devices are not GPUs. For example, Intel, which acquired Habana Labs, is interested in using this subsystem for machine learning accelerators. \n Source: opennet.ru<\/a> <\/p>","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"excerpt":{"rendered":"