Changes between Version 67 and Version 68 of WikiStart


Ignore:
Timestamp:
Nov 15, 2013, 11:04:27 AM (12 years ago)
Author:
almaless
Comment:

--

Legend:

Unmodified
Added
Removed
Modified

  • WikiStart

    v67 v68  
    77ALMOS (Advanced Locality Management Operating System) is an open-source research operating system currently under development at the [http://www-soc.lip6.fr/ SoC] department of the [http://www.lip6.fr/ LIP6] Laboratory ([http://www.upmc.fr UPMC Sorbonne Universités]). This new research operating system has been started from scratch and it is targeting cc-NUMA multi/many-cores. It is intended to investigate the scalability of the different components of an operating system on a current and future many-cores.
    88
    9 ALMOS is currently used in several research projects and Master degree courses at LIP6 and UPMC. Its development is currently a work-package in the SHARP (Scalable Heterogeneous ARchitecture for Processing) MEDEA+ European project ([http://www.catrene.org/web/projects/projects_call456.php #CA109]) which is an HPC oriented project leaded by Bull. ALMOS has been originally developed as a part of the TSAR MEDEA+ European project ([http://www.catrene.org/web/downloads/profiles_medea/2A718-TSAR-profile-outMEDEA2%20%2823-6-10%29.pdf #2A718]) and its kernel allows some shared-memory parallel applications to scale up to +1024 cores on the [https://www-asim.lip6.fr/trac/tsar TSAR] (Tera-Scale ARchitecture) single-chip cc-NUMA many-core. Work is in progress to port ALMOS to AMD Opteron (Interlagos) 64-cores cc-NUMA machine.
     9ALMOS is currently used in several research projects and Master degree courses at LIP6 and UPMC. Its development is currently a work-package in the SHARP (Scalable Heterogeneous ARchitecture for Processing) MEDEA+ European project ([http://www.catrene.org/web/projects/projects_call456.php #CA109]) which is an HPC oriented project leaded by Bull. ALMOS has been originally developed as a part of the TSAR MEDEA+ European project ([http://www.catrene.org/web/downloads/profiles_medea/2A718-TSAR-profile-outMEDEA2%20%2823-6-10%29.pdf #2A718]) and its kernel allows several shared-memory parallel applications to scale up to +1024 cores on the [https://www-asim.lip6.fr/trac/tsar TSAR] (Tera-Scale ARchitecture) single-chip cc-NUMA many-core. Work is in progress to port ALMOS to AMD Opteron (Interlagos) 64-cores cc-NUMA machine.
    1010
    1111
    … …  
    1414In a cc-NUMA architecture, the locality of memory access impacts directly both the scalability and the power consumption (energy by moved bit). The main challenge is to enforce the locality of memory access made by threads of parallel applications as well as kernel-level threads. Although the locality enforcing needs a fine management of hardware resources (mainly cores and physical memory), ALMOS aims to hide the hardware topology and the management of its resources to the user's applications. This allows POSIX shared-memory parallel applications as well as legacy applications to benefit from the performances offered by a many-core.
    1515
    16 The design of ALMOS kernel has a distributed approach articulated around objects named cluster-managers. The goal is to reduce the contention on the resources and to increase the locality of the kernel's memory access when referring to its data-structures. A cluster-manager is a manager of hardware resources of a hardware NUMA node (mainly cores and physical memory). A cluster-manager can locally supply its homed threads with any type of memory-objects including physical pages and kernel-level data-objects. A cluster-manager contains a per-core managers. Each core-manager has its own events-manager and a multi-policies scheduler server. In this distributed scheme, the kernel has no global-state notion of resources. Instead it has a decentralized approach to discover from where to allocate the requested resources in respect to the locality of memory access. This decentralized mechanism and its related policies are implemented in the DQDT (Distributed Quaternary Decision Tree). The DQDT is a wait-free mechanism based on a set of resources-usage indicators and integrates the locality awareness in all kernel decisions regarding to the tasks/threads placement, memory allocation and cores load-balancing. It constitutes a common kernel-level framework to build strategies in order to manage the resources of other kernel sub-systems like files and I/O.
     16The design of ALMOS kernel has a distributed approach articulated around objects named cluster-managers. The goal is to reduce the contention on the resources and to increase the locality of the kernel's memory access when referring to its own data structures. A cluster-manager is a manager of hardware resources of a cc-NUMA node (mainly cores and physical memory). A cluster-manager can locally supply its homed threads with any type of memory-objects including physical pages and kernel-level data-objects. A cluster-manager contains a per-core managers. Each core-manager has its own events-manager and a multi-policies scheduler server. In this distributed scheme, the kernel has no notion of resources global state. Instead, it has a decentralized mechanism for discovering from where to allocate the requested resources, taking in account the locality of memory access of the requester thread. This decentralized mechanism and its related policies are built using the DQDT (Distributed Quaternary Decision Tree). The DQDT is a distributed wait-free infrastructure based on a set of resource usage indicators and integrates the locality awareness needed by all kernel decisions regarding the tasks/threads placement, memory allocation and cores load-balancing. It constitutes a common framework for building strategies in order to manage the resources of other kernel sub-systems like files and I/O.
    1717
    18 The kernel of ALMOS replicates its code in each cluster by creating a per-cluster kernel-process. The threads of a cluster-manager are attached to their local kernel-process of their cluster. The kernel of ALMOS has a hybrid notion of threads which is based on a new organization of processes virtual address space. This new organization enables the kernel to replicate both of the pages-tables and the application's code and therefore, enforcing the locality of the per-core TLBs and I-cache misses. ALMOS threading model is compatible with PThreads standard and it allows the kernel to provide a native support to PGAS (Partitioned Global Address Space) programming model found in HPC oriented languages like UPC (Unified Parallel C).
     18The kernel of ALMOS replicates its code in each cluster by creating a per-cluster kernel process. The threads of a cluster-manager are attached to their local kernel process of their cluster. The kernel of ALMOS has a hybrid notion of threads which is based on a new organization of processes virtual address space. This new organization enables the kernel to replicate both of the page tables and the application's code; therefore, enforcing the locality of the per-core TLBs and I-cache misses. ALMOS threading model is compatible with PThreads standard and it allows the kernel to provide a native support to PGAS (Partitioned Global Address Space) programming model found in HPC oriented languages like UPC (Unified Parallel C).
    1919
    2020
    … …  
    2727 * Ghassan Almaless and Franck Wajsburt. Does Shared-Memory, Highly Multi-Threaded, Single-Application Scale on Many-Cores? In ''Proceedings of the 4th USENIX Workshop on Hot Topics in Parallelism'', Berkeley, CA, 2012. [https://www.usenix.org/conference/hotpar12/does-shared-memory-highly-multi-threaded-single-application-scale-many-cores Paper].
    2828
    29  * Ghassan Almaless. ALMOS Operating System on TSAR Single-Chip cc-NUMA Many-Core. Seminar, 2012. [htdocs:almos_on_tsar_03122012.pdf Slides].
     29 * Ghassan Almaless. ALMOS Operating System on TSAR Single-Chip cc-NUMA Many-Core. Seminar, ALSoC/LIP6, 2012. [htdocs:almos_on_tsar_03122012.pdf Slides].
    3030
    3131 * Ghassan Almaless. ALMOS : un système d'exploitation pour manycores en mémoire partagée cohérente. In ''Proceedings of the 8th French conference on Operating Systems (CFSE), the French chapter of ACM-SIGOPS, GDR ARP'', Saint-Malo, France, 2011. [http://renpar.irisa.fr/cfse8/cfse8_01.pdf Paper].