You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Dmitry Pavlov (Jira)" <ji...@apache.org> on 2022/09/09 07:30:00 UTC

[jira] [Updated] (IGNITE-15922) Create numa-aware allocator for data regions

     [ https://issues.apache.org/jira/browse/IGNITE-15922?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Dmitry Pavlov updated IGNITE-15922:
-----------------------------------
    Labels: important ise  (was: important)

> Create numa-aware allocator for data regions
> --------------------------------------------
>
>                 Key: IGNITE-15922
>                 URL: https://issues.apache.org/jira/browse/IGNITE-15922
>             Project: Ignite
>          Issue Type: Improvement
>            Reporter: Ivan Daschinsky
>            Assignee: Ivan Daschinsky
>            Priority: Major
>              Labels: important, ise
>             Fix For: 2.13
>
>          Time Spent: 5h 40m
>  Remaining Estimate: 0h
>
> Sometimes it is worth to allocate data region on specific numa node or on interleaved node (i.e. Intel Optane can be set up as a little bit slower, but cheaper than and have more capacity, than DDR, and unified as NUMA nodes). 
> Need to write C++ wrapper around {{libnuma}} and extract common allocator interface and add as a configuration parameter to {{DataRegionConfiguration}}



--
This message was sent by Atlassian Jira
(v8.20.10#820010)