This web mail archive is frozen.
This page is part of a frozen web archive of this mailing list.
You can still navigate around this archive, but know that no new mails
have been added to it since July of 2016.
Click here to be taken to the new web archives of this list; it includes all the mails that are in this frozen archive plus all new mails that have been sent to the list since it was migrated to the new archives.
Gleb Natapov wrote:
> On Mon, Dec 05, 2005 at 10:23:14AM -0700, Galen M. Shipman wrote:
>>>Also there is a code commented out that enables memory hooks if
>>>leave_pinned is set. Why this code is disabled? Infiniband will
>>>not work correctly in such setup.
>>There is still some debate about what will be the default behavior with
>>the leave_pinned option but we should have something in the trunk soon.
> Can you explain to me what is the concern? Infiniband will not function
> properly otherwise (can be easely demonstrated).
Agred - we will make this the default on the release branch.
> Also memfree hooks decrease cache efficiency, the better solution would
> be to catch brk() system calls and remove memory from cache only then,
> but there is no way to do it for now.
We are look at other options, including catching brk/munmap system calls, and
will be experimenting w/ these on the trunk.