[openib-general] FW: [PATCH 1 of 3] mad: large RMPP support

Michael S. Tsirkin mst at mellanox.co.il
Thu Feb 9 00:11:45 PST 2006


Quoting r. Sean Hefty <sean.hefty at intel.com>:
> I did consider this, and I don't know when we'll start hitting issues allocating
> a single data buffer.  But we're going to ask for 10,000 256-byte buffers - over
> 2.5 MB of kernel memory in order to perform this single data transfer.  Is it
> likely that we can allocate that much memory, but not the 40k buffer?  I really
> don't know.  If the answer is yes, then I agree that using a linked list would
> be better.

It seems to me a linked list is a better short term solution.
Lets consider NodeInfo record.

It seems a MAD segment would include only 4 of these.
So on a modest 4K node cluster to get a list of all of them
requires 1K segments. Keeping these in an array would need
8K bytes on a 64 bit server.




-- 
Michael S. Tsirkin
Staff Engineer, Mellanox Technologies



More information about the general mailing list