couchdb-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Adam Kocoloski <kocol...@apache.org>
Subject Re: couchdb on 8 core redhat/fedora
Date Sun, 09 Aug 2009 19:13:09 GMT
On Aug 9, 2009, at 12:39 PM, Nitin Borwankar wrote:

> Jan Lehnardt wrote:
>>
>> On 9 Aug 2009, at 07:57, Nitin Borwankar wrote:
>>
>>> Paul Davis wrote:
>>>
>>> Hey Paul,
>>>
>>> Thanks much for that.  I was hoping it was something simple like  
>>> that.
>>> "Spidermonkey doesn't use any SMP features" -  does that mean it  
>>> will only run on one core and essentially become the bottleneck ?
>>
>> We don't parallelize view index creation yet, so this is not an  
>> additional problem for you. You can however build two views in  
>> parallel and make use of two cores that way.
>
> Hi Jan,
>
> Last night on the #couchdb irc channel there seemed to be agreement  
> that since multiple couchjs processes are spawned (this can be seen  
> during view creation) we get parallelism at the OS level.  Benoitc  
> made the original point. So it seems that as long as the proper  
> flags are used during configure/build, multi-core should be fine.  I  
> don't expect access to anything huger "just" 4 and 8 core boxes,   
> this should be fine for the foreseeable future.
>
> Nitin

Hi Nitin, Jan's right, if you're only building views from a single  
design doc you won't get much indexing speedup from multi-core at the  
moment.  We do spawn multiple couchjs processes (often one does the  
map and the other the reduce), but we don't map docs out to them  
simultaneously or anything.  Also, the Erlang process communicating  
with couchjs blocks and waits for the results when it sends data out.   
Best,

Adam

Mime
View raw message