hadoop-hdfs-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Akash Ashok <thehellma...@gmail.com>
Subject Re: Why LightWeightGSet ?
Date Mon, 13 Jun 2011 16:24:30 GMT
Thanks a lot Harsh. I got the point . But I have one small question. Why is
2 or 3 reduced from exponent based on whether its a 32bit or 64 bit JVM
respectively while calculating LightWeightGSet capacity. Any Idea ?

On Mon, Jun 13, 2011 at 7:30 PM, Harsh J <harsh@cloudera.com> wrote:

> Akash,
>
> You should be able to come to a conclusion here:
> https://issues.apache.org/jira/browse/HDFS-1114
>
> On Mon, Jun 13, 2011 at 7:24 PM, Akash Ashok <thehellmaker@gmail.com>
> wrote:
> > Hi,
> > I was just going thru the implementation of LightWeightGSet and  i
> figured
> > LightWeightGSet implementation is pretty much the same as that of a
> HashMap
> > execept for the fact that HashMap resizes itself if the number of element
> > exceeds the load factor and HashMap has a much stronger hash. Could some1
> > tell me as to why LightWeightGSet was written when HashMap itself could
> have
> > been used ??
> >
> > Cheers,
> > Akash A
> >
>
>
>
> --
> Harsh J
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message