On 15.02.2011 21:05, poisontonic wrote:>
> Hi, I'm using the latest version of 64-bit R for Windows: R x64 2.12.1
>
> I'm using it because I currently need to do hierarchical clustering on
a
> very large object (too big for MATLAB, which I normally use).
> When I try to cluster my distance matrix d (obtained using dist on my
design
> matrix):
> hc<- hclust(d, method = 'average')
>
> I get an errror:
> Error in hclust(d, method = "average") :
> cannot allocate memory block of size 3.8 Gb
>
> However, the memory limits appear to be 16GB:
>> memory.limit()
> [1] 16378
>
> Does anyone know why R cannot allocate a memory block of size 3.8 GB, even
> though this is well within its memory limits??
If the available space got too fragmented, there is not single 3.8 block
of memory available any more....
Uwe Ligges
> Any help would be greatly appreciated!
> Thanks alot,
>
> Ben