samtools sort running extremely slow

[tsucheta]

Hi,

I am a new samtools user. I am trying to sort a bam file of the size 5 GB

using command

./samtools sort myFile.bam myFile.sorted

Apparently it is stuck at the merging phase:

saying
[bam_sort_core] merging from 49 files...

It has already been 8 hours and still the program is running. My Top command shows extremely low cpu usage by this program. The machine is a 64 bit dual core with 8 GB chip in it.

Any reason why this should be so slow?

Thanks

Sucheta

[mixter]

Bump. I'm also a bit confused by this message, in my case it reports merging 6 files. I use 3 lanes labeled with 5 barcodes each. To be sure I put each barcode-specific alignment in a separate directory and rerun, but it seems the "merging" isn't about that, only the .bam file you specify is processed...

This is a bit intransparent. In case anyone wants to investigate further, the source for bam_sort_core can be found here:

http://samtools.sourcearchive.com/do...8c-source.html

And some of the code in question:

for (;;) {
            if (buf[k] == 0) buf[k] = (bam1_t*)calloc(1, sizeof(bam1_t));
            b = buf[k];
            if ((ret = bam_read1(fp, b)) < 0) break;
            mem += ret;
            ++k;
            if (mem >= max_mem) {
                  sort_blocks(n++, k, buf, prefix, header);
                  mem = 0; k = 0;
            }
      }

Without taking apart bam_read1, it seems that the same file, fp, is read multiple times...

[lh3]

Probably because your file system is slow.