sparsehash: FATAL ERROR: failed to allocate 12 groups

[Swati Mishra]

Hello ,

I am working with the ABySS 1.3.4 version on Illuminia sequencing data paired end.
I get the following error and I have no idea why does it shoot this error. I am very new to this assembler

abyss-pe v=-v name=sample1TCGA k=32 in='/N/dc2/scratch/radovlab/Microinversion/1/FASTA/584d3eda47                                                    ec27e5fcabcd9551aaad55.normal.fa'
ABYSS -k32 -q3 -v --coverage-hist=coverage.hist -s sample1TCGA-bubbles.fa  -o sample1TCGA-1.fa /N/dc2/scratch/radovl                                                    ab/Microinversion/1/FASTA/584d3eda47ec27e5fcabcd9551aaad55.normal.fa
ABySS 1.3.4
ABYSS -k32 -q3 -v --coverage-hist=coverage.hist -s sample1TCGA-bubbles.fa -o sample1TCGA-1.fa /N/dc2/scratch/radovla                                                    b/Microinversion/1/FASTA/584d3eda47ec27e5fcabcd9551aaad55.normal.fa
Reading `/N/dc2/scratch/radovlab/Microinversion/1/FASTA/584d3eda47ec27e5fcabcd9551aaad55.normal.fa'...
Read 100000 reads. Hash load: 750087 / 1073741824 = 0.000699 using 394 MB
Read 200000 reads. Hash load: 1398880 / 1073741824 = 0.0013 using 424 MB
Read 300000 reads. Hash load: 2104886 / 1073741824 = 0.00196 using 457 MB
Read 400000 reads. Hash load: 3128660 / 1073741824 = 0.00291 using 504 MB
Read 500000 reads. Hash load: 4165456 / 1073741824 = 0.00388 using 551 MB
Read 600000 reads. Hash load: 5094945 / 1073741824 = 0.00475 using 593 MB
Read 700000 reads. Hash load: 6034174 / 1073741824 = 0.00562 using 635 MB
Read 800000 reads. Hash load: 7006839 / 1073741824 = 0.00653 using 678 MB
Read 900000 reads. Hash load: 7983366 / 1073741824 = 0.00744 using 720 MB
Read 1000000 reads. Hash load: 8844520 / 1073741824 = 0.00824 using 757 MB
Read 1100000 reads. Hash load: 9730120 / 1073741824 = 0.00906 using 795 MB
Read 1200000 reads. Hash load: 10655157 / 1073741824 = 0.00992 using 835 MB
Read 1300000 reads. Hash load: 11617140 / 1073741824 = 0.0108 using 875 MB
Read 1400000 reads. Hash load: 12494946 / 1073741824 = 0.0116 using 912 MB
Read 1500000 reads. Hash load: 13306615 / 1073741824 = 0.0124 using 946 MB
Read 1600000 reads. Hash load: 14091398 / 1073741824 = 0.0131 using 978 MB
Read 1700000 reads. Hash load: 14908720 / 1073741824 = 0.0139 using 1.01 GB
Read 1800000 reads. Hash load: 15653691 / 1073741824 = 0.0146 using 1.04 GB
Read 1900000 reads. Hash load: 16393010 / 1073741824 = 0.0153 using 1.07 GB
Read 2000000 reads. Hash load: 17185009 / 1073741824 = 0.016 using 1.1 GB
Read 2100000 reads. Hash load: 17983556 / 1073741824 = 0.0167 using 1.14 GB
Read 2200000 reads. Hash load: 18828626 / 1073741824 = 0.0175 using 1.17 GB
Read 2300000 reads. Hash load: 19761823 / 1073741824 = 0.0184 using 1.21 GB
Read 2400000 reads. Hash load: 20720641 / 1073741824 = 0.0193 using 1.24 GB
Read 2500000 reads. Hash load: 21578044 / 1073741824 = 0.0201 using 1.28 GB
Read 2600000 reads. Hash load: 22375481 / 1073741824 = 0.0208 using 1.31 GB
Read 2700000 reads. Hash load: 23141197 / 1073741824 = 0.0216 using 1.34 GB
Read 2800000 reads. Hash load: 23957478 / 1073741824 = 0.0223 using 1.37 GB
Read 2900000 reads. Hash load: 24744392 / 1073741824 = 0.023 using 1.4 GB
Read 3000000 reads. Hash load: 25565538 / 1073741824 = 0.0238 using 1.44 GB
Read 3100000 reads. Hash load: 26445054 / 1073741824 = 0.0246 using 1.47 GB
Read 3200000 reads. Hash load: 27214111 / 1073741824 = 0.0253 using 1.5 GB
Read 3300000 reads. Hash load: 27990787 / 1073741824 = 0.0261 using 1.53 GB
Read 3400000 reads. Hash load: 28878999 / 1073741824 = 0.0269 using 1.57 GB
Read 3500000 reads. Hash load: 29771349 / 1073741824 = 0.0277 using 1.61 GB
Read 3600000 reads. Hash load: 30689600 / 1073741824 = 0.0286 using 1.64 GB
Read 3700000 reads. Hash load: 31589029 / 1073741824 = 0.0294 using 1.68 GB
Read 3800000 reads. Hash load: 32450990 / 1073741824 = 0.0302 using 1.72 GB
Read 3900000 reads. Hash load: 33268234 / 1073741824 = 0.031 using 1.75 GB
Read 4000000 reads. Hash load: 34126715 / 1073741824 = 0.0318 using 1.79 GB
Read 4100000 reads. Hash load: 35043399 / 1073741824 = 0.0326 using 1.82 GB
Read 4200000 reads. Hash load: 35928033 / 1073741824 = 0.0335 using 1.86 GB
Read 4300000 reads. Hash load: 36790204 / 1073741824 = 0.0343 using 1.9 GB
Read 4400000 reads. Hash load: 37660928 / 1073741824 = 0.0351 using 1.93 GB
Read 4500000 reads. Hash load: 38566544 / 1073741824 = 0.0359 using 1.97 GB
Read 4600000 reads. Hash load: 39474127 / 1073741824 = 0.0368 using 2 GB
Read 4700000 reads. Hash load: 40224618 / 1073741824 = 0.0375 using 2.03 GB
Read 4800000 reads. Hash load: 41038930 / 1073741824 = 0.0382 using 2.07 GB
Read 4900000 reads. Hash load: 41755852 / 1073741824 = 0.0389 using 2.09 GB
Read 5000000 reads. Hash load: 42389822 / 1073741824 = 0.0395 using 2.12 GB
Read 5100000 reads. Hash load: 43067806 / 1073741824 = 0.0401 using 2.15 GB
Read 5200000 reads. Hash load: 43844042 / 1073741824 = 0.0408 using 2.18 GB
Read 5300000 reads. Hash load: 44605046 / 1073741824 = 0.0415 using 2.21 GB
Read 5400000 reads. Hash load: 45314499 / 1073741824 = 0.0422 using 2.24 GB
Read 5500000 reads. Hash load: 46056555 / 1073741824 = 0.0429 using 2.27 GB
Read 5600000 reads. Hash load: 46828517 / 1073741824 = 0.0436 using 2.3 GB
Read 5700000 reads. Hash load: 47651670 / 1073741824 = 0.0444 using 2.33 GB
Read 5800000 reads. Hash load: 48473375 / 1073741824 = 0.0451 using 2.37 GB
Read 5900000 reads. Hash load: 49306676 / 1073741824 = 0.0459 using 2.4 GB
Read 6000000 reads. Hash load: 50168853 / 1073741824 = 0.0467 using 2.43 GB
Read 6100000 reads. Hash load: 51060822 / 1073741824 = 0.0476 using 2.47 GB
Read 6200000 reads. Hash load: 51940827 / 1073741824 = 0.0484 using 2.51 GB
Read 6300000 reads. Hash load: 52815732 / 1073741824 = 0.0492 using 2.54 GB
Read 6400000 reads. Hash load: 53354406 / 1073741824 = 0.0497 using 2.56 GB
Read 6500000 reads. Hash load: 54026698 / 1073741824 = 0.0503 using 2.59 GB
Read 6600000 reads. Hash load: 54597568 / 1073741824 = 0.0508 using 2.61 GB
Read 6700000 reads. Hash load: 55425202 / 1073741824 = 0.0516 using 2.65 GB
Read 6800000 reads. Hash load: 56268846 / 1073741824 = 0.0524 using 2.68 GB
Read 6900000 reads. Hash load: 57170934 / 1073741824 = 0.0532 using 2.71 GB
Read 7000000 reads. Hash load: 58011455 / 1073741824 = 0.054 using 2.75 GB
Read 7100000 reads. Hash load: 58812284 / 1073741824 = 0.0548 using 2.78 GB
Read 7200000 reads. Hash load: 59605636 / 1073741824 = 0.0555 using 2.81 GB
Read 7300000 reads. Hash load: 60425641 / 1073741824 = 0.0563 using 2.84 GB
Read 7400000 reads. Hash load: 61266106 / 1073741824 = 0.0571 using 2.87 GB
Read 7500000 reads. Hash load: 62042844 / 1073741824 = 0.0578 using 2.9 GB
Read 7600000 reads. Hash load: 62874508 / 1073741824 = 0.0586 using 2.93 GB
Read 7700000 reads. Hash load: 63711629 / 1073741824 = 0.0593 using 2.96 GB
Read 7800000 reads. Hash load: 64495393 / 1073741824 = 0.0601 using 2.99 GB
Read 7900000 reads. Hash load: 65282297 / 1073741824 = 0.0608 using 3.02 GB
Read 8000000 reads. Hash load: 66067281 / 1073741824 = 0.0615 using 3.05 GB
Read 8100000 reads. Hash load: 66905629 / 1073741824 = 0.0623 using 3.08 GB
Read 8200000 reads. Hash load: 67647802 / 1073741824 = 0.063 using 3.11 GB
Read 8300000 reads. Hash load: 68407535 / 1073741824 = 0.0637 using 3.13 GB
Read 8400000 reads. Hash load: 69233068 / 1073741824 = 0.0645 using 3.16 GB
Read 8500000 reads. Hash load: 70064155 / 1073741824 = 0.0653 using 3.19 GB
Read 8600000 reads. Hash load: 70935321 / 1073741824 = 0.0661 using 3.23 GB
Read 8700000 reads. Hash load: 71770623 / 1073741824 = 0.0668 using 3.26 GB
Read 8800000 reads. Hash load: 72564419 / 1073741824 = 0.0676 using 3.28 GB
Read 8900000 reads. Hash load: 73395172 / 1073741824 = 0.0684 using 3.31 GB
Read 9000000 reads. Hash load: 74192381 / 1073741824 = 0.0691 using 3.34 GB
Read 9100000 reads. Hash load: 74979384 / 1073741824 = 0.0698 using 3.37 GB
Read 9200000 reads. Hash load: 75797786 / 1073741824 = 0.0706 using 3.4 GB
Read 9300000 reads. Hash load: 76654345 / 1073741824 = 0.0714 using 3.43 GB
Read 9400000 reads. Hash load: 77555153 / 1073741824 = 0.0722 using 3.46 GB
Read 9500000 reads. Hash load: 78364138 / 1073741824 = 0.073 using 3.49 GB
Read 9600000 reads. Hash load: 79154983 / 1073741824 = 0.0737 using 3.52 GB
Read 9700000 reads. Hash load: 79990124 / 1073741824 = 0.0745 using 3.55 GB
Read 9800000 reads. Hash load: 80818307 / 1073741824 = 0.0753 using 3.58 GB
Read 9900000 reads. Hash load: 81638249 / 1073741824 = 0.076 using 3.61 GB
Read 10000000 reads. Hash load: 82403044 / 1073741824 = 0.0767 using 3.63 GB
Read 10100000 reads. Hash load: 83226130 / 1073741824 = 0.0775 using 3.66 GB
Read 10200000 reads. Hash load: 84040990 / 1073741824 = 0.0783 using 3.69 GB
Read 10300000 reads. Hash load: 84911650 / 1073741824 = 0.0791 using 3.72 GB
Read 10400000 reads. Hash load: 85755386 / 1073741824 = 0.0799 using 3.75 GB
Read 10500000 reads. Hash load: 86607569 / 1073741824 = 0.0807 using 3.78 GB
Read 10600000 reads. Hash load: 87486295 / 1073741824 = 0.0815 using 3.81 GB
Read 10700000 reads. Hash load: 88341337 / 1073741824 = 0.0823 using 3.84 GB
Read 10800000 reads. Hash load: 89216575 / 1073741824 = 0.0831 using 3.87 GB
Read 10900000 reads. Hash load: 90016781 / 1073741824 = 0.0838 using 3.89 GB
Read 11000000 reads. Hash load: 90806566 / 1073741824 = 0.0846 using 3.92 GB
Read 11100000 reads. Hash load: 91660866 / 1073741824 = 0.0854 using 3.95 GB
Read 11200000 reads. Hash load: 92531339 / 1073741824 = 0.0862 using 3.97 GB
Read 11300000 reads. Hash load: 93403033 / 1073741824 = 0.087 using 4 GB
Read 11400000 reads. Hash load: 94180221 / 1073741824 = 0.0877 using 4.03 GB
Read 11500000 reads. Hash load: 94982377 / 1073741824 = 0.0885 using 4.05 GB
Read 11600000 reads. Hash load: 95738733 / 1073741824 = 0.0892 using 4.08 GB
Read 11700000 reads. Hash load: 96507926 / 1073741824 = 0.0899 using 4.1 GB
Read 11800000 reads. Hash load: 97252461 / 1073741824 = 0.0906 using 4.13 GB
Read 11900000 reads. Hash load: 98010454 / 1073741824 = 0.0913 using 4.15 GB
Read 12000000 reads. Hash load: 98750771 / 1073741824 = 0.092 using 4.18 GB
Read 12100000 reads. Hash load: 99584254 / 1073741824 = 0.0927 using 4.2 GB
Read 12200000 reads. Hash load: 100438024 / 1073741824 = 0.0935 using 4.23 GB
Read 12300000 reads. Hash load: 101251616 / 1073741824 = 0.0943 using 4.26 GB
sparsehash: FATAL ERROR: failed to allocate 12 groups
make: *** [sample1TCGA-1.fa] Error 1

Please could you help me out with this error or gimme hint at it.
Thanks!

Swati

[Shaun Jackman]

Hi Swati,

It looks as though your ABySS executable is 32-bit. You'll need to use
a 64-bit executable to make use of more than 4 GB of memory. To check
use the command
file -L `which ABYSS`

$ file -L `which ABYSS`
/home/sjackman/local/bin/ABYSS: ELF 64-bit LSB executable, AMD x86-64,
version 1 (SYSV), for GNU/Linux 2.6.9, dynamically linked (uses shared
libs), not stripped

Cheers,
Shaun

On 7 October 2013 12:01, Swati Mishra <swat...@gmail.com> wrote:
> Hello ,
>
> I am working with the ABySS 1.3.4 version on Illuminia sequencing data
> paired end.
> I get the following error and I have no idea why does it shoot this error. I
> am very new to this assembler
>
> abyss-pe v=-v name=sample1TCGA k=32
> in='/N/dc2/scratch/radovlab/Microinversion/1/FASTA/584d3eda47
> ec27e5fcabcd9551aaad55.normal.fa'
> ABYSS -k32 -q3 -v --coverage-hist=coverage.hist -s sample1TCGA-bubbles.fa
> -o sample1TCGA-1.fa /N/dc2/scratch/radovl
> ab/Microinversion/1/FASTA/584d3eda47ec27e5fcabcd9551aaad55.normal.fa
> ABySS 1.3.4
> ABYSS -k32 -q3 -v --coverage-hist=coverage.hist -s sample1TCGA-bubbles.fa -o
> sample1TCGA-1.fa /N/dc2/scratch/radovla
> b/Microinversion/1/FASTA/584d3eda47ec27e5fcabcd9551aaad55.normal.fa
> Reading
> `/N/dc2/scratch/radovlab/Microinversion/1/FASTA/584d3eda47ec27e5fcabcd9551aaad55.normal.fa'...
> Read 100000 reads. Hash load: 750087 / 1073741824 = 0.000699 using 394 MB
> Read 200000 reads. Hash load: 1398880 / 1073741824 = 0.0013 using 424 MB
…

[Swati Mishra]

Hello Shaun,

Thank you for your reply

i get this when I run which ABYSS
/N/soft/mason/abyss/1.3.4/abyss-1.3.4_maxk_96/bin/ABYSS

I am running it on a server with 512 GB MEMORY

Please let me know
Thanks
Regards,
SWATI

[Swati Mishra]

Also I checked the version

This is the following information
GNU Make 3.81
Copyright (C) 2006  Free Software Foundation, Inc.
This is free software; see the source for copying conditions.
There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A
PARTICULAR PURPOSE.

This program built for x86_64-redhat-linux-gnu
So I am guessing my machine is already x86_64 information.

Thanks
Regards,
Swati

[Tony Raymond]

Hi Swati,

Shaun asked about the output of the following command not just 'which':

file -L /N/soft/mason/abyss/1.3.4/abyss-1.3.4_maxk_96/bin/ABYSS

We need to double check this file is actually a 64-bit executable.

Thanks,

Tony

--
You received this message because you are subscribed to the Google Groups "ABySS" group.
To unsubscribe from this group and stop receiving emails from it, send an email to abyss-users...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

[Swati Mishra]

Hello,

This is the output of the following command
 file -L /N/soft/mason/abyss/1.3.4/abyss-1.3.4_maxk_96/bin/ABYSS
/N/soft/mason/abyss/1.3.4/abyss-1.3.4_maxk_96/bin/ABYSS: ELF 64-bit LSB executable, x86-64, version 1 (SYSV), dynamically linked (uses shared libs), for GNU/Linux 2.6.18, not stripped

Thanks!
Regards,
Swati

[Tony Raymond]

Thanks Swati!

I have personally never seen this error when ABYSS runs out of memory, but others have:

https://groups.google.com/forum/#!msg/abyss-users/HKdbQB1ANsc/tGMds1l1l4sJ

What size of computer are you running this job on? You may need to run this on a bigger computer, or check that the max memory and virtual memory aren't being limited (check these with ulimit -a). Also, it looks like you are running ABySS with k=32, but you compiled it with maxk=96. If you reconfigure and recompile with "./configure --enable-maxk=32 …", the process would use less memory/kmer.

Cheers,

Tony

[Swati Mishra]

Hello,

I run it on a server with 512 GB memory which is a supercomputer and the the version of abyss is 1.3.4 .
I don't really understand what you mean by "ABySS with k=32, but you compiled it with maxk=96. If you reconfigure and recompile with "./configure --enable-maxk=32 …", the process would use less memory/kmer.". Please could you explain this in simple terms.

I am very new to this genome assembler. SO i couldn't comprehend.

Thanks!
Regards,
Swati

[Swati Mishra]

Hello,

I also did the following commad ulimit -a and I got the following output

core file size          (blocks, -c) 0
data seg size           (kbytes, -d) unlimited
scheduling priority             (-e) 0
file size               (blocks, -f) unlimited
pending signals                 (-i) 4135980
max locked memory       (kbytes, -l) unlimited
max memory size         (kbytes, -m) unlimited
open files                      (-n) 4096
pipe size            (512 bytes, -p) 8
POSIX message queues     (bytes, -q) 819200
real-time priority              (-r) 0
stack size              (kbytes, -s) unlimited
cpu time               (seconds, -t) unlimited
max user processes              (-u) 1024
virtual memory          (kbytes, -v) 4194304
file locks                      (-x) unlimited

Thanks
Regards,
Swati

[Tony Raymond]

Your virtual memory is being restricted to ~4GB. You'll need to set it to something larger. If you aren't concerned about the possibility of a thrashing machine just set it to unlimited, i.e. "ulimit -v unlimited"

Cheers,

Tony

[Tony Raymond]

Compiling ABYSS with smaller maxk makes ABYSS use less memory, but then you can't assemble at higher kmer values. If memory was a limiting factor, then you could recompile ABySS with a different maxk, but it doesn't sound like that's the issue here as per your other email.

Cheers,

Tony

[Tony Raymond]

Running "ulimit -v 512000000" would also be sufficient to allow ABYSS to run on your 512GB machine.

Tony

[Swati Mishra]

I get the following output

abcd@Mason: /N/soft/mason/abyss/1.3.4/abyss-1.3.4_maxk_96/share/doc/abyss> ulimit -v 512000000
-bash: ulimit: virtual memory: cannot modify limit: Operation not permitted

So I guess need to ask for permission or ask the admins to grant me to extend the abyss virtual memory usage.
So may be the virtual memory is for interactive sessions? May be I need to submit a PBS job for this task?
Thanks!

Regards,
Swati