memory allocation error in one pc and no in another!

[julie]

Hi,
I've ran rseg in my dev-pc without problem.
But when I've tested it in my production-pc, this message error appear :
Use Hideaki's empirical method: ERROR: could not allocate memory.

I've tested your dataset of barski and it runs correctly.
When I test my dataset (which runs correctly in my dev-pc), this error appears.
I've test with and without deadzones, with and without removing replicates. Always the same message.

I've checked my free memory : I have 16G of RAM and it's available for the application.
My pc is in Ubuntu 12.04, 16G RAM, 30G swap, 8 cores, and its config is the same that my dev-pc

May be can help you, I've noted this : when rseg begin to use Hideaki's method, the occupancy of the CPU decreases very quickly and the occupancy of the RAM increases considerably : it's completely used and the swap begin to be used.

Thanks.
Julie

[Song, Qiang]

Is there any other memory-consuming job running in your production PC at the same time?

[julie dubois]

No. There is no other job in same time. I have tested it just after boot and I have checked that no other job run.
Thanks

[julie]

Hi all,
After several test and research, I've detected that the same problem appears for an other application which is coded in C ! And It seems that the problem comes from a bad allocation memory in the program.

Can you check this problem ? or it's not possible?

Thanks
Julie

[Song, Qiang]

Hi Julie,

Could you please try this new version of rseg: http://smithlab.usc.edu/histone/rseg/rseg-0.4.8-rc.tar.gz ?

It should use less memory.

Song Qiang

[julie]

Thanks for this new version, but  when I use it, it fails before my previous error !!!

This is my command line :

../rseg/bin/rseg -c mouse-mm9-size.bed -d deadzones-k36-mm9.bed -i 20 -v -T -C H3K36me3/_Mikkelsen_3T3-L1_replicate1_T1_H3K36me3_mm9_MAQv0.6.3_PI0.bed


AND this is the messages print in my terminal by RSEG :

[PROCESSING] -T
[LOADING_DATA] chromosomes
[LOADING_DATA] reads
[LOADING_DATA] deadzones
[SELECTING BIN SIZE] bin size =  1550
[ESTIMATIN PARAMETERS]
rseg: ../common/TwoStateScaleResolveMixture.cpp:51: void initialize_values(const std::vector<double>&, const std::vector<double>&, Distro&, Distro&): Assertion `val_weight > 0' failed.
Abandon (core dumped)

Any idea of where do this error come from?

Thanks

[julie]

Sorry, it's an error! Effectively I try to run this new version with old options.
With the new options, it seems to run...

Thanks

On Thursday, September 27, 2012 10:04:57 AM UTC+2, julie wrote: