Hi Sjors, I module loaded Relion/2.1 and added "--pad 1" to continue my refinement. However, it gave me a warning that: " The following warnings were encountered upon command-line parsing: WARNING: Option --pad is not a valid RELION argument " Does it indicate that there might be something wrong in our compiling process? Chen ________________________________ From: Sjors Scheres <[log in to unmask]> Sent: Wednesday, January 24, 2018 4:06:22 AM To: Chen Sun; [log in to unmask] Subject: Re: [ccpem] Relion 3D autorefine Dear Chen, First, update your installation to the stable version: 2.1.0. Then, try adding "--pad 1" as an additional argument. I think that will solve your problems. HTH, Sjors On 01/24/2018 12:26 AM, Chen Sun wrote: > > Hi Adrian, > > > We have 128GB memory on our motherboard. However, the particle mrcs > files has 150GB. So is it ok if I exclude some particles in the final > refinement? > > > Chen > > ------------------------------------------------------------------------ > *From:* Adrian Koh <[log in to unmask]> > *Sent:* Tuesday, January 23, 2018 6:21:41 PM > *To:* Chen Sun > *Cc:* [log in to unmask] > *Subject:* RE: [ccpem] Relion 3D autorefine > > > Dear Chen, > > > > How much system memory (RAM) do you have? I suspect that might be the > bottleneck. > > > > Best Regards > > Adrian > > > > Sent from Mail <https://go.microsoft.com/fwlink/?LinkId=550986> for > Windows 10 > > > > *From: *Chen Sun <mailto:[log in to unmask]> > *Sent: *Tuesday, 23 January 2018 5:22 PM > *To: *[log in to unmask] <mailto:[log in to unmask]> > *Subject: *Re: [ccpem] Relion 3D autorefine > > > > Hi Shitaro, > > > > I have been using the 3 MPI for my autorefine jobs. And most of the > time I am the only one using the GPUs. That is why I am really confused. > > > > Chen > > *From:*Shintaro Aibara <[log in to unmask]> > *Sent:* Tuesday, January 23, 2018 3:27:56 PM > *To:* Chen Sun > *Cc:* [log in to unmask] > *Subject:* Re: [ccpem] Relion 3D autorefine > > > > I am imagining that you are running 5 MPI processes. Try dropping that > to 3. Having said that you may still crash because of other processes > like X taking up some memory. > > > > Best wishes, > > Shintaro > > > > On Tue, Jan 23, 2018 at 9:17 PM, Chen Sun <[log in to unmask] > <mailto:[log in to unmask]>> wrote: > > Dear Relion developers and users, > > > > I have been processing some large datasets with more than 100k > particles (512 pixel boxsize) with Relion/2.1b1. However, the last > iteration of my 3D refinement always crashed with the following > error message: > > > > "ERROR: out of memory in > /apps/relion/relion/src/gpu_utils/cuda_projector.cu > <http://cuda_projector.cu> at line 116 (error-code 2)" > > > > And everytime I restarted it, it will give me such warning at the > beginning: > > > > "With the current settings and hardware, you will be able to > > use an estimated image-size of 365 pixels during the last iteration... > > > > ...but your input box-size (image_size) is however 512. This means > that > > you will likely run out of memory on the GPU(s), and will have to > then re-start > > from the last completed iteration (i.e. continue from it) > *without* the use of GPUs." > > > > We are equipped with two 1080Ti 11GB GPUs which should be > sufficient for the memory that it requires. Does anyone have the > same problem? Is there any way to fix it? Any suggestions would be > appreciated. > > > > Thanks, > > Chen > > > > > > > > -- > > Yours Sincerely, > Shintaro Aibara > > > -- Sjors Scheres MRC Laboratory of Molecular Biology Francis Crick Avenue, Cambridge Biomedical Campus Cambridge CB2 0QH, U.K. tel: +44 (0)1223 267061 http://www2.mrc-lmb.cam.ac.uk/groups/scheres