I have a laptop with 8GB RAM and 6GB swap yet when I try to make graph-tool it always runs out of RAM while compiling graph_blockmodel.hh:236:2 If I try with GCC it gives usually is killed but sometimes restarts the computer, if I try with clang it restarts the computer.
I'm using Solus as my distro.
-- View this message in context: http://main-discussion-list-for-the-graph-tool-project.982480.n3.nabble.com/... Sent from the Main discussion list for the graph-tool project mailing list archive at Nabble.com.
Ni! Hi Padraig,
Some things which might help us help you:
a) gcc version you're using b) graph-tool version you're compiling c) how much ram do you actually have available when you start your compilation d) the actual error message coming out of gcc
cheers,
ale .~´
On Wed, Jun 28, 2017 at 12:34 PM, padraigmac padraig_mac@hotmail.com wrote:
I have a laptop with 8GB RAM and 6GB swap yet when I try to make graph-tool it always runs out of RAM while compiling graph_blockmodel.hh:236:2 If I try with GCC it gives usually is killed but sometimes restarts the computer, if I try with clang it restarts the computer.
I'm using Solus as my distro.
-- View this message in context: http://main-discussion-list- for-the-graph-tool-project.982480.n3.nabble.com/Can-t- compile-graph-tool-on-Solus-not-enough-RAM-tp4027271.html Sent from the Main discussion list for the graph-tool project mailing list archive at Nabble.com. _______________________________________________ graph-tool mailing list graph-tool@skewed.de https://lists.skewed.de/mailman/listinfo/graph-tool
Hi Ale,
I did mean to provide the error message but the last time I tried to run it it restarted the computer so I couldn't get it but I got it again there, it got a little further this time:
graph_blockmodel.hh:236:28: note: here case weight_type::REAL_EXPONENTIAL: ^~~~ {standard input}: Assembler messages: {standard input}:625641: Warning: end of file not at end of a line; newline inserted {standard input}:626860: Error: expected comma after name `_ZZZ11export_lsbmvENKUlPT_E_clIN10graph_tool10BlockStateIJN5boost18undirected_adaptorINS5_8adj_listImEEEENS5_29unchecked_vector_property_mapISt6vectorISt5tupleIJmmmEESaISD_EENS5_27typed_identity_property_mapImEEEESt17integral_constantIbLb0EESJ_IbLb1EENS5_3anyESM_SM_NSA_IiNS5_27adj_edge_index_property_mapImEEEENSA_I' in .size directive g++: internal compiler error: Killed (program cc1plus) Please submit a full bug report, with preprocessed source if appropriate.
I'm using gcc version 7.10 trying to compile graph-tool version 2.22/
When I started the compilation I had 7.0G available according to free -h. However the while the swap was free it doesn't say anything under available there if that's a problem. Swap is on and I can hibernate.
Thanks,
Pádraig
Alexandre Hannud Abdo wrote
Ni! Hi Padraig,
Some things which might help us help you:
a) gcc version you're using b) graph-tool version you're compiling c) how much ram do you actually have available when you start your compilation d) the actual error message coming out of gcc
cheers,
ale .~´
On Wed, Jun 28, 2017 at 12:34 PM, padraigmac <
padraig_mac@
> wrote:
I have a laptop with 8GB RAM and 6GB swap yet when I try to make graph-tool it always runs out of RAM while compiling graph_blockmodel.hh:236:2 If I try with GCC it gives usually is killed but sometimes restarts the computer, if I try with clang it restarts the computer.
I'm using Solus as my distro.
-- View this message in context: http://main-discussion-list- for-the-graph-tool-project.982480.n3.nabble.com/Can-t- compile-graph-tool-on-Solus-not-enough-RAM-tp4027271.html Sent from the Main discussion list for the graph-tool project mailing list archive at Nabble.com. _______________________________________________ graph-tool mailing list
graph-tool@
graph-tool mailing list
graph-tool@
-- View this message in context: http://main-discussion-list-for-the-graph-tool-project.982480.n3.nabble.com/... Sent from the Main discussion list for the graph-tool project mailing list archive at Nabble.com.
On 28.06.2017 14:28, padraigmac wrote:
Hi Ale,
I did mean to provide the error message but the last time I tried to run it it restarted the computer so I couldn't get it but I got it again there, it got a little further this time:
graph_blockmodel.hh:236:28: note: here case weight_type::REAL_EXPONENTIAL: ^~~~ {standard input}: Assembler messages: {standard input}:625641: Warning: end of file not at end of a line; newline inserted {standard input}:626860: Error: expected comma after name `_ZZZ11export_lsbmvENKUlPT_E_clIN10graph_tool10BlockStateIJN5boost18undirected_adaptorINS5_8adj_listImEEEENS5_29unchecked_vector_property_mapISt6vectorISt5tupleIJmmmEESaISD_EENS5_27typed_identity_property_mapImEEEESt17integral_constantIbLb0EESJ_IbLb1EENS5_3anyESM_SM_NSA_IiNS5_27adj_edge_index_property_mapImEEEENSA_I' in .size directive g++: internal compiler error: Killed (program cc1plus) Please submit a full bug report, with preprocessed source if appropriate.
I'm using gcc version 7.10 trying to compile graph-tool version 2.22/
When I started the compilation I had 7.0G available according to free -h. However the while the swap was free it doesn't say anything under available there if that's a problem. Swap is on and I can hibernate.
I recommend you try the current git version, which should require less memory for that part to compile.
Best, Tiago
Got further using the current git version but still couldn't compile it:
CXX graph_similarity.lo {standard input}: Assembler messages: {standard input}:17833566: Warning: end of file not at end of a line; newline inserted {standard input}:17833641: Error: unknown pseudo-op: `.' {standard input}: Error: open CFI at the end of file; missing .cfi_endproc directive g++: internal compiler error: Killed (program cc1plus) Please submit a full bug report, with preprocessed source if appropriate.
-- View this message in context: http://main-discussion-list-for-the-graph-tool-project.982480.n3.nabble.com/... Sent from the Main discussion list for the graph-tool project mailing list archive at Nabble.com.
Just to update, I increased the Swap and it compiled okay. Took 15GB at it's peak, at one point I had 330MB memory free!
-- View this message in context: http://main-discussion-list-for-the-graph-tool-project.982480.n3.nabble.com/... Sent from the Main discussion list for the graph-tool project mailing list archive at Nabble.com.
On 30.06.2017 11:03, padraigmac wrote:
Just to update, I increased the Swap and it compiled okay. Took 15GB at it's peak, at one point I had 330MB memory free!
This is very strange; I never observed anything close to this.
Are you sure you are not compiling in parallel?
Sorry for the incredibly slow response, I took a screen grab of top while it was compiling at one stage, see here http://http://imgur.com/a/gBbki . It didn't seem to compile in parallel no.
-- View this message in context: http://main-discussion-list-for-the-graph-tool-project.982480.n3.nabble.com/... Sent from the Main discussion list for the graph-tool project mailing list archive at Nabble.com.
On 19.07.2017 11:00, padraigmac wrote:
Sorry for the incredibly slow response, I took a screen grab of top while it was compiling at one stage, see here http://http://imgur.com/a/gBbki . It didn't seem to compile in parallel no.
It's using 7 gigs in this case, not 15.
I noticed you were using graph-tool 2.22. You should try the git version, which has been modified to use less memory during compilation.
That must have been when I tried it first before turning swap on then, I thought I took other ones but I guess not. I did compile the master version in the end after you suggested it above but I didn't take any screen grabs that time. What I'll do is try and compile it again later and if I get any memory spikes I'll screenshot it and post it here.
-- View this message in context: http://main-discussion-list-for-the-graph-tool-project.982480.n3.nabble.com/... Sent from the Main discussion list for the graph-tool project mailing list archive at Nabble.com.
It ran out of memory the first time it got to graph_similarity.lo but the second time it finished it barely, here is a screen shot http://imgur.com/86FKunK , it got to 7.9G in swap (+7.3G in Mem) and then went on to the next one.
-- View this message in context: http://main-discussion-list-for-the-graph-tool-project.982480.n3.nabble.com/... Sent from the Main discussion list for the graph-tool project mailing list archive at Nabble.com.
-
Alexandre Hannud Abdo -
padraigmac -
Tiago de Paula Peixoto