Thanks a lot Lars, it builds perfectly. Now over to the end-user to validate it.
Assistance was much appreciated! Regards, Ricardo Gregorio Research and Systems Administrator -----Original Message----- From: [email protected] <[email protected]> On Behalf Of Lars Viklund Sent: 17 February 2020 12:04 To: [email protected] Subject: Re: [easybuild] Falcon --- Memory address map error --- Sanity check Hi Ricardo, The easyconfig I made for this can be found in a pull request against the easybuild-easyconfigs repository at https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Feasybuilders%2Feasybuild-easyconfigs%2Fpull%2F9873&data=01%7C01%7Cricardo.gregorio%40rothamsted.ac.uk%7C2997a2630a5d491dfb5508d7b3a1a014%7Cb688362589414342b0e37b8cc8392f64%7C1&sdata=otKL07MnrpcGVvU7hObloB6BY59YsJIVnD45rhEQCc4%3D&reserved=0 They are targetting the 2019b toolchain and as such have a dependency on a slightly more recent Python version. As the Python module now has scipy and numpy separated into the SciPy-bundle module, I've included an easyconfig for that as well. You can give your EasyBuild "--from-pr=9873" to get the easyconfigs from the PR. // Lars On 2/17/20 12:56 PM, Ricardo Gregorio wrote: > Hi Lars, > > Would you mind sharing the steps you have done to successfully build Falcon, > as I am still having issues???? > > I have made a copy of FALCON-1.8.8-intel-2017b.eb and try to build as follows: > > $ eb --optarch=march=skylake-avx512 --job-cores=8 > FALCON-1.8.8-intel-2019b.eb --try-toolchain=intel,2019b -r > > Regards, > Ricardo Gregorio > Research and Systems Administrator > > > -----Original Message----- > From: [email protected] > <[email protected]> On Behalf Of Lars Viklund > Sent: 13 February 2020 20:21 > To: [email protected] > Subject: Re: [easybuild] Falcon --- Memory address map error --- > Sanity check > > I made a copy of the easyconfig and built with intel/2019b instead. > > Having done so, the sanity checks are passing. This doesn't necessarily mean > that the actual software works, but it's probably better, assuming you have a > license for the Intel compilers in 2019b. > > // Lars > > On 2020-02-13 18:23, Lars Viklund wrote: >> "Good" news! >> >> I managed to reproduce this on a Skylake Xeon machine in a container >> with CentOS 7.6. >> >> https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgis >> t >> .github.com%2Fzao%2F95148a6da079fa8ea6fd9433c478d5d3&data=01%7C01 >> % >> 7Cricardo.gregorio%40rothamsted.ac.uk%7C99d4775a1ac64eaeb34008d7b0c26 >> 7 >> 7d%7Cb688362589414342b0e37b8cc8392f64%7C1&sdata=xerczbh9rvOCc8Csy >> k >> kk%2Bld3L6iV9Wtty5Oo%2FpkuAS4%3D&reserved=0 >> >> I don't have any solution yet, but at least it's not your particular >> environment that triggers the problem. >> >> // Lars >> >> On 2/13/20 12:19 PM, Ricardo Gregorio wrote: >>> Hi Lars, >>> >>> Thanks for replying.... >>> >>> We are trying to build on... >>> >>> CentOS 7.5 >>> Intel Gold (skylake) >>> >>> >>> Regards, >>> Ricardo Gregorio >>> Research and Systems Administrator >>> >>> >>> -----Original Message----- >>> From: [email protected] >>> <[email protected]> On Behalf Of Lars Viklund >>> Sent: 12 February 2020 17:59 >>> To: [email protected] >>> Subject: Re: [easybuild] Falcon --- Memory address map error --- >>> Sanity check >>> >>> Hi! >>> >>> I gave that easyconfig a try on a few different machines and can't >>> seem to get it to misbehave like you demonstrate. >>> >>> They were: >>> >>> * Ubuntu 18.04, 8700K >>> * Ubuntu 16.04, Broadwell, >>> * CentOS 7.6 (container on 8700K). >>> >>> Could you please let us know what OS release and processor are you >>> building this on? >>> >>> With kind regards, >>> Lars Viklund, HPC2N >>> >>> On 2/12/20 11:54 AM, [email protected] wrote: >>>> Hi all >>>> >>>> Hope someone has come across this before and can assist... >>>> >>>> we are trying to build Falcon-1.8.8-intel-2017b.eb >>>> [https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Feasybuilders%2Feasybuild-easyconfigs%2Fblob%2Fmaster%2Feasybuild%2Feasyconfigs%2Ff%2FFALCON%2FFALCON-1.8.8-intel-2017b.eb&data=01%7C01%7Cricardo.gregorio%40rothamsted.ac.uk%7C2997a2630a5d491dfb5508d7b3a1a014%7Cb688362589414342b0e37b8cc8392f64%7C1&sdata=q9BbMCdUWW9jGZHh3T%2Ffxnfx0UGp7tp3GqXTyW6t1I4%3D&reserved=0]. >>>> >>>> However it is failing during the sanity check stage and throwing >>>> some Memory address mapping errors >>>> >>>> +++++++++++++++++++++++ >>>> Extract of error message >>>> +++++++++++++++++++++++ >>>> >>>> *** Error in `/.../eb/software/Python/2.7.14-intel-2017b/bin/python': >>>> double free or corruption (out): 0x0000000000c3c510 *** ======= >>>> Backtrace: ========= /usr/lib64/libc.so.6(+0x81499)[0x2ad5c6c56499] >>>> /.../eb/software/FALCON/1.8.8-intel-2017b/lib/python2.7/site-packag >>>> e >>>> s/ >>>> falcon_kit-0.7-py2.7-linux-x86_64.egg/ext_falcon.so(find_best_aln_r >>>> a >>>> ng >>>> e+0xb14) >>>> [0x2ad5d3f3e294] >>>> /.../eb/software/FALCON/1.8.8-intel-2017b/lib/python2.7/site-packag >>>> e >>>> s/ >>>> falcon_kit-0.7-py2.7-linux-x86_64.egg/ext_falcon.so(generate_consen >>>> s >>>> us >>>> +0x1bd) >>>> [0x2ad5d3f3f01d] >>>> /.../eb/software/Python/2.7.14-intel-2017b/lib/python2.7/lib-dynloa >>>> d / _ctypes.so(ffi_call_unix64+0x4c)[0x2ad5d3d2881c] >>>> /.../eb/software/Python/2.7.14-intel-2017b/lib/python2.7/lib-dynloa >>>> d / _ctypes.so(ffi_call+0x579)[0x2ad5d3d27b29] >>>> /.../eb/software/Python/2.7.14-intel-2017b/lib/python2.7/lib-dynloa >>>> d / _ctypes.so(_ctypes_callproc+0x5ef)[0x2ad5d3d1d5cf] >>>> /.../eb/software/Python/2.7.14-intel-2017b/lib/python2.7/lib-dynloa >>>> d >>>> / >>>> _ctypes.so(+0xd926)[0x2ad5d3d17926] >>>> /.../eb/software/Python/2.7.14-intel-2017b/lib/libpython2.7.so. >>>> 1.0(PyObject_Call+0x5a)[0x2ad5c5a37eba] >>>> ... >>>> /usr/lib64/libc.so.6(__libc_start_main+0xf5)[0x2ad5c6bf7445] >>>> >>>> >>>> >>>> /.../eb/software/Python/2.7.14-intel-2017b/bin/python[0x4012e9] >>>> ======= Memory map: ======== >>>> 00400000-00418000 r-xp 00000000 00:28 >>>> 9014510081 /.../ >>>> eb/software/Python/2.7.14-intel-2017b/bin/python2.7 >>>> 00618000-00619000 r--p 00018000 00:28 >>>> 9014510081 /.../ >>>> eb/software/Python/2.7.14-intel-2017b/bin/python2.7 >>>> 00619000-0061a000 rw-p 00019000 00:28 >>>> 9014510081 /.../ >>>> eb/software/Python/2.7.14-intel-2017b/bin/python2.7 >>>> 009d7000-00ee7000 rw-p 00000000 00:00 0 [heap] >>>> 2ad5c57a8000-2ad5c57ca000 r-xp 00000000 fd:00 >>>> 790170 /usr/ >>>> lib64/ld-2.17.so >>>> 2ad5c57ca000-2ad5c57cc000 rw-p 00000000 00:00 0 >>>> 2ad5c57cc000-2ad5c57dd000 r-xp 00000000 00:28 >>>> 7731022159 /.../ >>>> eb/software/bzip2/1.0.6-GCCcore-6.4.0/lib/libbz2.so.1.0.6 >>>> 2ad5c57dd000-2ad5c57de000 r--p 00010000 00:28 >>>> 7731022159 /.../ >>>> eb/software/bzip2/1.0.6-GCCcore-6.4.0/lib/libbz2.so.1.0.6 >>>> 2ad5c57de000-2ad5c57df000 rw-p 00011000 00:28 >>>> 7731022159 /.../ >>>> eb/software/bzip2/1.0.6-GCCcore-6.4.0/lib/libbz2.so.1.0.6 >>>> 2ad5c57df000-2ad5c57e0000 rwxp 00000000 00:00 0 >>>> 2ad5c57e0000-2ad5c57e3000 rw-p 00000000 00:00 0 >>>> 2ad5c57e3000-2ad5c57f9000 r-xp 00000000 00:28 >>>> 7730996980 /.../ >>>> eb/software/GCCcore/6.4.0/lib64/libgcc_s.so.1 >>>> 2ad5c57f9000-2ad5c57fa000 r--p 00015000 00:28 >>>> 7730996980 /.../ >>>> eb/software/GCCcore/6.4.0/lib64/libgcc_s.so.1 >>>> 2ad5c57fa000-2ad5c57fb000 rw-p 00016000 00:28 >>>> 7730996980 /.../ >>>> eb/software/GCCcore/6.4.0/lib64/libgcc_s.so.1 >>>> 2ad5c57fb000-2ad5c58c1000 rw-p 00000000 00:00 0 >>>> 2ad5c58f1000-2ad5c59b1000 rw-p 00000000 00:00 0 >>>> 2ad5c59b1000-2ad5c59c7000 r-xp 00000000 00:28 >>>> 7730852699 /.../ >>>> eb/software/zlib/1.2.11-GCCcore-6.4.0/lib/libz.so.1.2.11 >>>> 2ad5c59c7000-2ad5c59c8000 r--p 00015000 00:28 >>>> 7730852699 /.../ >>>> eb/software/zlib/1.2.11-GCCcore-6.4.0/lib/libz.so.1.2.11 >>>> 2ad5c59c8000-2ad5c59c9000 rw-p 00016000 00:28 >>>> 7730852699 /.../ >>>> eb/software/zlib/1.2.11-GCCcore-6.4.0/lib/libz.so.1.2.11 >>>> 2ad5c59c9000-2ad5c59ca000 r--p 00021000 fd:00 >>>> 790170 /usr/ >>>> lib64/ld-2.17.so >>>> 2ad5c59ca000-2ad5c59cb000 rw-p 00022000 fd:00 >>>> 790170 /usr/ >>>> lib64/ld-2.17.so >>>> ... >>>> 2ad5d47ae000-2ad5d486f000 rw-p 00000000 00:00 0 >>>> 2ad5d8000000-2ad5d8021000 rw-p 00000000 00:00 0 >>>> 2ad5d8021000-2ad5dc000000 ---p 00000000 00:00 0 >>>> 7ffdeefc9000-7ffdeefef000 rw-p 00000000 00:00 0 [stack] >>>> 7ffdeeff5000-7ffdeeff7000 r-xp 00000000 00:00 0 [vdso] >>>> ffffffffff600000-ffffffffff601000 r-xp 00000000 00:00 0 [vsyscall] >>> Rothamsted Research is a company limited by guarantee, registered in >>> England at Harpenden, Hertfordshire, AL5 2JQ under the registration >>> number 2393175 and a not for profit charity number 802038. > Rothamsted Research is a company limited by guarantee, registered in England > at Harpenden, Hertfordshire, AL5 2JQ under the registration number 2393175 > and a not for profit charity number 802038. Rothamsted Research is a company limited by guarantee, registered in England at Harpenden, Hertfordshire, AL5 2JQ under the registration number 2393175 and a not for profit charity number 802038.
