Need Help?

OMP/MPI SCALING

Go back to main page

H2O-32.inp

View full archive

RevisionStatusSummaryAuthorCommit Message
18082 (0)rybkinjradjust tolerance in embedding test
18081 (-1)OKRuntime varied by 13.6% (threshold: 30.0%)alazzarofix bug in the previous commit when nprocs>1
18080 (-2)alazzaroclose file for restart
18079 (-3)agloessELPA: Add support for ELPA-2017. Adjust renamed dummy arguments and accommodate for new module enumerating ELPA-kernels as PARAMETERs. (Hans Pabst)
18078 (-4)OKRuntime varied by 10.6% (threshold: 30.0%)rybkinjradded test dependence on another test in embedding
18077 (-5)rybkinjrchanged the test reference
18076 (-6)rybkinjrchanged the test failing with farming
18075 (-7)OKRuntime varied by 16.4% (threshold: 30.0%)rybkinjrfixed constant potential and added a test
18074 (-8)jhutterEva Perlt (Bonn): new B97 paramtrization + short range D3 correction
18073 (-9)agloessVDW-corr: Adjust tolerance.
18072 (-10)OKRuntime varied by 12.1% (threshold: 30.0%)rybkinjrenable restart with embedding
18071 (-11)jhutterBug fix: Grimme D3 C9 term: use original parameter
18070 (-12)OKRuntime varied by 16.3% (threshold: 30.0%)nholmberCDFT: Tidy input descriptions
18069 (-13)OKRuntime varied by 24.1% (threshold: 30.0%)mkrackAdd cell optimisation constraint and 3 new regression tests
18068 (-14)OKRuntime varied by 8.0% (threshold: 30.0%)sandermattemd: fix input description
18067 (-15)mkrackDrop ACML and advance to GFortran 6.4
18066 (-16)mkrackDrop ACML and advance to GFortran 6.4
18065 (-17)nholmberWavelet: Prevent array access when index out of range
18064 (-18)sandermattlbfgs: fix geometry optimisations with a trust radius of exactly 0
18063 (-19)OKRuntime varied by 13.7% (threshold: 30.0%)agloessTOOLCHAIN: Correct description and trigger usage of new current-toolchain (r18061).
18062 (-20)OKRuntime varied by 10.0% (threshold: 30.0%)rybkinjrenable HF/MP2 embedded in DFT and a corresponding test
18061 (-21)agloessTOOLCHAIN: Partially revert r18060.
18060 (-22)agloessTOOLCHAIN: Add AMD-Ryzen support by OpenBLAS upgrade.
18059 (-23)OKRuntime varied by 21.1% (threshold: 30.0%)hbani(implicit ps) apply unit conversion on default values.
18058 (-24)FAILEDRuntime varied by 30.6% (threshold: 30.0%)alazzaroDBCSR: change report from elements to bytes
18057 (-25)FAILEDRuntime varied by 31.0% (threshold: 30.0%)rybkinjradd libint specification in TEST_DIRS for embedding test
18056 (-26)OKRuntime varied by 14.9% (threshold: 30.0%)rybkinjrembedding debug and test
18055 (-27)mkrackUpdate PSI ssmp regression tester
18054 (-28)hbaniclarify description.
18053 (-29)mkrackUpdate PSI psmp regresssion tester
18052 (-30)mkrackUpdate PSI pdbg regression tester
18051 (-31)OKRuntime varied by 14.5% (threshold: 30.0%)rybkinjrformatting
18050 (-32)rybkinjrdelete embed for now
18049 (-33)rybkinjranother fix
18048 (-34)rybkinjrfixed test
18047 (-35)rybkinjrbug fix and test restore in embedding
18046 (-36)OKRuntime varied by 19.0% (threshold: 30.0%)ibethuneFix uninitialised value
18045 (-37)OKRuntime varied by 26.7% (threshold: 30.0%)oschuettlibint: remove outdated note
18044 (-38)OKRuntime varied by 14.7% (threshold: 30.0%)alazzaroDBCSR: fix an overflow in the final statistics
18043 (-39)OKRuntime varied by 17.5% (threshold: 30.0%)jhutterLRI: distant pair approximation (basic implementation)
18042 (-40)OKRuntime varied by 16.4% (threshold: 30.0%)nholmberCDFT: One more tolerance adjustment
18041 (-41)nholmberCDFT: Increase tolerances
18040 (-42)OKRuntime varied by 9.1% (threshold: 30.0%)rybkinjrremoved the nasty test until full debugging
18039 (-43)OKRuntime varied by 29.2% (threshold: 30.0%)nholmberCDFT: Adjust tolerances
18038 (-44)nholmberCDFT: Add spin channel specific constraints
18037 (-45)OKRuntime varied by 27.9% (threshold: 30.0%)rybkinjrimprove parallel DFT embedding, added a test
18036 (-46)alazzarofix a typo in the previous commit
18035 (-47)alazzarointroducing a flag to disable timings: GLOBAL%TIMINGS%TIMINGS_LEVEL. Default 1=all timings are reported. Set to 0 to disable it
18034 (-48)OKRuntime varied by 10.2% (threshold: 30.0%)pseewaldERI MME: increase test tolerance
18033 (-49)OKRuntime varied by 17.6% (threshold: 30.0%)pseewaldERI MME: more stable tests for numerical stress
18032 (-50)OKRuntime varied by 10.5% (threshold: 30.0%)pseewaldERI MME: simplify loop to avoid false valgrind positive
18031 (-51)pseewaldERI MME: increase test tolerance for numerical stress
18030 (-52)agloessERI MME: Remove unused debug parameter.
18029 (-53)OKRuntime varied by 17.1% (threshold: 30.0%)pseewaldERI MME: efficient implementation for non-ortho cells.
18028 (-54)alazzaromove regtest-ot-1 vib tests in a specific directory, do not run when mpiranks>4
18027 (-55)alazzaroadjust tol. for 8 ranks
18026 (-56)alazzarofix one more valgrind problems with uninitialised values, by specifying ALLOCATABLE arrays
18025 (-57)agloessDBCSR: (Re)enable assert for missing CUDA kernels.
18024 (-58)OKRuntime varied by 7.1% (threshold: 30.0%)agloessPrettify.
18023 (-59)alazzarofix valgrind problems with uninitialised values, by replacing POINTER wtih ALLOCATABLE arrays
18022 (-60)agloessDBCSR: Revised CUDA-kernels and parameter heuristic. Minimize stack-data traffic. Add new P100 parameter sets - raw kernel speedup 1.0 < s < 1.8 (larger for small block sizes).
18021 (-61)jhutterLRIGPW: bug fixes for memory leaks and OpenMP dependencies
18020 (-62)OKRuntime varied by 9.4% (threshold: 30.0%)alazzaroenable tests for single rank
18019 (-63)OKRuntime varied by 14.8% (threshold: 30.0%)alazzarofix conventions
18018 (-64)alazzaromove regtests-5 tests without refs in a separate directory
18017 (-65)rybkinjrDFT embedding added
18016 (-66)alazzaroDBCSR: improve performance in make_images
18015 (-67)nholmberCDFT: Enable SVD without SCALAPACK, reenable tests for serial binaries
18014 (-68)nholmberCDFT: Disable new tests for serial binaries
18013 (-69)nholmberCDFT: Fix undefined symbol
18012 (-70)nholmberCDFT: Add support for diagonalization and fractionally occupied orbitals
18011 (-71)OKRuntime varied by 16.6% (threshold: 30.0%)jhutterLRIGPW: Critical section, reset one regtest
18010 (-72)jhutterLRIGPW: Allocation bug fix for OpenMP
18009 (-73)jhutterLRIGPW: screening for three index integral contraction, OpenMP code
18008 (-74)OKRuntime varied by 20.2% (threshold: 30.0%)jhutterLRIGPW: refactoring of KS build
18007 (-75)jhutterremove debug statement
18006 (-76)jhutterLRIGPW: refactoring of memory usage, reduce memory footprint
18005 (-77)OKRuntime varied by 8.7% (threshold: 30.0%)jhutterSome optimizations on abc_contract
18004 (-78)ibethuneAdapt auto-regtest for non-Linux OS
18003 (-79)OKRuntime varied by 11.9% (threshold: 30.0%)jhutterFist + peridoic displacement field: bug fix forces
18002 (-80)OKRuntime varied by 20.0% (threshold: 30.0%)jhutterRIGPW: minor adjustments
18001 (-81)OKRuntime varied by 15.8% (threshold: 30.0%)jhutterRIGPW: forces and parallelization
18000 (-82)OKRuntime varied by 12.1% (threshold: 30.0%)sandermattfix -- appearing in filenames, by removing empty middle_name strings
17999 (-83)sandermattenable forces printing during geo_opt and cell_opt
17998 (-84)OKRuntime varied by 7.8% (threshold: 30.0%)jhutterPrettify
17997 (-85)jhutterMore towards a complete RIGPW method implementation
17996 (-86)OKRuntime varied by 13.7% (threshold: 30.0%)jhutterMC cluster moves bug fix + 1 new regtest (Himanshu Goel)
17995 (-87)OKRuntime varied by 11.2% (threshold: 30.0%)hbanifix typo.
17994 (-88)OKRuntime varied by 10.5% (threshold: 30.0%)oschuettFix conventions
17993 (-89)oschuettPIMD: Fix overflow
17992 (-90)OKRuntime varied by 12.1% (threshold: 30.0%)jhutterTowards RI-GPW method, LRI basis input only with BASIS_SET possible.
17991 (-91)OKRuntime varied by 17.0% (threshold: 30.0%)oschuettDocker: Simplify and use CSCS-compatible mpich version
17990 (-92)jhutterFix Bug #46 (Alrik Stegmaier): increase string length for XC reference
17989 (-93)hbaniclarify input description.
17988 (-94)OKRuntime varied by 15.1% (threshold: 30.0%)oschuettAdd Dockerfiles
17987 (-95)hforbertIncreasing PILE sampling Efficency [Felix Uhl / Fabien Brieuc]
17986 (-96)OKRuntime varied by 16.4% (threshold: 30.0%)hbanilone keywords.
17985 (-97)OKRuntime varied by 8.9% (threshold: 30.0%)hbanifix access to uninitialized environment.
17984 (-98)OKRuntime varied by 15.3% (threshold: 30.0%)jhutterName special k-points in band structure calculation
17983 (-99)jhutterAll electron crystal basis sets (Bredow, Bonn)

View full archive

Page last updated: 2017-10-17T11:24:31