openff-bespokefit
openff-bespokefit copied to clipboard
openforcefield
Optimizations failing on large macromolecule
Description
During the BespokeFit followup workshop, @rvkrishnan30 pointed out that the S1A ligand takes >24 hours during ForceBalance. I haven't been able to reproduce this yet because my XTB calculations are slow for some reason but I will soon.
Reproduction
I'm attempting to reproduce in this gist
Output
@rvkrishnan30 any details you could provide on this would be great :)
Software versions
- Which operating system and version are you using?
- How did you install BespokeFit?
- What is the output of running
conda list?
Hi @Yoshanuikabundi , as suggested during the follow-up workshop, I loosened the convergence criteria for ForceBalance to check if I get any speed improvement without loss of accuracy.
By changing both convergence objective and convergence step from 0.01 to 1.00, the speed improved from ~40 hrs to ~19 hrs without much loss in accuracy as from the K values and the resultant torsion plots.
I am sure even this can be further improved in terms of speed without issues and suspect something is taking much longer than it has to slowing down the entire FB fitting process. I see this kind of a behaviour with a few other unrelated ligand molecules too.
Could you please suggest if there is a better way to debug and narrow down to what's causing the issue, and also possibly what other parameters can be tuned to avoid this behaviour.
Below is my example input file passed to ForceBalance.
$options ffdir forcefield penalty_type L1 jobtype optimize forcefield force-field.offxml
maxstep 10 convergence_step 1.00 convergence_objective 1.00 convergence_gradient 0.01 criteria 2 eig_lowerbound 0.01 finite_difference_h 0.01 penalty_additive 1.0
trust0 -0.25 mintrust 0.05 error_tolerance 1.0 adaptive_factor 0.2 adaptive_damping 1.0 normalize_weights False constrain_charge false
priors ProperTorsions/Proper/k : 6.0 /priors
$end
$target name ligand_fragment_around_22_24 weight 1.0
type TorsionProfile_SMIRNOFF mol2 ligand_fragment_around_22_24.sdf pdb ligand_fragment_around_22_24.pdb coords scan.xyz writelevel 2 attenuate 1
energy_denom 1.0 energy_upper 10.0 $end
@rvkrishnan30 I'm getting to this today, but I'm unable to reproduce the slow runtime on this molecule - it just ran using an openeye-less env in a little over 1 hour (20 minutes fragmentation, 20 minutes QC generation, 20 minutes parameter optimization) on my macbook air.
Could you upload the complete forcebalance inputs (sdf, pdb, scan.xyz, force-field.offxml) so I can see if that reproduces?
My whole environment is:
Full env
(bespokefit) jw@mba$ conda list
# packages in environment at /Users/jeffreywagner/conda/envs/bespokefit:
#
# Name Version Build Channel
amberlite 16.0 pypi_0 pypi
ambertools 21.11 py39hf80593e_0 conda-forge
amberutils 21.0 pypi_0 pypi
amqp 5.1.1 pyhd8ed1ab_0 conda-forge
anyio 3.6.2 pyhd8ed1ab_0 conda-forge
appnope 0.1.3 pyhd8ed1ab_0 conda-forge
argcomplete 1.12.3 pyhd8ed1ab_0 conda-forge
argon2-cffi 21.3.0 pyhd8ed1ab_0 conda-forge
argon2-cffi-bindings 21.2.0 py39ha30fb19_3 conda-forge
arpack 3.7.0 hefb7bc6_2 conda-forge
arrow-cpp 9.0.0 py39h7d8c460_10_cpu conda-forge
asttokens 2.1.0 pyhd8ed1ab_0 conda-forge
astunparse 1.6.3 pyhd8ed1ab_0 conda-forge
async-timeout 4.0.2 pyhd8ed1ab_0 conda-forge
attrs 22.1.0 pyh71513ae_1 conda-forge
aws-c-cal 0.5.11 hd2e2f4b_0 conda-forge
aws-c-common 0.6.2 h0d85af4_0 conda-forge
aws-c-event-stream 0.2.7 hb9330a7_13 conda-forge
aws-c-io 0.10.5 h35aa462_0 conda-forge
aws-checksums 0.1.11 h0010a65_7 conda-forge
aws-sdk-cpp 1.8.186 h7c85d8e_4 conda-forge
backcall 0.2.0 pyh9f0ad1d_0 conda-forge
backports 1.0 py_2 conda-forge
backports.functools_lru_cache 1.6.4 pyhd8ed1ab_0 conda-forge
basis_set_exchange 0.9 pyhd8ed1ab_0 conda-forge
beautifulsoup4 4.11.1 pyha770c72_0 conda-forge
billiard 3.6.4.0 py39ha30fb19_3 conda-forge
bleach 5.0.1 pyhd8ed1ab_0 conda-forge
blosc 1.21.1 h97e831e_3 conda-forge
boost 1.74.0 py39ha1f3e3e_5 conda-forge
boost-cpp 1.74.0 h8b082ac_8 conda-forge
brotli 1.0.9 hb7f2c08_8 conda-forge
brotli-bin 1.0.9 hb7f2c08_8 conda-forge
brotlipy 0.7.0 py39ha30fb19_1005 conda-forge
bzip2 1.0.8 h0d85af4_4 conda-forge
c-ares 1.18.1 h0d85af4_0 conda-forge
ca-certificates 2022.9.24 h033912b_0 conda-forge
cached-property 1.5.2 hd8ed1ab_1 conda-forge
cached_property 1.5.2 pyha770c72_1 conda-forge
cachetools 5.2.0 pyhd8ed1ab_0 conda-forge
cairo 1.16.0 h904041c_1014 conda-forge
celery 5.2.7 pyhd8ed1ab_0 conda-forge
certifi 2022.9.24 pyhd8ed1ab_0 conda-forge
cffi 1.15.1 py39h131948b_2 conda-forge
charset-normalizer 2.1.1 pyhd8ed1ab_0 conda-forge
chemper 1.0.1 pyhd8ed1ab_0 conda-forge
click 8.1.3 unix_pyhd8ed1ab_2 conda-forge
click-didyoumean 0.3.0 pyhd8ed1ab_0 conda-forge
click-option-group 0.5.3 pyhd8ed1ab_1 conda-forge
click-plugins 1.1.1 py_0 conda-forge
click-repl 0.2.0 pyhd8ed1ab_0 conda-forge
colorama 0.4.6 pyhd8ed1ab_0 conda-forge
commonmark 0.9.1 py_0 conda-forge
conda 22.9.0 py39h6e9494a_2 conda-forge
conda-package-handling 1.9.0 py39ha30fb19_1 conda-forge
contourpy 1.0.6 py39h92daf61_0 conda-forge
cryptography 38.0.3 py39h7eb6a14_0 conda-forge
curl 7.86.0 h57eb407_1 conda-forge
cycler 0.11.0 pyhd8ed1ab_0 conda-forge
cython 0.29.32 py39h7a8716b_1 conda-forge
dataclasses 0.8 pyhc8e2a94_3 conda-forge
debugpy 1.6.3 py39h7a8716b_1 conda-forge
decorator 5.1.1 pyhd8ed1ab_0 conda-forge
defusedxml 0.7.1 pyhd8ed1ab_0 conda-forge
deprecated 1.2.13 pyh6c4a22f_0 conda-forge
entrypoints 0.4 pyhd8ed1ab_0 conda-forge
executing 1.2.0 pyhd8ed1ab_0 conda-forge
expat 2.5.0 hf0c8a7f_0 conda-forge
fastapi 0.87.0 pyhd8ed1ab_0 conda-forge
fftw 3.3.10 nompi_h4fa670e_105 conda-forge
flit-core 3.8.0 pyhd8ed1ab_0 conda-forge
fmt 9.1.0 hb8565cd_0 conda-forge
font-ttf-dejavu-sans-mono 2.37 hab24e00_0 conda-forge
font-ttf-inconsolata 3.000 h77eed37_0 conda-forge
font-ttf-source-code-pro 2.038 h77eed37_0 conda-forge
font-ttf-ubuntu 0.83 hab24e00_0 conda-forge
fontconfig 2.14.1 h5bb23bf_0 conda-forge
fonts-conda-ecosystem 1 0 conda-forge
fonts-conda-forge 1 0 conda-forge
fonttools 4.38.0 py39ha30fb19_1 conda-forge
forcebalance 1.9.3 py39h6dc771e_2 conda-forge
freetype 2.12.1 h3f81eb7_0 conda-forge
future 0.18.2 pyhd8ed1ab_6 conda-forge
geometric 1.0 pyhd8ed1ab_0 conda-forge
gettext 0.21.1 h8a4c099_0 conda-forge
gflags 2.2.2 hb1e8313_1004 conda-forge
glog 0.6.0 h8ac2a54_0 conda-forge
greenlet 2.0.1 py39h7a8716b_0 conda-forge
h11 0.12.0 pyhd8ed1ab_0 conda-forge
h2 4.1.0 pyhd8ed1ab_0 conda-forge
h5py 3.1.0 nompi_py39hdc2b67d_100 conda-forge
hdf4 4.2.15 h7aa5921_5 conda-forge
hdf5 1.10.6 nompi_hc5d9132_1114 conda-forge
hpack 4.0.0 pyh9f0ad1d_0 conda-forge
httpcore 0.15.0 pyhd8ed1ab_0 conda-forge
httpx 0.23.0 py39h6e9494a_2 conda-forge
hyperframe 6.0.1 pyhd8ed1ab_0 conda-forge
icu 70.1 h96cf925_0 conda-forge
idna 3.4 pyhd8ed1ab_0 conda-forge
importlib-metadata 5.0.0 pyha770c72_1 conda-forge
importlib_resources 5.10.0 pyhd8ed1ab_0 conda-forge
ipykernel 6.17.1 pyh736e0ef_0 conda-forge
ipython 8.6.0 pyhd1c38e8_1 conda-forge
ipython_genutils 0.2.0 py_1 conda-forge
ipywidgets 8.0.2 pyhd8ed1ab_1 conda-forge
jedi 0.18.1 pyhd8ed1ab_2 conda-forge
jinja2 3.1.2 pyhd8ed1ab_1 conda-forge
jpeg 9e hac89ed1_2 conda-forge
jsonschema 4.17.0 pyhd8ed1ab_0 conda-forge
jupyter_client 7.4.6 pyhd8ed1ab_0 conda-forge
jupyter_core 5.0.0 py39h6e9494a_0 conda-forge
jupyter_server 1.23.2 pyhd8ed1ab_0 conda-forge
jupyterlab_pygments 0.2.2 pyhd8ed1ab_0 conda-forge
jupyterlab_widgets 3.0.3 pyhd8ed1ab_0 conda-forge
khronos-opencl-icd-loader 2022.09.30 hb7f2c08_2 conda-forge
kiwisolver 1.4.4 py39h92daf61_1 conda-forge
kombu 5.2.4 py39h6e9494a_2 conda-forge
krb5 1.19.3 hb49756b_0 conda-forge
lcms2 2.14 h90f4b2a_0 conda-forge
lerc 4.0.0 hb486fe8_0 conda-forge
libabseil 20220623.0 cxx17_h844d122_5 conda-forge
libarchive 3.5.2 hde4784d_3 conda-forge
libblas 3.9.0 16_osx64_openblas conda-forge
libbrotlicommon 1.0.9 hb7f2c08_8 conda-forge
libbrotlidec 1.0.9 hb7f2c08_8 conda-forge
libbrotlienc 1.0.9 hb7f2c08_8 conda-forge
libcblas 3.9.0 16_osx64_openblas conda-forge
libcrc32c 1.1.2 he49afe7_0 conda-forge
libcurl 7.86.0 h57eb407_1 conda-forge
libcxx 14.0.6 hccf4f1f_0 conda-forge
libdeflate 1.14 hb7f2c08_0 conda-forge
libedit 3.1.20191231 h0678c8f_2 conda-forge
libev 4.33 haf1e3a3_1 conda-forge
libevent 2.1.10 h815e4d9_4 conda-forge
libffi 3.4.2 h0d85af4_5 conda-forge
libgfortran 5.0.0 9_5_0_h97931a8_26 conda-forge
libgfortran5 11.3.0 h082f757_26 conda-forge
libglib 2.74.1 h4c723e1_1 conda-forge
libgoogle-cloud 2.3.0 hb6a50ef_1 conda-forge
libgrpc 1.49.1 h834a566_1 conda-forge
libiconv 1.17 hac89ed1_0 conda-forge
liblapack 3.9.0 16_osx64_openblas conda-forge
libmamba 1.0.0 h2bf831e_2 conda-forge
libmambapy 1.0.0 py39he069e75_2 conda-forge
libnetcdf 4.8.1 nompi_hb4d10b0_100 conda-forge
libnghttp2 1.47.0 h7cbc4dc_1 conda-forge
libopenblas 0.3.21 openmp_h429af6e_3 conda-forge
libpng 1.6.38 ha978bb4_0 conda-forge
libprotobuf 3.21.9 hbc0c0cd_0 conda-forge
libsodium 1.0.18 hbcb3906_1 conda-forge
libsolv 0.7.22 hd9580d2_0 conda-forge
libsqlite 3.39.4 ha978bb4_0 conda-forge
libssh2 1.10.0 h7535e13_3 conda-forge
libthrift 0.16.0 h08c06f4_2 conda-forge
libtiff 4.4.0 hdb44e8a_4 conda-forge
libutf8proc 2.8.0 hb7f2c08_0 conda-forge
libwebp-base 1.2.4 h775f41a_0 conda-forge
libxcb 1.13 h0d85af4_1004 conda-forge
libxml2 2.10.3 hb9e07b5_0 conda-forge
libxslt 1.1.37 h5d22bc9_0 conda-forge
libzip 1.9.2 h3ad4413_1 conda-forge
libzlib 1.2.13 hfd90126_4 conda-forge
llvm-openmp 15.0.4 h61d9ccf_0 conda-forge
lxml 4.9.1 py39hfbce9ca_1 conda-forge
lz4-c 1.9.3 he49afe7_1 conda-forge
lzo 2.10 haf1e3a3_1000 conda-forge
mamba 1.0.0 py39ha435c47_2 conda-forge
markupsafe 2.1.1 py39ha30fb19_2 conda-forge
matplotlib-base 3.6.2 py39hb2f573b_0 conda-forge
matplotlib-inline 0.1.6 pyhd8ed1ab_0 conda-forge
mctc-lib 0.3.1 h8dc6bf7_0 conda-forge
mdtraj 1.9.7 py39h5456c6e_4 conda-forge
mistune 2.0.4 pyhd8ed1ab_0 conda-forge
mmpbsa-py 16.0 pypi_0 pypi
mock 4.0.3 pyhd8ed1ab_4 conda-forge
msgpack-python 1.0.4 py39h92daf61_1 conda-forge
munkres 1.1.4 pyh9f0ad1d_0 conda-forge
nbclassic 0.4.8 pyhd8ed1ab_0 conda-forge
nbclient 0.7.0 pyhd8ed1ab_0 conda-forge
nbconvert 7.2.5 pyhd8ed1ab_0 conda-forge
nbconvert-core 7.2.5 pyhd8ed1ab_0 conda-forge
nbconvert-pandoc 7.2.5 pyhd8ed1ab_0 conda-forge
nbformat 5.7.0 pyhd8ed1ab_0 conda-forge
ncurses 6.3 h96cf925_1 conda-forge
nest-asyncio 1.5.6 pyhd8ed1ab_0 conda-forge
netcdf-fortran 4.5.3 nompi_hdf192a6_105 conda-forge
networkx 2.8.8 pyhd8ed1ab_0 conda-forge
nglview 3.0.3 pyh8a188c0_0 conda-forge
notebook 6.5.2 pyha770c72_1 conda-forge
notebook-shim 0.2.2 pyhd8ed1ab_0 conda-forge
numexpr 2.8.3 py39hecff1ad_1 conda-forge
numpy 1.23.4 py39hdfa1d0c_1 conda-forge
ocl_icd_wrapper_apple 1.0.0 hbcb3906_0 conda-forge
openff-bespokefit 0.1.2 pyhd8ed1ab_0 conda-forge
openff-forcefields 2.0.0 pyh6c4a22f_0 conda-forge
openff-fragmenter-base 0.1.2 pyhd8ed1ab_1 conda-forge
openff-qcsubmit 0.3.2 pyhd8ed1ab_0 conda-forge
openff-toolkit-base 0.10.6 pyhd8ed1ab_0 conda-forge
openff-utilities 0.1.7 pyh1a96a4e_0 conda-forge
openjpeg 2.5.0 h5d0d7b0_1 conda-forge
openmm 7.7.0 py39h8d72adf_0_khronos conda-forge
openssl 1.1.1s hfd90126_0 conda-forge
orc 1.8.0 ha9d861c_0 conda-forge
packaging 21.3 pyhd8ed1ab_0 conda-forge
packmol 20.010 h508aa58_0 conda-forge
packmol-memgen 1.2.1rc0 pypi_0 pypi
pandas 1.5.1 py39hecff1ad_1 conda-forge
pandoc 2.19.2 h694c41f_1 conda-forge
pandocfilters 1.5.0 pyhd8ed1ab_0 conda-forge
parmed 3.4.3 py39h7a8716b_3 conda-forge
parquet-cpp 1.5.1 2 conda-forge
parso 0.8.3 pyhd8ed1ab_0 conda-forge
pcre2 10.40 h1c4e4bc_0 conda-forge
pdb4amber 20.1 pypi_0 pypi
perl 5.32.1 2_h0d85af4_perl5 conda-forge
pexpect 4.8.0 pyh1a96a4e_2 conda-forge
pickleshare 0.7.5 py_1003 conda-forge
pillow 9.2.0 py39h35d4919_3 conda-forge
pint 0.20.1 pyhd8ed1ab_0 conda-forge
pip 22.3.1 pyhd8ed1ab_0 conda-forge
pixman 0.40.0 hbcb3906_0 conda-forge
pkgutil-resolve-name 1.3.10 pyhd8ed1ab_0 conda-forge
platformdirs 2.5.2 pyhd8ed1ab_1 conda-forge
plotly 5.11.0 pyhd8ed1ab_0 conda-forge
prometheus_client 0.15.0 pyhd8ed1ab_0 conda-forge
prompt-toolkit 3.0.32 pyha770c72_0 conda-forge
prompt_toolkit 3.0.32 hd8ed1ab_0 conda-forge
psutil 5.9.4 py39ha30fb19_0 conda-forge
pthread-stubs 0.4 hc929b4f_1001 conda-forge
ptyprocess 0.7.0 pyhd3deb0d_0 conda-forge
pure_eval 0.2.2 pyhd8ed1ab_0 conda-forge
py-cpuinfo 9.0.0 pyhd8ed1ab_0 conda-forge
pyarrow 9.0.0 py39hb941c77_10_cpu conda-forge
pybind11-abi 4 hd8ed1ab_3 conda-forge
pycairo 1.21.0 py39h41776c8_2 conda-forge
pycosat 0.6.4 py39ha30fb19_1 conda-forge
pycparser 2.21 pyhd8ed1ab_0 conda-forge
pydantic 1.10.2 py39ha30fb19_1 conda-forge
pygments 2.13.0 pyhd8ed1ab_0 conda-forge
pymbar 3.1.0 py39h7cc1f47_1 conda-forge
pyopenssl 22.1.0 pyhd8ed1ab_0 conda-forge
pyparsing 3.0.9 pyhd8ed1ab_0 conda-forge
pyrsistent 0.19.2 py39ha30fb19_0 conda-forge
pysocks 1.7.1 pyha2e5f31_6 conda-forge
pytables 3.6.1 py39hd07922a_3 conda-forge
python 3.9.13 h57e37ff_0_cpython conda-forge
python-constraint 1.4.0 py_0 conda-forge
python-dateutil 2.8.2 pyhd8ed1ab_0 conda-forge
python-fastjsonschema 2.16.2 pyhd8ed1ab_0 conda-forge
python_abi 3.9 2_cp39 conda-forge
pytraj 2.0.6 pypi_0 pypi
pytz 2022.6 pyhd8ed1ab_0 conda-forge
pyyaml 6.0 py39ha30fb19_5 conda-forge
pyzmq 24.0.1 py39hed8f129_1 conda-forge
qcelemental 0.25.1 pyhd8ed1ab_1 conda-forge
qcengine 0.25.0 pyhd8ed1ab_0 conda-forge
qcportal 0.15.8 pyhd8ed1ab_0 conda-forge
rdkit 2022.09.1 py39h67dd817_0 conda-forge
re2 2022.06.01 hb486fe8_0 conda-forge
readline 8.1.2 h3899abd_0 conda-forge
redis-py 4.3.4 pyhd8ed1ab_0 conda-forge
redis-server 7.0.0 h25ffcba_0 conda-forge
regex 2022.10.31 py39ha30fb19_0 conda-forge
reportlab 3.5.68 py39hf37cc50_1 conda-forge
reproc 14.2.3 h0d85af4_0 conda-forge
reproc-cpp 14.2.3 he49afe7_0 conda-forge
requests 2.28.1 pyhd8ed1ab_1 conda-forge
rfc3986 1.5.0 pyhd8ed1ab_0 conda-forge
rich 12.6.0 pyhd8ed1ab_0 conda-forge
ruamel_yaml 0.15.80 py39ha30fb19_1008 conda-forge
sander 16.0 pypi_0 pypi
scipy 1.9.3 py39h8a15683_2 conda-forge
send2trash 1.8.0 pyhd8ed1ab_0 conda-forge
setuptools 65.5.1 pyhd8ed1ab_0 conda-forge
six 1.16.0 pyh6c4a22f_0 conda-forge
smirnoff99frosst 1.1.0 pyh44b312d_0 conda-forge
snappy 1.1.9 h225ccf5_2 conda-forge
sniffio 1.3.0 pyhd8ed1ab_0 conda-forge
soupsieve 2.3.2.post1 pyhd8ed1ab_0 conda-forge
sqlalchemy 1.4.44 py39ha30fb19_0 conda-forge
sqlite 3.39.4 h9ae0607_0 conda-forge
stack_data 0.6.1 pyhd8ed1ab_0 conda-forge
starlette 0.21.0 pyhd8ed1ab_0 conda-forge
tenacity 8.1.0 pyhd8ed1ab_0 conda-forge
terminado 0.17.0 pyhd1c38e8_0 conda-forge
tinycss2 1.2.1 pyhd8ed1ab_0 conda-forge
tk 8.6.12 h5dbffcc_0 conda-forge
toolz 0.12.0 pyhd8ed1ab_0 conda-forge
tornado 6.2 py39ha30fb19_1 conda-forge
torsiondrive 1.1.0 pyhd8ed1ab_0 conda-forge
tqdm 4.64.1 pyhd8ed1ab_0 conda-forge
traitlets 5.5.0 pyhd8ed1ab_0 conda-forge
typing-extensions 4.4.0 hd8ed1ab_0 conda-forge
typing_extensions 4.4.0 pyha770c72_0 conda-forge
tzdata 2022f h191b570_0 conda-forge
unicodedata2 15.0.0 py39ha30fb19_0 conda-forge
urllib3 1.26.11 pyhd8ed1ab_0 conda-forge
uvicorn 0.19.0 py39h6e9494a_0 conda-forge
vine 5.0.0 pyhd8ed1ab_1 conda-forge
wcwidth 0.2.5 pyh9f0ad1d_2 conda-forge
webencodings 0.5.1 py_1 conda-forge
websocket-client 1.4.2 pyhd8ed1ab_0 conda-forge
wheel 0.38.4 pyhd8ed1ab_0 conda-forge
widgetsnbextension 4.0.3 pyhd8ed1ab_0 conda-forge
wrapt 1.14.1 py39ha30fb19_1 conda-forge
xmltodict 0.13.0 pyhd8ed1ab_0 conda-forge
xorg-kbproto 1.0.7 h35c211d_1002 conda-forge
xorg-libice 1.0.10 h0d85af4_0 conda-forge
xorg-libsm 1.2.3 h0d85af4_1000 conda-forge
xorg-libx11 1.7.2 h0d85af4_0 conda-forge
xorg-libxau 1.0.9 h35c211d_0 conda-forge
xorg-libxdmcp 1.1.3 h35c211d_0 conda-forge
xorg-libxext 1.3.4 h0d85af4_1 conda-forge
xorg-libxt 1.2.1 h0d85af4_2 conda-forge
xorg-xextproto 7.3.0 h35c211d_1002 conda-forge
xorg-xproto 7.0.31 h35c211d_1007 conda-forge
xtb 6.5.1 h877ed2f_0 conda-forge
xtb-python 20.2 py39h701faf5_5 conda-forge
xz 5.2.6 h775f41a_0 conda-forge
yaml 0.2.5 h0d85af4_2 conda-forge
yaml-cpp 0.7.0 hf0c8a7f_2 conda-forge
zeromq 4.3.4 he49afe7_1 conda-forge
zipp 3.10.0 pyhd8ed1ab_0 conda-forge
zlib 1.2.13 hfd90126_4 conda-forge
zstd 1.5.2 hfa58983_4 conda-forge
Here are my inputs, for me this finishes in about 500 seconds using ForceBalance 1.9.2 and 1.9.3.
@rvkrishnan30 I'm getting to this today, but I'm unable to reproduce the slow runtime on this molecule - it just ran using an openeye-less env in a little over 1 hour (20 minutes fragmentation, 20 minutes QC generation, 20 minutes parameter optimization) on my macbook air.
Could you upload the complete forcebalance inputs (sdf, pdb, scan.xyz, force-field.offxml) so I can see if that reproduces?
My whole environment is:
Full env
@j-wags Thank you for testing this and sharing your inputs. This looks like a more reasonable timescale for ForceBalance runs. With your inputs, for me it took around 700 seconds and I am happy with it.
However, I wonder why it is taking 20 hrs in my actual case. Some differences I see between our inputs are:
- penalty L1 vs. L2
- convergence objective 0.01 vs. 0.1
- qdata with vs. without GRADIENTS in it
- more importantly the fragment molecule processed: I have all terminal groups -OH, -CH3 and -OCH3 groups in the fragment while you don't. I wonder if that's affecting the "referencing all energies to snapshot " step at the beginning of FB and also the MM optimizations during the run thereby slowing the whole process.
Attached are my inputs (sdf, pdb, scan.xyz, qdata.txt, and forcefield) and I hope this will help reproduce my case.
S1A_venkat_Nov2022.tar.gz
Ahh, yeah. It looks like the structure at targets/3gid_D_S1A/input.sdf in your inputs has 2 hydroxyls, 3 methyls, and 4 methoxys that mine doesn't have. Especially with the hydrogens on the methyl/methoxy groups, that would add a lot more particles to optimize in the minimizations during the torsion fitting.
So fundamentally, we will eventually find macrocycles large enough to ruin things as long as we keep requiring that the entire ring gets included in the fragment we use for fitting. But the macrocycle here seems to be manageable, it's just that the substituents are somehow hanging on in your input, even after fragmentation.
Could you share your bespokefit run command, the output of running conda list, and input SDF/SMILES for the entire run? It looks like we're getting different results from the fragmentation step, so I can start debugging from there.
Could you share your bespokefit run command, the output of running conda list, and input SDF/SMILES for the entire run? It looks like we're getting different results from the fragmentation step, so I can start debugging from there.
I have a slightly different implementation of bespokefit without the use of qcsubmit as all data are locally generated. But the workflow and settings should predominantly be the same. Having said that, I am not sure why we end up with different fragments. With Ambertools WBO based fragmentation, I end up with a slightly smaller fragment than the parent but still have the -OH, -CH3 groups. (Attached are the original and fragment sdf files)
I wonder why it picks up these groups though they should not affect the WBO at the torsion we are talking about.
Ahh, yeah. It looks like the structure at
targets/3gid_D_S1A/input.sdfin your inputs has 2 hydroxyls, 3 methyls, and 4 methoxys that mine doesn't have. Especially with the hydrogens on the methyl/methoxy groups, that would add a lot more particles to optimize in the minimizations during the torsion fitting.
I'm not convinced that just having an extra 13 heavy atoms in the system should be enough to make the runtime 60x longer: your optimisation stage took 20 mins, ours takes ~20 hours.
