I’m attempting to speed up model fitting by specifying the “threads” argument in brms. However, when I do so I receive the following error (at least, I think this is the only/relevant error. I find the output from compilation baffling, so I’m not sure what’s relevant):
C:/rtools43/ucrt64/bin/../lib/gcc/x86_64-w64-mingw32/13.2.0/../../../../x86_64-w64-mingw32/bin/ld.exe
: src/cmdstan/main_threads.o:main.cpp:(.text$_ZN4stan2io15stan_csv_reader13read_metadataERSiRNS0_17stan_csv_metadataEPSo[_ZN4stan2io15stan_csv_reader13read_metadataERSiRNS0_17stan_csv_metadataEPSo]+0x15d): undefined reference to `std::istream::seekg(std::fpos<int>)'
C:/rtools43/ucrt64/bin/../lib/gcc/x86_64-w64-mingw32/13.2.0/../../../../x86_64-w64-mingw32/bin/ld.exe: src/cmdstan/main_threads.o:main.cpp:(.text$_ZN4stan2io15stan_csv_reader15read_adaptationERSiRNS0_19stan_csv_adaptationEPSo[_ZN4stan2io15stan_csv_reader15read_adaptationERSiRNS0_19stan_csv_adaptationEPSo]+0x24e): undefined reference to `std::istream::seekg(std::fpos<int>)'
C:/rtools43/ucrt64/bin/../lib/gcc/x86_64-w64-mingw32/13.2.0/../../../../x86_64-w64-mingw32/bin/ld.exe: src/cmdstan/main_threads.o:main.cpp:(.text$_ZN4stan2io15stan_csv_reader12read_samplesERSiRN5Eigen6MatrixIdLin1ELin1ELi0ELin1ELin1EEERNS0_15stan_csv_timingEPSo[_ZN4stan2io15stan_csv_reader12read_samplesERSiRN5Eigen6MatrixIdLin1ELin1ELi0ELin1ELin1EEERNS0_15stan_csv_timingEPSo]+0x1d8): undefined reference to `std::istream::seekg(std::fpos<int>)'
collect2.exe: error: ld returned 1 exit status
mingw32-make: *** [make/program:59: C:\Users\ugv4\AppData\Local\Temp\2\RtmpKgtsyt\model-6274408520a8.exe] Error 1
Error: An error occured during compilation! See the message above for more information.
I’m able to replicate this error with a simple model, as well as with my own data.
brm(Sepal.Width ~ Sepal.Length, data = iris, chains = 3, threads = 3, cores = 3, backend = 'cmdstanr')
If I remove the threads = 3 argument the model compiles and runs fine.
I’ve installed cmdstan using the Conda installation instructions for Windows 10 and am using a conda environment with reticulate to work with the package in R. I tried using rebuild_cmdstan(), but that seemed to just bork my cmdstan installation and no models would compile, so I had to reinstall cmdstan with conda.
SessionInfo:
R version 4.3.2 (2023-10-31 ucrt)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 10 x64 (build 19045)
Matrix products: default
locale:
[1] LC_COLLATE=English_United States.utf8 LC_CTYPE=English_United States.utf8 LC_MONETARY=English_United States.utf8
[4] LC_NUMERIC=C LC_TIME=English_United States.utf8
time zone: America/New_York
tzcode source: internal
attached base packages:
[1] stats graphics grDevices datasets utils methods base
other attached packages:
[1] brms_2.20.4 Rcpp_1.0.11 cmdstanr_0.6.1 dplyr_1.1.4 stringr_1.5.1 sf_1.0-14 geostan_0.5.3
loaded via a namespace (and not attached):
[1] DBI_1.1.3 deldir_2.0-2 gridExtra_2.3 s2_1.1.4 inline_0.3.19 rlang_1.1.2 magrittr_2.0.3
[8] matrixStats_1.1.0 e1071_1.7-13 compiler_4.3.2 loo_2.6.0 reshape2_1.4.4 png_0.1-8 callr_3.7.3
[15] vctrs_0.6.4 pkgconfig_2.0.3 wk_0.9.0 crayon_1.5.2 fastmap_1.1.1 ellipsis_0.3.2 backports_1.4.1
[22] utf8_1.2.4 threejs_0.3.3 promises_1.2.1 markdown_1.11 ps_1.7.5 xfun_0.41 jsonlite_1.8.7
[29] later_1.3.1 parallel_4.3.2 prettyunits_1.2.0 R6_2.5.1 dygraphs_1.1.1.6 stringi_1.8.2 StanHeaders_2.26.28
[36] reticulate_1.34.0 boot_1.3-28.1 rstan_2.32.3 knitr_1.45 zoo_1.8-12 base64enc_0.1-3 bayesplot_1.10.0
[43] httpuv_1.6.12 Matrix_1.6-1.1 igraph_1.5.1 tidyselect_1.2.0 rstudioapi_0.15.0 abind_1.4-5 miniUI_0.1.1.1
[50] codetools_0.2-19 processx_3.8.2 pkgbuild_1.4.2 plyr_1.8.9 lattice_0.21-9 tibble_3.2.1 withr_2.5.2
[57] shiny_1.8.0 bridgesampling_1.1-2 posterior_1.5.0 coda_0.19-4 units_0.8-4 spData_2.3.0 proxy_0.4-27
[64] RcppParallel_5.1.7 xts_0.13.1 pillar_1.9.0 tensorA_0.36.2 KernSmooth_2.23-22 DT_0.30 checkmate_2.3.0
[71] renv_1.0.3 stats4_4.3.2 shinyjs_2.1.0 distributional_0.3.2 generics_0.1.3 rprojroot_2.0.4 sp_2.1-2
[78] truncnorm_1.0-9 ggplot2_3.4.4 rstantools_2.3.1.1 munsell_0.5.0 scales_1.2.1 gtools_3.9.5 xtable_1.8-4
[85] class_7.3-22 glue_1.6.2 tools_4.3.2 signs_0.1.2 shinystan_2.6.0 colourpicker_1.3.0 mvtnorm_1.2-4
[92] grid_4.3.2 spdep_1.3-1 crosstalk_1.2.1 QuickJSR_1.0.8 colorspace_2.1-0 nlme_3.1-163 cli_3.6.1
[99] fansi_1.0.5 Brobdingnag_1.2-9 gtable_0.3.4 digest_0.6.33 classInt_0.4-10 htmlwidgets_1.6.3 farver_2.1.1
[106] htmltools_0.5.7 lifecycle_1.0.4 here_1.0.1 mime_0.12 shinythemes_1.2.0 MASS_7.3-60