Error in irlba(mat, nv = 50) : BLAS/LAPACK routine 'DLASCL' gave error code -4

[levinhein]

Hi. How can I resolve this irlba error? When nv > 40 it always fail. Re-installing the irlba via install_github("bwlewis/irlba", force = TRUE) also did not change anything.

library(irlba)
mat <- matrix(data = rnorm(n = 3000*2000), nrow = 3000, ncol = 2000)
test <- irlba(mat, nv = 50)
test <- irlba(mat, nv = 40)
test <- irlba(mat, nv = 30)

[bwlewis]

Thanks for the reproducible code. Unfortunately, I can't reproduce the error yet (it's working OK for me). Can you report the output of the R function: sessionInfo()

…

On 6/27/23, Jan Vincent B. Beltran ***@***.***> wrote: Hi. How can I resolve this irlba error? When nv > 40 it always fail. Re-installing the irlba via `install_github("bwlewis/irlba", force = TRUE) ` also did not change anything. ``` library(irlba) mat <- matrix(data = rnorm(n = 3000*2000), nrow = 3000, ncol = 2000) test <- irlba(mat, nv = 50) test <- irlba(mat, nv = 40) test <- irlba(mat, nv = 30) ```  -- Reply to this email directly or view it on GitHub: #68 You are receiving this because you are subscribed to this thread. Message ID: ***@***.***>

[levinhein]

Dear @bwlewis,

Below is the result of sessionInfo(). I'm really stuck for more than a month already and it is depressing. I hope you can help me. We have already re-installed R (including r-core and r-core-base) as well as upgraded gfortran and re-installed irlba based on other's previous experiences, but all unsuccessful. :(

> sessionInfo()
R version 4.3.0 (2023-04-21)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 23.04

Matrix products: default
BLAS/LAPACK: /usr/lib/x86_64-linux-gnu/libmkl_rt.so;  LAPACK version 3.8.0

locale:
 [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C               LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8     LC_MONETARY=en_US.UTF-8   
 [6] LC_MESSAGES=en_US.UTF-8    LC_PAPER=en_US.UTF-8       LC_NAME=C                  LC_ADDRESS=C               LC_TELEPHONE=C            
[11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       

time zone: Asia/Taipei
tzcode source: system (glibc)

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] irlba_2.3.5.1  Matrix_1.5-4.1 devtools_2.4.5 usethis_2.2.1 

loaded via a namespace (and not attached):
 [1] miniUI_0.1.1.1    compiler_4.3.0    crayon_1.5.2      promises_1.2.0.1  Rcpp_1.0.10       stringr_1.5.0     callr_3.7.3       later_1.3.1      
 [9] fastmap_1.1.1     lattice_0.21-8    mime_0.12         R6_2.5.1          curl_5.0.1        htmlwidgets_1.6.2 desc_1.4.2        profvis_0.3.8    
[17] rprojroot_2.0.3   shiny_1.7.4       rlang_1.1.1       cachem_1.0.8      stringi_1.7.12    httpuv_1.6.11     fs_1.6.2          pkgload_1.3.2    
[25] memoise_2.0.1     cli_3.6.1         magrittr_2.0.3    ps_1.7.5          grid_4.3.0        digest_0.6.31     processx_3.8.1    rstudioapi_0.14  
[33] xtable_1.8-4      remotes_2.4.2     lifecycle_1.0.3   prettyunits_1.1.1 vctrs_0.6.3       glue_1.6.2        urlchecker_1.0.1  sessioninfo_1.2.2
[41] pkgbuild_1.4.2    purrr_1.0.1       tools_4.3.0       ellipsis_0.3.2    htmltools_0.5.5  
> 

[bwlewis]

Working on a solution...

…

On 6/28/23, Jan Vincent B. Beltran ***@***.***> wrote: Dear @bwlewis, Below is the result of sessionInfo(). I'm really stuck for more than a month already and it is depressing. I hope you can help me. We have already re-installed R (including r-core and r-core-base) as well as upgraded gfortran and re-installed irlba based on other's previous experiences, but all unsuccessful. :(  ``` > sessionInfo() R version 4.3.0 (2023-04-21) Platform: x86_64-pc-linux-gnu (64-bit) Running under: Ubuntu 23.04 Matrix products: default BLAS/LAPACK: /usr/lib/x86_64-linux-gnu/libmkl_rt.so; LAPACK version 3.8.0 locale: [1] LC_CTYPE=en_US.UTF-8 LC_NUMERIC=C LC_TIME=en_US.UTF-8 LC_COLLATE=en_US.UTF-8 LC_MONETARY=en_US.UTF-8 [6] LC_MESSAGES=en_US.UTF-8 LC_PAPER=en_US.UTF-8 LC_NAME=C LC_ADDRESS=C LC_TELEPHONE=C [11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C time zone: Asia/Taipei tzcode source: system (glibc) attached base packages: [1] stats graphics grDevices utils datasets methods base other attached packages: [1] irlba_2.3.5.1 Matrix_1.5-4.1 devtools_2.4.5 usethis_2.2.1 loaded via a namespace (and not attached): [1] miniUI_0.1.1.1 compiler_4.3.0 crayon_1.5.2 promises_1.2.0.1 Rcpp_1.0.10 stringr_1.5.0 callr_3.7.3 later_1.3.1 [9] fastmap_1.1.1 lattice_0.21-8 mime_0.12 R6_2.5.1 curl_5.0.1 htmlwidgets_1.6.2 desc_1.4.2 profvis_0.3.8 [17] rprojroot_2.0.3 shiny_1.7.4 rlang_1.1.1 cachem_1.0.8 stringi_1.7.12 httpuv_1.6.11 fs_1.6.2 pkgload_1.3.2 [25] memoise_2.0.1 cli_3.6.1 magrittr_2.0.3 ps_1.7.5 grid_4.3.0 digest_0.6.31 processx_3.8.1 rstudioapi_0.14 [33] xtable_1.8-4 remotes_2.4.2 lifecycle_1.0.3 prettyunits_1.1.1 vctrs_0.6.3 glue_1.6.2 urlchecker_1.0.1 sessioninfo_1.2.2 [41] pkgbuild_1.4.2 purrr_1.0.1 tools_4.3.0 ellipsis_0.3.2 htmltools_0.5.5 > ``` -- Reply to this email directly or view it on GitHub: #68 (comment) You are receiving this because you were mentioned. Message ID: ***@***.***>

[lijxug]

Hi, I've encountered this problem too, any idea why and how to solve it?

[lijxug]

Setting nv to 40 does fix the problem. But why...

[ZCai-02]

Dear @bwlewis,

I also encountered this issue when running a Seurat project with the RunPCA() function. After searching for solutions, I still could not find a proper one. I finally tested the irlba package with the above codes, but the terminal also crushed when running with nv = 50. Meanwhile, nv=40 could sometimes (only with a small probability) give a result, but it went to ERROR when running it twice or more.

To note, this test is based on an R4.0.5 environment created by conda in VS code, in which the R (interactive) terminal would crush, and then back to the PowerShell with the exit code: -1073741819. By the way, other versions of R (v4.1.3, v4.3.3) are also crushed.

A wired issue is that the same environment run by another PC (Intel i10-10750H) is out of trouble. However, the crushed one has an AMD CPU (AI9 HX370). I`m not sure whether this difference will contribute to this issue.

Thanks!