** On entry to DLASCLS parameter number 4 had an illegal value

[albert-ying]

I got this message when running irlba(large_mat, nv = 100). Is it something I can ignore or what is happening here?

Also, the R seems to be stuck after outputting this message, not sure whether these two things are related or not.

Thank you so much

[bwlewis]

That looks like a bug. Can you tell me which version of the package (CRAN or GitHub) and the output of sessionInfo()?

[albert-ying]

I cannot replicate this error now. It seems that it only happens sometimes when nv is big. I uses cran version

[yannk-lm]

Hello,

I'm getting the same issue using irlba 2.3.3.
This is the exact error:

BLAS/LAPACK routine 'DLASCLSLeft' gave error code -4

Output of my sessionInfo() : sessionInfo.txt

When running with nv = 20 I get rapidly the error which is not the case with nv = 5.
Of note, for some reason, running time with nv = 5 was way too long so I cancelled the run before the end.

Thank you for your help.

[bwlewis]

Do you have any NA's in your matrix (see #14)?

(Also, can you try the current CRAN version, 2.3.5?).

[yannk-lm]

Hello,

I, indeed, have NA's in my matrix and even with the 2.3.5 CRAN version I still get the same error.