SzymonNowakowski

It is worth mentioning that the data that crashes `svd` `LAPACK` routine is a part of [Insurance data set](https://www.kaggle.com/c/prudential-life-insurance-assessment/data), isolated for the purpose of reproducing the bug.

I found and extracted another case for this *possible* bug while working with [Insurance data set](https://www.kaggle.com/c/prudential-life-insurance-assessment/data). It is a full rank 5344x255 matrix. To reproduce it, please run: ```{r code}...

I found and extracted another case for this *possible* bug while working with [Insurance data set](https://www.kaggle.com/c/prudential-life-insurance-assessment/data) some more. It is a full rank 5344x258 matrix this time. This is the...

No, on the Windows machine I don't have OpenBLAS. The Ubuntu machine is currently down, I will check & let you know when it gets up. ``` > La_version() [1]...

But I had the same error on another machine with an older R version (3.6.3) in Ubuntu with OpenBLAS: ``` > svd(crashes_svd) Error in La.svd(x, nu, nv) : error code...

Thank you. I installed OpenBlas on the Ubuntu machine and I confirm that this workaround works for all 3 isolated cases. For me it is sufficient as I can push...

Unfortunately, OpenBLAS creates problems of its own it seems. I have isolated another matrix which **passes SVD on a machine without OpenBLAS**, but **fails SVD with OpenBLAS**. Please check this...

Hi Weslley, on the two Ubuntu machines I got the exact same output as you did (`0` followed by the line `Singular values:` and then singular values followed starting with...

Thanks for those comparisons. You are working at the levels too deep for me to follow. Thus I just want to add one more piece of information from my side,...

Thank you so much for your work with this. So basically you are saying that (apart from my oldish Windows 7 machine) on LAPACK 3.10.0 it is a non-issue. After...