Closed GajowyJ closed 1 year ago
Hi,
Thanks for using rindow-math.
Maybe the openblas library version is not compatible with the rindow-openblas extension. The Windows version of the DLL in the openblas library does not have a function to determine the version. The DLLs in the openblas library are completely incompatible with different versions.
Read the rindow-openblas README.
If you are using Windows, you must Download the version of OpenBLAS binaries that correspond to the rindow_openblas binaries. The compatible OpenBLAS Library release number is included in the filename of the rindow-openblas pre-built archive file. If you use the wrong OpenBLAS release number DLL, it will not work properly.
The rindow-openblas extension you are probably using is;
You need openblas version 0.3.20.
Exactly the same problem occurs with clbrast library
Thx,
In the cmake file of my OpenBLAS installation there is
SET(OpenBLAS_VERSION "0.3.21")
Should I downgrade to 0.3.20 ? Yes, I should! After downgrade, the problem disappeared! Thank you for support!
Hi! I succesfully set up the enviroment and started play with the library. I hope to use it my scientific research - matrix operations consumes 99.8% of time of my algorithms execution :|. One of the most time consuming operation is cross product of matrices. My matrices are filled with float numbers and unfortunatelly I can not run the Rindow for floats. It works perfectly for int64, however if I switch type to float is hangs completly (for float32) or produce
** On entry to SSYR2K parameter number 1 had an illegal value
message (for float64). Is it the known issue? Can I workaroud it in any way? My env, Win10+PHP 8.1Here is the test script: