@acer Thanks a lot. From the times when Threads does work now, it does not offer any benefit. Rafal Ablamowicz

@acer I don't believe and I don't know if they are thread-safe or not. I don't know how to check if they are tread safe or how to make them thread safe. Is there any Maple documentation that says how to do that?

I have removed for now the option remember from two of my procedures in eClifford which have to do with computing the sign of the product. However, that would only may affect computations if I changed the signature say from [4,4,4] to [6,6,0] in the same worksheet yet I do not do that in any of these two worksheets: the signature stays the same so that should not be the issue. So far I see no errors.

These erroneous computations started only now when using the Threads package. So, maybe the miscomputing is the result of what you call not being "thread-safe".

Thanks,

Rafal

Download Parallel_computations_with_eCLifford_-_error_in_Threads_3.mw

Download Parallel_computations_with_eCLifford_-_error_in_Threads_2.mw

I hope I have managed to upload two Maple files: In fact, it is the same file executed twice without any changes. One shows that the differences err1 through err88 are all zero, yet in the second worksheet the are all NON-zero. This implies that there is an error in each computation with the Threads.

The nature of the difference between the correct computation result shown in out1 and those computed in Threads where each monomial in the difference has a factor of 2 seems to indicate an error in sign: instead of the difference of two quantities being zero, the value is doubled. 

Thanks,

Rafal Ablamowicz
rablamowicz@gmail.com

PS If anyone would like to be able to execute these worksheet, please email me privately and I would then email you a library file with my compiled package eClifford.