Dmitry I. Lyakh
Dmitry I. Lyakh
The problematic matrix and other details are below. To clarify, I am using OpenBLAS from github (https://github.com/xianyi/OpenBLAS.git) which contains the lapack-netlib directory (version 3.9.0) where I believe the zgesvdx implementation...
The specific OpenBLAS commit I am using is af8a619e1fbb4e41c566453baeb3b4e523c92337. This is in case you need to compare their copied netlib-lapack source with the main netlib-lapack repo.
On a related note, for a different but also very small matrix (4-by-4) we are getting "internal error occurred in DBDSVDX". I am wondering what exactly this means? Is it...
cuTT does not allow dimension extents of 1.
Can you rerun the failing case with the environment variable CUTENSORNET_LOG_LEVEL=5 set and attach the log file?
Is it a multi-GPU node? Do you mind rerunning it again with the the following environment variables set and attaching the log again? CUDA_VISIBLE_DEVICES=0 CUTENSORNET_LOG_LEVEL=5 CUTENSOR_LOG_LEVEL=5
Is this the Sycamore-53 depth-20 circuit? Do you have any transformations done on it which would introduce hyper-edges? Do you enable slicing? I see very large intermediate tensors there and...
Make sure you are building it for the right CPU architecture, especially if the platform you are building on has cross-compilation (many Cray platforms).
What do you mean exactly? Addition and subtraction of entire tensor networks or something else?
You will need to use the C++ compiler (g++) even if you only want to use the C API.