(remove completed reference task) |
|||
(One intermediate revision by one other user not shown) | |||
Line 1: | Line 1: | ||
+ | Issues with <tt>windscreen.B</tt>: |
||
− | '''TODO''' |
||
+ | * it is '''not''' in valid Matrix Market format, because the array format is supposed to have only the number rows (22692) and columns (1) written in one line after the comment section, and not also the number of nonzeros as in the coordinate format (e.g. <tt>windscreen.K</tt> and <tt>windscreen.M</tt>), |
||
− | |||
+ | * <tt>mmread.m</tt> will read this file without reporting the error, but adding an empty line immediately after the comment section (which is valid by itself) will cause <tt>mmread.m</tt> to report the error, |
||
− | * The citation information for the background on the benchmark is missing. - C. Himpe |
||
+ | * <tt>scipy.io.mmread</tt> raises an exception when reading this file (removing the number of nonzeros makes it readable), |
||
− | ** Y. Yue recommends citing https://doi.org/10.1002/nme.2058 - P. Mlinarić |
||
+ | * the array format is used, even though there is only one nonzero element, |
||
− | * Add a category for algebraic systems? - P. Mlinarić |
||
+ | * the complex format is used, but all imaginary parts are exactly zero. |
||
+ | - P. Mlinarić |
Latest revision as of 16:03, 31 August 2022
Issues with windscreen.B:
- it is not in valid Matrix Market format, because the array format is supposed to have only the number rows (22692) and columns (1) written in one line after the comment section, and not also the number of nonzeros as in the coordinate format (e.g. windscreen.K and windscreen.M),
- mmread.m will read this file without reporting the error, but adding an empty line immediately after the comment section (which is valid by itself) will cause mmread.m to report the error,
- scipy.io.mmread raises an exception when reading this file (removing the number of nonzeros makes it readable),
- the array format is used, even though there is only one nonzero element,
- the complex format is used, but all imaginary parts are exactly zero.
- P. Mlinarić