dhermes / berkeley-m273-s2016
100%

DEFAULT BRANCH: master
Repo Added
Files 9
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst

LAST BUILD ON BRANCH master
branch: SELECT
CHANGE BRANCH
x

121

push

travis-ci

dhermes 
Adding notes about stability in the high-precision DG1 notebook.

1148 of 1148 relevant lines covered (100.0%)

1.0 hits per line

Relevant lines Covered
1148 RELEVANT LINES 1148 COVERED LINES
1.0 HITS PER LINE
Source Files on master
Coverage ∆ File Lines Relevant Covered Missed Hits/Line

Recent builds

Builds Branch Commit Type Ran Committer Via Coverage
121 master Adding notes about stability in the high-precision DG1 notebook. push dhermes travis-ci pending completion  
120 master Adding note about CFL condition for high-precision DG1. push dhermes travis-ci pending completion  
119 master Adding notebook using the high-precision DG1 solver. This isn't really a revelation, but we want to get the speeds up before trying to do convergence plots for higher order. push dhermes travis-ci pending completion  
118 master Another lint fix (protected access outside of a class). push dhermes travis-ci pending completion  
114 master Adding high-precision Gauss-Lobatto helper to DG1. push dhermes travis-ci pending completion  
113 master Adding start of a high-precision DG1 helper. push dhermes travis-ci pending completion  
112 master Using NumPy's MATMUL in dg1. The comment explains that CBLAS is used when it can be [1] but the typical MATMUL is just naive [2] hence the flops occur based on the numerical type. [1]: https://github.com/numpy/numpy/blob/89d95a0660ffd5890916edb7... push dhermes travis-ci pending completion  
111 master Adding back some NumPy features to dg1. These were discarded when trying to make the module more general but now that MathProvider expects to produce NumPy arrays we can add back the nice behavior. push dhermes travis-ci pending completion  
110 master Dials back the generality of dg1.MathProvider. Instead assumes that a NumPy array will always be the container and just farm out to helpers for **math** rather than for data manipulation. push dhermes travis-ci pending completion  
109 master Beginning stages of factoring out NumPy dependence in DG1 code. This is so we can use an extended precision library instead so the errors can be measured beyond machine precision (this is a pain for high p). push dhermes travis-ci pending completion  
See All Builds (42)