CUDA 8 SSSP Tip:: When nvgraphSssp (part of the nvGraph library, new for CUDA 8) returns the “Not Converged” status, you *can* still call nvgraphGetVertexData to find the negative weight cycle(s) by examining costs and looking at which vertices have spiraling down in cost.

(Yes, nvgraphSssp works correctly in graphs with edges with negative weights, but no negative weight cycles.)

The image that’s currently on the top of my page is a picture I took. By my troth, I cannot recall the physical object which was the source for that image.

After great study, it appears to be some kind of rock.

Jewelry?

Bookend?

A very impressive Halloween costume?

#!/usr/bin/python
from nltk.corpus import wordnet 
shorts = "aceimnorsuvwxz" 
longs = "bdfhklt" 
for i1 in shorts: 
    for i2 in longs: 
        for i3 in shorts: 
            for i4 in shorts: 
                for i5 in shorts: 
                    wordinq = i1+i2+i3+i4+i5 
                    if i1=='a' or i3=='a' or i4=='a' or i5=='a': 
                        synsets = wordnet.synsets(wordinq) 
                        for synset in synsets: 
                            print wordinq+':', synset.definition 
                            break;

…is how to cheat at 3rd grade homework.