Speedups to CPU shortest distance / shortest path #24
Conversation
| } | ||
| } | ||
| } | ||
| g.addGrad(std::move(arcGrads)); | ||
| } | ||
|
|
||
| std::vector<int> topSort(const Graph& g) { |
There was a problem hiding this comment.
is it possible to throw an error if there is a loop in the graph.
There was a problem hiding this comment.
This top sort is much more efficient for graphs that have a lot of edges but it comes at the cost of not being able to do cycle detection. The idea is you don't have to explore every edge but you have to assume the graph is acyclic to make this assumption.
| @@ -125,6 +143,21 @@ class HDSpan { | |||
| return *this; | |||
| }; | |||
|
|
|||
| HDSpan& operator=(HDSpan&& other) { | |||
| if (this->data_ == other.data_) { | |||
There was a problem hiding this comment.
nit: I think it might be better to have "==" operator defined for HDSpan where we check of data_, device_, and size_ are the same. We can then do
if (this == other)
There was a problem hiding this comment.
The == operator is currently defined to check for semantic equality (like the elements in the two spans are the same). This is similar to std::vector.
| } | ||
| capacity_ = size; | ||
| auto newData = allocAndCopy(data_); | ||
| free(); |
There was a problem hiding this comment.
shouldn't we set size/capacity to zero in the call to free(). Then L191 would be of no effect.
There was a problem hiding this comment.
Its a little complicated at the moment :. The free is really just a convenience function for freeing the memory which first checks the device. The clear function actually sets the size to zero and frees the pointer. I could change "free" to "delete" or something like that if it would be less confusing?
TImings on my laptop using
./benchmarks/benchmark_functionsAfter
Timing forwardScoreLinearForward (cpu) ... 9.1351 msec
Timing forwardScoreLinearBackward (cpu) ... 9.085 msec
Timing forwardScoreRandDAGForward (cpu) ... 5.7167 msec
Timing forwardScoreRandDAGBackward (cpu) ... 6.4635 msec
Timing viterbiPathLinearForward (cpu) ... 1.7872 msec
Timing viterbiPathLinearBackward (cpu) ... 1.7198 msec
Timing viterbiPathRandDAGForward (cpu) ... 2.0626 msec
Timing viterbiPathRandDAGBackward (cpu) ... 0.1356 msec
Before
Timing forwardScoreLinearForward (cpu) ... 14.57 msec
Timing forwardScoreLinearBackward (cpu) ... 16.383 msec
Timing forwardScoreRandDAGForward (cpu) ... 5.8892 msec
Timing forwardScoreRandDAGBackward (cpu) ... 11.59 msec
Timing viterbiPathLinearForward (cpu) ... 4.2249 msec
Timing viterbiPathLinearBackward (cpu) ... 1.7191 msec
Timing viterbiPathRandDAGForward (cpu) ... 1.7667 msec
Timing viterbiPathRandDAGBackward (cpu) ... 0.087886 msec