You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: docs/src/writing_good_rules.md
+1-1Lines changed: 1 addition & 1 deletion
Original file line number
Diff line number
Diff line change
@@ -122,7 +122,7 @@ It is very easy to check gradients or derivatives with a computer algebra system
122
122
In principle, a perfect AD system only needs rules for basic operations and can infer the rules for more complicated functions automatically.
123
123
In practice, performance needs to be considered as well.
124
124
125
-
Some functions use `ccall` internally for performance reasons, for example [`^`](https://github.com/JuliaLang/julia/blob/v1.5.3/base/math.jl#L886).
125
+
Some functions use `ccall` internally, for example [`^`](https://github.com/JuliaLang/julia/blob/v1.5.3/base/math.jl#L886).
126
126
These functions can not be differentiated through by AD systems, and need custom rules.
127
127
128
128
Other functions can in principle be differentiated through by an AD system, but there exists a mathematical insight that can dramatically improve the computation of the derivative.
0 commit comments