commit d84a54c747ab02fb320f4c0eb75d19927b8cbf73
parent 98512605f8d30ca54a9a5fd2ccf0df30867ef0ea
Author: Ivan Gankevich <igankevich@ya.ru>
Date: Mon, 13 Nov 2017 15:47:02 +0300
Absence.
Diffstat:
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/arma-thesis.org b/arma-thesis.org
@@ -125,7 +125,7 @@ software.
3. Since the formula which is used in the method is explicit, there is no need
in data exchange between parallel processes, which allows to scale
performance to a large number of processor cores.
-4. Finally, ARMA model is itself more efficient than LH model due to vicinity of
+4. Finally, ARMA model is itself more efficient than LH model due to absence of
trigonometric functions in its formula: In fact, wavy surface is computed as
a sum of large number of polynomials, for which there is low-level FMA (Fused
Multiply-Add) assembly instruction, and memory access pattern allows for
@@ -2281,7 +2281,7 @@ multi-dimensional derivatives were available.
AR and MA models outperforms LH model in benchmarks and does not require GPU to
do so. From computational point of view their strengths are
-- vicinity of transcendental mathematical functions, and
+- absence of transcendental mathematical functions, and
- simple algorithm for both AR and MA model, performance of which depends on the
performance of multi-dimensional array library and FFT library.
Providing main functionality via low-level libraries makes performance of the
