Add packetized versions of i0e and i1e special functions.

  - In particular refactor the i0e and i1e code so scalar and vectorized path share code.
  - Move chebevl to GenericPacketMathFunctions.


A brief benchmark with building Eigen with FMA, AVX and AVX2 flags

Before:

CPU: Intel Haswell with HyperThreading (6 cores)
Benchmark                  Time(ns)        CPU(ns)     Iterations
-----------------------------------------------------------------
BM_eigen_i0e_double/1            57.3           57.3     10000000
BM_eigen_i0e_double/8           398            398        1748554
BM_eigen_i0e_double/64         3184           3184         218961
BM_eigen_i0e_double/512       25579          25579          27330
BM_eigen_i0e_double/4k       205043         205042           3418
BM_eigen_i0e_double/32k     1646038        1646176            422
BM_eigen_i0e_double/256k   13180959       13182613             53
BM_eigen_i0e_double/1M     52684617       52706132             10
BM_eigen_i0e_float/1             28.4           28.4     24636711
BM_eigen_i0e_float/8             75.7           75.7      9207634
BM_eigen_i0e_float/64           512            512        1000000
BM_eigen_i0e_float/512         4194           4194         166359
BM_eigen_i0e_float/4k         32756          32761          21373
BM_eigen_i0e_float/32k       261133         261153           2678
BM_eigen_i0e_float/256k     2087938        2088231            333
BM_eigen_i0e_float/1M       8380409        8381234             84
BM_eigen_i1e_double/1            56.3           56.3     10000000
BM_eigen_i1e_double/8           397            397        1772376
BM_eigen_i1e_double/64         3114           3115         223881
BM_eigen_i1e_double/512       25358          25361          27761
BM_eigen_i1e_double/4k       203543         203593           3462
BM_eigen_i1e_double/32k     1613649        1613803            428
BM_eigen_i1e_double/256k   12910625       12910374             54
BM_eigen_i1e_double/1M     51723824       51723991             10
BM_eigen_i1e_float/1             28.3           28.3     24683049
BM_eigen_i1e_float/8             74.8           74.9      9366216
BM_eigen_i1e_float/64           505            505        1000000
BM_eigen_i1e_float/512         4068           4068         171690
BM_eigen_i1e_float/4k         31803          31806          21948
BM_eigen_i1e_float/32k       253637         253692           2763
BM_eigen_i1e_float/256k     2019711        2019918            346
BM_eigen_i1e_float/1M       8238681        8238713             86


After:

CPU: Intel Haswell with HyperThreading (6 cores)
Benchmark                  Time(ns)        CPU(ns)     Iterations
-----------------------------------------------------------------
BM_eigen_i0e_double/1            15.8           15.8     44097476
BM_eigen_i0e_double/8            99.3           99.3      7014884
BM_eigen_i0e_double/64          777            777         886612
BM_eigen_i0e_double/512        6180           6181         100000
BM_eigen_i0e_double/4k        48136          48140          14678
BM_eigen_i0e_double/32k      385936         385943           1801
BM_eigen_i0e_double/256k    3293324        3293551            228
BM_eigen_i0e_double/1M     12423600       12424458             57
BM_eigen_i0e_float/1             16.3           16.3     43038042
BM_eigen_i0e_float/8             30.1           30.1     23456931
BM_eigen_i0e_float/64           169            169        4132875
BM_eigen_i0e_float/512         1338           1339         516860
BM_eigen_i0e_float/4k         10191          10191          68513
BM_eigen_i0e_float/32k        81338          81337           8531
BM_eigen_i0e_float/256k      651807         651984           1000
BM_eigen_i0e_float/1M       2633821        2634187            268
BM_eigen_i1e_double/1            16.2           16.2     42352499
BM_eigen_i1e_double/8           110            110        6316524
BM_eigen_i1e_double/64          822            822         851065
BM_eigen_i1e_double/512        6480           6481         100000
BM_eigen_i1e_double/4k        51843          51843          10000
BM_eigen_i1e_double/32k      414854         414852           1680
BM_eigen_i1e_double/256k    3320001        3320568            212
BM_eigen_i1e_double/1M     13442795       13442391             53
BM_eigen_i1e_float/1             17.6           17.6     41025735
BM_eigen_i1e_float/8             35.5           35.5     19597891
BM_eigen_i1e_float/64           240            240        2924237
BM_eigen_i1e_float/512         1424           1424         485953
BM_eigen_i1e_float/4k         10722          10723          65162
BM_eigen_i1e_float/32k        86286          86297           8048
BM_eigen_i1e_float/256k      691821         691868           1000
BM_eigen_i1e_float/1M       2777336        2777747            256


This shows anywhere from a 50% to 75% improvement on these operations.

I've also benchmarked without any of these flags turned on, and got similar
performance to before (if not better).

Also tested packetmath.cpp + special_functions to ensure no regressions.

1 files changed, 91 insertions, 0 deletions

diff --git a/Eigen/src/Core/arch/Default/GenericPacketMathFunctions.h b/Eigen/src/Core/arch/Default/GenericPacketMathFunctions.h
index 518db2207..0a4b66089 100644
--- a/Eigen/src/Core/arch/Default/GenericPacketMathFunctions.h
+++ b/Eigen/src/Core/arch/Default/GenericPacketMathFunctions.h
@@ -570,6 +570,97 @@ struct ppolevl<Packet, 0> {
   }
 };
 
+/* chbevl (modified for Eigen)
+ *
+ *     Evaluate Chebyshev series
+ *
+ *
+ *
+ * SYNOPSIS:
+ *
+ * int N;
+ * Scalar x, y, coef[N], chebevl();
+ *
+ * y = chbevl( x, coef, N );
+ *
+ *
+ *
+ * DESCRIPTION:
+ *
+ * Evaluates the series
+ *
+ *        N-1
+ *         - '
+ *  y  =   >   coef[i] T (x/2)
+ *         -            i
+ *        i=0
+ *
+ * of Chebyshev polynomials Ti at argument x/2.
+ *
+ * Coefficients are stored in reverse order, i.e. the zero
+ * order term is last in the array.  Note N is the number of
+ * coefficients, not the order.
+ *
+ * If coefficients are for the interval a to b, x must
+ * have been transformed to x -> 2(2x - b - a)/(b-a) before
+ * entering the routine.  This maps x from (a, b) to (-1, 1),
+ * over which the Chebyshev polynomials are defined.
+ *
+ * If the coefficients are for the inverted interval, in
+ * which (a, b) is mapped to (1/b, 1/a), the transformation
+ * required is x -> 2(2ab/x - b - a)/(b-a).  If b is infinity,
+ * this becomes x -> 4a/x - 1.
+ *
+ *
+ *
+ * SPEED:
+ *
+ * Taking advantage of the recurrence properties of the
+ * Chebyshev polynomials, the routine requires one more
+ * addition per loop than evaluating a nested polynomial of
+ * the same degree.
+ *
+ */
+template <typename Packet, int N>
+struct generic_cheb_recurrence {
+  EIGEN_DEVICE_FUNC
+  static EIGEN_STRONG_INLINE Packet run(Packet x, const typename unpacket_traits<Packet>::type coef[]) {
+    EIGEN_STATIC_ASSERT((N > 2), YOU_MADE_A_PROGRAMMING_MISTAKE);
+    return pmadd(
+        generic_cheb_recurrence<Packet, N - 1>::run(x, coef), x,
+        psub(pset1<Packet>(coef[N - 1]), generic_cheb_recurrence<Packet, N -
+             2>::run(x, coef)));
+  }
+};
+
+template <typename Packet>
+struct generic_cheb_recurrence<Packet, 2> {
+  EIGEN_DEVICE_FUNC
+  static EIGEN_STRONG_INLINE Packet run(Packet x, const typename unpacket_traits<Packet>::type coef[]) {
+    return pmadd(pset1<Packet>(coef[0]), x, pset1<Packet>(coef[1]));
+  }
+};
+
+template <typename Packet>
+struct generic_cheb_recurrence<Packet, 1> {
+  EIGEN_DEVICE_FUNC
+  static EIGEN_STRONG_INLINE Packet run(Packet x, const typename unpacket_traits<Packet>::type coef[]) {
+    EIGEN_UNUSED_VARIABLE(x);
+    return pset1<Packet>(coef[0]);
+  }
+};
+
+template <typename Packet, int N>
+struct pchebevl {
+  EIGEN_DEVICE_FUNC
+  static EIGEN_STRONG_INLINE Packet run(Packet x, const typename unpacket_traits<Packet>::type coef[]) {
+    const Packet half = pset1<Packet>(0.5);
+    return pmul(half, psub(
+        generic_cheb_recurrence<Packet, N>::run(x, coef),
+        generic_cheb_recurrence<Packet, N - 2>::run(x, coef)));
+  }
+};
+
 } // end namespace internal
 } // end namespace Eigen