thirdparty/ppsspp
2
0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

VFPU: Compute sines and cosines in double precision.

Browse source 
Let's see if we can do without the special checks in double precision.
This commit is contained in:
Henrik Rydgård 2020-10-10 10:06:28 +02:00
parent 0ee996b2a0
commit 16654d37ea
2 changed files with 35 additions and 42 deletions
Download patch file Download diff file Expand all files Collapse all files

14
unittest/UnitTest.cpp
View file

@ -23,6 +23,9 @@
//
// TODO: Make a test of nice unittest asserts and count successes etc.
// Or just integrate with an existing testing framework.
//
// To use, set command line parameter to one or more of the tests below, or "all".
// Search for "availableTests".
#include <cstdio>
#include <cstdlib>
@ -308,10 +311,19 @@ bool TestVFPUSinCos() {
EXPECT_APPROX_EQ_FLOAT(sine, 1.0f);
EXPECT_APPROX_EQ_FLOAT(cosine, 0.0f);
for (float angle = -10.0f; angle < 10.0f; angle++) {
vfpu_sincos(-1.0f, sine, cosine);
EXPECT_EQ_FLOAT(sine, -1.0f);
EXPECT_EQ_FLOAT(cosine, 0.0f);
vfpu_sincos(-2.0f, sine, cosine);
EXPECT_EQ_FLOAT(sine, 0.0f);
EXPECT_EQ_FLOAT(cosine, -1.0f);
for (float angle = -10.0f; angle < 10.0f; angle += 0.1f) {
vfpu_sincos(angle, sine, cosine);
EXPECT_APPROX_EQ_FLOAT(sine, sinf(angle * M_PI_2));
EXPECT_APPROX_EQ_FLOAT(cosine, cosf(angle * M_PI_2));
printf("sine: %f==%f cosine: %f==%f\n", sine, sinf(angle * M_PI_2), cosine, cosf(angle * M_PI_2));
}
return true;
}