[FP16] Add end to end tests of all the F16x8 intrinisics.

test/test_fp16.c

@@ -0,0 +1,154 @@
+#include <assert.h>
+#include <emscripten.h>
+#include <math.h>
+#include <stdio.h>
+#include <wasm_simd128.h>
+
+// Not used below, but helpful for debugging.
+void print_f16x8(v128_t v1) {
+  printf("[%g, %g, %g, %g, %g, %g, %g, %g]\n",
+         wasm_f16x8_extract_lane(v1, 0),
+         wasm_f16x8_extract_lane(v1, 1),
+         wasm_f16x8_extract_lane(v1, 2),
+         wasm_f16x8_extract_lane(v1, 3),
+         wasm_f16x8_extract_lane(v1, 4),
+         wasm_f16x8_extract_lane(v1, 5),
+         wasm_f16x8_extract_lane(v1, 6),
+         wasm_f16x8_extract_lane(v1, 7));
+}
+
+void assert_lanes_eq_uint16(v128_t v,
+                            uint16_t l0,
+                            uint16_t l1,
+                            uint16_t l2,
+                            uint16_t l3,
+                            uint16_t l4,
+                            uint16_t l5,
+                            uint16_t l6,
+                            uint16_t l7) {
+  assert(wasm_u16x8_extract_lane(v, 0) == l0);
+  assert(wasm_u16x8_extract_lane(v, 1) == l1);
+  assert(wasm_u16x8_extract_lane(v, 2) == l2);
+  assert(wasm_u16x8_extract_lane(v, 3) == l3);
+  assert(wasm_u16x8_extract_lane(v, 4) == l4);
+  assert(wasm_u16x8_extract_lane(v, 5) == l5);
+  assert(wasm_u16x8_extract_lane(v, 6) == l6);
+  assert(wasm_u16x8_extract_lane(v, 7) == l7);
+}
+
+void assert_lanes_eq_float(v128_t v,
+                           float l0,
+                           float l1,
+                           float l2,
+                           float l3,
+                           float l4,
+                           float l5,
+                           float l6,
+                           float l7) {
+  assert(wasm_f16x8_extract_lane(v, 0) == l0);
+  assert(wasm_f16x8_extract_lane(v, 1) == l1);
+  assert(wasm_f16x8_extract_lane(v, 2) == l2);
+  assert(wasm_f16x8_extract_lane(v, 3) == l3);
+  assert(wasm_f16x8_extract_lane(v, 4) == l4);
+  assert(wasm_f16x8_extract_lane(v, 5) == l5);
+  assert(wasm_f16x8_extract_lane(v, 6) == l6);
+  assert(wasm_f16x8_extract_lane(v, 7) == l7);
+}
+
+void assert_all_lanes_eq(v128_t v, float value) {
+  assert_lanes_eq_float(
+    v, value, value, value, value, value, value, value, value);
+}
+
+void assert_all_lanes_eq_uint16(v128_t v, uint16_t value) {
+  assert_lanes_eq_uint16(
+    v, value, value, value, value, value, value, value, value);
+}
+
+int main() {
+  v128_t a, b, c;
+  a = wasm_f16x8_splat(2.0f);
+  assert_all_lanes_eq(a, 2.0f);
+
+  a = wasm_f16x8_splat(1.0f);
+  a = wasm_f16x8_replace_lane(a, 0, 99.0f);
+  a = wasm_f16x8_replace_lane(a, 7, 99.0f);
+  assert_lanes_eq_float(a, 99.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 99.0f);
+
+  a = wasm_f16x8_abs(wasm_f16x8_splat(-2.0f));
+  assert_all_lanes_eq(a, 2.0f);
+
+  a = wasm_f16x8_neg(wasm_f16x8_splat(2.0f));
+  assert_all_lanes_eq(a, -2.0f);
+
+  a = wasm_f16x8_sqrt(wasm_f16x8_splat(4.0f));
+  assert_all_lanes_eq(a, 2.0f);
+
+  a = wasm_f16x8_ceil(wasm_f16x8_splat(1.2f));
+  assert_all_lanes_eq(a, 2.0f);
+
+  a = wasm_f16x8_floor(wasm_f16x8_splat(1.2f));
+  assert_all_lanes_eq(a, 1.0f);
+
+  a = wasm_f16x8_trunc(wasm_f16x8_splat(1.2f));
+  assert_all_lanes_eq(a, 1.0f);
+
+  a = wasm_f16x8_nearest(wasm_f16x8_splat(1.5f));
+  assert_all_lanes_eq(a, 2.0f);
+
+  a = wasm_f16x8_eq(wasm_f16x8_splat(2.0f), wasm_f16x8_splat(3.0f));
+  assert_all_lanes_eq_uint16(a, 0);
+
+  a = wasm_f16x8_ne(wasm_f16x8_splat(2.0f), wasm_f16x8_splat(3.0f));
+  assert_all_lanes_eq_uint16(a, 0xFFFF);
+
+  a = wasm_f16x8_lt(wasm_f16x8_splat(2.0f), wasm_f16x8_splat(3.0f));
+  assert_all_lanes_eq_uint16(a, 0xFFFF);
+
+  a = wasm_f16x8_gt(wasm_f16x8_splat(2.0f), wasm_f16x8_splat(3.0f));
+  assert_all_lanes_eq_uint16(a, 0);
+
+  a = wasm_f16x8_le(wasm_f16x8_splat(2.0f), wasm_f16x8_splat(3.0f));
+  assert_all_lanes_eq_uint16(a, 0xFFFF);
+
+  a = wasm_f16x8_ge(wasm_f16x8_splat(2.0f), wasm_f16x8_splat(3.0f));
+  assert_all_lanes_eq_uint16(a, 0);
+
+  a = wasm_f16x8_add(wasm_f16x8_splat(2.0f), wasm_f16x8_splat(3.0f));
+  assert_all_lanes_eq(a, 5.0f);
+
+  a = wasm_f16x8_sub(wasm_f16x8_splat(5.0f), wasm_f16x8_splat(3.0f));
+  assert_all_lanes_eq(a, 2.0f);
+
+  a = wasm_f16x8_mul(wasm_f16x8_splat(5.0f), wasm_f16x8_splat(3.0f));
+  assert_all_lanes_eq(a, 15.0f);
+
+  a = wasm_f16x8_div(wasm_f16x8_splat(10.0f), wasm_f16x8_splat(5.0f));
+  assert_all_lanes_eq(a, 2.0f);
+
+  a = wasm_f16x8_min(wasm_f16x8_splat(10.0f), wasm_f16x8_splat(5.0f));
+  assert_all_lanes_eq(a, 5.0f);
+
+  a = wasm_f16x8_max(wasm_f16x8_splat(10.0f), wasm_f16x8_splat(5.0f));
+  assert_all_lanes_eq(a, 10.0f);
+
+  a = wasm_f16x8_pmin(wasm_f16x8_splat(10.0f), wasm_f16x8_splat(5.0f));
+  assert_all_lanes_eq(a, 5.0f);
+
+  a = wasm_f16x8_pmax(wasm_f16x8_splat(10.0f), wasm_f16x8_splat(5.0f));
+  assert_all_lanes_eq(a, 10.0f);
+
+  // TODO needed in binaryen
+  // wasm_i16x8_trunc_sat_f16x8;
+  // wasm_u16x8_trunc_sat_f16x8
+  // wasm_f16x8_convert_i16x8
+  // wasm_f16x8_convert_u16x8
+
+  a = wasm_f16x8_relaxed_madd(
+    wasm_f16x8_splat(2), wasm_f16x8_splat(3), wasm_f16x8_splat(4));
+  assert_all_lanes_eq(a, 10.0f);
+
+  a = wasm_f16x8_relaxed_nmadd(
+    wasm_f16x8_splat(2), wasm_f16x8_splat(3), wasm_f16x8_splat(4));
+  assert_all_lanes_eq(a, -2.0f);
+}

test/test_other.py

@@ -15141,3 +15141,15 @@ def test_extra_struct_info(self):
   def test_fs_writev_partial_write(self):
     self.set_setting('FORCE_FILESYSTEM')
     self.do_run_in_out_file_test('fs/test_writev_partial_write.c')
+
+  @requires_v8
+  @parameterized({
+    '': [[]],
+    'O2': [['-O2']]
+  })
+  def test_fp16(self, opts):
+    self.v8_args += ['--experimental-wasm-fp16']
+    # TODO Remove this. Liftoff is currently broken for this test.
+    self.v8_args += ['--no-liftoff']
+    self.emcc_args = ['-msimd128', '-mfp16', '-sENVIRONMENT=shell'] + opts
+    self.do_runf('test_fp16.c')