/* { dg-final { check-function-bodies "**" "" "-DCHECK_ASM" } } */ #include "test_sme2_acle.h" /* ** dot_0_z0_z0: ** mov (w8|w9|w10|w11), #?0 ** bfdot za\.s\[\1, 0, vgx2\], {z0\.h - z1\.h}, {z0\.h - z1\.h} ** ret */ TEST_ZA_XN (dot_0_z0_z0, svbfloat16x2_t, svdot_za32_bf16_vg1x2 (0, z0, z0), svdot_za32_vg1x2 (0, z0, z0)) /* ** dot_w0_z0_z0: ** mov (w8|w9|w10|w11), w0 ** bfdot za\.s\[\1, 0, vgx2\], {z0\.h - z1\.h}, {z0\.h - z1\.h} ** ret */ TEST_ZA_XN (dot_w0_z0_z0, svbfloat16x2_t, svdot_za32_bf16_vg1x2 (w0, z0, z0), svdot_za32_vg1x2 (w0, z0, z0)) /* ** dot_w8_z0_z4: ** bfdot za\.s\[w8, 0, vgx2\], {z0\.h - z1\.h}, {z4\.h - z5\.h} ** ret */ TEST_ZA_XN (dot_w8_z0_z4, svbfloat16x2_t, svdot_za32_bf16_vg1x2 (w8, z0, z4), svdot_za32_vg1x2 (w8, z0, z4)) /* ** dot_w8_z4_z18: ** bfdot za\.s\[w8, 0, vgx2\], {z4\.h - z5\.h}, {z18\.h - z19\.h} ** ret */ TEST_ZA_XN (dot_w8_z4_z18, svbfloat16x2_t, svdot_za32_bf16_vg1x2 (w8, z4, z18), svdot_za32_vg1x2 (w8, z4, z18)) /* Leave the assembler to check for correctness for misaligned registers. */ /* ** dot_w8_z0_z23: ** ... ** bfdot za\.s\[w8, 0, vgx2\], {z0\.h - z1\.h}, [^\n]+ ** ret */ TEST_ZA_XN (dot_w8_z0_z23, svbfloat16x2_t, svdot_za32_bf16_vg1x2 (w8, z0, z23), svdot_za32_vg1x2 (w8, z0, z23)) /* ** dot_w8_z23_z0: ** ... ** bfdot za\.s\[w8, 0, vgx2\], [^\n]+, {z0\.h - z1\.h} ** ret */ TEST_ZA_XN (dot_w8_z23_z0, svbfloat16x2_t, svdot_za32_bf16_vg1x2 (w8, z23, z0), svdot_za32_vg1x2 (w8, z23, z0)) /* ** dot_w8_z18_z28: ** bfdot za\.s\[w8, 0, vgx2\], {z18\.h - z19\.h}, {z28\.h - z29\.h} ** ret */ TEST_ZA_XN (dot_w8_z18_z28, svbfloat16x2_t, svdot_za32_bf16_vg1x2 (w8, z18, z28), svdot_za32_vg1x2 (w8, z18, z28)) /* ** dot_w8_z28_z4: ** bfdot za\.s\[w8, 0, vgx2\], {z28\.h - z29\.h}, {z4\.h - z5\.h} ** ret */ TEST_ZA_XN (dot_w8_z28_z4, svbfloat16x2_t, svdot_za32_bf16_vg1x2 (w8, z28, z4), svdot_za32_vg1x2 (w8, z28, z4)) /* ** dot_w8p1_z4_z0: ** bfdot za\.s\[w8, 1, vgx2\], {z4\.h - z5\.h}, {z0\.h - z1\.h} ** ret */ TEST_ZA_XN (dot_w8p1_z4_z0, svbfloat16x2_t, svdot_za32_bf16_vg1x2 (w8 + 1, z4, z0), svdot_za32_vg1x2 (w8 + 1, z4, z0)) /* ** dot_w8p2_z4_z0: ** bfdot za\.s\[w8, 2, vgx2\], {z4\.h - z5\.h}, {z0\.h - z1\.h} ** ret */ TEST_ZA_XN (dot_w8p2_z4_z0, svbfloat16x2_t, svdot_za32_bf16_vg1x2 (w8 + 2, z4, z0), svdot_za32_vg1x2 (w8 + 2, z4, z0)) /* ** dot_w11p4_z4_z0: ** bfdot za\.s\[w11, 4, vgx2\], {z4\.h - z5\.h}, {z0\.h - z1\.h} ** ret */ TEST_ZA_XN (dot_w11p4_z4_z0, svbfloat16x2_t, svdot_za32_bf16_vg1x2 (w11 + 4, z4, z0), svdot_za32_vg1x2 (w11 + 4, z4, z0)) /* ** dot_w8p7_z4_z0: ** bfdot za\.s\[w8, 7, vgx2\], {z4\.h - z5\.h}, {z0\.h - z1\.h} ** ret */ TEST_ZA_XN (dot_w8p7_z4_z0, svbfloat16x2_t, svdot_za32_bf16_vg1x2 (w8 + 7, z4, z0), svdot_za32_vg1x2 (w8 + 7, z4, z0)) /* ** dot_w8p8_z4_z4: ** add (w8|w9|w10|w11), w8, #?8 ** bfdot za\.s\[\1, 0, vgx2\], {z4\.h - z5\.h}, {z4\.h - z5\.h} ** ret */ TEST_ZA_XN (dot_w8p8_z4_z4, svbfloat16x2_t, svdot_za32_bf16_vg1x2 (w8 + 8, z4, z4), svdot_za32_vg1x2 (w8 + 8, z4, z4)) /* ** dot_w8m1_z4_z0: ** sub (w8|w9|w10|w11), w8, #?1 ** bfdot za\.s\[\1, 0, vgx2\], {z4\.h - z5\.h}, {z0\.h - z1\.h} ** ret */ TEST_ZA_XN (dot_w8m1_z4_z0, svbfloat16x2_t, svdot_za32_bf16_vg1x2 (w8 - 1, z4, z0), svdot_za32_vg1x2 (w8 - 1, z4, z0)) /* ** dot_single_0_z1_z0: ** mov (w8|w9|w10|w11), #?0 ** bfdot za\.s\[\1, 0, vgx2\], {z1\.h - z2\.h}, z0\.h ** ret */ TEST_ZA_SINGLE (dot_single_0_z1_z0, svbfloat16x2_t, svbfloat16_t, svdot_single_za32_bf16_vg1x2 (0, z1, z0), svdot_za32_vg1x2 (0, z1, z0)) /* ** dot_single_w0_z1_z0: ** mov (w8|w9|w10|w11), w0 ** bfdot za\.s\[\1, 0, vgx2\], {z1\.h - z2\.h}, z0\.h ** ret */ TEST_ZA_SINGLE (dot_single_w0_z1_z0, svbfloat16x2_t, svbfloat16_t, svdot_single_za32_bf16_vg1x2 (w0, z1, z0), svdot_za32_vg1x2 (w0, z1, z0)) /* ** dot_single_w8_z1_z0: ** bfdot za\.s\[w8, 0, vgx2\], {z1\.h - z2\.h}, z0\.h ** ret */ TEST_ZA_SINGLE (dot_single_w8_z1_z0, svbfloat16x2_t, svbfloat16_t, svdot_single_za32_bf16_vg1x2 (w8, z1, z0), svdot_za32_vg1x2 (w8, z1, z0)) /* ** dot_single_w8p1_z1_z0: ** bfdot za\.s\[w8, 1, vgx2\], {z1\.h - z2\.h}, z0\.h ** ret */ TEST_ZA_SINGLE (dot_single_w8p1_z1_z0, svbfloat16x2_t, svbfloat16_t, svdot_single_za32_bf16_vg1x2 (w8 + 1, z1, z0), svdot_za32_vg1x2 (w8 + 1, z1, z0)) /* ** dot_single_w8p2_z20_z0: ** bfdot za\.s\[w8, 2, vgx2\], {z20\.h - z21\.h}, z0\.h ** ret */ TEST_ZA_SINGLE (dot_single_w8p2_z20_z0, svbfloat16x2_t, svbfloat16_t, svdot_single_za32_bf16_vg1x2 (w8 + 2, z20, z0), svdot_za32_vg1x2 (w8 + 2, z20, z0)) /* ** dot_single_w11p4_z27_z0: ** bfdot za\.s\[w11, 4, vgx2\], {z27\.h - z28\.h}, z0\.h ** ret */ TEST_ZA_SINGLE (dot_single_w11p4_z27_z0, svbfloat16x2_t, svbfloat16_t, svdot_single_za32_bf16_vg1x2 (w11 + 4, z27, z0), svdot_za32_vg1x2 (w11 + 4, z27, z0)) /* ** dot_single_w8p7_z1_z0: ** bfdot za\.s\[w8, 7, vgx2\], {z1\.h - z2\.h}, z0\.h ** ret */ TEST_ZA_SINGLE (dot_single_w8p7_z1_z0, svbfloat16x2_t, svbfloat16_t, svdot_single_za32_bf16_vg1x2 (w8 + 7, z1, z0), svdot_za32_vg1x2 (w8 + 7, z1, z0)) /* ** dot_single_w8p8_z1_z0: ** add (w8|w9|w10|w11), w8, #?8 ** bfdot za\.s\[\1, 0, vgx2\], {z1\.h - z2\.h}, z0\.h ** ret */ TEST_ZA_SINGLE (dot_single_w8p8_z1_z0, svbfloat16x2_t, svbfloat16_t, svdot_single_za32_bf16_vg1x2 (w8 + 8, z1, z0), svdot_za32_vg1x2 (w8 + 8, z1, z0)) /* ** dot_single_w0m1_z1_z0: ** sub (w8|w9|w10|w11), w0, #?1 ** bfdot za\.s\[\1, 0, vgx2\], {z1\.h - z2\.h}, z0\.h ** ret */ TEST_ZA_SINGLE (dot_single_w0m1_z1_z0, svbfloat16x2_t, svbfloat16_t, svdot_single_za32_bf16_vg1x2 (w0 - 1, z1, z0), svdot_za32_vg1x2 (w0 - 1, z1, z0)) /* ** dot_single_w8_z0_z15: ** str d15, \[sp, #?-16\]! ** bfdot za\.s\[w8, 0, vgx2\], {z0\.h - z1\.h}, z15\.h ** ldr d15, \[sp\], #?16 ** ret */ TEST_ZA_SINGLE_Z15 (dot_single_w8_z0_z15, svbfloat16x2_t, svbfloat16_t, svdot_single_za32_bf16_vg1x2 (w8, z0, z15), svdot_za32_vg1x2 (w8, z0, z15)) /* ** dot_single_w8_z20_z16: ** mov (z[0-7]).d, z16.d ** bfdot za\.s\[w8, 0, vgx2\], {z20\.h - z21\.h}, \1\.h ** ret */ TEST_ZA_SINGLE (dot_single_w8_z20_z16, svbfloat16x2_t, svbfloat16_t, svdot_single_za32_bf16_vg1x2 (w8, z20, z16), svdot_za32_vg1x2 (w8, z20, z16))