Function VecCeil

TSSimd.h:1340–1373

	}

	inline vec_float VecCeil(const vec_float& v)
	{
		#if defined(TERATHON_SSE)

			#if defined(TERATHON_SSE4)

				return (_mm_round_ps(v, _MM_FROUND_TO_POS_INF | _MM_FROUND_NO_EXC));

			#else

				const vec_float one = VecLoadVectorConstant<0x3F800000>();
				const vec_float two23 = VecLoadVectorConstant<0x4B000000>();

				vec_float result = _mm_sub_ps(_mm_add_ps(_mm_add_ps(_mm_sub_ps(v, two23), two23), two23), two23);
				result = _mm_add_ps(result, _mm_and_ps(one, _mm_cmplt_ps(result, v)));

				vec_float mask = _mm_cmplt_ps(two23, _mm_andnot_ps(VecFloatGetMinusZero(), v));
				return (_mm_or_ps(_mm_andnot_ps(mask, result), _mm_and_ps(mask, v)));

			#endif

		#elif defined(TERATHON_NEON)

			const vec_uint32 one = vdupq_n_u32(0x3F800000);
			const vec_float two23 = vdupq_n_f32(8388608.0F);

			vec_float result = vsubq_f32(vaddq_f32(vaddq_f32(vsubq_f32(v, two23), two23), two23), two23);
			result = vaddq_f32(result, vreinterpretq_f32_u32(vandq_u32(one, vcltq_f32(result, v))));

			vec_uint32 mask = vcltq_f32(two23, vabsq_f32(v));
			return (vbslq_f32(mask, v, result));

		#endif
	}

	inline vec_float VecCeilScalar(const vec_float& v)
	{