cudev: add feature to convert FP32(float) from/to FP16(half) on GPU
* add feature of Fp16 on GPU (cudev) * add test * leave template function as unimplemented to raise error
This commit is contained in:
parent
f2e9588c93
commit
2974b049e7
@ -855,6 +855,8 @@ private:
|
||||
CV_EXPORTS void printCudaDeviceInfo(int device);
|
||||
CV_EXPORTS void printShortCudaDeviceInfo(int device);
|
||||
|
||||
CV_EXPORTS void convertFp16Cuda(InputArray _src, OutputArray _dst, Stream& stream = Stream::Null());
|
||||
|
||||
//! @} cudacore_init
|
||||
|
||||
}} // namespace cv { namespace cuda {
|
||||
|
@ -510,6 +510,17 @@ namespace
|
||||
|
||||
gridTransformUnary_< ConvertToPolicy<scalar_type> >(globPtr<T>(src), globPtr<D>(dst), op, stream);
|
||||
}
|
||||
|
||||
template <typename T, typename D>
|
||||
void convertScaleHalf(const GpuMat& src, const GpuMat& dst, Stream& stream)
|
||||
{
|
||||
typedef typename VecTraits<T>::elem_type src_elem_type;
|
||||
typedef typename VecTraits<D>::elem_type dst_elem_type;
|
||||
typedef typename LargerType<src_elem_type, float>::type larger_elem_type;
|
||||
typedef typename LargerType<float, dst_elem_type>::type scalar_type;
|
||||
|
||||
gridTransformUnary_< ConvertToPolicy<scalar_type> >(globPtr<T>(src), globPtr<D>(dst), saturate_cast_fp16_func<T,D>(), stream);
|
||||
}
|
||||
}
|
||||
|
||||
void cv::cuda::GpuMat::convertTo(OutputArray _dst, int rtype, Stream& stream) const
|
||||
@ -583,4 +594,36 @@ void cv::cuda::GpuMat::convertTo(OutputArray _dst, int rtype, double alpha, doub
|
||||
funcs[sdepth][ddepth](reshape(1), dst.reshape(1), alpha, beta, stream);
|
||||
}
|
||||
|
||||
void cv::cuda::convertFp16Cuda(InputArray _src, OutputArray _dst, Stream& stream)
|
||||
{
|
||||
GpuMat src = _src.getGpuMat();
|
||||
int ddepth = 0;
|
||||
|
||||
switch(src.depth())
|
||||
{
|
||||
case CV_32F:
|
||||
ddepth = CV_16S;
|
||||
break;
|
||||
case CV_16S:
|
||||
ddepth = CV_32F;
|
||||
break;
|
||||
default:
|
||||
CV_Error(Error::StsUnsupportedFormat, "Unsupported input depth");
|
||||
return;
|
||||
}
|
||||
int type = CV_MAKE_TYPE(CV_MAT_DEPTH(ddepth), src.channels());
|
||||
_dst.create(src.size(), type);
|
||||
GpuMat dst = _dst.getGpuMat();
|
||||
|
||||
typedef void (*func_t)(const GpuMat& src, const GpuMat& dst, Stream& stream);
|
||||
static const func_t funcs[] =
|
||||
{
|
||||
0, 0, 0,
|
||||
convertScaleHalf<float, short>, 0, convertScaleHalf<short, float>,
|
||||
0, 0,
|
||||
};
|
||||
|
||||
funcs[ddepth](src.reshape(1), dst.reshape(1), stream);
|
||||
}
|
||||
|
||||
#endif
|
||||
|
@ -668,6 +668,27 @@ template <typename T, typename D> struct saturate_cast_func : unary_function<T,
|
||||
}
|
||||
};
|
||||
|
||||
// Convert Fp16 dummy
|
||||
template <typename T, typename D> struct saturate_cast_fp16_func;
|
||||
|
||||
// Convert Fp16 from Fp32
|
||||
template <> struct saturate_cast_fp16_func<float, short> : unary_function<float, short>
|
||||
{
|
||||
__device__ __forceinline__ short operator ()(float v) const
|
||||
{
|
||||
return cast_fp16<float, short>(v);
|
||||
}
|
||||
};
|
||||
|
||||
// Convert Fp16 to Fp32
|
||||
template <> struct saturate_cast_fp16_func<short, float> : unary_function<short, float>
|
||||
{
|
||||
__device__ __forceinline__ float operator ()(short v) const
|
||||
{
|
||||
return cast_fp16<short, float>(v);
|
||||
}
|
||||
};
|
||||
|
||||
// Threshold Functors
|
||||
|
||||
template <typename T> struct ThreshBinaryFunc : unary_function<T, T>
|
||||
|
@ -270,6 +270,17 @@ template <> __device__ __forceinline__ uint saturate_cast<uint>(double v)
|
||||
#endif
|
||||
}
|
||||
|
||||
template <typename T, typename D> __device__ __forceinline__ D cast_fp16(T v);
|
||||
|
||||
template <> __device__ __forceinline__ float cast_fp16<short, float>(short v)
|
||||
{
|
||||
return __half2float(v);
|
||||
}
|
||||
|
||||
template <> __device__ __forceinline__ short cast_fp16<float, short>(float v)
|
||||
{
|
||||
return (short)__float2half_rn(v);
|
||||
}
|
||||
//! @}
|
||||
|
||||
}}
|
||||
|
@ -49,6 +49,7 @@ using namespace cv::cudev;
|
||||
using namespace cvtest;
|
||||
|
||||
typedef ::testing::Types<uchar, ushort, short, int, float> AllTypes;
|
||||
typedef ::testing::Types<short, float> Fp16Types;
|
||||
|
||||
////////////////////////////////////////////////////////////////////////////////
|
||||
// CvtTest
|
||||
@ -75,9 +76,75 @@ public:
|
||||
}
|
||||
};
|
||||
|
||||
// dummy class
|
||||
template <typename T>
|
||||
class CvFp16Test : public ::testing::Test
|
||||
{
|
||||
public:
|
||||
void test_gpumat()
|
||||
{
|
||||
}
|
||||
};
|
||||
|
||||
template <>
|
||||
class CvFp16Test <short> : public ::testing::Test
|
||||
{
|
||||
public:
|
||||
void test_gpumat()
|
||||
{
|
||||
const Size size = randomSize(100, 400);
|
||||
const int type = DataType<float>::type;
|
||||
|
||||
Mat src = randomMat(size, type), dst, ref;
|
||||
|
||||
GpuMat_<float> g_src(src);
|
||||
GpuMat g_dst;
|
||||
|
||||
// Fp32 -> Fp16
|
||||
convertFp16Cuda(g_src, g_dst);
|
||||
convertFp16Cuda(g_dst.clone(), g_dst);
|
||||
// Fp16 -> Fp32
|
||||
convertFp16(src, dst);
|
||||
convertFp16(dst, ref);
|
||||
|
||||
g_dst.download(dst);
|
||||
EXPECT_MAT_NEAR(dst, ref, 0.0);
|
||||
}
|
||||
};
|
||||
|
||||
template <>
|
||||
class CvFp16Test <float> : public ::testing::Test
|
||||
{
|
||||
public:
|
||||
void test_gpumat()
|
||||
{
|
||||
const Size size = randomSize(100, 400);
|
||||
const int type = DataType<float>::type;
|
||||
|
||||
Mat src = randomMat(size, type), dst, ref;
|
||||
|
||||
GpuMat_<float> g_src(src);
|
||||
GpuMat g_dst;
|
||||
|
||||
// Fp32 -> Fp16
|
||||
convertFp16Cuda(g_src, g_dst);
|
||||
convertFp16(src, ref);
|
||||
|
||||
g_dst.download(dst);
|
||||
EXPECT_MAT_NEAR(dst, ref, 0.0);
|
||||
}
|
||||
};
|
||||
|
||||
TYPED_TEST_CASE(CvtTest, AllTypes);
|
||||
|
||||
TYPED_TEST(CvtTest, GpuMat)
|
||||
{
|
||||
CvtTest<TypeParam>::test_gpumat();
|
||||
}
|
||||
|
||||
TYPED_TEST_CASE(CvFp16Test, Fp16Types);
|
||||
|
||||
TYPED_TEST(CvFp16Test, GpuMat)
|
||||
{
|
||||
CvFp16Test<TypeParam>::test_gpumat();
|
||||
}
|
||||
|
Loading…
Reference in New Issue
Block a user