Skip to content

Fix bugs in 2D FFTs and add tests #587

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
cliffburdick merged 1 commit into from
Mar 11, 2024
Merged

Fix bugs in 2D FFTs and add tests #587

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
14 changes: 7 additions & 7 deletions include/matx/transforms/fft.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -235,23 +235,23 @@ template <typename OutTensorType, typename InTensorType> class matxFFTPlan_t {
else if (fft_rank == 2) {
if (params.transform_type == CUFFT_C2R ||
params.transform_type == CUFFT_Z2D) {
params.n[0] = o.Size(RANK-1);
params.n[1] = o.Size(RANK-2);
params.n[1] = o.Size(RANK-1);
params.n[0] = o.Size(RANK-2);
}
else {
params.n[1] = i.Size(RANK-1);
params.n[0] = i.Size(RANK-2);
}

params.batch = (RANK == 2) ? 1 : i.Size(RANK - 3);
params.inembed[1] = o.Size(RANK-1);
params.onembed[1] = i.Size(RANK-1);
params.inembed[1] = i.Size(RANK-1);
params.onembed[1] = o.Size(RANK-1);
params.istride = i.Stride(RANK-1);
params.ostride = o.Stride(RANK-1);
params.idist = (RANK<=2) ? 1 : (int) i.Stride(RANK-3);
params.odist = (RANK<=2) ? 1 : (int) o.Stride(RANK-3);

if constexpr (is_complex_half_v<T1> || is_complex_half_v<T1>) {
if constexpr (is_complex_half_v<T1> || is_half_v<T1>) {
if ((params.n[0] & (params.n[0] - 1)) != 0 ||
(params.n[1] & (params.n[1] - 1)) != 0) {
MATX_THROW(matxInvalidDim,
Expand Down Expand Up @@ -367,7 +367,7 @@ template <typename OutTensorType, typename InTensorType> class matxFFTPlan_t {
if constexpr (is_complex_half_v<T2>) {
return CUFFT_C2C;
}
else if constexpr (is_half_v<T1>) {
else if constexpr (is_half_v<T2>) {
return CUFFT_R2C;
}
}
Expand Down Expand Up @@ -1057,7 +1057,7 @@ __MATX_INLINE__ void ifft2_impl(OutputTensor o, const InputTensor i,
}

// Get parameters required by these tensors
auto params = detail::matxFFTPlan_t<decltype(in), decltype(out)>::GetFFTParams(out, in, 2);
auto params = detail::matxFFTPlan_t<decltype(out), decltype(in)>::GetFFTParams(out, in, 2);
params.stream = stream;

// Get cache or new FFT plan if it doesn't exist
Expand Down
152 changes: 152 additions & 0 deletions test/00_transform/FFT.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -640,6 +640,65 @@ TYPED_TEST(FFTTestComplexTypes, FFT2D16C2C)
MATX_EXIT_HANDLER();
}

TYPED_TEST(FFTTestComplexTypes, FFT2D16x32C2C)
{
MATX_ENTER_HANDLER();
const index_t fft_dim[] = {16, 32};
this->pb->template InitAndRunTVGenerator<TypeParam>(
"00_transforms", "fft_operators", "fft_2d", {fft_dim[0], fft_dim[1]});

tensor_t<TypeParam, 2> av{{fft_dim[0], fft_dim[1]}};
tensor_t<TypeParam, 2> avo{{fft_dim[0], fft_dim[1]}};
this->pb->NumpyToTensorView(av, "a_in");

(avo = fft2(av)).run();
cudaStreamSynchronize(0);

MATX_TEST_ASSERT_COMPARE(this->pb, avo, "a_out", this->thresh);
MATX_EXIT_HANDLER();
}

TYPED_TEST(FFTTestComplexTypes, FFT2D16BatchedC2C)
{
MATX_ENTER_HANDLER();
const index_t batch_size = 10;
const index_t fft_dim = 16;
this->pb->template InitAndRunTVGenerator<TypeParam>(
"00_transforms", "fft_operators", "fft_2d_batched",
{batch_size, fft_dim, fft_dim});

tensor_t<TypeParam, 3> av{{batch_size, fft_dim, fft_dim}};
tensor_t<TypeParam, 3> avo{{batch_size, fft_dim, fft_dim}};
this->pb->NumpyToTensorView(av, "a_in");

(avo = fft2(av)).run();
cudaStreamSynchronize(0);

MATX_TEST_ASSERT_COMPARE(this->pb, avo, "a_out", this->thresh);
MATX_EXIT_HANDLER();
}

TYPED_TEST(FFTTestComplexTypes, FFT2D16BatchedStridedC2C)
{
MATX_ENTER_HANDLER();
const index_t batch_size = 10;
const index_t fft_dim = 16;
this->pb->template InitAndRunTVGenerator<TypeParam>(
"00_transforms", "fft_operators", "fft_2d_batched_strided",
{fft_dim, batch_size, fft_dim});

tensor_t<TypeParam, 3> av{{fft_dim, batch_size, fft_dim}};
tensor_t<TypeParam, 3> avo{{fft_dim, batch_size, fft_dim}};
this->pb->NumpyToTensorView(av, "a_in");

const int32_t axes[] = {0, 2};
(avo = fft2(av, axes)).run();
cudaStreamSynchronize(0);

MATX_TEST_ASSERT_COMPARE(this->pb, avo, "a_out", this->thresh);
MATX_EXIT_HANDLER();
}

TYPED_TEST(FFTTestComplexTypes, IFFT2D16C2C)
{
MATX_ENTER_HANDLER();
Expand All @@ -658,6 +717,99 @@ TYPED_TEST(FFTTestComplexTypes, IFFT2D16C2C)
MATX_EXIT_HANDLER();
}

TYPED_TEST(FFTTestComplexTypes, IFFT2D16x32C2C)
{
MATX_ENTER_HANDLER();
const index_t fft_dim[] = {16, 32};
this->pb->template InitAndRunTVGenerator<TypeParam>(
"00_transforms", "fft_operators", "ifft_2d", {fft_dim[0], fft_dim[1]});

tensor_t<TypeParam, 2> av{{fft_dim[0], fft_dim[1]}};
tensor_t<TypeParam, 2> avo{{fft_dim[0], fft_dim[1]}};
this->pb->NumpyToTensorView(av, "a_in");

(avo = ifft2(av)).run();
cudaStreamSynchronize(0);

MATX_TEST_ASSERT_COMPARE(this->pb, avo, "a_out", this->thresh);
MATX_EXIT_HANDLER();
}

TYPED_TEST(FFTTestComplexNonHalfTypes, FFT2D16R2C)
{
MATX_ENTER_HANDLER();
const index_t fft_dim = 16;
using rtype = typename TypeParam::value_type;
this->pb->template InitAndRunTVGenerator<rtype>(
"00_transforms", "fft_operators", "rfft_2d", {fft_dim, fft_dim});

tensor_t<rtype, 2> av{{fft_dim, fft_dim}};
tensor_t<TypeParam, 2> avo{{fft_dim, fft_dim / 2 + 1}};
this->pb->NumpyToTensorView(av, "a_in");

(avo = fft2(av)).run();
cudaStreamSynchronize(0);

MATX_TEST_ASSERT_COMPARE(this->pb, avo, "a_out", this->thresh);
MATX_EXIT_HANDLER();
}

TYPED_TEST(FFTTestComplexNonHalfTypes, FFT2D16x32R2C)
{
MATX_ENTER_HANDLER();
const index_t fft_dim[] = {16, 32};
using rtype = typename TypeParam::value_type;
this->pb->template InitAndRunTVGenerator<rtype>(
"00_transforms", "fft_operators", "rfft_2d", {fft_dim[0], fft_dim[1]});

tensor_t<rtype, 2> av{{fft_dim[0], fft_dim[1]}};
tensor_t<TypeParam, 2> avo{{fft_dim[0], fft_dim[1] / 2 + 1}};
this->pb->NumpyToTensorView(av, "a_in");

(avo = fft2(av)).run();
cudaStreamSynchronize(0);

MATX_TEST_ASSERT_COMPARE(this->pb, avo, "a_out", this->thresh);
MATX_EXIT_HANDLER();
}

TYPED_TEST(FFTTestComplexNonHalfTypes, IFFT2D16C2R)
{
MATX_ENTER_HANDLER();
const index_t fft_dim = 16;
using rtype = typename TypeParam::value_type;
this->pb->template InitAndRunTVGenerator<TypeParam>(
"00_transforms", "fft_operators", "irfft_2d", {fft_dim, fft_dim});

tensor_t<TypeParam, 2> av{{fft_dim, fft_dim / 2 + 1}};
tensor_t<rtype, 2> avo{{fft_dim, fft_dim}};
this->pb->NumpyToTensorView(av, "a_in");

(avo = ifft2(av)).run();
cudaStreamSynchronize(0);

MATX_TEST_ASSERT_COMPARE(this->pb, avo, "a_out", this->thresh);
MATX_EXIT_HANDLER();
}

TYPED_TEST(FFTTestComplexNonHalfTypes, IFFT2D16x32C2R)
{
MATX_ENTER_HANDLER();
const index_t fft_dim[] = {16, 32};
using rtype = typename TypeParam::value_type;
this->pb->template InitAndRunTVGenerator<TypeParam>(
"00_transforms", "fft_operators", "irfft_2d", {fft_dim[0], fft_dim[1]});

tensor_t<TypeParam, 2> av{{fft_dim[0], fft_dim[1] / 2 + 1}};
tensor_t<rtype, 2> avo{{fft_dim[0], fft_dim[1]}};
this->pb->NumpyToTensorView(av, "a_in");

(avo = ifft2(av)).run();
cudaStreamSynchronize(0);

MATX_TEST_ASSERT_COMPARE(this->pb, avo, "a_out", this->thresh);
MATX_EXIT_HANDLER();
}

TYPED_TEST(FFTTestComplexNonHalfTypes, FFT1D1024C2CShort)
{
Expand Down
36 changes: 34 additions & 2 deletions test/test_vectors/generators/00_transforms.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -319,15 +319,47 @@ def fft_2d(self) -> Dict[str, np.ndarray]:
(self.size[0], self.size[1]), self.dtype)
return {
'a_in': seq,
'a_out': np.fft.fft2(seq, (self.size[1], self.size[1]))
'a_out': np.fft.fft2(seq, (self.size[0], self.size[1]))
}

def fft_2d_batched(self) -> Dict[str, np.ndarray]:
seq = matx_common.randn_ndarray(
(self.size[0], self.size[1], self.size[2]), self.dtype)
return {
'a_in': seq,
'a_out': np.fft.fft2(seq, (self.size[1], self.size[2]))
}

def fft_2d_batched_strided(self) -> Dict[str, np.ndarray]:
seq = matx_common.randn_ndarray(
(self.size[0], self.size[1], self.size[2]), self.dtype)
return {
'a_in': seq,
'a_out': np.fft.fft2(seq, (self.size[0], self.size[2]), axes=(0, 2))
}

def ifft_2d(self) -> Dict[str, np.ndarray]:
seq = matx_common.randn_ndarray(
(self.size[0], self.size[1]), self.dtype)
return {
'a_in': seq,
'a_out': np.fft.ifft2(seq, (self.size[1], self.size[1]))
'a_out': np.fft.ifft2(seq, (self.size[0], self.size[1]))
}

def rfft_2d(self) -> Dict[str, np.ndarray]:
seq = matx_common.randn_ndarray(
(self.size[0], self.size[1]), self.dtype)
return {
'a_in': seq,
'a_out': np.fft.rfft2(seq, (self.size[0], self.size[1]))
}

def irfft_2d(self) -> Dict[str, np.ndarray]:
seq = matx_common.randn_ndarray(
(self.size[0], self.size[1]), self.dtype)
return {
'a_in': seq,
'a_out': np.fft.irfft2(seq, (self.size[0], self.size[1]))
}


Expand Down