mirror of
https://github.com/KhronosGroup/OpenCL-CTS.git
synced 2026-03-19 06:09:01 +00:00
3871149208
Before this change, `max_error` was initialized to `0.0f`, which caused issues when the actual maximum `error` was also `0.0f`. In such cases, `max_val` would never be updated, leading to misleading test output showing `NaN` values: ``` degrees: Max error 0.000000 ulps at 0: *nan vs 0x1.9802318e8abefp+21 ``` This fix initializes `max_error` to `-INFINITY` to ensure `max_val` is always updated at least once. Additionally, the log output now includes the input value for better debugging: ``` degrees: Max error 0.000000 ulps at 0, input 0x1.c7bffcp+15: *0x1.9802318e8abefp+21 vs 0x1.9802318e8abefp+21 ``` This makes the test output more informative and eliminates confusing `NaN` values in the summary.
402 lines
13 KiB
C++
402 lines
13 KiB
C++
//
|
|
// Copyright (c) 2023 The Khronos Group Inc.
|
|
//
|
|
// Licensed under the Apache License, Version 2.0 (the "License");
|
|
// you may not use this file except in compliance with the License.
|
|
// You may obtain a copy of the License at
|
|
//
|
|
// http://www.apache.org/licenses/LICENSE-2.0
|
|
//
|
|
// Unless required by applicable law or agreed to in writing, software
|
|
// distributed under the License is distributed on an "AS IS" BASIS,
|
|
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
// See the License for the specific language governing permissions and
|
|
// limitations under the License.
|
|
//
|
|
#include <stdio.h>
|
|
#include <string.h>
|
|
#include <sys/types.h>
|
|
#include <sys/stat.h>
|
|
|
|
#include <vector>
|
|
|
|
#include "harness/deviceInfo.h"
|
|
#include "harness/stringHelpers.h"
|
|
#include "harness/typeWrappers.h"
|
|
|
|
#include "test_base.h"
|
|
|
|
#ifndef M_PI
|
|
#define M_PI 3.14159265358979323846264338327950288
|
|
#endif
|
|
|
|
// clang-format off
|
|
const char *unary_fn_code_pattern =
|
|
"%s\n" /* optional pragma */
|
|
"__kernel void test_fn(__global %s%s *src, __global %s%s *dst)\n"
|
|
"{\n"
|
|
" int tid = get_global_id(0);\n"
|
|
"\n"
|
|
" dst[tid] = %s(src[tid]);\n"
|
|
"}\n";
|
|
|
|
const char *unary_fn_code_pattern_v3 =
|
|
"%s\n" /* optional pragma */
|
|
"__kernel void test_fn(__global %s *src, __global %s *dst)\n"
|
|
"{\n"
|
|
" int tid = get_global_id(0);\n"
|
|
"\n"
|
|
" vstore3(%s(vload3(tid,src)), tid, dst);\n"
|
|
"}\n";
|
|
// clang-format on
|
|
|
|
#define MAX_ERR 2.0f
|
|
|
|
namespace {
|
|
|
|
template <typename T>
|
|
int verify_degrees(const T *const inptr, const T *const outptr, int n)
|
|
{
|
|
float error, max_error = -INFINITY;
|
|
double r, max_val = NAN;
|
|
int max_index = 0;
|
|
|
|
for (int i = 0, j = 0; i < n; i++, j++)
|
|
{
|
|
r = (180.0 / M_PI) * conv_to_dbl(inptr[i]);
|
|
|
|
error = UlpFn(outptr[i], r);
|
|
|
|
if (fabsf(error) > max_error)
|
|
{
|
|
max_error = error;
|
|
max_index = i;
|
|
max_val = r;
|
|
if (fabsf(error) > MAX_ERR)
|
|
{
|
|
if (std::is_same<T, half>::value)
|
|
log_error(
|
|
"%d) Error @ %a: *%a vs %a (*%g vs %g) ulps: %f\n", i,
|
|
conv_to_flt(inptr[i]), r, conv_to_flt(outptr[i]), r,
|
|
conv_to_flt(outptr[i]), error);
|
|
else
|
|
log_error(
|
|
"%d) Error @ %a: *%a vs %a (*%g vs %g) ulps: %f\n", i,
|
|
inptr[i], r, outptr[i], r, outptr[i], error);
|
|
return 1;
|
|
}
|
|
}
|
|
}
|
|
|
|
if (std::is_same<T, half>::value)
|
|
log_info("degrees: Max error %f ulps at %d, input %a: *%a vs %a (*%g "
|
|
"vs %g)\n",
|
|
max_error, max_index, conv_to_flt(inptr[max_index]), max_val,
|
|
conv_to_flt(outptr[max_index]), max_val,
|
|
conv_to_flt(outptr[max_index]));
|
|
else
|
|
log_info("degrees: Max error %f ulps at %d, input %a: *%a vs %a (*%g "
|
|
"vs %g)\n",
|
|
max_error, max_index, conv_to_flt(inptr[max_index]), max_val,
|
|
outptr[max_index], max_val, outptr[max_index]);
|
|
|
|
return 0;
|
|
}
|
|
|
|
template <typename T>
|
|
int verify_radians(const T *const inptr, const T *const outptr, int n)
|
|
{
|
|
float error, max_error = -INFINITY;
|
|
double r, max_val = NAN;
|
|
int max_index = 0;
|
|
|
|
for (int i = 0, j = 0; i < n; i++, j++)
|
|
{
|
|
r = (M_PI / 180.0) * conv_to_dbl(inptr[i]);
|
|
|
|
error = UlpFn(outptr[i], r);
|
|
|
|
if (fabsf(error) > max_error)
|
|
{
|
|
max_error = error;
|
|
max_index = i;
|
|
max_val = r;
|
|
if (fabsf(error) > MAX_ERR)
|
|
{
|
|
if (std::is_same<T, half>::value)
|
|
log_error(
|
|
"%d) Error @ %a: *%a vs %a (*%g vs %g) ulps: %f\n", i,
|
|
conv_to_flt(inptr[i]), r, conv_to_flt(outptr[i]), r,
|
|
conv_to_flt(outptr[i]), error);
|
|
else
|
|
log_error(
|
|
"%d) Error @ %a: *%a vs %a (*%g vs %g) ulps: %f\n", i,
|
|
inptr[i], r, outptr[i], r, outptr[i], error);
|
|
return 1;
|
|
}
|
|
}
|
|
}
|
|
|
|
if (std::is_same<T, half>::value)
|
|
log_info("radians: Max error %f ulps at %d, input %a: *%a vs %a (*%g "
|
|
"vs %g)\n",
|
|
max_error, max_index, conv_to_flt(inptr[max_index]), max_val,
|
|
conv_to_flt(outptr[max_index]), max_val,
|
|
conv_to_flt(outptr[max_index]));
|
|
else
|
|
log_info("radians: Max error %f ulps at %d, input %a: *%a vs %a (*%g "
|
|
"vs %g)\n",
|
|
max_error, max_index, conv_to_flt(inptr[max_index]), max_val,
|
|
outptr[max_index], max_val, outptr[max_index]);
|
|
|
|
return 0;
|
|
}
|
|
|
|
template <typename T>
|
|
int verify_sign(const T *const inptr, const T *const outptr, int n)
|
|
{
|
|
double r = 0;
|
|
for (int i = 0; i < n; i++)
|
|
{
|
|
if (conv_to_dbl(inptr[i]) > 0.0f)
|
|
r = 1.0;
|
|
else if (conv_to_dbl(inptr[i]) < 0.0f)
|
|
r = -1.0;
|
|
else
|
|
r = 0.0;
|
|
if (r != conv_to_dbl(outptr[i])) return -1;
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
}
|
|
|
|
template <typename T>
|
|
int test_unary_fn(cl_device_id device, cl_context context,
|
|
cl_command_queue queue, int n_elems,
|
|
const std::string &fnName, VerifyFuncUnary<T> verifyFn)
|
|
{
|
|
clMemWrapper streams[2];
|
|
std::vector<T> input_ptr, output_ptr;
|
|
|
|
std::vector<clProgramWrapper> programs;
|
|
std::vector<clKernelWrapper> kernels;
|
|
|
|
int err, i;
|
|
MTdataHolder d = MTdataHolder(gRandomSeed);
|
|
|
|
assert(BaseFunctionTest::type2name.find(sizeof(T))
|
|
!= BaseFunctionTest::type2name.end());
|
|
auto tname = BaseFunctionTest::type2name[sizeof(T)];
|
|
|
|
programs.resize(kTotalVecCount);
|
|
kernels.resize(kTotalVecCount);
|
|
|
|
int num_elements = n_elems * (1 << (kTotalVecCount - 1));
|
|
|
|
input_ptr.resize(num_elements);
|
|
output_ptr.resize(num_elements);
|
|
|
|
for (i = 0; i < 2; i++)
|
|
{
|
|
streams[i] = clCreateBuffer(context, CL_MEM_READ_WRITE,
|
|
sizeof(T) * num_elements, NULL, &err);
|
|
test_error(err, "clCreateBuffer failed");
|
|
}
|
|
|
|
std::string pragma_str;
|
|
if (std::is_same<T, float>::value)
|
|
{
|
|
for (int j = 0; j < num_elements; j++)
|
|
{
|
|
input_ptr[j] = get_random_float((float)(-100000.f * M_PI),
|
|
(float)(100000.f * M_PI), d);
|
|
}
|
|
}
|
|
else if (std::is_same<T, double>::value)
|
|
{
|
|
pragma_str = "#pragma OPENCL EXTENSION cl_khr_fp64 : enable\n";
|
|
for (int j = 0; j < num_elements; j++)
|
|
{
|
|
input_ptr[j] =
|
|
get_random_double(-100000.0 * M_PI, 100000.0 * M_PI, d);
|
|
}
|
|
}
|
|
else if (std::is_same<T, half>::value)
|
|
{
|
|
pragma_str = "#pragma OPENCL EXTENSION cl_khr_fp16 : enable\n";
|
|
for (int j = 0; j < num_elements; j++)
|
|
{
|
|
input_ptr[j] = conv_to_half(get_random_float(
|
|
(float)(-10000.f * M_PI), (float)(10000.f * M_PI), d));
|
|
}
|
|
}
|
|
|
|
err = clEnqueueWriteBuffer(queue, streams[0], true, 0,
|
|
sizeof(T) * num_elements, &input_ptr.front(), 0,
|
|
NULL, NULL);
|
|
test_error(err, "clEnqueueWriteBuffer failed\n");
|
|
|
|
for (i = 0; i < kTotalVecCount; i++)
|
|
{
|
|
std::string kernelSource;
|
|
const char vecSizeNames[][3] = { "", "2", "4", "8", "16", "3" };
|
|
|
|
if (i >= kVectorSizeCount)
|
|
{
|
|
std::string str = unary_fn_code_pattern_v3;
|
|
kernelSource = str_sprintf(str, pragma_str.c_str(), tname.c_str(),
|
|
tname.c_str(), fnName.c_str());
|
|
}
|
|
else
|
|
{
|
|
std::string str = unary_fn_code_pattern;
|
|
kernelSource = str_sprintf(str, pragma_str.c_str(), tname.c_str(),
|
|
vecSizeNames[i], tname.c_str(),
|
|
vecSizeNames[i], fnName.c_str());
|
|
}
|
|
|
|
/* Create kernels */
|
|
const char *programPtr = kernelSource.c_str();
|
|
err =
|
|
create_single_kernel_helper(context, &programs[i], &kernels[i], 1,
|
|
(const char **)&programPtr, "test_fn");
|
|
|
|
err = clSetKernelArg(kernels[i], 0, sizeof streams[0], &streams[0]);
|
|
err |= clSetKernelArg(kernels[i], 1, sizeof streams[1], &streams[1]);
|
|
if (err != CL_SUCCESS)
|
|
{
|
|
log_error("clSetKernelArgs failed\n");
|
|
return -1;
|
|
}
|
|
|
|
// Line below is troublesome...
|
|
size_t threads = (size_t)num_elements / ((g_arrVecSizes[i]));
|
|
err = clEnqueueNDRangeKernel(queue, kernels[i], 1, NULL, &threads, NULL,
|
|
0, NULL, NULL);
|
|
if (err != CL_SUCCESS)
|
|
{
|
|
log_error("clEnqueueNDRangeKernel failed\n");
|
|
return -1;
|
|
}
|
|
|
|
cl_uint dead = 42;
|
|
memset_pattern4(&output_ptr[0], &dead, sizeof(T) * num_elements);
|
|
err = clEnqueueReadBuffer(queue, streams[1], true, 0,
|
|
sizeof(T) * num_elements, &output_ptr[0], 0,
|
|
NULL, NULL);
|
|
if (err != CL_SUCCESS)
|
|
{
|
|
log_error("clEnqueueReadBuffer failed\n");
|
|
return -1;
|
|
}
|
|
|
|
if (verifyFn((T *)&input_ptr.front(), (T *)&output_ptr.front(),
|
|
n_elems * (i + 1)))
|
|
{
|
|
log_error("%s %s%d test failed\n", fnName.c_str(), tname.c_str(),
|
|
((g_arrVecSizes[i])));
|
|
err = -1;
|
|
}
|
|
else
|
|
{
|
|
log_info("%s %s%d test passed\n", fnName.c_str(), tname.c_str(),
|
|
((g_arrVecSizes[i])));
|
|
}
|
|
|
|
if (err) break;
|
|
}
|
|
|
|
return err;
|
|
}
|
|
|
|
cl_int DegreesTest::Run()
|
|
{
|
|
cl_int error = CL_SUCCESS;
|
|
if (is_extension_available(device, "cl_khr_fp16"))
|
|
{
|
|
error = test_unary_fn<half>(device, context, queue, num_elems,
|
|
fnName.c_str(), verify_degrees<half>);
|
|
test_error(error, "DegreesTest::Run<cl_half> failed");
|
|
}
|
|
|
|
error = test_unary_fn<float>(device, context, queue, num_elems,
|
|
fnName.c_str(), verify_degrees<float>);
|
|
test_error(error, "DegreesTest::Run<float> failed");
|
|
|
|
if (is_extension_available(device, "cl_khr_fp64"))
|
|
{
|
|
error = test_unary_fn<double>(device, context, queue, num_elems,
|
|
fnName.c_str(), verify_degrees<double>);
|
|
test_error(error, "DegreesTest::Run<double> failed");
|
|
}
|
|
|
|
return error;
|
|
}
|
|
|
|
cl_int RadiansTest::Run()
|
|
{
|
|
cl_int error = CL_SUCCESS;
|
|
if (is_extension_available(device, "cl_khr_fp16"))
|
|
{
|
|
error = test_unary_fn<half>(device, context, queue, num_elems,
|
|
fnName.c_str(), verify_radians<half>);
|
|
test_error(error, "RadiansTest::Run<cl_half> failed");
|
|
}
|
|
|
|
error = test_unary_fn<float>(device, context, queue, num_elems,
|
|
fnName.c_str(), verify_radians<float>);
|
|
test_error(error, "RadiansTest::Run<float> failed");
|
|
|
|
if (is_extension_available(device, "cl_khr_fp64"))
|
|
{
|
|
error = test_unary_fn<double>(device, context, queue, num_elems,
|
|
fnName.c_str(), verify_radians<double>);
|
|
test_error(error, "RadiansTest::Run<double> failed");
|
|
}
|
|
|
|
return error;
|
|
}
|
|
|
|
cl_int SignTest::Run()
|
|
{
|
|
cl_int error = CL_SUCCESS;
|
|
if (is_extension_available(device, "cl_khr_fp16"))
|
|
{
|
|
error = test_unary_fn<half>(device, context, queue, num_elems,
|
|
fnName.c_str(), verify_sign<half>);
|
|
test_error(error, "SignTest::Run<cl_half> failed");
|
|
}
|
|
|
|
error = test_unary_fn<float>(device, context, queue, num_elems,
|
|
fnName.c_str(), verify_sign<float>);
|
|
test_error(error, "SignTest::Run<float> failed");
|
|
|
|
if (is_extension_available(device, "cl_khr_fp64"))
|
|
{
|
|
error = test_unary_fn<double>(device, context, queue, num_elems,
|
|
fnName.c_str(), verify_sign<double>);
|
|
test_error(error, "SignTest::Run<double> failed");
|
|
}
|
|
|
|
return error;
|
|
}
|
|
|
|
REGISTER_TEST(degrees)
|
|
{
|
|
return MakeAndRunTest<DegreesTest>(device, context, queue, num_elements,
|
|
"degrees");
|
|
}
|
|
|
|
REGISTER_TEST(radians)
|
|
{
|
|
return MakeAndRunTest<RadiansTest>(device, context, queue, num_elements,
|
|
"radians");
|
|
}
|
|
|
|
REGISTER_TEST(sign)
|
|
{
|
|
return MakeAndRunTest<SignTest>(device, context, queue, num_elements,
|
|
"sign");
|
|
}
|