Relax TestGpuNnapi accuracy check with a tolerance. am: 97b7166ad0 am: a2cd69d871 am: 627387d9e8
Original change: https://googleplex-android-review.googlesource.com/c/platform/packages/modules/NeuralNetworks/+/15756398
Change-Id: I92001b7a8a4aeafb8470d10247464c940d8d592b
diff --git a/runtime/test/TestGpuNnapi.cpp b/runtime/test/TestGpuNnapi.cpp
index 2782e9d..61809c4 100644
--- a/runtime/test/TestGpuNnapi.cpp
+++ b/runtime/test/TestGpuNnapi.cpp
@@ -103,6 +103,7 @@
// - ElementType: The corresponding C++ type. Use sizeof(ElementType) to get the element size.
// - kIsQuantized: Whether the data type is a quantized type or not.
// - kClearData: The CLEAR_DATA used in the compute shader.
+// - kTolerance: The absolute tolerance used to check the computation result.
template <Type dataType>
struct TestTypeHelper;
template <>
@@ -111,6 +112,7 @@
static constexpr bool kIsQuantized = false;
// One float32 of value (1.0) packed into uint32_t
static constexpr uint32_t kClearData = 0x3f800000;
+ static constexpr double kTolerance = 1e-6;
};
template <>
struct TestTypeHelper<Type::TENSOR_FLOAT16> {
@@ -118,6 +120,7 @@
static constexpr bool kIsQuantized = false;
// Two float16 of value (1.0) packed into uint32_t
static constexpr uint32_t kClearData = 0x3c003c00;
+ static constexpr double kTolerance = 1e-3;
};
template <>
struct TestTypeHelper<Type::TENSOR_QUANT8_ASYMM> {
@@ -125,6 +128,7 @@
static constexpr bool kIsQuantized = true;
// Four uint8_t of value (1) packed into uint32_t
static constexpr uint32_t kClearData = 0x01010101;
+ static constexpr double kTolerance = 0;
};
template <>
struct TestTypeHelper<Type::TENSOR_QUANT8_ASYMM_SIGNED> {
@@ -132,6 +136,7 @@
static constexpr bool kIsQuantized = true;
// Four int8_t of value (1) packed into uint32_t
static constexpr uint32_t kClearData = 0x01010101;
+ static constexpr double kTolerance = 0;
};
bool isExtensionSupported(const std::vector<VkExtensionProperties>& supportedExtensions,
@@ -960,12 +965,13 @@
auto [nnapiSuccess, nnapiSyncFd] = nnapi->run(gpuSyncFd);
ASSERT_TRUE(nnapiSuccess);
- checkResults<dataType>(std::move(nnapiSyncFd));
+ const double tolerance = TestTypeHelper<dataType>::kTolerance;
+ checkResults<dataType>(std::move(nnapiSyncFd), tolerance);
}
}
template <Type dataType>
- void checkResults(base::unique_fd syncFd) {
+ void checkResults(base::unique_fd syncFd, double tolerance) {
using ElementType = typename TestTypeHelper<dataType>::ElementType;
// Lock the buffer with the sync fence
@@ -981,12 +987,14 @@
for (uint32_t i = 0; i < kOperandLength; i++) {
const ElementType actual = reinterpret_cast<ElementType*>(data)[i];
- // We expect bit-exact here because the arithmetic is trivial, and all intermediate
- // and final results can be exactly represented by the primary data type.
- if (actual != expected) {
+ // We expect the absolute difference in double is within the tolerance.
+ const double expected_f64 = static_cast<double>(expected);
+ const double actual_f64 = static_cast<double>(actual);
+ const double diff = std::abs(expected_f64 - actual_f64);
+ if (diff > tolerance) {
// Print at most kMaxNumberOfPrintedErrors errors by EXPECT_EQ
if (numberOfErrors < kMaxNumberOfPrintedErrors) {
- EXPECT_EQ(actual, expected)
+ EXPECT_NEAR(actual_f64, expected_f64, tolerance)
<< "When comparing element [" << kOperandLength / kOperandSizeX << ", "
<< kOperandLength % kOperandSizeX << "]";
}
