diff options
| author | Mike Kelly <mike.kelly@arm.com> | 2020-10-08 12:19:01 +0100 |
|---|---|---|
| committer | Mike Kelly <mike.kelly@arm.com> | 2020-10-12 15:44:56 +0100 |
| commit | fc6ab47673057d77b06507e9a32af82a90b28540 (patch) | |
| tree | 9366a8b3eae826149f5ea1ae080fcfa90b60e3f8 /tests | |
| parent | fbab30b63ee557f26467835554ec544f0bf150d5 (diff) | |
| download | armnn-fc6ab47673057d77b06507e9a32af82a90b28540.tar.gz | |
IVGCVSW-5335 Documentation for fast_math
* Changed documentation for fast_math to add warning about possibly reduction in precision.
Signed-off-by: Mike Kelly <mike.kelly@arm.com>
Change-Id: If954471efc6aef702e8490585571815d9e19b3fc
Diffstat (limited to 'tests')
| -rw-r--r-- | tests/ExecuteNetwork/ExecuteNetwork.cpp | 3 |
1 files changed, 2 insertions, 1 deletions
diff --git a/tests/ExecuteNetwork/ExecuteNetwork.cpp b/tests/ExecuteNetwork/ExecuteNetwork.cpp index c15e33fc13..58f1bd3783 100644 --- a/tests/ExecuteNetwork/ExecuteNetwork.cpp +++ b/tests/ExecuteNetwork/ExecuteNetwork.cpp @@ -139,7 +139,8 @@ int main(int argc, const char* argv[]) "Infers output tensor shape from input tensor shape and validate where applicable (where supported by " "parser)") ("enable-fast-math", po::bool_switch()->default_value(false), - "Enable fast_math computation of Convolution2D operator where applicable (where supported by backend)"); + "Enables fast_math options in backends that support it. Using the fast_math flag can lead to " + "performance improvements but may result in reduced or different precision."); } catch (const std::exception& e) { |