diff options
| author | Mike Kelly <mike.kelly@arm.com> | 2020-10-14 11:48:21 +0100 |
|---|---|---|
| committer | TeresaARM <teresa.charlinreyes@arm.com> | 2020-10-14 12:05:00 +0000 |
| commit | f9f33a04626756b73e6fd5c89092fd4bcb504b16 (patch) | |
| tree | d9162d3186993bd01c24cfc99bb384bb4e7b7090 /src | |
| parent | c456c31a5871a4ae1a8bbc687557ff304078aacc (diff) | |
| download | armnn-f9f33a04626756b73e6fd5c89092fd4bcb504b16.tar.gz | |
IVGCVSW-5335 Added Documentation for fast_math
* Added Documentation for fast_math to CLBackendModelContext
* Added Documentation for fast_math to NeonBackendModelContext
Signed-off-by: Mike Kelly <mike.kelly@arm.com>
Change-Id: I43a0568ae6914e074a80130a051e5d9bb849f2ba
Diffstat (limited to 'src')
| -rw-r--r-- | src/backends/cl/ClBackendModelContext.hpp | 5 | ||||
| -rw-r--r-- | src/backends/neon/NeonBackendModelContext.hpp | 5 |
2 files changed, 10 insertions, 0 deletions
diff --git a/src/backends/cl/ClBackendModelContext.hpp b/src/backends/cl/ClBackendModelContext.hpp index 59f7f8ff92..577649aafb 100644 --- a/src/backends/cl/ClBackendModelContext.hpp +++ b/src/backends/cl/ClBackendModelContext.hpp @@ -9,6 +9,11 @@ namespace armnn { +/// The ClBackendModelContext is used to pass in CL specific backend ModelOptions. The supported backend ModelOptions +/// are: +/// - "FastMathEnabled"\n +/// Using the fast_math flag can lead to performance improvements in fp32 and fp16 layers but may result in\n +/// results with reduced or different precision. The fast_math flag will not have any effect on int8 performance. class ClBackendModelContext : public IBackendModelContext { public: diff --git a/src/backends/neon/NeonBackendModelContext.hpp b/src/backends/neon/NeonBackendModelContext.hpp index 938d8af1cd..430ae45a6f 100644 --- a/src/backends/neon/NeonBackendModelContext.hpp +++ b/src/backends/neon/NeonBackendModelContext.hpp @@ -9,6 +9,11 @@ namespace armnn { +/// The NeonBackendModelContext is used to pass in Neon specific backend ModelOptions. The supported backend +/// ModelOptions are: +/// - "FastMathEnabled"\n +/// Using the fast_math flag can lead to performance improvements in fp32 and fp16 layers but may result in\n +/// results with reduced or different precision. The fast_math flag will not have any effect on int8 performance. class NeonBackendModelContext : public IBackendModelContext { public: |