diff options
author | Louis Verhaard <louis.verhaard@arm.com> | 2020-11-25 14:10:30 +0100 |
---|---|---|
committer | Louis Verhaard <louis.verhaard@arm.com> | 2020-11-26 08:13:50 +0100 |
commit | 933f55ea6f686d0cf390f4767e87a391686c3df8 (patch) | |
tree | 370321021ef2553df76e6e46b127cac07ec9d8be /API.md | |
parent | 34b9dc15b27219bd6485eb5104506d647e1f6d29 (diff) | |
download | ethos-u-vela-933f55ea6f686d0cf390f4767e87a391686c3df8.tar.gz |
MLBEDSW-3599: Added API for finding block configs
Added public API function npu_find_block_configs.
Change-Id: Ib0925a62d7c5d19a9b9fbd8d808943c2ea2df02f
Signed-off-by: Louis Verhaard <louis.verhaard@arm.com>
Diffstat (limited to 'API.md')
-rw-r--r-- | API.md | 7 |
1 files changed, 7 insertions, 0 deletions
@@ -40,6 +40,13 @@ these basic NPU operations. Note that the compiler is responsible for all address planning, i.e. it needs to supply addresses of all input and output tensors, weights, and biases. +### Finding block configs + +For all NPU operations, a block config must be set, which is the unit of work in +which the NPU generates the output. There are restrictions to the size of block +configs. Function `npu_find_block_configs` can be used to find valid block +configs for an operation. + ### Encoding of weights and biases All weights that are used in the NPU operations must be encoded using |