diff options
| author | Patrik Gustavsson <patrik.gustavsson@arm.com> | 2021-04-08 09:04:00 +0200 |
|---|---|---|
| committer | Patrik Gustavsson <patrik.gustavsson@arm.com> | 2021-04-08 11:21:35 +0200 |
| commit | ee99bb124b088430b97d205df9fc90a1e9412e0c (patch) | |
| tree | bfb1a6799f9e9ff9d0f387471277b8a26edbab70 /ethosu/vela/npu_performance.py | |
| parent | 95b279f1454d58a93238851cb5ff394c7782ad32 (diff) | |
| download | ethos-u-vela-ee99bb124b088430b97d205df9fc90a1e9412e0c.tar.gz | |
MLBEDSW-4334 Non-linear format decision in graph opt.
Check if non linear tensor format can be used is
refactored.
-Flag avoid_NHCWB16 replaced with needs_linear_format
-Checking restrictions located to one function in graph optimiser.
Signed-off-by: Patrik Gustavsson <patrik.gustavsson@arm.com>
Change-Id: Iec5c7996a1a6039cad052197f1ae56f7c0290440
Diffstat (limited to 'ethosu/vela/npu_performance.py')
| -rw-r--r-- | ethosu/vela/npu_performance.py | 3 |
1 files changed, 2 insertions, 1 deletions
diff --git a/ethosu/vela/npu_performance.py b/ethosu/vela/npu_performance.py index e315f1f1..c83f8f52 100644 --- a/ethosu/vela/npu_performance.py +++ b/ethosu/vela/npu_performance.py @@ -389,7 +389,8 @@ def estimate_memory_transfer_efficiency( elem_size = tensor.dtype.size_in_bytes() is_ifm = direction == BandwidthDirection.Read tens = tensor.clone() - if not tens.avoid_NHCWB16: + + if not tensor.needs_linear_format: tens.set_format(TensorFormat.NHCWB16, arch) strides = tens.get_strides(shape4D=shape4D) |