diff options
| author | Dwight Lidman <dwight.lidman@arm.com> | 2020-05-06 14:09:17 +0200 |
|---|---|---|
| committer | Tim Hall <tim.hall@arm.com> | 2020-06-18 17:53:52 +0100 |
| commit | 0538a77dfe65b66efedc4e4b92d851a5aeb46ea4 (patch) | |
| tree | 73afa3d08b717f33533de0eb6c0c5ebe6754bfe7 | |
| parent | 10349e70e994130d9e1ac8b72f32575025b2b453 (diff) | |
| download | ethos-u-vela-0538a77dfe65b66efedc4e4b92d851a5aeb46ea4.tar.gz | |
MLBEDSW-1970: Add stride 3 support
This patch adds support for strides of size 3.
It removes some obsolete code for a corner case that
no longer exists.
It also changes the setting of the bitfield in
NPU_SET_KERNEL_STRIDE so that it matches the specification.
Change-Id: I7dabcf72b7826ca0b3c98e9d23209027204079a8
Signed-off-by: Dwight Lidman <dwight.lidman@arm.com>
| -rw-r--r-- | ethosu/vela/architecture_features.py | 20 | ||||
| -rw-r--r-- | ethosu/vela/register_command_stream_generator.py | 11 | ||||
| -rw-r--r-- | ethosu/vela/supported_operators.py | 4 |
3 files changed, 13 insertions, 22 deletions
diff --git a/ethosu/vela/architecture_features.py b/ethosu/vela/architecture_features.py index 4a03d0ef..51c632e0 100644 --- a/ethosu/vela/architecture_features.py +++ b/ethosu/vela/architecture_features.py @@ -348,15 +348,7 @@ Note the difference between ArchitectureFeatures and CompilerOptions + ifm_odd_2x_height_enable ) // upscaling - if kernel.stride.y == 1: - ifm_block_height = round_up(ifm_block_height, self.ofm_ublock.height) - elif kernel.stride.y == 2: - if (self.ofm_ublock.height == 2) and (ifm_block_height % 4 == 2): - ifm_block_height = ifm_block_height + 2 - else: - ifm_block_height = round_up(ifm_block_height, self.ofm_ublock.height) - else: - assert False + ifm_block_height = round_up(ifm_block_height, self.ofm_ublock.height) # Width ifm_odd_2x_width_enable = 0 @@ -367,15 +359,7 @@ Note the difference between ArchitectureFeatures and CompilerOptions + ifm_odd_2x_width_enable ) // upscaling - if kernel.stride.x == 1: - ifm_block_width = round_up(ifm_block_width, self.ofm_ublock.width) - elif kernel.stride.x == 2: - if (self.ofm_ublock.width == 2) and (ifm_block_width % 4 == 2): - ifm_block_width = ifm_block_width + 2 - else: - ifm_block_width = round_up(ifm_block_width, self.ofm_ublock.width) - else: - assert False + ifm_block_width = round_up(ifm_block_width, self.ofm_ublock.width) return Block(ifm_block_width, ifm_block_height, ifm_block_depth) diff --git a/ethosu/vela/register_command_stream_generator.py b/ethosu/vela/register_command_stream_generator.py index a2b2f4d0..120cf8b1 100644 --- a/ethosu/vela/register_command_stream_generator.py +++ b/ethosu/vela/register_command_stream_generator.py @@ -569,8 +569,15 @@ def generate_register_command_stream(nng, sg, arch, verbose=False): emit.cmd0_with_param(cmd0.NPU_SET_IFM_PAD_BOTTOM, explicit_padding[2]) emit.cmd0_with_param(cmd0.NPU_SET_IFM_PAD_RIGHT, explicit_padding[3]) - stride = primary_op.attrs["strides"][2] - 1 - stride |= (primary_op.attrs["strides"][1] - 1) << 1 + # set kernel x stride low bit + stride = primary_op.attrs["strides"][2] - 1 & 1 + # set kernel y stride low bit + stride |= (primary_op.attrs["strides"][1] - 1 & 1) << 1 + # set kernel x stride extension bits + stride |= (primary_op.attrs["strides"][2] - 1 >> 1) << 6 + # set kernel y stride extension bits + stride |= (primary_op.attrs["strides"][1] - 1 >> 1) << 9 + if npu_block_type == NpuBlockType.Pooling: k_height, k_width = primary_op.attrs["ksize"][1:3] diff --git a/ethosu/vela/supported_operators.py b/ethosu/vela/supported_operators.py index fb3061f7..1e11d788 100644 --- a/ethosu/vela/supported_operators.py +++ b/ethosu/vela/supported_operators.py @@ -142,7 +142,7 @@ class SupportedOperators: def check_convolution_restrictions(self, op): # check stride - if op.attrs["stride_w"] > 2 or op.attrs["stride_h"] > 2: + if op.attrs["stride_w"] > 3 or op.attrs["stride_h"] > 3: return False # check dilation @@ -182,7 +182,7 @@ class SupportedOperators: def check_pooling_restrictions(self, op): # check stride - if op.attrs["stride_w"] > 2 or op.attrs["stride_h"] > 2: + if op.attrs["stride_w"] > 3 or op.attrs["stride_h"] > 3: return False # check data type |