Modify short sweep to run only required test cases.

tenstorrent · Oct 15, 2024 · 6345745 · 6345745
1 parent 4a37127
commit 6345745
Show file tree

Hide file tree

Showing 2 changed files with 4 additions and 93 deletions.
diff --git a/tests/sweep_framework/sweep_utils/conv2d_common.py b/tests/sweep_framework/sweep_utils/conv2d_common.py
@@ -171,23 +171,7 @@ def run_full(
 
 def run_short(
     input_specs,
-    transpose_mcast,
-    output_layout,
-    enable_act_double_buffer,
-    enable_split_reader,
-    enable_subblock_padding,
-    activations_dtype,
-    weights_dtype,
-    math_fidelity,
-    fp32_accum,
-    packer_l1_acc,
-    override_sharding_config,
-    core_grid,
-    use_shallow_conv_variant,
-    deallocate_activation,
-    enable_auto_formatting,
     device,
-    padded_input_channels=None,
 ) -> list:
     [
         batch_size,
@@ -205,6 +189,8 @@ def run_short(
         has_bias,
         dilation,
     ] = input_specs
+    print(input_specs)
+
     conv_input_shape = [batch_size, input_channels, input_height, input_width]
     conv_weight_shape = [output_channels, input_channels // groups, kernel_height, kernel_width]
     conv_bias_shape = [1, 1, 1, output_channels]
@@ -226,39 +212,13 @@ def run_short(
         groups=groups,
     )
 
-    tt_weight_tensor = ttnn.from_torch(
-        torch_weight_tensor, weights_dtype if weights_dtype != ttnn.bfloat8_b else ttnn.float32
-    )
+    tt_weight_tensor = ttnn.from_torch(torch_weight_tensor, ttnn.bfloat16)
     tt_bias_tensor = None
     if has_bias:
-        tt_bias_tensor = ttnn.from_torch(
-            torch_bias_tensor, weights_dtype if weights_dtype != ttnn.bfloat8_b else ttnn.float32
-        )
+        tt_bias_tensor = ttnn.from_torch(torch_bias_tensor, ttnn.bfloat16)
 
     tt_input_tensor = ttnn.from_torch(torch_input_tensor, ttnn.bfloat16)
 
-    conv_config = ttnn.Conv2dConfig(
-        dtype=activations_dtype,
-        weights_dtype=weights_dtype,
-        math_fidelity=math_fidelity,
-        shard_layout=None,
-        deallocate_activation=deallocate_activation,
-        fp32_dest_acc_enabled=fp32_accum,
-        packer_l1_accum_enabled=packer_l1_acc,
-        override_sharding_config=override_sharding_config,
-        output_layout=output_layout,
-        enable_act_double_buffer=enable_act_double_buffer,
-        enable_split_reader=enable_split_reader,
-        enable_subblock_padding=enable_subblock_padding,
-    )
-
-    if override_sharding_config:
-        if len(core_grid) == 2:
-            conv_config.core_grid = ttnn.CoreRangeSet({ttnn.CoreRange(core_grid[0], core_grid[1])})
-        elif len(core_grid) == 4:
-            conv_config.core_grid = ttnn.CoreRangeSet(
-                {ttnn.CoreRange(core_grid[0], core_grid[1]), ttnn.CoreRange(core_grid[2], core_grid[3])}
-            )
     start_time = start_measuring_time()
     [tt_output_tensor_on_device, out_height, out_width, weights_device, bias_device] = ttnn.conv2d(
         input_tensor=tt_input_tensor,
@@ -274,7 +234,6 @@ def run_short(
         batch_size=batch_size,
         input_height=input_height,
         input_width=input_width,
-        conv_config=conv_config,
         groups=groups,
     )
 

diff --git a/tests/sweep_framework/sweeps/conv2d/short/conv2d_short_sweep.py b/tests/sweep_framework/sweeps/conv2d/short/conv2d_short_sweep.py
@@ -421,22 +421,6 @@
             [1, 960, 960, 7, 7, 3, 3, 1, 1, 1, 1, 960, False, 1],
             [1, 960, 960, 7, 7, 5, 5, 1, 1, 2, 2, 960, False, 1],
         ],
-        "transpose_mcast": [False],
-        "output_layout": [ttnn.TILE_LAYOUT],
-        "enable_act_double_buffer": [False],
-        "enable_split_reader": [False],
-        "enable_subblock_padding": [False],
-        "activations_dtype": [ttnn.bfloat8_b],
-        "weights_dtype": [ttnn.bfloat8_b],
-        "math_fidelity": [ttnn.MathFidelity.LoFi],
-        "fp32_accum": [True],
-        "packer_l1_acc": [True],
-        "override_sharding_config": [False],
-        "core_grid": [None],
-        "use_shallow_conv_variant": [False],
-        "deallocate_activation": [False],
-        "enable_auto_formatting": [False],
-        "padded_input_channels": [None],
     },
 }
 
@@ -447,42 +431,10 @@ def invalidate_vector(test_vector) -> Tuple[bool, Optional[str]]:
 
 def run(
     input_specs,
-    transpose_mcast,
-    output_layout,
-    enable_act_double_buffer,
-    enable_split_reader,
-    enable_subblock_padding,
-    activations_dtype,
-    weights_dtype,
-    math_fidelity,
-    fp32_accum,
-    packer_l1_acc,
-    override_sharding_config,
-    core_grid,
-    use_shallow_conv_variant,
-    deallocate_activation,
-    enable_auto_formatting,
-    padded_input_channels=None,
     *,
     device,
 ) -> list:
     return run_short(
         input_specs,
-        transpose_mcast,
-        output_layout,
-        enable_act_double_buffer,
-        enable_split_reader,
-        enable_subblock_padding,
-        activations_dtype,
-        weights_dtype,
-        math_fidelity,
-        fp32_accum,
-        packer_l1_acc,
-        override_sharding_config,
-        core_grid,
-        use_shallow_conv_variant,
-        deallocate_activation,
-        enable_auto_formatting,
         device,
-        padded_input_channels,
     )