-
Notifications
You must be signed in to change notification settings - Fork 97
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Merge branch 'main' into sminakov/tensor-constr-rem
- Loading branch information
Showing
54 changed files
with
2,112 additions
and
144 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
102 changes: 102 additions & 0 deletions
102
tests/sweep_framework/sweeps/eltwise/binary/add/add_llama.py
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,102 @@ | ||
| # SPDX-FileCopyrightText: © 2024 Tenstorrent Inc. | ||
|
|
||
| # SPDX-License-Identifier: Apache-2.0 | ||
|
|
||
| from typing import Optional, Tuple | ||
| from functools import partial | ||
|
|
||
| import torch | ||
| import ttnn | ||
| from tests.tt_eager.python_api_testing.sweep_tests.generation_funcs import gen_func_with_cast_tt | ||
|
|
||
| from tests.ttnn.utils_for_testing import check_with_pcc, start_measuring_time, stop_measuring_time | ||
| from models.utility_functions import torch_random | ||
|
|
||
|
|
||
| # Parameters provided to the test vector generator are defined here. | ||
| # They are defined as dict-type suites that contain the arguments to the run function as keys, and lists of possible inputs as values. | ||
| # Each suite has a key name (in this case "suite_1" and "suite_2") which will associate the test vectors to this specific suite of inputs. | ||
| # Developers can create their own generator functions and pass them to the parameters as inputs. | ||
|
|
||
|
|
||
| parameters = { | ||
| "nightly": { | ||
| "input_shape": [ | ||
| {"self": [1, 1, 32, 1024], "other": [1, 1, 32, 1024], "input_dtype": "ttnn.bfloat16"}, | ||
| {"self": [1, 1, 32, 4096], "other": [1, 1, 32, 4096], "input_dtype": "ttnn.bfloat16"}, | ||
| ], | ||
| "input_a_layout": [ttnn.TILE_LAYOUT], | ||
| "input_b_layout": [ttnn.TILE_LAYOUT], | ||
| }, | ||
| } | ||
|
|
||
|
|
||
| # Invalidate vector is called during the generation phase where each vector will be passed in. | ||
| # If invalidated, the vector will still be stored but will be skipped. | ||
| # Returns False, None if the vector is valid, and True, str with a reason for invalidation if it is invalid. | ||
| def invalidate_vector(test_vector) -> Tuple[bool, Optional[str]]: | ||
| if test_vector["input_a_layout"] == ttnn.ROW_MAJOR_LAYOUT or test_vector["input_b_layout"] == ttnn.ROW_MAJOR_LAYOUT: | ||
| return True, "Row Major layout is not supported" | ||
| return False, None | ||
|
|
||
|
|
||
| # This is the run instructions for the test, defined by the developer. | ||
| # The run function must take the above-defined parameters as inputs. | ||
| # The runner will call this run function with each test vector, and the returned results from this function will be stored. | ||
| # If you defined a mesh_device_fixture above, the object you yielded will be passed into this function as 'device'. Otherwise, it will be the default ttnn device opened by the infra. | ||
| def run( | ||
| input_shape, | ||
| input_a_layout, | ||
| input_b_layout, | ||
| *, | ||
| device, | ||
| ) -> list: | ||
| torch.manual_seed(0) | ||
| if input_shape["input_dtype"] == "ttnn.bfloat16": | ||
| input_dtype = ttnn.bfloat16 | ||
| elif input_shape["input_dtype"] == "ttnn.float32": | ||
| input_dtype = ttnn.float32 | ||
| elif input_shape["input_dtype"] == "ttnn.int32": | ||
| input_dtype = ttnn.int32 | ||
|
|
||
| torch_input_tensor_a = gen_func_with_cast_tt( | ||
| partial(torch_random, low=-100, high=100, dtype=torch.float32), input_dtype | ||
| )(input_shape["self"]) | ||
| torch_input_tensor_b = gen_func_with_cast_tt( | ||
| partial(torch_random, low=-100, high=100, dtype=torch.float32), input_dtype | ||
| )(input_shape["other"]) | ||
|
|
||
| sharded_config = ttnn.create_sharded_memory_config_( | ||
| shape=input_shape["self"], | ||
| core_grid=ttnn.CoreGrid(y=4, x=8), | ||
| strategy=ttnn.ShardStrategy.WIDTH, | ||
| orientation=ttnn.ShardOrientation.ROW_MAJOR, | ||
| use_height_and_width_as_shard_shape=False, | ||
| halo=0, | ||
| tile_layout=True, | ||
| ) | ||
| golden_function = ttnn.get_golden_function(ttnn.add) | ||
| torch_output_tensor = golden_function(torch_input_tensor_a, torch_input_tensor_b) | ||
|
|
||
| input_tensor_a = ttnn.from_torch( | ||
| torch_input_tensor_a, | ||
| dtype=input_dtype, | ||
| layout=input_a_layout, | ||
| device=device, | ||
| memory_config=sharded_config, | ||
| ) | ||
|
|
||
| input_tensor_b = ttnn.from_torch( | ||
| torch_input_tensor_b, | ||
| dtype=input_dtype, | ||
| layout=input_b_layout, | ||
| device=device, | ||
| memory_config=sharded_config, | ||
| ) | ||
|
|
||
| start_time = start_measuring_time() | ||
| result = ttnn.add(input_tensor_a, input_tensor_b, memory_config=sharded_config) | ||
| output_tensor = ttnn.to_torch(result) | ||
| e2e_perf = stop_measuring_time(start_time) | ||
|
|
||
| return [check_with_pcc(torch_output_tensor, output_tensor, 0.999), e2e_perf] |
103 changes: 103 additions & 0 deletions
103
tests/sweep_framework/sweeps/eltwise/binary/multiply/mul_no_act_llama.py
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,103 @@ | ||
| # SPDX-FileCopyrightText: © 2024 Tenstorrent Inc. | ||
|
|
||
| # SPDX-License-Identifier: Apache-2.0 | ||
|
|
||
| from typing import Optional, Tuple | ||
| from functools import partial | ||
|
|
||
| import torch | ||
| import ttnn | ||
| from tests.tt_eager.python_api_testing.sweep_tests.generation_funcs import gen_func_with_cast_tt | ||
|
|
||
| from tests.ttnn.utils_for_testing import check_with_pcc, start_measuring_time, stop_measuring_time | ||
| from models.utility_functions import torch_random | ||
|
|
||
|
|
||
| # Parameters provided to the test vector generator are defined here. | ||
| # They are defined as dict-type suites that contain the arguments to the run function as keys, and lists of possible inputs as values. | ||
| # Each suite has a key name (in this case "suite_1" and "suite_2") which will associate the test vectors to this specific suite of inputs. | ||
| # Developers can create their own generator functions and pass them to the parameters as inputs. | ||
|
|
||
|
|
||
| parameters = { | ||
| "nightly": { | ||
| "input_spec": [ | ||
| {"self": [1, 1, 32, 3584], "other": [1, 1, 32, 3584], "input_dtype": "ttnn.bfloat16", "y": 2}, | ||
| {"self": [1, 1, 32, 14336], "other": [1, 1, 32, 14336], "input_dtype": "ttnn.bfloat16", "y": 4}, | ||
| ], | ||
| "input_a_layout": [ttnn.TILE_LAYOUT], | ||
| "input_b_layout": [ttnn.TILE_LAYOUT], | ||
| }, | ||
| } | ||
|
|
||
|
|
||
| # Invalidate vector is called during the generation phase where each vector will be passed in. | ||
| # If invalidated, the vector will still be stored but will be skipped. | ||
| # Returns False, None if the vector is valid, and True, str with a reason for invalidation if it is invalid. | ||
| def invalidate_vector(test_vector) -> Tuple[bool, Optional[str]]: | ||
| if test_vector["input_a_layout"] == ttnn.ROW_MAJOR_LAYOUT or test_vector["input_b_layout"] == ttnn.ROW_MAJOR_LAYOUT: | ||
| return True, "Row Major layout is not supported" | ||
| return False, None | ||
|
|
||
|
|
||
| # This is the run instructions for the test, defined by the developer. | ||
| # The run function must take the above-defined parameters as inputs. | ||
| # The runner will call this run function with each test vector, and the returned results from this function will be stored. | ||
| # If you defined a mesh_device_fixture above, the object you yielded will be passed into this function as 'device'. Otherwise, it will be the default ttnn device opened by the infra. | ||
| def run( | ||
| input_spec, | ||
| input_a_layout, | ||
| input_b_layout, | ||
| *, | ||
| device, | ||
| ) -> list: | ||
| torch.manual_seed(0) | ||
| if input_spec["input_dtype"] == "ttnn.bfloat16": | ||
| input_dtype = ttnn.bfloat16 | ||
| elif input_spec["input_dtype"] == "ttnn.float32": | ||
| input_dtype = ttnn.float32 | ||
| elif input_spec["input_dtype"] == "ttnn.int32": | ||
| input_dtype = ttnn.int32 | ||
|
|
||
| torch_input_tensor_a = gen_func_with_cast_tt( | ||
| partial(torch_random, low=-100, high=100, dtype=torch.float32), input_dtype | ||
| )(input_spec["self"]) | ||
| torch_input_tensor_b = gen_func_with_cast_tt( | ||
| partial(torch_random, low=-100, high=100, dtype=torch.float32), input_dtype | ||
| )(input_spec["other"]) | ||
|
|
||
| golden_function = ttnn.get_golden_function(ttnn.multiply) | ||
| torch_output_tensor = golden_function(torch_input_tensor_a, torch_input_tensor_b) | ||
|
|
||
| sharded_config = ttnn.create_sharded_memory_config_( | ||
| shape=input_spec["self"], | ||
| core_grid=ttnn.CoreGrid(y=input_spec["y"], x=8), | ||
| strategy=ttnn.ShardStrategy.WIDTH, | ||
| orientation=ttnn.ShardOrientation.ROW_MAJOR, | ||
| use_height_and_width_as_shard_shape=False, | ||
| halo=0, | ||
| tile_layout=True, | ||
| ) | ||
|
|
||
| input_tensor_a = ttnn.from_torch( | ||
| torch_input_tensor_a, | ||
| dtype=input_dtype, | ||
| layout=input_a_layout, | ||
| device=device, | ||
| memory_config=sharded_config, | ||
| ) | ||
|
|
||
| input_tensor_b = ttnn.from_torch( | ||
| torch_input_tensor_b, | ||
| dtype=input_dtype, | ||
| layout=input_b_layout, | ||
| device=device, | ||
| memory_config=sharded_config, | ||
| ) | ||
|
|
||
| start_time = start_measuring_time() | ||
| result = ttnn.multiply(input_tensor_a, input_tensor_b, memory_config=sharded_config) | ||
| output_tensor = ttnn.to_torch(result) | ||
| e2e_perf = stop_measuring_time(start_time) | ||
|
|
||
| return [check_with_pcc(torch_output_tensor, output_tensor, 0.999), e2e_perf] |
Oops, something went wrong.