leslie-fang-intel

The running script as:
```
import torch
import torch.nn as nn
import torch._dynamo as torchdynamo
import copy

class Mod(torch.nn.Module):
    def __init__(self) -> None:
        super().__init__()

leslie-fang-intel

guards is: {
Guard(name='self', source=<GuardSource.LOCAL: 0>, create_fn=<function GuardBuilder.NN_MODULE at 0x7f01161ab160>, is_volatile=False, guard_types=['ID_MATCH'], code_list=['___check_obj_id(self, 139643301645088)'], obj_weakref=<weakref at 0x7f00fa06a180; to 'Mod' at 0x7f013d63cb20>, guarded_class_weakref=<weakref at 0x7f01362b6360; to 'type' at 0x5572acee40e0 (Mod)>), 
Guard(name='x', source=<GuardSource.LOCAL: 0>, create_fn=<function GuardBuilder.TENSOR_MATCH at 0x7f01161ab8b0>, is_volatile=False, guard_types=['TENSOR_MATCH'], code_list=None, obj_weakref=<weakref at 0x7f0116fac2c0; to 'Tensor' at 0x7f0116ed6810>, guarded_class_weakref=<weakref at 0x7f0119143900; to 'torch._C._TensorMeta' at 0x5572ab714830 (Tensor)>), 
Guard(name='self.linear', source=<GuardSource.LOCAL_NN_MODULE: 2>, create_fn=<function GuardBuilder.NN_MODULE at 0x7f01161ab160>, is_volatile=False, guard_types=None, code_list=None, obj_weakref=None, guarded_class_weakref=None), 
Guard(name='torch', source=<GuardSource.GLOBAL: 1>, cr

leslie-fang-intel

-----------------------------------------------  ------------  ------------  ------------  ------------  ------------  ------------ 
                                           Name    Self CPU %      Self CPU   CPU total %     CPU total  CPU time avg    # of Calls 
-----------------------------------------------  ------------  ------------  ------------  ------------  ------------  ------------ 
        quantized::conv_relu_int8_packed_weight        13.77%      69.356ms        20.19%     101.737ms       3.083ms            33 
             quantized::conv_int8_packed_weight        13.59%      68.488ms        19.10%      96.245ms       4.812ms            20 
                           graph_1_kernel_cpp_1         9.96%      50.175ms         9.96%      50.175ms      50.175ms             1 
                           graph_1_kernel_cpp_2         9.32%      46.958ms         9.32%      46.958ms      46.958ms             1 
                           graph_1_kernel_cpp_3         6.54%      32.936ms         6.54%    

leslie-fang-intel

kernel_cpp_0 = async_compile.cpp('''
#include <ATen/record_function.h>
#include "/tmp/torchinductor_root/dm/cdmaihqxwe73zkb3he2zizktpq5uujetg2db26c3r4lgsmlx3b4c.h"
extern "C" void kernel(const float* __restrict__ in_ptr0,
                       const float* __restrict__ in_ptr1,
                       const long* __restrict__ in_ptr2,
                       unsigned char* __restrict__ out_ptr0)
{
    RECORD_FUNCTION("graph_1_kernel_cpp_0", c10::ArrayRef<c10::IValue>({}));
    #pragma omp parallel num_threads(28)

leslie-fang-intel

import torch
import torch._dynamo as torchdynamo
import copy

class Mod(torch.nn.Module):
    def __init__(self) -> None:
        super().__init__()
        self.relu = torch.nn.ReLU()

    def forward(self, x):

leslie-fang-intel

from ctypes import c_void_p, c_long
import torch
import random
from torch import empty_strided, as_strided, device
from torch._inductor.codecache import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels

aten = torch.ops.aten
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
async_compile = AsyncCompile()

leslie-fang-intel

import torch
import torch._dynamo as torchdynamo
from torch.ao.quantization import (
    get_default_qconfig,
    QConfigMapping,
)
from torch.ao.quantization._quantize_pt2e import prepare_pt2e, convert_pt2e
from torch._inductor.compile_fx import compile_fx

def test_single_conv():

leslie-fang-intel

from ctypes import c_void_p, c_long
import torch
import math
import random
from torch import empty_strided, as_strided, device
from torch._inductor.codecache import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels

aten = torch.ops.aten
assert_size_stride = torch._C._dynamo.guards.assert_size_stride

leslie-fang-intel

# Running CMD is: clear &&  TORCHDYNAMO_DYNAMIC_SHAPES=1 python test_rn50.py 2>&1 | tee test.log

import torch
import torch._dynamo as torchdynamo
import torchvision.models as models
import copy
from torch._inductor.compile_fx import compile_fx

import logging
torch._dynamo.config.log_level = logging.DEBUG

leslie-fang-intel

from ctypes import c_void_p, c_long
import torch
import math
import random
from torch import empty_strided, as_strided, device
from torch._inductor.codecache import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels

aten = torch.ops.aten
assert_size_stride = torch._C._dynamo.guards.assert_size_stride