cinn(dynamic): support run exp sub subgraph with dynamic shape graph

[6clc]

PR types

Others

PR changes

Others

Description

Pcard-78120

• 修改broadcast的compute，使得output shape和input shape 一致的计算支持动态形状
• 联调bucket机制，在不进行op schedule、group schedule的情况下可以跑通流程
• 增加exp sub动态形状的子图单测。单测对应的host kernel和 cuda kernel如下

|| function fn_exp_subtract (kernel_args, kernel_args_num, kernel_stream)
|| {
||   int32 S0 = cinn_get_value_in_cuda_kernel_args(kernel_args, 3)
||   int32 S1 = cinn_get_value_in_cuda_kernel_args(kernel_args, 4)
||   if (true) {
||     cinn_call_cuda_kernel(fn_exp_subtract__COND_true__kernel, kernel_args, kernel_args_num, 1, 1, 1, 1, 1, 1, kernel_stream)
||   }
|| }

cuda kernel

|| function fn_exp_subtract__COND_true__kernel (_var_0, _var_1, _var, S0, S1)
|| {
||   {
||     serial for (i, 0, S0)
||     {
||       serial for (j, 0, 128)
||       {
||         var[i, j] = cinn_nvgpu_exp_fp32(var_0[i, j])
||       }
||     }
||     serial for (i, 0, S1)
||     {
||       serial for (j, 0, 128)
||       {
||         var_1[i, j] = (var[i, j] - var_0[i, j])
||       }
||     }
||   }
|| }

[paddle-bot]

你的PR提交成功，感谢你对开源项目的贡献!
请关注后续CI自动化测试结果，详情请参考Paddle-CI手册。
Your PR has been submitted. Thanks for your contribution!
Please wait for the result of CI firstly. See Paddle CI Manual for details.

[paddle-bot]

✅ This PR's description meets the template requirements!
Please wait for other CI results.

[zhhsplendid]

LGTM

[zyfncg]

2022 -> 2023

[6clc]

收到，下个pr中修改。

[zhangbo9674]

LGTM for cinn_instruction