The mkldnn matmul_v2 op does not support broadcast when the dims length of two inputs are not equal.

[baoachun]

Steps to reproduce:

1. Pull this pr add mkldnn matmul_v2 ut #38625, and cmake with -DWITH_TESTING=ON, then build paddle.
2. ctest -R test_mkldnn_matmulv2 -V

@lidanqing-intel @jakpiase @jczaja
Note: If we add fp32 operators in the future, we must write the unit test according to the new unit test framework, for example #38625.

[paddle-bot-old]

您好，我们已经收到了您的问题，会安排技术人员尽快解答您的问题，请耐心等待。请您再次检查是否提供了清晰的问题描述、复现代码、环境&版本、报错信息等。同时，您也可以通过查看官网API文档、常见问题、历史Issue、AI社区来寻求解答。祝您生活愉快～

Hi! We've received your issue and please be patient to get responded. We will arrange technicians to answer your questions as soon as possible. Please make sure that you have posted enough message to demo your request. You may also check out the API，FAQ，Github Issue and AI community to get the answer.Have a nice day!

[jakpiase]

Hi @baoachun, I will start working at fixing this issue today

[baoachun]

Hi @baoachun, I will start working at fixing this issue today

Got it, thanks a lot!

[baoachun]

Hi @jakpiase, after merging your repair code, there is still a problem that the output shape is incorrect.

[jakpiase]

Hi @baoachun, I have fixed the wrong output formats in the next commit, could you confirm that now it passes all tests?

[baoachun]

Hi @jakpiase, all test cases have passed, you can also verify by yourself.

[jakpiase]

Hi @baoachun, I have verified it by myself, but I just wanted to have your confirmation, since last time my last fix did not cover wrong output formats