[Bug]: suspicious bugs in class LinearKsplit2 in llm_mlp.cpp

[troore]

OpenVINO Version

Master Branch

Operating System

Ubuntu 20.04 (LTS)

Device used for inference

CPU

Framework

None

Model used

TinyLlama-1.1B-Chat-v1.0

Issue description

In file openvino/src/plugins/intel_cpu/src/nodes/llm_mlp.cpp, and in class LinearKsplit2, there are 3 suspicious bugs in two member functions:

1. in setup,

...
auto valid_nthr = m_threads_num / 2;
auto blkN_per_thread = (num_blk_N) / valid_nthr;
...

when m_threads_num is 1, the next line would cause division 0 exception.
Possible fixing:

auto valid_nthr = (m_threads_num+1) / 2;

2. still in setup,

...
auto K_splits = 2;
...
    for (int ik = 0; ik < K_splits; ik++) {
        ...
        auto& work = works[ithr + ik];
        ...
    }
...

when m_threads_num is 1, works[ithr + ik] would cause memory access out of bounds.
Possible fixing:

auto K_splits = std::min(m_threads_num, 2);

3. in run,

auto peer_ithr = (ithr & 1) ? (ithr - 1) : (ithr + 1);
auto* p_peerC = works[peer_ithr].m_C.template ptr<float>();

when ithr is 0, peer_ithr is 1, also causing works[peer_thr] access out of bounds when nthr is 1.
Possible fixing:

auto peer_ithr = (nthr == 1) ? 0 : ((ithr & 1) ? (ithr - 1) : (ithr + 1));

Step-by-step reproduction

No response

Relevant log output

Issue submission checklist

• I'm reporting an issue. It's not a question.
• I checked the problem with the documentation, FAQ, open issues, Stack Overflow, etc., and have not found a solution.
• There is reproducer code and related data files such as images, videos, models, etc.

[github-actions]

Thank you for looking into this issue! Please let us know if you have any questions or require any help.

[AbhiRam162105]

.take

[github-actions]

Thanks for being interested in this issue. It looks like this ticket is already assigned to a contributor. Please communicate with the assigned contributor to confirm the status of the issue.