[Feature]: [AutoDeploy] tighten the perf thresholds of test_perf.py

[MrGeva]

🚀 The feature, motivation and pitch

track the CI perf results at: https://urm.nvidia.com/artifactory/sw-tensorrt-generic/llm-artifacts/LLM/main/L0_MergeRequest_PR/15279/test-results/results-H100_PCIe-PyTorch-Perf-1.tar.gz
change the build number in the path to look at more builds. check if the results are stable and can be tightened.

Alternatives

No response

Additional context

No response

Before submitting a new issue...

• Make sure you already searched for relevant issues, and checked the documentation and examples for answers to frequently asked questions.