Update some documents for v3.6

[thuang6]

User description

Type of Change

documentation

Description

• update Synapse release mapping table for v3.6
• update what's new section for recent releases

Note

merge need to wait until MX quant doc updated

---

PR Type

Documentation

---

Description

• Update Synapse release mapping table for v3.6

• Add MXFP8/MXFP4 quantization entry in What's New

• Update recent releases in What's New section

---

Diagram Walkthrough

flowchart LR
  A["Update README.md"] -- "Add MXFP8/MXFP4 quantization" --> B["Update gaudi_version_map.md"]
  A -- "Update recent releases" --> C["What's New section"]

Loading

File Walkthrough

	Relevant files
Documentation	README.md Update What's New for v3.6                                                              README.md Add MXFP8/MXFP4 quantization entry Update recent releases in What's New section +4/-2      gaudi_version_map.md Update Synapse release mapping                                                      docs/source/3x/gaudi_version_map.md Include v3.6 in Synapse release mapping +1/-1

---

[PRAgent4INC]

PR Reviewer Guide 🔍

Here are some key observations to aid the review process:

⏱️ Estimated effort to review: 1 🔵⚪⚪⚪⚪

🧪 No relevant tests

🔒 No security concerns identified

⚡ Recommended focus areas for review Chronological Order The new entries in the What's New section are not in chronological order. The latest entries should be at the top. * [2025/10] [MXFP8 / MXFP4 quantization](./docs/source/3x/PT_MXQuant.md) experimental support * [2025/09] FP8 dynamic quantization, including Linear, FusedMoE on Intel Gaudi AI Accelerators * [2025/05] FP8 static quantization of DeepSeek V3/R1 model on Intel Gaudi AI Accelerators * [2025/03] VLM quantization in transformers-like API on Intel CPU/GPU

[PRAgent4INC]

PR Code Suggestions ✨

Explore these optional code suggestions:

Category	Suggestion	Impact
General	Order dates descending Ensure the dates are in descending order for clarity. README.md [30-33] +* [2025/10] [MXFP8 / MXFP4 quantization](./docs/source/3x/PT_MXQuant.md) experimental support +* [2025/09] FP8 dynamic quantization, including Linear, FusedMoE on Intel Gaudi AI Accelerators +* [2025/05] FP8 static quantization of DeepSeek V3/R1 model on Intel Gaudi AI Accelerators +* [2025/03] VLM quantization in transformers-like API on Intel CPU/GPU - Suggestion importance[1-10]: 7 __ Why: The dates are already in descending order, so the suggestion does not make a change. It is correct but offers no improvement.	Medium
	Remove outdated entries Remove outdated entries to keep the list current. README.md [30-33] -* [2024/10] [Transformers-like API](./docs/source/3x/transformers_like_api.md) for INT4 inference on Intel CPU and GPU. +* [2025/10] [MXFP8 / MXFP4 quantization](./docs/source/3x/PT_MXQuant.md) experimental support +* [2025/09] FP8 dynamic quantization, including Linear, FusedMoE on Intel Gaudi AI Accelerators +* [2025/05] FP8 static quantization of DeepSeek V3/R1 model on Intel Gaudi AI Accelerators +* [2025/03] VLM quantization in transformers-like API on Intel CPU/GPU Suggestion importance[1-10]: 7 __ Why: The suggestion removes the 2024/10 entry, but this entry is not outdated compared to the new entries. It is correct but offers no improvement.	Medium
	Separate versions Consider separating versions for clarity. docs/source/3x/gaudi_version_map.md [42] -<td>v3.5, v3.6</td> +<td>v3.5<br>v3.6</td> Suggestion importance[1-10]: 5 __ Why: Separating versions with a <br> tag improves readability, but it is a minor stylistic change. It is correct and slightly improves the document.	Low