Add zero point support to dp4a 2-bit dequantization in the WebGPU MatMulNbits

[HectorSVC]

Add zero point support to dp4a 2-bit dequantization in the WebGPU MatMulNBits kernel. Previously, the dp4a path for 2-bit quantization used a hardcoded 256-entry LUT assuming zero_point=2, and was blocked from running when custom zero points were provided.

1. dp4a_matmul_common.wgsl.template — Core LUT & dequantization function
   Added a 1024-entry LUT (4 sections × 256 entries) when has_zero_points is true. Each section corresponds to a zero point value (0–3), pre-computing pack4xI8(value - zero_point) for every possible byte input.
   Added a new DequantizedFrom2BitsTo8Bits(in: u32, zero: i32) overload that indexes the LUT as zero * 256 + byte_value.
   Original 256-entry LUT and parameterless function preserved for the !has_zero_points path.

2. dp4a_matmul.wgsl.template — Large-M tiled kernel (workgroup=256)
   loadSHMB for n_bits==2: reads zero point via mm_read_zero() and passes it to DequantizedFrom2BitsTo8Bits(b_value, zero) when has_zero_points.
   LoadDequantizationTable: expanded to 4 calls (local_idx + 0/256/512/768) to load all 1024 entries when has_zero_points.

3. dp4a_matmul_small_m.wgsl.template — Small-M kernel (workgroup=128)
   LoadDequantizationTable: expanded to 8 calls to load 1024 entries when has_zero_points.
   DequantizedFrom2BitsTo8Bits calls pass zero when has_zero_points.
   Bug fix: corrected off-by-one local_idx+127 → local_idx+128 in the non-zero-point path.

4. matmul_nbits.cc — Kernel dispatch logic
   Removed the guard !(has_zero_points && nbits == 2) that previously blocked the dp4a path for 2-bit with custom zero points.
   Updated comment to document the new 1024-entry LUT support.

[HectorSVC]

reflect the review comments in PR: #27285

[guschmue]

/azp run Linux QNN CI Pipeline,Win_TRT_Minimal_CUDA_Test_CI,Windows ARM64 QNN CI Pipeline,Windows GPU Doc Gen CI Pipeline

[azure-pipelines]

Azure Pipelines successfully started running 4 pipeline(s).

[Copilot]

Pull request overview

This PR enables the WebGPU DP4A implementation of MatMulNBits for 2-bit quantized weights when custom zero_points are provided, by extending the dequantization LUT logic in the DP4A WGSL shaders and removing the previous dispatch guard that blocked this path.

Changes:

• Extend the Q2 DP4A dequantization LUT from 256 entries to 1024 entries when has_zero_points is true, and route dequantization through the zero-point-aware lookup.
• Update both DP4A kernels (large-M and small-M variants) to load the expanded LUT and pass zero points into Q2 dequantization.
• Remove the C++ dispatch guard that prevented selecting the DP4A path for (nbits=2, has_zero_points=true), and update the in-code comment accordingly.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 1 comment.

File	Description
onnxruntime/contrib_ops/webgpu/quantization/matmul_nbits.cc	Allows DP4A dispatch for 2-bit with zero points by removing the previous guard and updating the comment.
onnxruntime/contrib_ops/webgpu/quantization/dp4a_matmul_common.wgsl.template	Adds a 1024-entry LUT (4×256) and a zero-point-aware Q2 dequantization function.
onnxruntime/contrib_ops/webgpu/quantization/dp4a_matmul.wgsl.template	Large-M DP4A kernel loads all 1024 LUT entries when needed and passes per-block zero point into Q2 dequantization.
onnxruntime/contrib_ops/webgpu/quantization/dp4a_matmul_small_m.wgsl.template	Small-M DP4A kernel loads all 1024 LUT entries when needed, uses zero-point-aware Q2 dequantization, and fixes a LUT-load offset bug in the non-zero-point path.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 5 out of 5 changed files in this pull request and generated 2 comments.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 7 out of 7 changed files in this pull request and generated no new comments.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.