Increased core count for paged SDPA for Qwen

[atupe-tt]

Problem description

Improve the decode perf for Qwen on TG

What's changed

Increased the core count for paged SDPA (decode)

Checklist

• 
• 
• 
• New/Existing tests provide coverage for changes

Model tests

If your changes cover model-related code, you should run tests corresponding to affected models and platforms (Single card, T3K, Galaxy). "Choose your pipeline" workflows facilitate running multiple kinds of tests in a single run. Each offers models-mandatory and models-extended presets.
The former includes a minimal set of tests, to be run always. The latter extends that with additional ones - use your best judgement in deciding which is the most appropriate for your PR.

• 

  • models-mandatory preset (runs: Device perf regressions and Frequent model and ttnn tests)
  • models-extended preset (runs: the mandatory tests, plus Demo and Model perf tests)
  • other selection - specify runs

• 

  • models-mandatory preset (runs: Unit tests)
  • models-extended preset (runs: the mandatory tests, plus Demo and Model perf tests)
  • other selection - specify runs

• 

  • models-mandatory preset (runs: Quick tests)
  • models-extended preset (runs: the mandatory tests, plus Demo and Model perf tests)
  • other selection - specify runs

[Copilot]

Pull request overview

This PR improves decode performance for Qwen on TG by increasing the core count for paged SDPA (Scaled Dot-Product Attention) decode operations from 32 to 48 cores. The change aligns the Qwen-specific model configuration with the base Llama model configuration in model_config.py, which already uses these optimized settings.

Changes:

• Increased compute grid size from (8, 4) to (8, 6) for PAGED_SDPA_DECODE_PROGCFG
• Updated core count from 32 to 48 to match the new grid size (8 × 6 = 48)

[yalrawwashTT]

Approving to unblock, pending CI run of galaxy demo