Support GroupedGemm types hhs,hss

[mengzcai]

• enable groupedgemm hhs,hss in gtest
• add logic yamls with tdm3

Motivation

Add GroupedGemm support for HHS and HSS data types on gfx1250 with TDM3 kernels.

Technical Details

• Logic YAMLs: Added TDM3 kernel configurations for GroupedGemm HHS and HSS types on gfx1250
• gtest: Enabled GroupedGemm HHS/HSS in matmul_gtest.yaml by updating gpu_arch_exclude regex from '1[1-2]\d{2}' to
  '11\d{2}|120\d{1}', allowing gfx1250 to run these tests

Test Plan

• Verified GroupedGemm HHS/HSS kernels are generated from the new logic YAMLs
• Ran matmul_groupedgemm and matmul_groupedgemm_f32 gtests on gfx1250

Test Result

• no CI validation currently
• test hipblaslt-test on local

Submission Checklist

• Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests.