Add CuTeDSL math tests for sqrt operation

[Pritiks23]

Add sqrt unary math operation test

Description

Add coverage for cute.math.sqrt by introducing a new unary operation test.

This change:

• Adds a _sqrt_ops_kernel that applies cute.math.sqrt elementwise.
• Adds a _sqrt_ops_host JIT wrapper to launch the kernel.
• Adds test_sqrt_ops to verify correctness against expected square root values using CUDA tensors.

The test ensures that cute.math.sqrt correctly lowers and executes through CuTe's math operation interface.

Testing

• Added a CUDA unit test covering float32 sqrt operations.
• Verified output correctness against expected values:
  • sqrt(1.0) = 1.0
  • sqrt(4.0) = 2.0
  • sqrt(9.0) = 3.0
  • sqrt(16.0) = 4.0