Add CodSpeed performance benchmarking

[codspeed-hq]

Summary

This PR integrates CodSpeed for continuous performance benchmarking of scipp. It adds pytest-codspeed benchmarks that mirror the existing ASV benchmarks and sets up a GitHub Actions workflow to run them automatically on every push to main and on pull requests.

Changes

• New benchmark files (benchmarks/test_variable.py, benchmarks/test_bin.py, benchmarks/test_binned.py): 29 pytest-codspeed benchmarks covering:
  • Variable operations (copy, inplace/non-inplace arithmetic, numpy assignment)
  • 1D and 2D binning with various parameter sizes
  • Binned data operations (constituents, size, sum, concat)
  • Grouping operations
• CI workflow (.github/workflows/codspeed.yml): GitHub Actions workflow that installs scipp from PyPI and runs the benchmarks using CodSpeedHQ/action@v4 in simulation mode with OIDC authentication
• README badge: Added a CodSpeed badge linking to the project dashboard

Next steps

• After merging, CodSpeed will automatically track performance across commits and flag regressions on pull requests
• The existing ASV benchmarks remain untouched and can continue to be used independently
• Benchmark coverage can be expanded over time by adding more test_* functions to the benchmarks/ directory