UPSTREAM PR #18002: Clarify that steps also apply to linux

[loci-dev]

Mirrored from ggml-org/llama.cpp#18002

Added note that the windows PowerShell commands in the file work on Linux as well.

[loci-agentic-ai]

Explore the complete analysis inside the Version Insights

Performance Analysis Summary: PR #554

Analysis Type: Documentation-only change
Files Modified: 1 documentation file (docs/backend/OPENCL.md)
Code Changes: None

This PR adds a clarification note to the OpenCL backend documentation indicating that PowerShell setup commands also work on Linux. No source code, headers, build configurations, or binaries were modified. All performance metrics show 0% change across all functions and binaries. No impact on inference performance or tokens per second.

[loci-agentic-ai]

Explore the complete analysis inside the Version Insights

Performance Analysis Summary - PR #554

Analysis Type: Documentation-only change
Files Modified: 1 documentation file (docs/backend/OPENCL.md)
Code Changes: None

This PR adds clarification text to the OpenCL backend documentation indicating that Windows PowerShell setup commands also work on Linux. No source code, build configuration, or binary files were modified.

Performance Impact: Zero measurable change across all binaries and functions. All analyzed functions (llama_decode, main, llama_new_context_with_model, llama_sampler_sample, ggml_backend_sched_graph_compute) show identical throughput and response time values between versions. Power consumption remains unchanged across all binaries including libllama.so, libggml-cpu.so, and llama-run.