SDK Stability, Simulation Performance, and Host Utility Improvements

[Xeratec]

Changelog

This PR introduces several SDK, host, and simulation improvements focused on correctness, performance, and maintainability.

Added

• Support for ASIC targets in the runner script.
• getc and flush support for host-side stdio.

Changed

• Minor utility refactoring and general SDK cleanup.
• Cleaned up the picolibc CMake build flow.
• Increased default UART baud rate to 9600.
• Improved default frequency estimation accuracy (1.56%).
• Reduced .bss size and skipped UART initialization in RTL to accelerate simulations.

Fixed

• Correct handling of CLINT sleep behavior and interrupt handler stack.
• Properly set busy flag in interrupt handler.
• Fixed overflow in 32-bit clint_get_core_freq.
• Improved RTL simulation speed and overall stability.

Checklist

1. The PR is rebased on the latest devel commit and pointing to devel.
2. Your PR reviewed and approved.
3. The documentation is updated.
4. All checks are passing.

[viv-eth]

Few minor nitpicks, but overall LGTM! Thank you for your contribution @Xeratec .

[Xeratec]

I implemented the requested changes in 8f2e466

IMPORTANT: There are currently some assumptions on the number of clusters in the SoC. I know this is not ideal, but I prefer to fix this in my work for the multi ABI/ISA support where I need to do some file restructuring anyway and properly test on a fully heterogeneous SoC.