Add fine-grained pipeline cut-points for INSIDE config

[gamzeisl]

Previously, the INSIDE pipe configuration placed all NumPipeRegs registers at a single mid-point inside each unit, requiring retiming.

This PR introduces multiple cut-points located at roughly ideal boundaries in each unit's datapath, allowing NumPipeRegs registers to be distributed where they most effectively break critical paths.

Units updated: fpnew_noncomp, fpnew_cast_multi, fpnew_fma_multi, fpnew_mxdotp_multi, fpnew_sdotp_multi.

The updated DEFAULT_SNITCH_PIPE localparam in fpnew_pkg.sv reflects the recommended register counts for the Snitch configuration.

[Copilot]

Pull request overview

This PR refines the INSIDE pipeline configuration to support multiple, unit-specific cut-points, allowing NumPipeRegs to be distributed across more optimal datapath boundaries (reducing reliance on synthesis retiming). It also updates the recommended Snitch pipeline preset and documentation to reflect the new behavior.

Changes:

• Introduces finer-grained internal pipeline stage allocation for INSIDE across several units (FMA, casts, non-comp, dotp variants).
• Refactors MXDOTP normalization into multiple sub-stages/modules to enable additional cut-points.
• Adds DEFAULT_SNITCH_PIPE in fpnew_pkg.sv and updates docs to describe the new placement strategy and include a distribution figure.

Reviewed changes

Copilot reviewed 8 out of 9 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
src/mxdotp/fpnew_mxdotp_multi_modules.sv	Refactors MXDOTP helper modules and splits normalization to support additional internal cut-points.
src/fpnew_mxdotp_multi.sv	Updates MXDOTP top-level pipeline segmentation and reconnects to new normalization stages.
src/fpnew_sdotp_multi.sv	Adds multiple internal cut-points and new internal pipeline segments for SDOTP.
src/fpnew_fma_multi.sv	Adds multiple internal cut-points and re-stages FMA datapath around them.
src/fpnew_cast_multi.sv	Adds internal cut-point pipelines around normalization/rounding and retimes control/data.
src/fpnew_noncomp.sv	Adds an internal mid-pipeline stage for INSIDE and updates handshake/busy tracking.
src/fpnew_pkg.sv	Introduces DEFAULT_SNITCH_PIPE preset for recommended INSIDE register counts.
docs/README.md	Updates documentation for PipeConfig behavior and adds a distribution figure reference.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 8 out of 9 changed files in this pull request and generated 3 comments.

[Copilot]

Pull request overview

Copilot reviewed 8 out of 9 changed files in this pull request and generated 2 comments.