[hardware] fix latency stall bug in VMFPU

[Navaneeth-KunhiPurayil]

Fixed

• Latency of some instructions were not specified leading to wrong results in the _blackscholes benchmark from rivec. When low latency operation/instruction follows a higher latency operation, it needs to be stalled for high latency operation to complete to avoid low latency operation results to be committed to VRF ahead of the higher latency operation into the wrong destination register.

Checklist

• Automated tests pass
• Changelog updated
• Code style guideline is observed

[hopang-0221]

Hi Navaneeth, can you explain why you use the latency_stall instead of the registered latency_stall_{d,q}? The same for vinsn_issue_lat and vinsn_processing_lat signals.

[Navaneeth-KunhiPurayil]

This is because it is computed from an already registered signal vinsn_issue_q and vinsn_processing_q while previously it was using _d signals

[hopang-0221]

LGTM!