US Patent Application 18450079. STATIC POWER REDUCTION IN CACHES USING DETERMINISTIC NAPS simplified abstract

From WikiPatents
Jump to navigation Jump to search

STATIC POWER REDUCTION IN CACHES USING DETERMINISTIC NAPS

Organization Name

TEXAS INSTRUMENTS INCORPORATED

Inventor(s)

Oluleye Olorode of Pflugerville TX (US)

Mehrdad Nourani of Plano TX (US)

STATIC POWER REDUCTION IN CACHES USING DETERMINISTIC NAPS - A simplified explanation of the abstract

This abstract first appeared for US patent application 18450079 titled 'STATIC POWER REDUCTION IN CACHES USING DETERMINISTIC NAPS

Simplified Explanation

The patent application describes a dNap architecture that efficiently manages cache lines to reduce power consumption without affecting performance.

  • The dNap architecture accurately transitions cache lines to a full power state before they are accessed, eliminating delays caused by waking up drowsy lines.
  • Only cache lines that are predicted to be accessed in the immediate future are fully powered, while others are put in a drowsy mode to save power.
  • This approach significantly reduces leakage power without impacting cache performance.
  • The hardware overhead required for implementing this architecture is minimal, especially for caches with higher associativities.
  • The dNap architecture achieves up to 92% static/leakage power savings with minimal hardware overhead and no performance tradeoff.


Original Abstract Submitted

Disclosed embodiments relate to a dNap architecture that accurately transitions cache lines to full power state before an access to them. This ensures that there are no additional delays due to waking up drowsy lines. Only cache lines that are determined by the DMC to be accessed in the immediate future are fully powered while others are put in drowsy mode. As a result, we are able to significantly reduce leakage power with no cache performance degradation and minimal hardware overhead, especially at higher associativities. Up to 92% static/leakage power savings are accomplished with minimal hardware overhead and no performance tradeoff.