PTAB

IPR2015-00965

Apple Inc v. ZiiLabs Inc

Log In
Key Events
Petition
petition Intelligence

1. Case Identification

2. Patent Overview

  • Title: Rendering System With Mini-Patch Retrieval From Local Texture Storage
  • Brief Description: The ’584 patent describes a method and system for computer graphics processing intended to improve efficiency when retrieving texture data from memory. The invention focuses on retrieving a two-dimensional (2D) block of texture data, called a "mini-patch," as a single "wide word" from one memory location to reduce memory page breaks and access overhead during rendering.

3. Grounds for Unpatentability

Ground 1: Claims 1, 2, 4, and 6 are obvious over Winser in view of Blinn.

  • Prior Art Relied Upon: Winser (Patent 5,544,292) and Blinn (Jim Blinn, The Truth About Texture Mapping, IEEE Computer Graphics & Applications, 1990).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner asserted that Winser taught a highly analogous graphics rendering system that retrieved 2D patches of texture data (e.g., a 2x2 patch) using parallel read cycles. However, during prosecution of the ’584 patent, the Patent Owner distinguished Winser by arguing it did not retrieve the four pixels from a single memory location. Petitioner contended that Blinn, which was not considered during prosecution, supplied this missing element. Blinn directly addressed memory page faults—the same problem cited in the ’584 patent—by teaching the organization of texture data into 2D "tiles" in memory, such that a single memory "access" retrieves an entire 2x2 array of pixels as a single word from one location.
    • Motivation to Combine: A Person of Ordinary Skill in the Art (POSITA) would combine Winser and Blinn to improve performance and efficiency. Both references concern efficient memory access for 2D texture data. Petitioner argued a POSITA would find it logical to apply Blinn’s memory tiling solution to the rendering system of Winser to simplify its hardware and overcome the exact memory paging issues Blinn described, thereby achieving a more efficient readout of texel values.
    • Expectation of Success: The combination would have yielded a predictable result. Petitioner argued that modifying Winser’s addressing circuitry to incorporate Blinn's more efficient single-access method for tiled data was a routine engineering task, particularly since the underlying texture mapping techniques were well-known at the time.

Ground 2: Claims 1, 2, 4, and 6 are obvious over Winser in view of Blinn, further in view of Apgar.

  • Prior Art Relied Upon: Winser (Patent 5,544,292), Blinn (Jim Blinn, The Truth About Texture Mapping, 1990), and Apgar (Apgar and Bersack, A Display System for the Stellar Graphics Supercomputer Model GS1000, 1988).
  • Core Argument for this Ground:
    • Prior Art Mapping: This ground reinforced the arguments from Ground 1, with Apgar providing additional teachings on key claim elements. Petitioner asserted Apgar explicitly taught the computation of pixel depth ("Z coordinates") alongside color values during the rendering of primitives. Apgar also disclosed retrieving 2D blocks of pixels (e.g., 4x4 blocks) via a single memory access to achieve higher efficiency ("more useful pixels per memory access"), further supporting the rationale for retrieving a 2D patch from a single location.
    • Motivation to Combine: A POSITA would recognize that Winser, Blinn, and Apgar all address complementary problems in computer graphics rendering. It would have been obvious to integrate Apgar's express teaching of pixel depth calculation into the Winser/Blinn system to create a more functionally complete rendering pipeline. The addition of Apgar’s teachings would be a straightforward enhancement to the already-improved system.
    • Expectation of Success: Petitioner asserted that integrating the teachings would be a routine task with a high expectation of success. A POSITA would understand that implementing pixel depth calculation would not interfere with the 2D texture word retrieval method taught by the combination of Winser and Blinn, as the processes are distinct but complementary parts of a standard rendering pipeline.

4. Key Claim Construction Positions

  • "retrieving at least four pixels of a stored texture from a single memory location..." (Claims 1 and 4): Petitioner argued this phrase, central to patentability, should be construed as fetching a set of texture data for at least four texels via a single memory "access" from one location in memory. This construction was positioned to directly counter the distinction the patentee made during prosecution to overcome Winser and to highlight how Blinn and Apgar teach this exact functionality.
  • "said pixels... being grouped in... a pattern which is more than one pixel wide and more than one pixel high" (Claim 1): Petitioner proposed this means the retrieved word corresponds to a 2D set of texture pixels. Petitioner noted that while prosecution history arguments might suggest a narrower construction requiring "the actual physical grouping of the pixels in memory," its invalidity grounds were asserted to be valid under either the broader or narrower construction, as Blinn and Apgar taught physical grouping in memory tiles or blocks.

5. Relief Requested

  • Petitioner requests institution of an inter partes review and cancellation of claims 1, 2, 4, and 6 of the ’584 patent as unpatentable.