PTAB

IPR2019-00322

Qualcomm Inc. v. Apple Inc.

Log In

1. Case Identification

  • Patent #: 8,443,216
  • Filed: November 12, 2018
  • Petitioner(s): QUALCOMM INC. and QUALCOMM TECHNOLOGIES, INC.
  • Patent Owner(s): APPLE INC.
  • Challenged Claims: 1-3, 6, 8-10, and 13

2. Patent Overview

  • Title: Hardware Automatic Performance State Transitions in System on Processor Sleep and Wake Events
  • Brief Description: The ’216 patent discloses a system for managing power in an integrated circuit, such as a system-on-chip (SOC). The technology involves transitioning the performance states of different "performance domains" and their components using a programmable power management unit (PMU) in response to a processor entering or exiting a sleep state.

3. Grounds for Unpatentability

Ground 1: Anticipation over Mandelblat - Claims 1-3, 6, 8-10, and 13 are anticipated by Mandelblat under 35 U.S.C. §102.

  • Prior Art Relied Upon: Mandelblat (Application # 2007/0043965).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that Mandelblat discloses all elements of the challenged claims. Mandelblat describes a multi-component processor (e.g., Core 1 902 and Dynamically Sizeable Memory 905) where each component is in its own performance domain. It further discloses a Power Management Logic (PML 906) that functions as the claimed power management unit. This PML establishes performance states for each domain, such as by changing the memory size or controlling the core’s sleep/wake state. Crucially, Petitioner asserted Mandelblat shows the PML transitioning a second performance domain (the memory) to a new performance state (e.g., expanding cache size) in direct response to a first performance domain (the processor core) transitioning to a different state (e.g., waking from sleep). This mapping was argued to meet the limitations of independent claims 1 and 8.

Ground 2: Obviousness over Mandelblat and Kurts - Claims 1-3, 6, 8-10, and 13 are obvious over Mandelblat in view of Kurts under 35 U.S.C. §103.

  • Prior Art Relied Upon: Mandelblat (Application # 2007/0043965) and Kurts (Patent 7,363,523).
  • Core Argument for this Ground:
    • Prior Art Mapping: Mandelblat taught managing performance by changing resource allocation (e.g., dynamic memory sizing) based on the processor's general sleep/wake state. Kurts addressed the same power management problem but focused on more granular control of C-state transitions, such as waking a processor into a low-frequency mode (LFM) to reduce latency before resuming a high-frequency mode. Petitioner contended that while Mandelblat discloses the broad structure of the claims, Kurts supplies specific teachings on transitioning between different active performance states upon waking.
    • Motivation to Combine: A Person of Ordinary Skill in the Art (POSITA) would combine these references to improve the power management system of Mandelblat. Both references address power management in processors using ACPI-based concepts and were assigned to the same entity (Intel). A POSITA would be motivated to integrate Kurts’s efficient state transition techniques (e.g., waking to an LFM) with Mandelblat’s resource-based power management (e.g., dynamic cache sizing) to create a more optimized and efficient system.
    • Expectation of Success: The combination involved applying known and compatible power management techniques to achieve the predictable result of improved power savings and reduced latency upon waking from a sleep state.

Ground 3: Obviousness over Kurts and Kang - Claims 1-3, 6, 8-10, and 13 are obvious over Kurts in view of Kang under §103.

  • Prior Art Relied Upon: Kurts (Patent 7,363,523) and Kang (Patent 7,369,815).
  • Core Argument for this Ground:
    • Prior Art Mapping: Kurts taught a power management unit controlling processor transitions between sleep and wake states (C-states), including waking to an intermediate low-frequency state. Kang taught a complementary power-saving technique: partitioning a processor into multiple, independently controllable "collapsible power domains" that can be powered down entirely when not needed. Petitioner argued that combining these references would result in the claimed invention.
    • Motivation to Combine: A POSITA would combine the teachings of Kurts and Kang to achieve more granular and effective power savings. Both references are in the same field of processor power management. A POSITA would see a clear benefit in applying Kurts's state transition management within the architectural framework of Kang's independently powered domains. This would allow not only for C-state management but also for completely powering down unused sections of the chip, a known method for improving power efficiency. Petitioner also noted the examiner of the parent ’812 patent considered Kang but not in combination with Kurts.
    • Expectation of Success: Combining processor state control (Kurts) with independent power domain control (Kang) was a well-understood design choice in the field, and a POSITA would have expected the combination to work predictably to reduce overall power consumption.

4. Key Claim Construction Positions

  • "performance domain": Petitioner argued for adopting the patent owner's proposed construction of "one or more components that may be controlled as a unit or independently for performance configuration purposes." This construction supports the argument that components in different domains can have their performance states changed independently, a key feature taught by the prior art combinations.
  • "power management unit": Petitioner agreed that a PMU could be "hardware or the combination of hardware and software," as construed by the district court. This broad construction allows prior art references disclosing software-assisted or logic-based power controllers (like Mandelblat's PML) to qualify as the claimed PMU.
  • "establish a . . . performance state": Petitioner argued this term should be construed as to "set the one or more performance characteristics to the appropriate values for the performance state." This construction, consistent with the district court's finding, does not require an actual transition to occur, only that the PMU is configured to set the values, which aligns with the capabilities described in the prior art.

5. Relief Requested

  • Petitioner requested the Board institute an inter partes review of claims 1-3, 6, 8-10, and 13 of the ’216 patent and find them unpatentable and cancelled.