DCT

1:19-cv-00070

Advanced Micro Devices Inc v. MediaTek Inc

Log In

I. Executive Summary and Procedural Information

  • Parties & Counsel:
  • Case Identification: 1:19-cv-00070, D. Del., 01/10/2019
  • Venue Allegations: Venue is alleged to be proper as Defendant MediaTek USA Inc. is a Delaware corporation, Defendant MediaTek Inc. is a foreign corporation that may be sued in any district, and both defendants have allegedly committed acts of infringement within the District of Delaware.
  • Core Dispute: Plaintiff alleges that Defendant’s graphics-capable integrated circuits, used in consumer electronics, infringe patents related to parallel pipeline graphics processing architectures and unified shader technology.
  • Technical Context: The lawsuit concerns the fundamental architecture of Graphics Processing Units (GPUs), technology critical for rendering complex 2D and 3D graphics in a wide range of consumer and professional devices.
  • Key Procedural History: The parties were previously involved in an International Trade Commission (ITC) investigation where the '506 patent was found valid and infringed by certain of Defendant's products, resulting in a limited exclusion order. In the same action, the '133 patent was found valid but not infringed. The complaint also notes that at the time of filing, Inter Partes Review (IPR) proceedings had been instituted against the '506 patent but no final written decisions had been issued. A subsequent IPR certificate, provided with the case materials, indicates that claims 1-9 of the '506 patent were later cancelled, which, if upheld, would significantly impact the scope of the present litigation concerning that patent.

Case Timeline

Date Event
2002-11-18 Priority Date for U.S. Patent No. 7,796,133
2002-11-27 Priority Date for U.S. Patent No. 7,633,506
2009-12-15 Issue Date for U.S. Patent No. 7,633,506
2010-09-14 Issue Date for U.S. Patent No. 7,796,133
2017-01-01 Alleged Notice of Patents via ITC Investigation
2019-01-10 Complaint Filing Date

II. Technology and Patent(s)-in-Suit Analysis

U.S. Patent No. 7,633,506: "Parallel pipeline graphics system" (Issued Dec. 15, 2009)

The Invention Explained

  • Problem Addressed: The patent's background section describes that as graphics processing moves toward wider memory bus architectures (e.g., 256-bit), the efficient organization and use of data becomes problematic, creating a performance bottleneck due to the coarse granularity of data "words" ('506 Patent, col. 2:35-45).
  • The Patented Solution: The invention proposes a graphics system architecture that divides the processing workload across multiple parallel pipelines within the GPU's "back-end" ('506 Patent, Abstract). In a key embodiment, the display screen is conceptually divided into "tiles," and each parallel pipeline is assigned to process the geometric data that falls within its designated screen tiles, thereby parallelizing the rendering task ('506 Patent, col. 4:27-33; Fig. 3).
  • Technical Importance: This tile-based, parallel-pipeline approach provided a method for scaling GPU performance to handle increasingly complex graphics and higher screen resolutions by mitigating memory bandwidth limitations ('506 Patent, col. 2:35-45).

Key Claims at a Glance

  • The complaint asserts infringement of claims 1-21 ('Compl. ¶30). Independent claim 1 is representative and recites:
    • A graphics chip with a "front-end" to receive instructions and output "geometry".
    • A "back-end" to receive and process the geometry for a frame buffer.
    • The back-end comprises "multiple parallel pipelines".
    • Geometry is located in a portion of an output screen defined by a "tile".
    • Each parallel pipeline comprises a "unified shader" programmable to perform both "color shading and texture shading".

U.S. Patent No. 7,796,133: "Unified shader" (Issued Sep. 14, 2010)

The Invention Explained

  • Problem Addressed: The patent describes conventional graphics systems that used separate hardware units for processing pixel color and for calculating texture coordinates ('133 Patent, col. 1:11-15). These units had different, fixed levels of numerical precision, which limited the flexibility of shading algorithms and complicated hardware and driver design ('133 Patent, col. 1:45-67).
  • The Patented Solution: The invention discloses a "unified shader" architecture that combines color and texture shading functions into a single, programmable unit that uses the same high precision for both types of operations ('133 Patent, Abstract; col. 3:19-30). This architecture allows complex operations to be created by looping results back through the same shader hardware, as shown conceptually in the control logic of Figure 4 ('133 Patent, Fig. 4).
  • Technical Importance: This unified approach was a key step toward the more flexible, programmable GPUs common today, freeing graphics programmers from the constraints of fixed-function pipelines and enabling more sophisticated visual effects.

Key Claims at a Glance

  • The complaint asserts infringement of claims 1, 3, 8, 14-15, 20-25, 27-31, and 33-40 ('Compl. ¶44). Independent claim 1 recites:
    • An "input interface" for receiving a packet from a rasterizer.
    • A "shading processing mechanism" configured to produce a resultant value by performing both "texture operations and color operations", comprising "at least one ALU/memory pair" operative for both types of operations.
    • An "output interface" to send the resultant value to a frame buffer.
  • The complaint notes that the right to assert additional claims is reserved.

III. The Accused Instrumentality

  • Product Identification: The complaint identifies the accused instrumentalities as "MediaTek Accused Products," which are graphics-capable integrated circuits (Compl. ¶21). Exemplary product families cited include those containing ARM Utgard, Midgard, and Bifrost series GPU designs, as well as Imagination Series 5 through 9 GPU designs (Compl. ¶24 a-d). Specific examples include the MediaTek MT6570, Helio P10, Helio P30, and Helio P60 integrated circuits (Compl. ¶24 a-c).
  • Functionality and Market Context: These integrated circuits function as Graphics Processing Units (GPUs) or Accelerated Processing Units (APUs) (Compl. ¶17). They are incorporated into a variety of consumer products, such as televisions, and are marketed, sold, and imported throughout the United States (Compl. ¶4). The complaint alleges these products incorporate the patented graphics architectures to render images (Compl. ¶18, 21).

IV. Analysis of Infringement Allegations

The complaint references claim chart exhibits (Exhibits C-H) that were not publicly filed with the pleading; therefore, the infringement allegations are summarized below in prose based on the complaint's narrative. No probative visual evidence provided in complaint.

  • '506 Patent Infringement Allegations: The complaint alleges that the MediaTek Accused Products directly infringe the '506 Patent by implementing the claimed parallel pipeline graphics system (Compl. ¶27). The infringement theory suggests that the accused GPUs contain a back-end architecture with multiple parallel processing pipelines. It is further alleged that these pipelines are assigned to process geometry based on which "tile" of the screen the geometry falls into, and that each pipeline utilizes a unified shader capable of performing both color and texture shading, thereby meeting the limitations of at least claim 1 (Compl. ¶31).
  • '133 Patent Infringement Allegations: The complaint alleges that the MediaTek Accused Products infringe the '133 Patent by incorporating the claimed unified shader architecture (Compl. ¶41). The core of this allegation is that the shaders within the accused GPUs are "unified" as claimed, meaning they have a shading mechanism with an input interface, an output interface, and at least one ALU/memory pair that is operative to perform both high-precision texture and color operations, mapping to the elements of at least claim 1 of the '133 Patent (Compl. ¶45).
  • Identified Points of Contention:
    • Technical Questions: A central technical question, particularly for the '133 Patent, will be whether the accused ARM and Imagination-based GPU architectures actually implement the specific "ALU/memory pair" structure as claimed, or if they achieve a similar result through a different, non-infringing design. The ITC's prior finding of non-infringement on this patent suggests this will be a significant evidentiary hurdle for the plaintiff.
    • Scope Questions: For the '506 Patent, a dispute may arise over the scope of the term "multiple parallel pipelines" operating on screen "tiles". The analysis will question whether the workload division in the accused products functions in the specific manner required by the patent's claims and specification.

V. Key Claim Terms for Construction

  • Term from '506 Patent, Claim 1: "unified shader"

    • Context and Importance: The definition of this term is critical because infringement requires the accused parallel pipelines to each contain a "unified shader." Practitioners may focus on this term to determine if the functionality of the shaders in the accused products, while performing both color and texture operations, aligns with the specific structural and functional meaning of "unified" as contemplated by the patent.
    • Intrinsic Evidence for Interpretation:
      • Evidence for a Broader Interpretation: The specification describes the unified shader functionally as combining "the functions of a traditional color shader and a traditional texture address shader... into a single, unified shader" ('506 Patent, col. 6:49-52).
      • Evidence for a Narrower Interpretation: The description further elaborates that the "unified shader performs both color shading and texture address shading" and that an operation may "loop back into the shader and be combined with any other operation" ('506 Patent, col. 6:54-62). This could be used to argue for a narrower construction requiring this specific loop-back capability.

  • Term from '133 Patent, Claim 1: "at least one ALU/memory pair operative to perform both texture operations and color operations"

    • Context and Importance: This term defines the core structural element of the invention. Its construction will be paramount, as the infringement case for the '133 patent hinges on whether the accused GPUs contain this specific hardware configuration.
    • Intrinsic Evidence for Interpretation:
      • Evidence for a Broader Interpretation: A party might argue the term should be read broadly to cover any architecture where a single arithmetic logic unit (ALU) and its associated memory are used to execute both types of shading instructions.
      • Evidence for a Narrower Interpretation: The specification's detailed embodiment describes "Four identical SRAM/ALU pairs" that process data in a specific, time-skewed, four-clock cycle ('133 Patent, col. 5:23-44, Fig. 2). A party could argue that the term "ALU/memory pair" is implicitly limited to this more complex, multi-pair, synchronized architecture.

VI. Other Allegations

  • Indirect Infringement: The complaint alleges both induced infringement (§ 271(b)) and contributory infringement (§ 271(c)) (Compl. ¶22-23, 28-29, 42-43). The factual basis alleged is that MediaTek sells the accused integrated circuits to its customers with the knowledge and intent that they will be incorporated into end-user products and used in an infringing manner (Compl. ¶22, 28).
  • Willful Infringement: Willfulness is alleged for both patents based on MediaTek having received actual notice "at least as early as January 2017 by way of the ITC Investigation" (Compl. ¶38, 52). The complaint alleges that MediaTek continued its infringing conduct despite an objectively high likelihood of infringement (Compl. ¶39, 53).

VII. Analyst’s Conclusion: Key Questions for the Case

  • A key evidentiary question will be one of technical proof: Can AMD produce new or more compelling evidence than it did at the ITC to demonstrate that the specific ARM and Imagination GPU architectures in MediaTek’s products meet the limitations of the '133 patent, particularly the "ALU/memory pair" element?
  • The case will also turn on a question of claim construction: Will the court construe the term "unified shader" broadly based on its combined function, or narrowly, requiring the specific structural embodiments detailed in the patents? The outcome of this construction will likely determine infringement for both patents.
  • Finally, a central issue will be the impact of prior proceedings: How will the district court and the jury weigh the ITC's prior, conflicting findings—infringement of the '506 patent versus non-infringement of the '133 patent—and the subsequent IPR cancellation of several asserted '506 patent claims, in their assessment of liability and damages?