DCT

2:22-cv-00134

ATI Tech ULC v. TCL Smart Device Vietnam Co Ltd

Log In
Key Events
Amended Complaint

I. Executive Summary and Procedural Information

  • Parties & Counsel:
  • Case Identification: 2:22-cv-00134, E.D. Tex., 03/09/2023
  • Venue Allegations: Plaintiff alleges venue is proper in the Eastern District of Texas based on Defendants' business activities in the district. Allegations for TCL include marketing and sales activities, such as product demonstrations at the SXSW festival in Austin and product availability for in-store pickup at retailers like BestBuy in Longview, Texas. Allegations for Realtek include its use of authorized distributors with sales offices and warehouses located in Texas.
  • Core Dispute: Plaintiff alleges that Defendants’ smart televisions and the graphics processing components within them, such as systems-on-a-chip (SoCs), infringe five patents related to graphics processing architectures, texture decompression, and task management.
  • Technical Context: The technology relates to the architecture of graphics processing units (GPUs), which are fundamental components for rendering images in a vast array of consumer electronics, most notably smart televisions.
  • Key Procedural History: The complaint notes that U.S. Patent No. 7,742,053 has been previously upheld as valid by the U.S. Court of Appeals for the Federal Circuit in ATI Techs. ULC v. Iancu, 920 F.3d 1362 (Fed. Cir. 2019). This ruling may affect arguments related to the validity of that patent. The complaint also references a co-pending International Trade Commission action involving the parties.

Case Timeline

Date Event
2001-08-24 Earliest conception/reduction to practice date alleged for the ’053 & ’454 Patents
2006-08-31 Earliest priority date for the ’628 Patent
2009-09-03 Priority date for the ’381 Patent
2010-06-22 U.S. Patent No. 7,742,053 issues
2010-11-23 Priority date for the ’547 Patent
2013-06-18 U.S. Patent No. 8,468,547 issues
2014-06-24 U.S. Patent No. 8,760,454 issues
2014-10-07 U.S. Patent No. 8,854,381 issues
2018-10-25 Plaintiff allegedly informed TCL of the ’454 Patent
2019-11-01 Plaintiff allegedly informed TCL of the ’053 Patent
2021-11-23 U.S. Patent No. 11,184,628 issues
2023-03-09 Complaint filed

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

U.S. Patent No. 7,742,053 - "Multi-Thread Graphics Processing System" (issued June 22, 2010)

The Invention Explained

  • Problem Addressed: Prior art graphics processors required separate, dedicated hardware units ("shaders") to perform vertex operations (calculating the geometry of an object) and pixel operations (calculating the color and texture of on-screen pixels), which had to be performed separately and sequentially (Compl. ¶6). This created inefficiencies in processing graphics data.
  • The Patented Solution: The invention describes a graphics processing system that unifies these operations. It uses a common memory device to store both "pixel command threads" and "vertex command threads" and employs an "arbiter" to select which type of thread to send to a shared pool of command processing engines ('053 Patent, Abstract; Fig. 4). This allows for a more flexible and efficient allocation of processing resources based on the immediate needs of the graphics task.
  • Technical Importance: This unified architecture approach allows graphics processors to be smaller, more power-efficient, and better able to handle varying workloads where the balance between geometry and pixel processing changes dynamically (Compl. ¶74).

Key Claims at a Glance

  • The complaint asserts independent claim 1 and reserves the right to assert others (Compl. ¶¶108, 110).
  • Claim 1 Elements:
    • A graphics processing system comprising:
    • at least one memory device comprising a first portion operative to store a plurality of pixel command threads and a second portion operative to store a plurality of vertex command threads;
    • an arbiter, coupled to the at least one memory device, operable to select a command thread from either of the plurality of pixel command threads and the plurality of vertex command threads; and
    • a plurality of command processing engines, coupled to the arbiter, each operable to receive and process the command thread.

U.S. Patent No. 8,760,454 - "Graphics processing architecture employing a unified shader" (issued June 24, 2014)

The Invention Explained

  • Problem Addressed: Similar to the '053 Patent, this patent addresses the inefficiencies of conventional graphics architectures that used separate, dedicated hardware for vertex and pixel processing ('454 Patent, col. 2:6-21).
  • The Patented Solution: The invention discloses a "unified shader" architecture. This architecture features a central processor unit coupled with a general-purpose register block that can hold data for both vertex and pixel calculations. A "sequencer" directs the processor to execute either vertex or pixel operations on the data held in the registers, allowing a single processor core to flexibly handle both types of tasks ('454 Patent, Abstract; col. 3:5-12).
  • Technical Importance: This design allows a single, programmable processing unit to replace two separate, fixed-function units, leading to greater efficiency and adaptability in graphics hardware (Compl. ¶74).

Key Claims at a Glance

  • The complaint asserts independent claim 2 and reserves the right to assert others (Compl. ¶¶126, 128).
  • Claim 2 Elements:
    • A unified shader, comprising:
    • a general purpose register block for maintaining data;
    • a processor unit;
    • a sequencer, coupled to the general purpose register block and the processor unit, the sequencer maintaining instructions operative to cause the processor unit to execute vertex calculation and pixel calculation operations on selected data maintained in the general purpose register block; and
    • wherein the processor unit executes instructions that generate a pixel color in response to selected data from the general purpose register block and generates vertex position and appearance data in response to selected data from the general purpose register block.

U.S. Patent No. 11,184,628 - "Texture Decompression Techniques" (issued November 23, 2021)

Technology Synopsis

This patent describes a method for decompressing texture data, a form of image compression used in graphics. The method involves receiving a compressed block of texture data and unpacking it into two or more "disjoint subsets." The color for each pixel (texel) within a subset is then determined by interpolating between color endpoints defined for that specific subset, improving image quality over existing schemes ('628 Patent, Abstract).

Asserted Claims

Claim 7 (independent) is asserted (Compl. ¶147).

Accused Features

The complaint alleges that the ARM Mali-G31 GPU in the accused products infringes by supporting "Adaptive Scalable Texture Compression (ASTC)," which allegedly operates by partitioning texture blocks and interpolating colors within those partitions (Compl. ¶149).

U.S. Patent No. 8,468,547 - "Method and System for Synchronizing Thread Wavefront Data and Events" (issued June 18, 2013)

Technology Synopsis

This patent addresses the problem of "wavefront stalls," where a group of processing threads must wait for data from a previous group to be written to memory before it can proceed. The invention describes a system using an "event synchronizer" that manages the release of processing events. It ensures that an event signaling the completion of a task is only released after the data associated with that task has been stored in memory, preventing subsequent tasks from starting prematurely ('547 Patent, Abstract).

Asserted Claims

Claim 16 (independent) is asserted (Compl. ¶164).

Accused Features

The complaint alleges that the ARM Mali-G31 GPU contains a "unified shader architecture" and a "Synchronization Circuit" that together perform the claimed functions of executing a wavefront, generating an associated event, and synchronizing the release of that event with the storage of the wavefront's output data (Compl. ¶¶166-167).

U.S. Patent No. 8,854,381 - "Processing Unit That Enables Asynchronous Task Dispatch" (issued October 7, 2014)

Technology Synopsis

The patent describes a computing system that avoids inefficient "context switching" when handling tasks of different priorities. The system uses a plurality of "engines" that receive tasks from a scheduling module and load the associated state data. A shader core can then execute multiple tasks concurrently, for instance executing a second task while a first task is still in process, based on their respective state data, improving efficiency ('381 Patent, Abstract).

Asserted Claims

Claim 15 (independent) is asserted (Compl. ¶182).

Accused Features

The complaint alleges the accused products, which contain both a CPU and a GPU (such as the Mali-G31), infringe. It maps the GPU's task management units to the claimed "plurality of engines" and the GPU's shader cores to the claimed "shader core" that can execute multiple tasks concurrently (Compl. ¶¶183-185).

III. The Accused Instrumentality

Product Identification

The accused instrumentalities are TCL's 3-Series, 4-Series, 5-Series, 6-Series, and 8-Series smart televisions (running Roku, Android, or Google TV platforms) and the Realtek SoCs and other graphics components incorporated within them (Compl. ¶81). The complaint identifies the TCL 50S535 television, containing a Realtek RTD2873 SoC with an ARM Mali-G31 GPU, as an exemplary infringing product (Compl. ¶¶109, 111).

Functionality and Market Context

  • The complaint alleges that the accused TCL televisions incorporate graphics processing systems that perform the infringing functions (Compl. ¶111). Specifically, it focuses on the functionality of the ARM Mali-G31 GPU, which is alleged to implement a "Bifrost architecture" (Compl. ¶111). This GPU is accused of having a "Frontend Circuit" that manages and schedules "vertex/tiling/compute workloads and...fragment workloads" for execution by its processing cores (Compl. ¶¶41, 130, 131). The complaint includes a screenshot from a BestBuy webpage showing the TCL 50S535 television available for purchase and pickup in Longview, Texas (Compl. p. 9).
  • TCL is alleged to be a major player in the U.S. television market, ranking "No. 2 in the U.S." by sales volume in 2021 (Compl. ¶11). Realtek is described as a "world-leading IC provider" that develops "Integrated 4K Smart LCD TV Controllers" for the North American market (Compl. ¶¶42, 46).

IV. Analysis of Infringement Allegations

U.S. Patent No. 7,742,053 Infringement Allegations

Claim Element (from Independent Claim 1) Alleged Infringing Functionality Complaint Citation Patent Citation
at least one memory device comprising a first portion operative to store a plurality of pixel command threads and a second portion operative to store a plurality of vertex command threads The ARM Mali-G31 GPU allegedly includes "a pair of queues, one for vertex/tiling/compute workloads and one for fragment workloads," which store vertex and pixel command threads, respectively. ¶112 col. 3:4-10
an arbiter...operable to select a command thread from either of the plurality of pixel command threads and the plurality of vertex command threads The Mali GPU's "Frontend Circuit" (which includes various management units) is alleged to select command threads from either the pixel (fragment) or vertex command queues for processing. ¶41 col. 3:11-15
a plurality of command processing engines...each operable to receive and process the command thread The Mali GPU's "programmable Execution Core (EC)" is alleged to consist of one or more "Execution Engines (EEs)" that receive and process the selected command threads from the Frontend Circuit. ¶41 col. 3:16-19
  • Identified Points of Contention:
    • Scope Questions: A central question may be whether the various components of the accused "Frontend Circuit" collectively perform the function of the claimed "arbiter." The defense may argue that the distributed logic of the accused GPU does not correspond to the single "arbiter" element described in the patent.
    • Technical Questions: The analysis may turn on whether the "queues" for workloads in the Mali GPU function as the claimed "memory device" for storing distinct "command threads." The court may need to determine if a "workload" submitted to a queue is equivalent to a "command thread" as contemplated by the patent.

U.S. Patent No. 8,760,454 Infringement Allegations

Claim Element (from Independent Claim 2) Alleged Infringing Functionality Complaint Citation Patent Citation
a general purpose register block for maintaining data The complaint alleges the Mali-G31 GPU contains a general purpose register block. ¶130 col. 5:5-7
a processor unit The complaint alleges the Mali-G31 GPU contains a processor unit. ¶129 col. 3:60-62
a sequencer, coupled to the general purpose register block and the processor unit, the sequencer maintaining instructions operative to cause the processor unit to execute vertex calculation and pixel calculation operations on selected data... The Mali GPU's "Frontend Circuit" is alleged to be the sequencer, which is "responsible for scheduling workloads onto the various processing units inside the GPU" and maintains instructions to cause the execution of both vertex and pixel calculations. ¶130 col. 5:8-15
wherein the processor unit executes instructions that generate a pixel color...and generates vertex position and appearance data... The complaint alleges the Mali GPU processes workloads from both its vertex and fragment queues, at times in parallel, thereby executing instructions to generate both pixel color and vertex data. The complaint includes a screenshot from Realtek's distributor, Symmetry Electronics, showing a warehouse in Fort Worth, Texas (Compl. p. 28). ¶131 col. 5:16-22
  • Identified Points of Contention:
    • Scope Questions: The dispute may focus on the term "sequencer." The defense may argue that the accused "Frontend Circuit," described as a collection of management and creator units, is not the single, specific "sequencer" element required by the claim.
    • Technical Questions: A key question will be whether the evidence shows that the accused "Frontend Circuit" actually "maintains instructions" in the manner claimed, or if it merely schedules tasks for execution by a processor that fetches instructions from a separate instruction store, potentially creating a mismatch with the claim language.

V. Key Claim Terms for Construction

Patent: ’053 Patent

  • The Term: "arbiter"
  • Context and Importance: The infringement theory hinges on mapping the accused Mali GPU's "Frontend Circuit"—a collection of units like the Job Manager, Compute Frontend, and Fragment Frontend—to the single claimed "arbiter." The definition of "arbiter" will determine if this mapping is plausible or if the claim requires a more specific, singular component that is absent in the accused device.
  • Intrinsic Evidence for Interpretation:
    • Evidence for a Broader Interpretation: The specification describes the arbiter functionally as something that "retrieves a first command thread" and "provides the retrieved command thread to the command processing engine" ('053 Patent, col. 2:48-51). This functional language could support an interpretation where any collection of hardware performing this selection and provision role meets the limitation.
    • Evidence for a Narrower Interpretation: The patent figures (e.g., Fig. 2, element 204; Fig. 4, element 306) consistently depict the "arbiter" as a distinct, singular block situated between the memory/stations and the processing engine. This could support an argument that the term requires a specific structural configuration, not just a distributed function.

Patent: ’454 Patent

  • The Term: "sequencer"
  • Context and Importance: Similar to "arbiter" in the '053 Patent, the infringement allegation maps the accused "Frontend Circuit" to the claimed "sequencer." The construction of this term is critical to determining whether a collection of scheduling and management units can be considered a "sequencer" that "maintain[s] instructions operative to cause the processor unit to execute" the claimed operations.
  • Intrinsic Evidence for Interpretation:
    • Evidence for a Broader Interpretation: The patent states the sequencer "maintaining instructions operative to cause the processor unit to execute" various operations ('454 Patent, Claim 2). This could be read functionally to cover a system that orchestrates execution, even if the instructions themselves are stored elsewhere but are managed or pointed to by the sequencer.
    • Evidence for a Narrower Interpretation: The detailed description links the sequencer to an "instruction store" and states it "determines whether the next instruction to be executed...is an arithmetic or logical instruction or a memory...instruction" ('454 Patent, col. 9:60-65). This suggests a more active role in handling instructions directly, which the defense may argue the accused scheduling circuit does not perform.

VI. Other Allegations

  • Indirect Infringement: The complaint alleges both induced and contributory infringement. Inducement is based on allegations that Defendants encourage customers to use the infringing products through marketing and user manuals (Compl. ¶115). Contributory infringement is based on the allegation that the accused graphics components are a material part of the patented inventions, are not staple articles of commerce, and have no substantial non-infringing uses (Compl. ¶116).
  • Willful Infringement: The complaint alleges willful infringement for all five patents. It asserts that Defendants had pre-suit knowledge based on specific notice letters sent to TCL and Realtek. For example, Plaintiff allegedly notified TCL of the ’053 Patent on November 1, 2019, and of the ’454 Patent on October 25, 2018 (Compl. ¶¶122, 141). Knowledge of the remaining patents is alleged based on pre-suit letters or the filing of the complaint itself (Compl. ¶¶158, 176, 194).

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

This case presents several critical questions for the court that blend claim interpretation with complex technical analysis of modern GPU architecture.

  • A core issue will be one of structural and functional correspondence: Can the collection of distributed task-management units in the accused ARM Mali-G31 GPU's "Frontend Circuit" be legally and factually equated to the singular "arbiter" of the ’053 Patent and the "sequencer" of the ’454 Patent, or do the claims require discrete components that are absent in the accused design?
  • A second key question will be one of definitional scope: Does the term "command thread" as used in the ’053 Patent, which arises from a particular graphics processing paradigm, read on the "workloads" that are submitted to queues in the accused modern GPU architecture?
  • Finally, an evidentiary question will be one of pre-suit knowledge: The complaint alleges specific dates on which Defendants were notified of certain patents. The evidence presented regarding this alleged notice will be central to the determination of willfulness and potential enhanced damages.