PTAB
IPR2015-01314
Samsung Electronics Co Ltd v. NVIDIA Corp
Key Events
Petition
Table of Contents
petition
1. Case Identification
- Case #: IPR2015-01314
- Patent #: 7,209,140
- Filed: May 29, 2015
- Petitioner(s): Samsung Electronics Co., Ltd., Samsung Electronics America, Inc., and Samsung Semiconductor, Inc.
- Patent Owner(s): NVIDIA Corporation
- Challenged Claims: 1-10, 12-14
2. Patent Overview
- Title: System, method and article of manufacture for a programmable vertex processing model with instruction set
- Brief Description: The ā140 patent relates to a programmable hardware graphics accelerator that implements a specific instruction set. The invention purports to increase the flexibility of operations that can be performed on graphics data as it passes through a graphics processing pipeline.
3. Grounds for Unpatentability
Ground 1: Obviousness over RenderMan, Mallory, and i860 Reference Manual - Claims 1-2, 5-6, and 13-14 are obvious over RenderMan, Mallory, and the i860 Reference Manual.
- Prior Art Relied Upon: RenderMan (The RenderMan Interface Version 3.1, Sep. 1989), Mallory ("New for PCs: Truevision ships DOS version of RenderPak," Newsbytes News, Apr. 1991), and the i860 Reference Manual (The Intel i860 Microprocessor Family Programmer's Reference Manual, 1991).
- Core Argument for this Ground:
- Prior Art Mapping: Petitioner argued that this combination taught every element of the challenged claims. RenderMan disclosed a high-level, C-like programmable shading language with an instruction set for complex graphics operations. Mallory, a news article, described a commercial product (the Horizon860 graphics card) that explicitly bundled the RenderMan software with an Intel i860 microprocessor, which functioned as a hardware graphics accelerator. The i860 Reference Manual detailed the native instruction set and graphics capabilities of that microprocessor. Petitioner contended that the combination of RenderMan's high-level instructions and the i860's native instructions, running on the dedicated hardware described by Mallory, met the claim limitations for a programmable graphics system executing a complete instruction set.
- Motivation to Combine: The motivation was exceptionally strong and explicit, as Mallory described a pre-existing commercial product that physically combined the core components. Petitioner argued this was not a hypothetical combination but a real-world implementation that demonstrated the desire and feasibility of pairing RenderMan's software flexibility with the i860's hardware acceleration.
- Expectation of Success: Success was not merely expected but was actually achieved and documented in the prior art, as evidenced by the commercial availability of the combined product.
Ground 2: Obviousness over Kochan, Kirk, AltiVec Manual, and AltiVec White Paper - Claims 1-10 and 12-14 are obvious over Kochan, Kirk, the AltiVec Manual, and the AltiVec White Paper.
- Prior Art Relied Upon: Kochan ("Programming in ANSI C," 1994), Kirk ("Graphics Gems III," 1992), the AltiVec Manual (Motorola Inc., Nov. 1998), and the AltiVec White Paper ("Motorola's AltiVec Technology," 1998).
- Core Argument for this Ground:
- Prior Art Mapping: Petitioner asserted that the AltiVec Manual and White Paper disclosed a hardware graphics accelerator in the form of the AltiVec high-performance vector parallel processing extension for PowerPC processors, along with its specific instruction set. Kochan taught the fundamentals of the C programming language, while Kirk provided C-language libraries and algorithms specifically for advanced graphics processing, including lighting and transformations. The combination allegedly disclosed a system where a programmable hardware accelerator (AltiVec) is programmed using a high-level language (C from Kochan) and specialized graphics libraries (from Kirk) to perform the graphics operations recited in the claims, including in parallel.
- Motivation to Combine: A POSITA would combine these references because the AltiVec White Paper explicitly identified C and C++ as the "preferred programming environment" for the AltiVec technology. To facilitate this, Motorola developed and provided standardized C/C++ language extensions. This provided a direct and compelling reason for a developer to apply the well-known C-language graphics techniques taught by Kochan and Kirk to program the powerful AltiVec hardware accelerator.
- Expectation of Success: A POSITA would have a high expectation of success, as the AltiVec hardware and its associated programming extensions were specifically designed by Motorola to be used with the C language for performance-driven applications like graphics.
4. Key Claim Construction Positions
- Petitioner dedicated a substantial portion of its argument to claim construction, asserting that the claims are unpatentable even under the broader constructions advanced by Patent Owner NVIDIA in related litigation. This strategy aimed to preempt disputes over claim scope.
- "instruction set": Petitioner proceeded with NVIDIA's proposed construction of "the complete set of instructions recognized by a given computer or provided by a given programming language." This broad interpretation allowed Petitioner to combine instructions from a high-level language (like RenderMan or C) with the native instructions of a hardware accelerator (like the i860 or AltiVec) to meet the claim limitation.
- "hardware graphics accelerator": Petitioner noted that NVIDIA had argued for a construction of "graphics processing unit (GPU)" in litigation. Petitioner contended the proper construction is broader ("hardware for processing graphics data") but argued the prior art i860 and AltiVec processors met either construction.
- "operation": Petitioner adopted NVIDIA's broad construction of "action or process" for the purpose of the petition, which allowed it to map functions from high-level programming languages and libraries to the claimed "operations."
5. Relief Requested
- Petitioner requests institution of an inter partes review and cancellation of claims 1-10 and 12-14 of Patent 7,209,140 as unpatentable.
Analysis metadata