CN117057305A - Wafer-level chip system architecture design method and device based on software and hardware cooperation - Google Patents
Wafer-level chip system architecture design method and device based on software and hardware cooperation Download PDFInfo
- Publication number
- CN117057305A CN117057305A CN202310805607.6A CN202310805607A CN117057305A CN 117057305 A CN117057305 A CN 117057305A CN 202310805607 A CN202310805607 A CN 202310805607A CN 117057305 A CN117057305 A CN 117057305A
- Authority
- CN
- China
- Prior art keywords
- simulation
- module
- architecture
- wafer
- hardware
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/30—Circuit design
- G06F30/39—Circuit design at the physical level
- G06F30/398—Design verification or optimisation, e.g. using design rule check [DRC], layout versus schematics [LVS] or finite element methods [FEM]
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/30—Circuit design
- G06F30/39—Circuit design at the physical level
- G06F30/392—Floor-planning or layout, e.g. partitioning or placement
Landscapes
- Engineering & Computer Science (AREA)
- Computer Hardware Design (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Architecture (AREA)
- Design And Manufacture Of Integrated Circuits (AREA)
Abstract
Description
技术领域Technical field
本发明涉及芯片系统体系结构设计技术领域,尤其涉及一种基于软硬件协同的晶圆级芯片系统体系结构设计方法及装置。The present invention relates to the technical field of chip system architecture design, and in particular to a wafer-level chip system architecture design method and device based on software and hardware collaboration.
背景技术Background technique
2017年,美国DARPA(Defense Advanced Research Projects Agency)在“电子复兴计划”中规划了名为“通用异构集成和IP重用战略”(CommonHeterogeneous Integrationand IPReuse Strategies,CHIPS)的Chiplet项目,参与方包括英特尔、美光、Cadence、Synopsys多类型企业等。它是一类满足特定功能的die,我们称它为模块芯片。Chiplet模式是通过die-to-die内部互联技术将多个模块芯片与底层基础芯片封装在一起,构成多功能的异构System in Packages(SiPs)芯片的模式。近年来,芯粒(Chiplet)已成为半导体产业的热门词。在摩尔定律奔向7纳米、3纳米、1纳米的物理极限之际,工艺制程缩减所需要的成本和开发时间均将大幅提升,后摩尔定律时代已经来临。Chiplet作为业界为了弥补硅工艺技术增长放缓所做的几项努力之一,被认为是延缓摩尔定律失效、放缓工艺进程时间、支撑半导体产业继续发展的最为有效的方案。In 2017, the U.S. DARPA (Defense Advanced Research Projects Agency) planned a Chiplet project called "Common Heterogeneous Integration and IP Reuse Strategies (CHIPS)" in the "Electronics Renaissance Plan". Participants include Intel, Micron, Cadence, Synopsys, etc. It is a type of die that meets specific functions. We call it a module chip. Chiplet mode is a mode that uses die-to-die internal interconnection technology to package multiple module chips and underlying basic chips together to form a multi-functional heterogeneous System in Packages (SiPs) chip. In recent years, chiplet has become a hot word in the semiconductor industry. As Moore's Law approaches the physical limits of 7 nanometers, 3 nanometers, and 1 nanometer, the cost and development time required for process reduction will increase significantly. The post-Moore's Law era has arrived. Chiplet is one of several efforts made by the industry to make up for the slowdown in the growth of silicon process technology. It is considered to be the most effective solution to delay the failure of Moore's Law, slow down the process time, and support the continued development of the semiconductor industry.
晶圆级芯片具有数量级倍增的资源规模,生产、加工、集成等过程中不可避免引入连线及功能单元的物理失效问题,系统级的鲁棒性与可用性要求架构必须可变;晶圆级芯片集成不同功能、混合力度的异构资源,不同任务与不同时刻会呈现服务高动态特性,系统服务的高灵活性与高效能必须动态演化架构。晶圆级芯片动态演化架构的设计空间探索是要解决的首要科学问题,目前尚无面向晶圆级芯片的晶上系统体系结构设计方法。Wafer-level chips have an order-of-magnitude doubling of resource scale. Physical failures of connections and functional units are inevitably introduced during production, processing, integration and other processes. System-level robustness and availability require that the architecture must be variable; wafer-level chips Integrating heterogeneous resources with different functions and mixed strengths, different tasks and different times will show highly dynamic characteristics of services. The high flexibility and high performance of system services must dynamically evolve the architecture. The design space exploration of the dynamic evolution architecture of wafer-level chips is the primary scientific problem to be solved. Currently, there is no on-chip system architecture design method for wafer-level chips.
发明内容Contents of the invention
针对目前尚无面向晶圆级芯片的晶上系统体系结构设计方法的问题,本发明提供一种基于软硬件协同的晶圆级芯片系统体系结构设计方法及装置,用于解决面向特定领域(比如加解密、深度学习等)的晶上系统优化体系结构设计。Aiming at the problem that there is currently no on-chip system architecture design method for wafer-level chips, the present invention provides a wafer-level chip system architecture design method and device based on software and hardware collaboration, which are used to solve the problem of wafer-level chip system architecture design for specific fields (such as Encryption, decryption, deep learning, etc.) on-chip system optimization architecture design.
为了实现上述目的,本发明采用以下技术方案:In order to achieve the above objects, the present invention adopts the following technical solutions:
本发明一方面提出一种基于软硬件协同的晶圆级芯片系统体系结构设计装置,包括晶圆级芯片体系结构迭代设计模块,编译仿真平台构建模块,仿真模型库设计模块,异构模拟仿真平台构建及仿真模块;On the one hand, the present invention proposes a wafer-level chip system architecture design device based on software and hardware collaboration, including a wafer-level chip architecture iterative design module, a compilation simulation platform building module, a simulation model library design module, and a heterogeneous simulation simulation platform. Build and simulate modules;
所述晶圆级芯片体系结构迭代设计模块用于完成整个系统级开发环境的架构设计,指导编译仿真平台和异构模拟仿真平台的设计,研究面向晶圆级芯片的体系结构迭代优化方法,设计从算法输入产生可执行程序或者配置代码给仿真平台,仿真平台进行仿真并输出性能结果,并通过迭代寻优进行优化设计的全流程;其中设计的架构包括软件架构和硬件架构两部分,并将软件架构交付编译仿真平台构建模块,将硬件架构交付仿真模型库设计模块和异构模拟仿真平台构建及仿真模块;The wafer-level chip architecture iterative design module is used to complete the architectural design of the entire system-level development environment, guide the design of the compilation simulation platform and heterogeneous simulation platform, study the iterative optimization method of the architecture for wafer-level chips, and design Generate executable programs or configuration codes from algorithm input to the simulation platform, which simulates and outputs performance results, and optimizes the entire design process through iterative optimization; the designed architecture includes two parts: software architecture and hardware architecture, and The software architecture is delivered to the compilation simulation platform building module, and the hardware architecture is delivered to the simulation model library design module and heterogeneous simulation platform building and simulation modules;
所述编译仿真平台构建模块用于根据晶圆级芯片体系结构迭代设计模块提供的软件架构进行软件编译器框架的搭建,得到编译仿真平台;并对软件架构进行处理,交付异构模拟仿真平台构建及仿真模块使用;The compilation simulation platform building module is used to build a software compiler framework based on the software architecture provided by the wafer-level chip architecture iterative design module to obtain a compilation simulation platform; and process the software architecture to deliver the construction of a heterogeneous simulation simulation platform and use of simulation modules;
所述仿真模型库设计模块用于完成仿真模型库的设计,具体根据晶圆级芯片体系结构迭代设计模块提供的硬件架构对晶圆级芯片的多个模型进行抽象和典型应用系统进行模型构建,并将构建的算核模型交付异构模拟仿真平台构建及仿真模块调用;The simulation model library design module is used to complete the design of the simulation model library. Specifically, it abstracts multiple models of wafer-level chips and builds models of typical application systems based on the hardware architecture provided by the wafer-level chip architecture iterative design module. And deliver the constructed computing model to the heterogeneous simulation platform construction and simulation module calling;
所述异构模拟仿真平台构建及仿真模块用于完成硬件仿真平台基础仿真框架的搭建,得到异构模拟仿真平台,具体根据晶圆级芯片体系结构迭代设计模块交付的硬件架构进行硬件仿真平台基础框架的构建,根据编译仿真平台构建模块交付的信息,通过调用仿真模型库设计模块的算核模型及基础仿真平台的运行,实现硬件部分的仿真。The heterogeneous simulation platform construction and simulation module are used to complete the construction of the basic simulation framework of the hardware simulation platform and obtain the heterogeneous simulation platform. Specifically, the hardware simulation platform foundation is based on the hardware architecture delivered by the wafer-level chip architecture iterative design module. The construction of the framework is based on the information delivered by the compilation simulation platform building module, and the simulation of the hardware part is realized by calling the computing model of the simulation model library design module and the operation of the basic simulation platform.
进一步地,还包括系统验证评估模块;所述系统验证评估模块用于基于异构仿真平台构建及仿真模块的仿真运行结果及性能功耗数据进行系统验证和评估。Further, it also includes a system verification and evaluation module; the system verification and evaluation module is used for system verification and evaluation based on the heterogeneous simulation platform construction and the simulation running results and performance power consumption data of the simulation module.
进一步地,所述软件架构包括应用程序和硬件资源检测信息。Further, the software architecture includes application programs and hardware resource detection information.
进一步地,在所述硬件架构中,晶圆级芯片体系结构迭代设计模块为仿真模型库设计模块交付算核库构建指导,为异构模拟仿真平台构建及仿真模块提供硬件仿真环境的整体架构指导。Further, in the hardware architecture, the wafer-level chip architecture iterative design module delivers calculation library construction guidance to the simulation model library design module, and provides overall architecture guidance for the hardware simulation environment for heterogeneous simulation simulation platform construction and simulation modules. .
进一步地,所述编译仿真平台构建模块具体用于接收晶圆级芯片体系结构迭代设计模块交付的硬件资源检测信息对软件编译环境进行资源映射和任务调度,并将晶圆级芯片体系结构迭代设计模块输入的应用程序翻译成所需的目标代码和可执行程序,交付异构模拟仿真平台构建及仿真模块使用。Further, the compilation simulation platform building module is specifically used to receive the hardware resource detection information delivered by the wafer-level chip architecture iterative design module, perform resource mapping and task scheduling on the software compilation environment, and perform the wafer-level chip architecture iterative design. The application program input by the module is translated into the required target code and executable program, and delivered to the heterogeneous simulation platform for construction and simulation module use.
进一步地,所述异构模拟仿真平台构建及仿真模块根据编译仿真平台构建模块交付的目标代码和可执行程序,通过调用仿真模型库设计模块的算核模型及基础仿真平台的运行,实现硬件部分的仿真。Further, the heterogeneous simulation platform construction and simulation module implements the hardware part by calling the calculation model of the simulation model library design module and the operation of the basic simulation platform according to the target code and executable program delivered by the compilation simulation platform construction module. simulation.
本发明另一方面提出一种基于软硬件协同的晶圆级芯片系统体系结构设计方法,包括:On the other hand, the present invention proposes a wafer-level chip system architecture design method based on software and hardware collaboration, including:
步骤1:用户选择其采用的算法集以及约束条件、约束目标;Step 1: The user selects the algorithm set, constraints, and constraint goals to be used;
步骤2:初始化时,智能抽取符合工艺约束的不同布局方案,并根据迭代流程智能选取优化方法;Step 2: During initialization, intelligently extract different layout solutions that meet process constraints, and intelligently select optimization methods based on the iterative process;
步骤3:如果算法集不为空,继续进行算法的自动编译和调度,否则返回平均执行时间以及吞吐量、功耗;Step 3: If the algorithm set is not empty, continue automatic compilation and scheduling of the algorithm, otherwise return the average execution time, throughput, and power consumption;
步骤4:编译仿真平台根据用户输入的算法,编译成异构模拟仿真平台能够识别的可执行代码;Step 4: The compilation and simulation platform compiles into executable code that can be recognized by the heterogeneous simulation platform based on the algorithm input by the user;
步骤5:异构模拟仿真平台接收步骤4产生的可执行代码,进行仿真执行,并反馈当前算法的执行时间以及吞吐量、功耗;Step 5: The heterogeneous simulation platform receives the executable code generated in step 4, performs simulation execution, and feeds back the execution time, throughput, and power consumption of the current algorithm;
步骤6:如果达到用户要求的优化目标,则产生最后的体系布局图。Step 6: If the optimization goal required by the user is achieved, the final system layout diagram is generated.
进一步地,所述步骤6包括:Further, the step 6 includes:
基于仿真运行结果及性能功耗数据进行系统验证和评估。System verification and evaluation are performed based on simulation running results and performance and power consumption data.
与现有技术相比,本发明具有的有益效果:Compared with the prior art, the present invention has the following beneficial effects:
本发明基于工艺特性和服务特性双重约束,以高鲁棒、高灵活、高效能等为目标,采用软硬件协同设计方法,研究复杂系统宏观逻辑结构到晶圆级电路围观物理结构的动态映射及应用演化,探索晶圆级芯片软硬件协同系统架构的设计方法学。Based on the dual constraints of process characteristics and service characteristics, this invention aims at high robustness, high flexibility, high efficiency, etc., and adopts software and hardware collaborative design methods to study the dynamic mapping of complex system macro logical structures to wafer-level circuit peripheral physical structures and Apply evolution to explore the design methodology of wafer-level chip software and hardware collaborative system architecture.
附图说明Description of the drawings
图1为本发明实施例基于软硬件协同的晶圆级芯片系统体系结构设计框图;Figure 1 is a block diagram of a wafer-level chip system architecture design based on software and hardware collaboration according to an embodiment of the present invention;
图2为本发明实施例基于软硬件协同的晶圆级芯片系统体系结构设计流程图。Figure 2 is a flow chart of a wafer-level chip system architecture design based on software and hardware collaboration according to an embodiment of the present invention.
具体实施方式Detailed ways
下面结合附图和具体的实施例对本发明做进一步的解释说明:The present invention will be further explained below in conjunction with the accompanying drawings and specific embodiments:
如图1所示,一种基于软硬件协同的晶圆级芯片系统体系结构设计装置,包括晶圆级芯片体系结构迭代设计模块(简称为模块1),编译仿真平台构建模块(简称为模块2),仿真模型库设计模块(简称为模块3),异构模拟仿真平台构建及仿真模块(简称为模块4);As shown in Figure 1, a wafer-level chip system architecture design device based on software and hardware collaboration includes a wafer-level chip architecture iterative design module (referred to as module 1), a compilation simulation platform building module (referred to as module 2) ), simulation model library design module (referred to as module 3), heterogeneous simulation platform construction and simulation module (referred to as module 4);
模块1主要完成整个系统级开发环境的架构设计,指导系统编译仿真平台和系统模拟仿真平台的设计,研究面向晶圆级芯片的体系结构迭代优化方法,设计从算法输入产生可执行程序或者配置代码给仿真平台,仿真平台进行仿真并输出性能结果,并通过迭代寻优进行优化设计的全流程,所设计的架构包括软件架构和硬件架构两部分,模块1将软件架构部分交付模块2,其中最主要的内容为面向应用的特定程序。将硬件架构部分交付模块3和模块4,在硬件架构中,模块1将为模块3交付算核库构建指导,为模块4提供硬件仿真环境的整体架构指导。Module 1 mainly completes the architectural design of the entire system-level development environment, guides the design of the system compilation simulation platform and system simulation simulation platform, studies the iterative optimization method of the architecture for wafer-level chips, and designs the generation of executable programs or configuration codes from algorithm input. The simulation platform performs simulation and outputs performance results, and optimizes the entire design process through iterative optimization. The designed architecture includes two parts: software architecture and hardware architecture. Module 1 delivers the software architecture part to module 2, of which the final The main content is application-specific procedures. The hardware architecture part is delivered to modules 3 and 4. In the hardware architecture, module 1 will deliver computing library construction guidance for module 3, and provide overall architecture guidance for the hardware simulation environment for module 4.
模块2根据模块1提供的软件架构进行软件编译器框架的搭建,得到编译仿真平台。模块2接收模块1交付的硬件资源检测信息对软件编译环境进行资源映射和任务调度,并将模块1输入的应用程序翻译成所需的目标代码和可执行程序,交付模块4使用,便于系统的模拟仿真;Module 2 builds the software compiler framework based on the software architecture provided by module 1 to obtain a compilation simulation platform. Module 2 receives the hardware resource detection information delivered by module 1, performs resource mapping and task scheduling on the software compilation environment, and translates the application program input by module 1 into the required target code and executable program, and delivers it to module 4 for use to facilitate the system. simulation simulation;
模块3和模块4共同实现硬件模拟仿真平台的设计,用于等效实际系统硬件实体。其中模块3完成仿真模型库的设计,模块3根据模块1提供的应用硬件架构,对晶圆级芯片的诸多模型(主要包括一些芯粒(chiplet),比如飞腾、申威等CPU die,NMS3210等互连die以及存储芯粒等)进行抽象和典型应用系统(比如加解密、深度学习等)进行模型构建,并将构建的算核模型交付模块4调用。Module 3 and module 4 jointly realize the design of the hardware simulation platform, which is used to equivalent the actual system hardware entity. Module 3 completes the design of the simulation model library. Based on the application hardware architecture provided by module 1, module 3 designs many models of wafer-level chips (mainly including some chiplets, such as Feiteng, Shenwei and other CPU dies, NMS3210, etc. Interconnect dies and storage chips, etc.) are abstracted and modeled with typical application systems (such as encryption and decryption, deep learning, etc.), and the constructed computing model is delivered to module 4 for calling.
模块4完成硬件仿真平台基础仿真框架的搭建,得到异构模拟仿真平台。模块4根据模块1交付的硬件架构进行硬件仿真平台基础框架的构建,支持基础仿真程序的运行,模块4根据模块2交付的目标代码和可执行程序,通过调用模块3的算核模型及基础仿真平台的运行,实现硬件部分的仿真,模块4的仿真运行结果及性能功耗等数据将交付模块5使用进行系统验证和评估。Module 4 completes the construction of the basic simulation framework of the hardware simulation platform and obtains a heterogeneous simulation platform. Module 4 builds the basic framework of the hardware simulation platform based on the hardware architecture delivered by module 1 and supports the operation of basic simulation programs. Module 4 uses the target code and executable program delivered by module 2 to call the computing model and basic simulation of module 3. The operation of the platform realizes the simulation of the hardware part. The simulation operation results and performance and power consumption data of module 4 will be delivered to module 5 for system verification and evaluation.
本发明另一方面还提出一种基于软硬件协同的晶圆级芯片系统体系结构设计方法,相应的流程如图2所示,包括:On the other hand, the present invention also proposes a wafer-level chip system architecture design method based on software and hardware collaboration. The corresponding process is shown in Figure 2, including:
步骤1:用户选择其采用的算法集(比如加解密应用领域种的md5、sha1、des等算法;深度学习中的CNN、RNN等算法)以及约束条件(比如8英寸晶圆)、约束目标(如平均执行时间最短)等;Step 1: The user selects the algorithm set he uses (such as md5, sha1, des and other algorithms in the field of encryption and decryption; CNN, RNN and other algorithms in deep learning) and constraints (such as 8-inch wafer), constraint goals ( Such as the shortest average execution time), etc.;
步骤2:初始化时,智能抽取符合工艺约束的不同布局方案,并可根据迭代流程,智能选取优化方法;Step 2: During initialization, different layout plans that meet process constraints are intelligently extracted, and optimization methods can be intelligently selected based on the iterative process;
步骤3:如果算法集不为空,继续进行算法的自动编译和调度,否则返回平均执行时间以及吞吐量、功耗等;Step 3: If the algorithm set is not empty, continue automatic compilation and scheduling of the algorithm, otherwise return the average execution time, throughput, power consumption, etc.;
步骤4:编译仿真平台根据用户输入的算法,编译成异构模拟仿真平台能够识别的可执行代码;Step 4: The compilation and simulation platform compiles into executable code that can be recognized by the heterogeneous simulation platform based on the algorithm input by the user;
步骤5:异构模拟仿真平台接收步骤4产生的可执行代码,进行仿真执行,并反馈当前算法的执行时间以及吞吐量、功耗等参数;Step 5: The heterogeneous simulation platform receives the executable code generated in step 4, performs simulation execution, and feeds back the execution time of the current algorithm as well as parameters such as throughput and power consumption;
步骤6:如果达到用户要求的优化目标,则产生最后的体系布局图。Step 6: If the optimization goal required by the user is achieved, the final system layout diagram is generated.
进一步地,所述步骤6包括:Further, the step 6 includes:
基于仿真运行结果及性能功耗数据进行系统验证和评估。System verification and evaluation are performed based on simulation running results and performance and power consumption data.
综上,本发明基于工艺特性和服务特性双重约束,以高鲁棒、高灵活、高效能等为目标,采用软硬件协同设计方法,研究复杂系统宏观逻辑结构到晶圆级电路围观物理结构的动态映射及应用演化,探索晶圆级芯片软硬件协同系统架构的设计方法学。In summary, the present invention is based on the dual constraints of process characteristics and service characteristics, aims at high robustness, high flexibility, and high efficiency, and uses software and hardware co-design methods to study the macroscopic logical structure of complex systems to the physical structure of wafer-level circuits. Dynamic mapping and application evolution, exploring the design methodology of wafer-level chip software and hardware collaborative system architecture.
以上所示仅是本发明的优选实施方式,应当指出,对于本技术领域的普通技术人员来说,在不脱离本发明原理的前提下,还可以做出若干改进和润饰,这些改进和润饰也应视为本发明的保护范围。What is shown above is only the preferred embodiment of the present invention. It should be pointed out that those skilled in the art can also make several improvements and modifications without departing from the principles of the present invention. These improvements and modifications can also be made. should be regarded as the protection scope of the present invention.
Claims (8)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310805607.6A CN117057305A (en) | 2023-07-03 | 2023-07-03 | Wafer-level chip system architecture design method and device based on software and hardware cooperation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310805607.6A CN117057305A (en) | 2023-07-03 | 2023-07-03 | Wafer-level chip system architecture design method and device based on software and hardware cooperation |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117057305A true CN117057305A (en) | 2023-11-14 |
Family
ID=88656128
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310805607.6A Pending CN117057305A (en) | 2023-07-03 | 2023-07-03 | Wafer-level chip system architecture design method and device based on software and hardware cooperation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117057305A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118821709A (en) * | 2024-09-20 | 2024-10-22 | 河南嵩山实验室产业研究院有限公司洛阳分公司 | Heuristic-based domain-specific wafer-level chip design optimization method, system and storage medium |
| CN120163113A (en) * | 2025-03-26 | 2025-06-17 | 西安交通大学 | A wafer-level chip system design space construction and fast parameter search method |
-
2023
- 2023-07-03 CN CN202310805607.6A patent/CN117057305A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118821709A (en) * | 2024-09-20 | 2024-10-22 | 河南嵩山实验室产业研究院有限公司洛阳分公司 | Heuristic-based domain-specific wafer-level chip design optimization method, system and storage medium |
| CN118821709B (en) * | 2024-09-20 | 2025-03-18 | 河南嵩山实验室产业研究院有限公司洛阳分公司 | Heuristic-based domain-specific wafer-level chip design optimization method, system and storage medium |
| CN120163113A (en) * | 2025-03-26 | 2025-06-17 | 西安交通大学 | A wafer-level chip system design space construction and fast parameter search method |
| CN120163113B (en) * | 2025-03-26 | 2025-11-11 | 西安交通大学 | Wafer-level chip system design space construction and rapid parameter searching method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Davidson et al. | The Celerity open-source 511-core RISC-V tiered accelerator fabric: Fast architectures and design methodologies for fast chips | |
| Gordon et al. | Novel computer architectures and quantum chemistry | |
| CN117057305A (en) | Wafer-level chip system architecture design method and device based on software and hardware cooperation | |
| Moon et al. | Evaluating spatial accelerator architectures with tiled matrix-matrix multiplication | |
| US12112202B2 (en) | Framework for application driven exploration and optimization of hardware engines | |
| Gerstlauer et al. | System-level abstraction semantics | |
| US20250216920A1 (en) | Compiler that generates configuration information for configuring an integrated circuit to mitigate inductive-induced voltage droop | |
| US20240020537A1 (en) | Methodology to generate efficient models and architectures for deep learning | |
| US20250315093A1 (en) | Integrated circuit that mitigates supply voltage deviation using compute unit group identifiers | |
| Liang et al. | FlexCL: A model of performance and power for OpenCL workloads on FPGAs | |
| Boutros et al. | Architecture and application co-design for beyond-FPGA reconfigurable acceleration devices | |
| CN115271078A (en) | Software stack with cooperation of supercomputer and quantum computer and working method | |
| Chamberlain et al. | Auto-Pipe: Streaming applications on architecturally diverse systems | |
| Russo et al. | Multiobjective end-to-end design space exploration of parameterized dnn accelerators | |
| Lohoff et al. | Interfacing neuromorphic hardware with machine learning frameworks-a review | |
| US20240020265A1 (en) | Operating a Cost Estimation Tool for Placing and Routing an Operation Unit Graph on a Reconfigurable Processor | |
| US20230385231A1 (en) | Low Latency Nodes Fusion in a Reconfigurable Data Processor | |
| Cristal et al. | LEGaTO: first steps towards energy-efficient toolset for heterogeneous computing | |
| Luchterhandt et al. | Towards a Rocket Chip Based Implementation of the RISC-V GPC Architecture. | |
| Zhao et al. | A verification method for array-based vision chip using a fixed-point neural network simulation tool | |
| US12260200B2 (en) | Defining and executing a section of code for an interpreted software language | |
| Lunnikivi et al. | Modular RTIC: Lightweight real time for customized architectures | |
| Xu et al. | Software-defined process-near-memory architecture using 3D hybrid bonding integration | |
| US12204488B2 (en) | High performance softmax for large models | |
| US12461889B2 (en) | Intelligent graph execution and orchestration engine for a reconfigurable data processor |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Country or region after: China Address after: 450000 Science Avenue 62, Zhengzhou High-tech Zone, Henan Province Applicant after: Information Engineering University of the Chinese People's Liberation Army Cyberspace Force Address before: No. 62 Science Avenue, High tech Zone, Zhengzhou City, Henan Province Applicant before: Information Engineering University of Strategic Support Force,PLA Country or region before: China |