CN109683877A - A kind of GPU software and hardware interaction TLM system based on SystemC - Google Patents
A kind of GPU software and hardware interaction TLM system based on SystemC Download PDFInfo
- Publication number
- CN109683877A CN109683877A CN201811516374.3A CN201811516374A CN109683877A CN 109683877 A CN109683877 A CN 109683877A CN 201811516374 A CN201811516374 A CN 201811516374A CN 109683877 A CN109683877 A CN 109683877A
- Authority
- CN
- China
- Prior art keywords
- module
- transaction
- model
- hardware
- gpu
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F8/00—Arrangements for software engineering
- G06F8/30—Creation or generation of source code
- G06F8/31—Programming languages or programming paradigms
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
- G06F9/5027—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T1/00—General purpose image data processing
- G06T1/20—Processor architectures; Processor configuration, e.g. pipelining
Abstract
The present invention relates to computer hardware modeling technique field more particularly to a kind of GPU software and hardware interaction TLM models based on SystemC.The GPU software and hardware interaction TLM model based on SystemC includes master agent layer model and host interface layer model, wherein master agent layer model provides the interaction approach of software-driven Yu host interface layer model, host interface layer model provides the interaction approach of master agent layer model Yu GPU hardware TLM model, interactive content includes issuing for OpenGL CM (Command Packet) order packet, the read-write of architecture register, interrupt processing, image and data texturing load obtain.The present invention realizes the GPU software and hardware model exchange method based on TLM model, solves software-driven and hardware model interaction problems in GPU modeling, has effectively helped RTL design exploitation and verifying.
Description
Technical field
The present invention relates to computer hardware modeling technique field more particularly to a kind of GPU software and hardware friendships based on SystemC
Mutual TLM system.
Background technique
At present China just the domestic graphics processor chip of independent development (Graphic Processing Unit,
GPU), but due to before this country always using the commercial or military GPU chip of external import, independently ground in GPU
System aspect lacks design experiences, belongs to the exploration development stage to core technology, at the same time, the rule of GPU chip hardware logic
Mould is huge, designs and realizes from system architecture document preparation to rtl, this will be a very long engineering, and even if rtl has been designed
The a large amount of time will be also needed at the verifying of rear functional performance etc., therefore, in order to improve independent research graphics processor project
Progress and reliability, there is an urgent need for a kind of methods of colleges and universities can shorten setting for entire design cycle and secondary graphics processor
Meter process, this method are exactly the modeling of (Transaction Level Modeling, TLM) transaction-level.TLM (things grade modeling
Method) it is a kind of advanced digital system model method, system model can be quickly constructed in the early stage of project, it
The details of communication details and function unit or communication construction between model is separated.Communication mechanism (such as bus or FIFO) quilt
It is modeled as channel, and is presented in the form of SystemC interface class to module.Things request is generally calling these channel models
Interface function when occur, and interface function encapsulates the low-level details of information exchange.In things level, TLM more emphasizes to count
According to the function of transmission itself -- the content of data and the start-stop point of transmission, and be related to implementing as far as possible less.
Summary of the invention
The problem of based on background technique, a kind of GPU software and hardware interaction based on SystemC provided by the invention
TLM modelling is able to solve the interaction problems of software-driven Yu GPU hardware model, for that can shift to an earlier date in hardware micro-structure TLM
Functional verification is carried out on model gets through software and hardware interaction channel.
Of the invention provides a kind of graphics processor GPU software and hardware interaction TLM system based on SystemC, feature
It is, including master agent layer model and host interface layer model, the master agent layer model and the host interface layer mould
Type interacts for realizing drive software with the software and hardware of GPU hardware model.
Optionally, the master agent layer model, for realizing the drive software based on C language and based on SystemC's
The data interaction of the host interface model, specifically includes: order issues module (1), module for reading and writing (2), interruption processing module
(3) and main memory read module (4), wherein the order issues module (1) for sending OpenGL CM (Command
Package) order packet, order packet carry OpenGL function code and relevant parameter information;The module for reading and writing (2) is for reading and writing
Architecture register, architecture register is used to characterize the working condition and mode of GPU.
Optionally, the host interface layer model, for realizing the GPU host interface based on SystemC transaction-level interface
The interaction of model and GPU PCIe hardware model, the host interface layer model include: host memory read-write transaction-level module
(5), architecture register read-write transaction-level module (6), interrupt status read and write transaction-level module (7), direct memory reading DMA
Configuration transaction grade module (8), wherein host memory read-write transaction-level module (5) provides host memory read-write transaction-level
Interface, architecture register read-write transaction-level module (6) provide architecture register read-write transaction-level interface, institute
It states interrupt status read-write transaction-level module (7) and provides interrupt status read-write transaction-level interface, the direct memory reading DMA is matched
It sets transaction-level module (8) and provides DMA configuration transaction grade interface, called for master agent layer model and GPU hardware model.
Optionally, order issues module (1) and is configured by the DMA configuration transaction grade interface and direct memory reading DMA
CM order packet is handed down to host interface layer model, dma descriptor packet by transaction-level module (8) connection in a manner of dma descriptor
Include DMA destination address, DMA source address, DMA path and DMA length.
Module for reading and writing (2) reads and writes transaction-level interface by the architecture register and architecture register reads and writes thing
Business grade module (6) connection for GP configuring U operating mode and obtains GPU working condition.
Interruption processing module (3) reads and writes transaction-level interface by the interrupt status and interrupt status reads and writes transaction-level module
(7) it connects, for obtaining and removing Host Interrupt.
Main memory read module (4) reads and writes transaction-level interface by the host memory and host memory reads and writes transaction-level module
(5) it connects, for reading and writing main memory data.
A kind of GPU software and hardware interaction TLM system based on SystemC provided by the invention, should the GPU based on SystemC
Software and hardware interaction TLM model includes master agent layer model and host interface layer model, and wherein master agent layer model provides
The interaction approach of software-driven and host interface layer model, it is hard with GPU that host interface layer model provides master agent layer model
The interaction approach of part TLM model, interactive content include issuing for OpenGL CM (Command Packet) order packet, system
The read-write of structure register, interrupt processing, image and data texturing load obtain.It is able to solve software-driven and GPU hardware
The interaction problems of model get through software and hardware interaction channel for that can shift to an earlier date the progress functional verification on hardware micro-structure TLM model.
Detailed description of the invention
Fig. 1 is a kind of GPU software and hardware interaction TLM model structure schematic diagram based on SystemC provided by the invention;
Specific embodiment
A kind of graphics processor GPU software and hardware interaction TLM system based on SystemC, as shown in Figure 1, including host generation
It is soft for realizing driving to manage layer model and host interface layer model, the master agent layer model and the host interface layer model
Part is interacted with the software and hardware of GPU hardware model.
Optionally, the master agent layer model, for realizing the drive software based on C language and based on SystemC's
The data interaction of the host interface model, specifically includes: order issues module (1), module for reading and writing (2), interruption processing module
(3) and main memory read module (4), wherein the order issues module (1) for sending OpenGL CM (Command
Package) order packet, order packet carry OpenGL function code and relevant parameter information;The module for reading and writing (2) is for reading and writing
Architecture register, architecture register is used to characterize the working condition and mode of GPU.
Optionally, the host interface layer model, for realizing the GPU host interface based on SystemC transaction-level interface
The interaction of model and GPUPCIe hardware model, the host interface layer model include: host memory read-write transaction-level module (5),
Architecture register reads and writes transaction-level module (6), interrupt status read-write transaction-level module (7), the configuration of direct memory reading DMA
Transaction-level module (8), wherein host memory read-write transaction-level module (5) provides host memory read-write transaction-level interface,
Architecture register read-write transaction-level module (6) provides architecture register read-write transaction-level interface, it is described in
Disconnected state read-write transaction-level module (7) provides interrupt status read-write transaction-level interface, and the direct memory reading DMA configures thing
Business grade module (8) provides DMA configuration transaction grade interface, calls for master agent layer model and GPU hardware model.
Optionally, it specifically includes:
Order issues module (1) and passes through the DMA configuration transaction grade interface and direct memory reading DMA configuration transaction grade mould
CM order packet is handed down to host interface layer model by block (8) connection in a manner of dma descriptor, and dma descriptor includes DMA mesh
Address, DMA source address, DMA path and DMA length.
Module for reading and writing (2) reads and writes transaction-level interface by the architecture register and architecture register reads and writes thing
Business grade module (6) connection for GP configuring U operating mode and obtains GPU working condition.
Interruption processing module (3) reads and writes transaction-level interface by the interrupt status and interrupt status reads and writes transaction-level module
(7) it connects, for obtaining and removing Host Interrupt.
Main memory read module (4) reads and writes transaction-level interface by the host memory and host memory reads and writes transaction-level module
(5) it connects, for reading and writing main memory data.
A kind of GPU software and hardware interaction TLM system based on SystemC provided by the invention, should the GPU based on SystemC
Software and hardware interaction TLM model includes master agent layer model and host interface layer model, and wherein master agent layer model provides
The interaction approach of software-driven and host interface layer model, it is hard with GPU that host interface layer model provides master agent layer model
The interaction approach of part TLM model, interactive content include issuing for OpenGL CM (Command Packet) order packet, system
The read-write of structure register, interrupt processing, image and data texturing load obtain.It is able to solve software-driven and GPU hardware
The interaction problems of model get through software and hardware interaction channel for that can shift to an earlier date the progress functional verification on hardware micro-structure TLM model.
Finally it should be noted that the above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations;Although
Present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that;It still may be used
To modify to the technical solution that foregoing embodiments are recorded or equivalent replacement of some of the technical features;And
These are modified or replaceed, the spirit and model of technical solution of various embodiments of the present invention that it does not separate the essence of the corresponding technical solution
It encloses.
Claims (4)
1. a kind of graphics processor GPU software and hardware interaction TLM system based on SystemC, which is characterized in that the system comprises
Master agent layer model and host interface layer model, the master agent layer model and the host interface layer model for realizing
Drive software is interacted with the software and hardware of GPU hardware model.
2. software and hardware interaction TLM system according to claim 1, which is characterized in that the master agent layer model is used for
The data interaction for realizing the drive software based on C language and the host interface model based on SystemC, specifically includes: life
Order issues module (1), module for reading and writing (2), interruption processing module (3) and main memory read module (4), wherein the order issues
Module (1) carries OpenGL function code and phase for sending OpenGL CM (Command Package) order packet, order packet
Close parameter information;The module for reading and writing (2) is for reading and writing architecture register, and architecture register is for characterizing GPU
Working condition and mode.
3. software and hardware interaction TLM system according to claim 2, which is characterized in that the host interface layer model is used for
Realize the interaction of GPU host interface model and GPU PCIe hardware model based on SystemC transaction-level interface, the host connects
Mouth layer model includes: host memory read-write transaction-level module (5), architecture register read-write transaction-level module (6), interrupts shape
State reads and writes transaction-level module (7), direct memory reading DMA configuration transaction grade module (8), wherein the host memory reads and writes thing
Business grade module (5) provides host memory read-write transaction-level interface, and architecture register read-write transaction-level module (6) mentions
Architecture register has been supplied to read and write transaction-level interface, interrupt status read-write transaction-level module (7) provides interrupt status
Transaction-level interface is read and write, the direct memory reading DMA configuration transaction grade module (8) provides DMA configuration transaction grade interface, supplies
Master agent layer model and GPU hardware model call.
4. software and hardware interaction TLM system according to claim 3, it is characterised in that:
Order issues module (1) and passes through the DMA configuration transaction grade interface and direct memory reading DMA configuration transaction grade module
(8) it connects, CM order packet is handed down to host interface layer model in a manner of dma descriptor, dma descriptor includes DMA purpose
Address, DMA source address, DMA path and DMA length.
Module for reading and writing (2) reads and writes transaction-level interface by the architecture register and architecture register reads and writes transaction-level
Module (6) connection for GP configuring U operating mode and obtains GPU working condition.
Interruption processing module (3) reads and writes transaction-level interface and interrupt status read-write transaction-level module (7) by the interrupt status
Connection, for obtaining and removing Host Interrupt.
Main memory read module (4) reads and writes transaction-level interface and host memory read-write transaction-level module (5) by the host memory
Connection, for reading and writing main memory data.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811516374.3A CN109683877B (en) | 2018-12-11 | 2018-12-11 | SystemC-based GPU software and hardware interaction TLM system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811516374.3A CN109683877B (en) | 2018-12-11 | 2018-12-11 | SystemC-based GPU software and hardware interaction TLM system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109683877A true CN109683877A (en) | 2019-04-26 |
CN109683877B CN109683877B (en) | 2022-03-15 |
Family
ID=66186543
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811516374.3A Active CN109683877B (en) | 2018-12-11 | 2018-12-11 | SystemC-based GPU software and hardware interaction TLM system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109683877B (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110941898A (en) * | 2019-11-18 | 2020-03-31 | 中国航空工业集团公司西安航空计算技术研究所 | Design model and design method of graphic processor |
CN110942478A (en) * | 2019-11-18 | 2020-03-31 | 中国航空工业集团公司西安航空计算技术研究所 | Texture integrity calculation method and calculation unit based on SystemC |
CN110955572A (en) * | 2019-11-21 | 2020-04-03 | 中国航空工业集团公司西安航空计算技术研究所 | GPU (graphics processing Unit) chip-oriented host interface unit performance simulation method and platform |
CN111008515A (en) * | 2019-11-18 | 2020-04-14 | 中国航空工业集团公司西安航空计算技术研究所 | TLM microstructure for GPU hardware sub-texture replacement storage algorithm |
CN111047504A (en) * | 2019-11-21 | 2020-04-21 | 中国航空工业集团公司西安航空计算技术研究所 | TLM microstructure for GPU sub-image processing based on SystemC |
CN112579254A (en) * | 2020-12-08 | 2021-03-30 | 海光信息技术股份有限公司 | Simulation method and device of graphic processor, electronic equipment and storage medium |
CN115719047A (en) * | 2022-11-14 | 2023-02-28 | 沐曦集成电路(上海)有限公司 | Joint simulation system based on waveform GPU |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1828551A (en) * | 2006-04-13 | 2006-09-06 | 电子科技大学 | High speed synergy emulation method for transaction-grade software and hardware |
US20090132991A1 (en) * | 2007-11-16 | 2009-05-21 | Nec Laboratories America, Inc | Partial order reduction for scalable testing in system level design |
CN101630343A (en) * | 2009-08-18 | 2010-01-20 | 中兴通讯股份有限公司 | Simulation method and simulation system |
CN106888271A (en) * | 2017-03-30 | 2017-06-23 | 北京萤芯科技有限公司 | A kind of software and hardware interactive interface applied in BLE5.0 protocol systems and system |
-
2018
- 2018-12-11 CN CN201811516374.3A patent/CN109683877B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1828551A (en) * | 2006-04-13 | 2006-09-06 | 电子科技大学 | High speed synergy emulation method for transaction-grade software and hardware |
US20090132991A1 (en) * | 2007-11-16 | 2009-05-21 | Nec Laboratories America, Inc | Partial order reduction for scalable testing in system level design |
CN101630343A (en) * | 2009-08-18 | 2010-01-20 | 中兴通讯股份有限公司 | Simulation method and simulation system |
CN106888271A (en) * | 2017-03-30 | 2017-06-23 | 北京萤芯科技有限公司 | A kind of software and hardware interactive interface applied in BLE5.0 protocol systems and system |
Non-Patent Citations (1)
Title |
---|
卢艳君: "基于SystemC和SystemVerilog的联合仿真平台设计", 《科学技术创新》 * |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110941898A (en) * | 2019-11-18 | 2020-03-31 | 中国航空工业集团公司西安航空计算技术研究所 | Design model and design method of graphic processor |
CN110942478A (en) * | 2019-11-18 | 2020-03-31 | 中国航空工业集团公司西安航空计算技术研究所 | Texture integrity calculation method and calculation unit based on SystemC |
CN111008515A (en) * | 2019-11-18 | 2020-04-14 | 中国航空工业集团公司西安航空计算技术研究所 | TLM microstructure for GPU hardware sub-texture replacement storage algorithm |
CN111008515B (en) * | 2019-11-18 | 2023-06-09 | 中国航空工业集团公司西安航空计算技术研究所 | TLM microstructure for GPU hardware sub-texture replacement storage algorithm |
CN110942478B (en) * | 2019-11-18 | 2023-09-19 | 中国航空工业集团公司西安航空计算技术研究所 | Texture integrity calculation method and calculation unit based on SystemC |
CN110941898B (en) * | 2019-11-18 | 2023-10-17 | 中国航空工业集团公司西安航空计算技术研究所 | Design model and design method of graphic processor |
CN110955572A (en) * | 2019-11-21 | 2020-04-03 | 中国航空工业集团公司西安航空计算技术研究所 | GPU (graphics processing Unit) chip-oriented host interface unit performance simulation method and platform |
CN111047504A (en) * | 2019-11-21 | 2020-04-21 | 中国航空工业集团公司西安航空计算技术研究所 | TLM microstructure for GPU sub-image processing based on SystemC |
CN111047504B (en) * | 2019-11-21 | 2023-06-13 | 中国航空工业集团公司西安航空计算技术研究所 | TLM microstructure for GPU sub-image processing based on SystemC |
CN112579254A (en) * | 2020-12-08 | 2021-03-30 | 海光信息技术股份有限公司 | Simulation method and device of graphic processor, electronic equipment and storage medium |
CN112579254B (en) * | 2020-12-08 | 2022-03-29 | 成都海光微电子技术有限公司 | Simulation method and device of graphic processor, electronic equipment and storage medium |
CN115719047A (en) * | 2022-11-14 | 2023-02-28 | 沐曦集成电路(上海)有限公司 | Joint simulation system based on waveform GPU |
Also Published As
Publication number | Publication date |
---|---|
CN109683877B (en) | 2022-03-15 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109683877A (en) | A kind of GPU software and hardware interaction TLM system based on SystemC | |
CN102339343B (en) | Method for generating system model and method transmitting data in same | |
CN102866971B (en) | Device, the system and method for transmission data | |
TWI278755B (en) | An apparatus and method for high performance volatile disk drive memory access using an integrated DMA engine | |
CN103034454B (en) | Flexible flash command | |
CN101231589B (en) | System and method for developing embedded software in-situ | |
CN105793829A (en) | Integrated component interconnect | |
CN101344870B (en) | FIFO control module with strong reusability and method for managing internal memory | |
CN110109626B (en) | NVMe SSD command processing method based on FPGA | |
EP3234789A1 (en) | Method, apparatus and system for integrating devices in a root complex | |
CN106325758B (en) | A kind of queue storage space management method and device | |
CN103198048B (en) | Processor communication | |
CN100464319C (en) | Device and method for implementing communication between processes | |
CN105335548B (en) | A kind of MCU emulation mode for ICE | |
CN109871182A (en) | Storage device and its operating method and the method for issuing order | |
CN108694135A (en) | For the method and apparatus that persistently cache stores data in page cache | |
CN109840878B (en) | GPU parameter management method based on SystemC | |
CN102207904B (en) | Device and method for being emulated to reconfigurable processor | |
US20080046228A1 (en) | Emulation System | |
US10303833B1 (en) | Parallelizing timing-based operations for circuit designs | |
WO2024027140A1 (en) | Data processing method and apparatus, and device, system and readable storage medium | |
US9690736B2 (en) | Managing state transitions of a data connector using a finite state machine | |
CN101876952B (en) | System and method for realizing software-hardware interaction between transmission layer and application layer of host | |
CN109741235A (en) | A kind of host interface means and modeling method based on GPU chip | |
CN105893036A (en) | Compatible accelerator extension method for embedded system |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |