-
公开(公告)号:WO2023064728A1
公开(公告)日:2023-04-20
申请号:PCT/US2022/077848
申请日:2022-10-10
Applicant: ADVANCED MICRO DEVICES, INC. , ATI TECHNOLOGIES ULC
Inventor: DONG, HaiKun , CHRISTIDIS, Kostantinos Danny , WANG, Ling-Ling , WU, MinHua , CONG, Gaojian , WANG, Rui
IPC: G06F11/00
Abstract: Systems, apparatuses, and methods for implementing duplicated registers for access by initiators across multiple semiconductor dies are disclosed. A system includes multiple initiators on multiple semiconductor dies of a chiplet processor. One of the semiconductor dies is the master die, and this master die has copies of registers which can be accessed by the multiple initiators on the multiple semiconductor dies. When a given initiator on a given secondary die generates a register access, the register access is routed to the master die and a particular duplicate copy of the register maintained for the given secondary die. From the point of view of software, the multiple semiconductor dies appear as a single die, and the multiple initiators appear as a single initiator. Multiple types of registers can be maintained by the master die, with a flush register being one of the register types.
-
公开(公告)号:WO2023014512A1
公开(公告)日:2023-02-09
申请号:PCT/US2022/037897
申请日:2022-07-21
Applicant: ADVANCED MICRO DEVICES, INC. , ATI TECHNOLOGIES ULC
Inventor: VENKATRAMANI, Rajagopalan , GADDI, Renato Dimatula , MARTINEZ, Liane , SANTOS, Warren Alexander , SURELL, Dennis Glenn Lozanta
IPC: G06F30/392 , G06F30/394 , G06F30/398 , G06F115/12 , G06F113/18
Abstract: A system and method for automatically generating placement of vias within redistribution layers of a semiconductor package are described. A user defines attributes to use for automatic via generation in redistribution layers of a semiconductor package (702). The circuitry of a processor of a computing device used by the user executes instructions of an automatic redistribution layer (RDL) via generator (704). The automatic via generator uses the attributes, data indicative of the RDL netlist of signal routes within the RDL, and RDL mask layout data representing the signal masks of the metal layers within the RDL. The processor generates placement of vias for in the RDL based on the attributes and an identification of overlapping regions between metal layers.
-
公开(公告)号:WO2022271800A1
公开(公告)日:2022-12-29
申请号:PCT/US2022/034486
申请日:2022-06-22
Applicant: ADVANCED MICRO DEVICES, INC. , ATI TECHNOLOGIES ULC
Inventor: SMITH, Wade K. , ASARO, Anthony
IPC: G06F12/1027 , G06F12/0891 , G06F2212/68
Abstract: A translation lookaside buffer (TLB) [110] receives mapping invalidation requests [105, 106] from one or more sources, such as one or more processing units [102, 104] of a processing system. The TLB includes one or more invalidation processing pipelines [112], wherein each processing pipeline includes multiple processing states arranged in a pipeline, so that a given stage executes its processing operations concurrent with other stages of the pipeline executing their processing operations.
-
公开(公告)号:WO2022137046A1
公开(公告)日:2022-06-30
申请号:PCT/IB2021/061867
申请日:2021-12-16
Applicant: ATI TECHNOLOGIES ULC
Inventor: RAHMAN, Arshad , PANCHACHARAMOORTHY, Rajeevan , IVANOVIC, Boris
IPC: G09G5/12
Abstract: Systems, apparatuses, and methods for performing asynchronous memory clock changes on multiple displays are disclosed. From time to time, a memory clock frequency change is desired for a memory subsystem storing frame buffer(s) used to drive pixels to multiple displays. For example, when the real-time memory bandwidth demand differs from the memory bandwidth available with the existing memory clock frequency, a control unit tracks the vertical blanking interval (VBI) timing of a first display. Also, the control unit causes a second display to enter into panel self-refresh (PSR) mode. Once the PSR mode of the second display overlaps with a VBI of the first display, a memory clock frequency change, including memory training, is initiated. After the memory clock frequency change, the displays are driven by the frame buffer(s) in the memory subsystem at an updated frequency.
-
公开(公告)号:WO2022123450A1
公开(公告)日:2022-06-16
申请号:PCT/IB2021/061430
申请日:2021-12-07
Applicant: ATI TECHNOLOGIES ULC [CA]/[CA]
Inventor: JIANG, Yinan , ZHANG, Min
Abstract: Virtual functions are implemented using a plurality of resources [220] and physical function circuitry [210] that executes a virtual function [215] using information stored in the plurality of resources. A processing unit [105] executes a host driver [205] that selectively enables access to the plurality of resources by the virtual function based on an operational state of the processing unit. In some cases, a state machine [300] that determines a state of the virtual function and the host driver that enables access to the plurality of resources by the virtual function based on the state of the virtual function executing on the processing unit. The subsets of the plurality of resources are used to implement a frame buffer [222], one or more context registers [225, 226], a doorbell [231], and one or more mailbox registers [235, 236].
-
公开(公告)号:WO2022103879A1
公开(公告)日:2022-05-19
申请号:PCT/US2021/058840
申请日:2021-11-10
Applicant: ADVANCED MICRO DEVICES, INC. , ATI TECHNOLOGIES ULC
Inventor: HAUKE, Jonathan David , CLARK, Adam
IPC: G06F1/3206 , G06F1/324 , G06F1/3296
Abstract: Systems, apparatuses, and methods for performing a software override of a power estimation mechanism are disclosed. A computing system includes a plurality of tuned parameters for generating an estimate of power consumption. The tuned parameters are generated based on post-silicon characterization of the system. After deployment, the system executes a plurality of different applications. When launching a particular application, the system loads a corresponding set of override parameters which are used to replace the plurality of tuned parameters. The system generates an estimate of power consumption using the set of override parameters rather than the previously determined tuned parameters. Then while executing the particular application, the system makes adjustments to power and frequency values for the various system components based on the estimate of power consumption.
-
公开(公告)号:WO2021242576A1
公开(公告)日:2021-12-02
申请号:PCT/US2021/033033
申请日:2021-05-18
Applicant: ADVANCED MICRO DEVICES, INC. , ATI TECHNOLOGIES ULC
Inventor: TOVEY, Steven J. , CHEN, Zhuo , OLDCORN, David Ronald
IPC: G06F9/48 , G06F8/30 , G06F9/3005 , G06F9/30076 , G06F9/3836 , G06F9/4881
Abstract: Techniques for generating a task graph for workload scheduling based on a task graph specification program are provided. The techniques include executing control flow instructions of the task graph specification program to traverse the task graph specification program; generating pass nodes of the task graph based on pass instructions of the task graph specification program; generating resource nodes and directed edges based on resource declarations of the task graph specification program; and outputting the task graph specification program to a command scheduler for scheduling.
-
8.
公开(公告)号:WO2021198810A1
公开(公告)日:2021-10-07
申请号:PCT/IB2021/051882
申请日:2021-03-05
Applicant: ATI TECHNOLOGIES ULC
Inventor: HARIRI, Arash , SAEEDI, Mehdi , IVANOVIC, Boris , SINES, Gabor
IPC: G06N3/063 , G06N3/04 , G06N3/08 , G06K9/6215 , G06N20/10 , H03M7/3048
Abstract: A processing device for executing a machine learning neural network operation includes memory and a processor. The processor is configured to receive input data at a layer of the machine learning neural network operation, receive a plurality of sorted filters to be applied to the input data, apply the plurality of sorted filters to the input data to produce a plurality of different feature maps, compress the plurality of different feature maps according to a similarity of the feature maps relative to each other and store the plurality of different feature maps in the memory.
-
公开(公告)号:WO2021061967A1
公开(公告)日:2021-04-01
申请号:PCT/US2020/052471
申请日:2020-09-24
Applicant: ADVANCED MICRO DEVICES, INC. , ATI TECHNOLOGIES ULC
Inventor: RAO, Murali , IP, Clarence , SCANLON, Joseph , DOCTOR, Mihir S. , STEWART, Norman , KRISHNAN, Guhan
Abstract: A processing system isolates at a physically or logically separate memory region of a processing unit boot code that is received from an external boot source for programming a boot memory of the processing unit until after the boot code is validated to protect against buffer overruns that could compromise the processing system. The processing unit includes a secure buffer region of memory that is physically or logically isolated from the remainder of the processing unit for receiving boot code from an external boot source such as a personal computer (PC) such that any buffer overruns at the secure buffer simply overwrite data stored at the secure buffer, and do not affect data or instructions that are executing at the processing unit.
-
公开(公告)号:WO2020240315A1
公开(公告)日:2020-12-03
申请号:PCT/IB2020/054448
申请日:2020-05-11
Applicant: ATI TECHNOLOGIES ULC
Inventor: KWONG, Tung Chuen , CHAN, Benjamin Koon Pan , PORPINO SOBREIRA MARQUES, David , IP, Clarence , YU, Hung Wilson
Abstract: Systems, apparatuses, and methods for implementing a safety monitor framework for a safety-critical inference application are disclosed. A system includes a safety-critical inference application, a safety monitor, and an inference accelerator engine. The safety monitor receives an input image, test data, and a neural network specification from the safety-critical inference application. The safety monitor generates a modified image by adding additional objects outside of the input image. The safety monitor provides the modified image and neural network specification to the inference accelerator engine which processes the modified image and provides outputs to the safety monitor. The safety monitor determines the likelihood of erroneous processing of the original input image by comparing the outputs for the additional objects with a known good result. The safety monitor complements the overall fault coverage of the inference accelerator engine and covers faults only observable at the network level.
-
-
-
-
-
-
-
-
-