专利快速检索-快速检索全球专利，免费商用专利数据库-IPRDB

11. 发明授权

US08572588B2 Thread-local memory reference promotion for translating CUDA code for execution by a general purpose processor 有权
标题翻译：用于翻译CUDA代码以供通用处理器执行的线程本地内存引用升级
公开(公告)号：US08572588B2
公开(公告)日：2013-10-29
申请号：US12415118
申请日：2009-03-31
申请人： Vinod Grover , Bastiaan Joannes Matheus Aarts , Michael Murphy
发明人： Vinod Grover , Bastiaan Joannes Matheus Aarts , Michael Murphy
IPC分类号： G06F9/45 , G06F9/30
CPC分类号： G06F9/45537 , G06F8/45 , G06F8/456 , G06F9/4843 , G06F11/261 , G06F12/0253
摘要： One embodiment of the present invention sets forth a technique for translating application programs written using a parallel programming model for execution on multi-core graphics processing unit (GPU) for execution by general purpose central processing unit (CPU). Portions of the application program that rely on specific features of the multi-core GPU are converted by a translator for execution by a general purpose CPU. The application program is partitioned into regions of synchronization independent instructions. The instructions are classified as convergent or divergent and divergent memory references that are shared between regions are replicated. Thread loops are inserted to ensure correct sharing of memory between various threads during execution by the general purpose CPU.
摘要翻译：本发明的一个实施例提出了一种用于翻译使用并行编程模型编写的应用程序的技术，用于在多核图形处理单元（GPU）上执行以由通用中央处理单元（CPU）执行。依赖于多核GPU的特定功能的应用程序的部分由翻译器转换，以供通用CPU执行。应用程序被划分为独立于同步的指令的区域。指令被分类为在区域之间共享的收敛或发散和不同的存储器引用。插入线程循环以确保在通用CPU执行期间在不同线程之间正确共享内存。

12. 发明申请

US20100218195A1 Software filtering in a transactional memory system 有权
标题翻译：事务存储系统中的软件过滤
公开(公告)号：US20100218195A1
公开(公告)日：2010-08-26
申请号：US12653471
申请日：2009-12-15
申请人： Ali-Reza Adl-Tabatabai , David Callahan , Jan Gray , Vinod Grover , Bratin Saha , Gad Sheaffer
发明人： Ali-Reza Adl-Tabatabai , David Callahan , Jan Gray , Vinod Grover , Bratin Saha , Gad Sheaffer
IPC分类号： G06F9/46
CPC分类号： G06F9/522 , G06F8/458 , G06F9/467
摘要： A method and apparatus for utilizing hardware mechanisms of a transactional memory system is herein described. Various embodiments relate to software-based filtering of operations from read and write barriers and read isolation barriers during transactional execution. Other embodiments relate to software-implemented read barrier processing to accelerate strong atomicity. Other embodiments are also described and claimed.
摘要翻译：这里描述了一种利用事务存储器系统的硬件机制的方法和装置。各种实施例涉及在事务执行期间从读写障碍和读隔离屏障的基于软件的操作过滤。其他实施例涉及软件实现的读屏障处理以加速强原子性。还描述和要求保护其他实施例。

13. 发明申请

US20100191930A1 TRANSACTIONAL MEMORY COMPATIBILITY MANAGEMENT 有权
标题翻译：交易记忆兼容性管理
公开(公告)号：US20100191930A1
公开(公告)日：2010-07-29
申请号：US12359492
申请日：2009-01-26
申请人： Dana Groff , Yosseff Levanoni , Stephen Toub , Michael McKenzie Magruder , Weirong Zhu , Timothy Lawrence Harris , Christopher William Dern , John Joseph Duffy , David Detlefs , Martin Abadi , Sukhdeep Singh Sodhi , Lingli Zhang , Alexander Dadiomov , Vinod Grover
发明人： Dana Groff , Yosseff Levanoni , Stephen Toub , Michael McKenzie Magruder , Weirong Zhu , Timothy Lawrence Harris , Christopher William Dern , John Joseph Duffy , David Detlefs , Martin Abadi , Sukhdeep Singh Sodhi , Lingli Zhang , Alexander Dadiomov , Vinod Grover
IPC分类号： G06F12/00 , G06F9/44
CPC分类号： G06F8/44
摘要： Transactional memory compatibility type attributes are associated with intermediate language code to specify, for example, that intermediate language code must be run within a transaction, or must not be run within a transaction, or may be run within a transaction. Attributes are automatically produced while generating intermediate language code from annotated source code. Default rules also generate attributes. Tools use attributes to statically or dynamically check for incompatibility between intermediate language code and a transactional memory implementation.
摘要翻译：事务性内存兼容类型属性与中间语言代码相关联，以指定例如中间语言代码必须在事务中运行，或者不能在事务中运行，或者可以在事务中运行。自动生成属性，同时从注释的源代码生成中间语言代码。默认规则也生成属性。工具使用属性来静态或动态地检查中间语言代码和事务内存实现之间的不兼容性。

14. 发明授权

US09678775B1 Allocating memory for local variables of a multi-threaded program for execution in a single-threaded environment 有权
公开(公告)号：US09678775B1
公开(公告)日：2017-06-13
申请号：US12393763
申请日：2009-02-26
申请人： Vinod Grover , John A. Stratton
发明人： Vinod Grover , John A. Stratton
IPC分类号： G06F9/455 , G06F9/48 , G06F11/26 , G06F12/02
CPC分类号： G06F9/45537 , G06F8/45 , G06F8/456 , G06F9/4843 , G06F11/261 , G06F12/0253
摘要： Computer code written to execute on a multi-threaded computing environment is transformed into code designed to execute on a single-threaded computing environment and simulate concurrent executing threads. Optimization techniques during the transformation process are utilized to identify local variables for scalar expansion. A first set of local variables is defined that includes those local variables in the code identified as “Downward exposed Defined” (DD). A second set of local variables is defined that includes those local variables in the code identified as “Upward exposed Use” (UU). The intersection of the first set and the second set identifies local variables for scalar expansion.

15. 发明授权

US09367306B2 Method for transforming a multithreaded program for general execution 有权
标题翻译：用于转换用于一般执行的多线程程序的方法
公开(公告)号：US09367306B2
公开(公告)日：2016-06-14
申请号：US13076258
申请日：2011-03-30
申请人： Jaydeep Marathe , Vinod Grover
发明人： Jaydeep Marathe , Vinod Grover
IPC分类号： G06F9/46 , G06F9/44 , G06F9/52
CPC分类号： G06F8/72 , G06F9/522
摘要： A technique is disclosed for executing a program designed for multi-threaded operation on a general purpose processor. Original source code for the program is transformed from a multi-threaded structure into a computationally equivalent single-threaded structure. A transform operation modifies the original source code to insert code constructs for serial thread execution. The transform operation also replaces synchronization barrier constructs in the original source code with synchronization barrier code that is configured to facilitate serialization. The transformed source code may then be conventionally compiled and advantageously executed on the general purpose processor.
摘要翻译：公开了一种用于在通用处理器上执行针对多线程操作设计的程序的技术。程序的原始源代码从多线程结构转换为计算等效的单线程结构。转换操作修改原始源代码以插入用于串行线程执行的代码结构。变换操作还用原始源代码中的同步屏障代码替代配置为便于序列化的同步屏障代码。然后可以在通用处理器上常规地编译和有利地执行变换的源代码。

16. 发明授权

US08776030B2 Partitioning CUDA code for execution by a general purpose processor 有权
标题翻译：将CUDA代码分区以供通用处理器执行
公开(公告)号：US08776030B2
公开(公告)日：2014-07-08
申请号：US12415075
申请日：2009-03-31
申请人： Vinod Grover , Bastiaan Joannes Matheus Aarts , Michael Murphy
发明人： Vinod Grover , Bastiaan Joannes Matheus Aarts , Michael Murphy
IPC分类号： G06F9/44
CPC分类号： G06F8/456
摘要： One embodiment of the present invention sets forth a technique for translating application programs written using a parallel programming model for execution on multi-core graphics processing unit (GPU) for execution by general purpose central processing unit (CPU). Portions of the application program that rely on specific features of the multi-core GPU are converted by a translator for execution by a general purpose CPU. The application program is partitioned into regions of synchronization independent instructions. The instructions are classified as convergent or divergent and divergent memory references that are shared between regions are replicated. Thread loops are inserted to ensure correct sharing of memory between various threads during execution by the general purpose CPU.
摘要翻译：本发明的一个实施例提出了一种用于翻译使用并行编程模型编写的应用程序的技术，用于在多核图形处理单元（GPU）上执行以由通用中央处理单元（CPU）执行。依赖于多核GPU的特定功能的应用程序的部分由翻译器转换，以供通用CPU执行。应用程序被划分为独立于同步的指令的区域。指令被分类为在区域之间共享的收敛或发散和不同的存储器引用。插入线程循环以确保在通用CPU执行期间在不同线程之间正确共享内存。

17. 发明授权

US08719514B2 Software filtering in a transactional memory system 有权
标题翻译：事务存储系统中的软件过滤
公开(公告)号：US08719514B2
公开(公告)日：2014-05-06
申请号：US12653471
申请日：2009-12-15
申请人： Ali-Reza Adl-Tabatabai , David Callahan , Jan Gray , Vinod Grover , Bratin Saha , Gad Sheaffer
发明人： Ali-Reza Adl-Tabatabai , David Callahan , Jan Gray , Vinod Grover , Bratin Saha , Gad Sheaffer
IPC分类号： G06F13/00
CPC分类号： G06F9/522 , G06F8/458 , G06F9/467
摘要： A method and apparatus for utilizing hardware mechanisms of a transactional memory system is herein described. Various embodiments relate to software-based filtering of operations from read and write barriers and read isolation barriers during transactional execution. Other embodiments relate to software-implemented read barrier processing to accelerate strong atomicity. Other embodiments are also described and claimed.
摘要翻译：这里描述了一种利用事务存储器系统的硬件机制的方法和装置。各种实施例涉及在事务执行期间从读写障碍和读隔离屏障的基于软件的操作过滤。其他实施例涉及软件实现的读屏障处理以加速强原子性。还描述和要求保护其他实施例。

18. 发明授权

US6148302A Method, apparatus, system and computer program product for initializing a data structure at its first active use 失效
标题翻译：用于在首次主动使用时初始化数据结构的方法，装置，系统和计算机程序产品
公开(公告)号：US6148302A
公开(公告)日：2000-11-14
申请号：US31229
申请日：1998-02-26
申请人： Boris Beylin , Vinod Grover
发明人： Boris Beylin , Vinod Grover
IPC分类号： G06F9/45 , G06F9/44 , G06F9/445 , G06F7/00 , G06F12/00
CPC分类号： G06F9/4428 , G06F9/445 , Y10S707/99942 , Y10S707/99943 , Y10S707/99944 , Y10S707/99945
摘要： Apparatus, methods, systems and computer program products are disclosed that provide an efficient mechanism for invoking a programmed operation at the first active use of the OOP object or data structure. The programmed operation can be used to initialize an object-oriented programming (OOP) object or data structure. The first active use of the data structure or OOP object is detected because the initial access mechanism is constrained to cause a misaligned memory access fault (trap) by attempting a non-byte access-mode memory access to an odd byte address. As the fault is processed, the access mechanism is converted so that the initial and subsequent non-byte access-mode memory accesses will succeed. In addition, the OOP object or data structure is initialized. Then the initial access attempt is repeated on the just initialized OOP object or data structure using the converted access mechanism. The use of the invention improves the performance of computers by reducing the overhead involved with particular computational operations.
摘要翻译：公开了装置，方法，系统和计算机程序产品，其提供了在第一主动使用OOP对象或数据结构时调用编程操作的有效机构。编程操作可用于初始化面向对象编程（OOP）对象或数据结构。检测到数据结构或OOP对象的第一个主动使用，因为初始访问机制被限制为通过尝试非字节访问模式存储器访问奇数字节地址而导致未对齐的存储器访问故障（陷阱）。当故障被处理时，转换访问机制，使得初始和随后的非字节访问模式存储器访问将成功。此外，OOP对象或数据结构已初始化。然后，使用转换的访问机制，在刚刚初始化的OOP对象或数据结构上重复初始访问尝试。本发明的使用通过减少特定计算操作涉及的开销来改善计算机的性能。

19. 发明授权

US09448779B2 Execution of retargetted graphics processor accelerated code by a general purpose processor 有权
标题翻译：重定向图形处理器的执行由通用处理器加速代码
公开(公告)号：US09448779B2
公开(公告)日：2016-09-20
申请号：US12408559
申请日：2009-03-20
申请人： Vinod Grover , Bastiaan Joannes Matheus Aarts , Michael Murphy , Jayant B. Kolhe , John Bryan Pormann , Douglas Saylor
发明人： Vinod Grover , Bastiaan Joannes Matheus Aarts , Michael Murphy , Jayant B. Kolhe , John Bryan Pormann , Douglas Saylor
IPC分类号： G06F9/45
CPC分类号： G06F9/45537 , G06F8/45 , G06F8/456 , G06F9/4843 , G06F11/261 , G06F12/0253
摘要： One embodiment of the present invention sets forth a technique for translating application programs written using a parallel programming model for execution on multi-core graphics processing unit (GPU) for execution by general purpose central processing unit (CPU). Portions of the application program that rely on specific features of the multi-core GPU are converted by a translator for execution by a general purpose CPU. The application program is partitioned into regions of synchronization independent instructions. The instructions are classified as convergent or divergent and divergent memory references that are shared between regions are replicated. Thread loops are inserted to ensure correct sharing of memory between various threads during execution by the general purpose CPU.
摘要翻译：本发明的一个实施例提出了一种用于翻译使用并行编程模型编写的应用程序的技术，用于在多核图形处理单元（GPU）上执行以由通用中央处理单元（CPU）执行。依赖于多核GPU的特定功能的应用程序的部分由翻译器转换，以供通用CPU执行。应用程序被划分为独立于同步的指令的区域。指令被分类为在区域之间共享的收敛或发散和不同的存储器引用。插入线程循环以确保在通用CPU执行期间在不同线程之间正确共享内存。

20. 发明授权

US09361079B2 Method for compiling a parallel thread execution program for general execution 有权
标题翻译：用于编译并行线程执行程序以进行一般执行的方法
公开(公告)号：US09361079B2
公开(公告)日：2016-06-07
申请号：US13361408
申请日：2012-01-30
申请人： Vinod Grover , Andrew Kerr , Sean Lee
发明人： Vinod Grover , Andrew Kerr , Sean Lee
IPC分类号： G06F9/44 , G06F9/45
CPC分类号： G06F8/53
摘要： A technique is disclosed for executing a compiled parallel application on a general purpose processor. The compiled parallel application comprises parallel thread execution code, which includes single-instruction multiple-data (SIMD) constructs, as well as references to intrinsic functions conventionally available in a graphics processing unit. The parallel thread execution code is transformed into an intermediate representation, which includes vector instruction constructs. The SIMD constructs are mapped to vector instructions available within the intermediate representation. Intrinsic functions are mapped to corresponding emulated runtime implementations. The technique advantageously enables parallel applications compiled for execution on a graphics processing unit to be executed on a general purpose central processing unit configured to support vector instructions.
摘要翻译：公开了一种用于在通用处理器上执行编译并行应用的技术。编译并行应用程序包括并行线程执行代码，其包括单指令多数据（SIMD）结构，以及对图形处理单元中常规可用的内在函数的引用。并行线程执行代码被转换为包含向量指令结构的中间表示。 SIMD构造被映射到中间表示中可用的向量指令。内在函数映射到相应的仿真运行时实现。该技术有利地使编译为在图形处理单元上执行的并行应用能够在被配置为支持向量指令的通用中央处理单元上执行。

你已经成功收藏专利！

检索式保存成功!

IPRDB

热门服务

关于我们

友情链接

联系方式