Patents by Inventor Shubhendu S. Mukherjee
Shubhendu S. Mukherjee has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).
-
Patent number: 11868262Abstract: A memory request, including an address, is accessed. The memory request also specifies a type of an operation (e.g., a read or write) associated with an instance (e.g., a block) of data. A group of caches is selected using a bit or bits in the address. A first hash of the address is performed to select a cache in the group. A second hash of the address is performed to select a set of cache lines in the cache. Unless the operation results in a cache miss, the memory request is processed at the selected cache. When there is a cache miss, a third hash of the address is performed to select a memory controller, and a fourth hash of the address is performed to select a bank group and a bank in memory.Type: GrantFiled: February 9, 2023Date of Patent: January 9, 2024Assignee: Marvell Asia Pte, Ltd.Inventors: Richard E. Kessler, David Asher, Shubhendu S Mukherjee, Wilson P. Snyder, II, David Carlson, Jason Zebchuk, Isam Akkawi
-
Patent number: 11748109Abstract: A system and corresponding method enforce strong load ordering in a processor. The system comprises an ordering ring that stores entries corresponding to in-flight memory instructions associated with a program order, scanning logic, and recovery logic. The scanning logic scans the ordering ring in response to execution or completion of a given load instruction of the in-flight memory instructions and detects an ordering violation in an event at least one entry of the entries indicates that a younger load instruction has completed and is associated with an invalidated cache line. In response to the ordering violation, the recovery logic allows the given load instruction to complete, flushes the younger load instruction, and restarts execution of the processor after the given load instruction in the program order, causing data returned by the given and younger load instructions to be returned consistent with execution according to the program order to satisfy strong load ordering.Type: GrantFiled: December 2, 2022Date of Patent: September 5, 2023Assignee: Marvell Asia Pte, Ltd.Inventors: David A. Carlson, Shubhendu S. Mukherjee, Wilson P. Snyder, II
-
Publication number: 20230185720Abstract: A memory request, including an address, is accessed. The memory request also specifies a type of an operation (e.g., a read or write) associated with an instance (e.g., a block) of data. A group of caches is selected using a bit or bits in the address. A first hash of the address is performed to select a cache in the group. A second hash of the address is performed to select a set of cache lines in the cache. Unless the operation results in a cache miss, the memory request is processed at the selected cache. When there is a cache miss, a third hash of the address is performed to select a memory controller, and a fourth hash of the address is performed to select a bank group and a bank in memory.Type: ApplicationFiled: February 9, 2023Publication date: June 15, 2023Inventors: Richard E. Kessler, David Asher, Shubhendu S. Mukherjee, Wilson P. Snyder, II, David Carlson, Jason Zebchuk, Isam Akkawi
-
Publication number: 20230099730Abstract: A system and corresponding method enforce strong load ordering in a processor. The system comprises an ordering ring that stores entries corresponding to in-flight memory instructions associated with a program order, scanning logic, and recovery logic. The scanning logic scans the ordering ring in response to execution or completion of a given load instruction of the in-flight memory instructions and detects an ordering violation in an event at least one entry of the entries indicates that a younger load instruction has completed and is associated with an invalidated cache line. In response to the ordering violation, the recovery logic allows the given load instruction to complete, flushes the younger load instruction, and restarts execution of the processor after the given load instruction in the program order, causing data returned by the given and younger load instructions to be returned consistent with execution according to the program order to satisfy strong load ordering.Type: ApplicationFiled: December 2, 2022Publication date: March 30, 2023Inventors: David A. Carlson, Shubhendu S. Mukherjee, Wilson P. Snyder, II
-
Patent number: 11615027Abstract: A memory request, including an address, is accessed. The memory request also specifies a type of an operation (e.g., a read or write) associated with an instance (e.g., a block) of data. A group of caches is selected using a bit or bits in the address. A first hash of the address is performed to select a cache in the group. A second hash of the address is performed to select a set of cache lines in the cache. Unless the operation results in a cache miss, the memory request is processed at the selected cache. When there is a cache miss, a third hash of the address is performed to select a memory controller, and a fourth hash of the address is performed to select a bank group and a bank in memory.Type: GrantFiled: November 18, 2021Date of Patent: March 28, 2023Assignee: Marvell Asia Pte, Ltd.Inventors: Richard E. Kessler, David Asher, Shubhendu S. Mukherjee, Wilson P. Snyder, II, David Carlson, Jason Zebchuk, Isam Akkawi
-
Patent number: 11550590Abstract: A system and corresponding method enforce strong load ordering in a processor. The system comprises an ordering ring that stores entries corresponding to in-flight memory instructions associated with a program order, scanning logic, and recovery logic. The scanning logic scans the ordering ring in response to execution or completion of a given load instruction of the in-flight memory instructions and detects an ordering violation in an event at least one entry of the entries indicates that a younger load instruction has completed and is associated with an invalidated cache line. In response to the ordering violation, the recovery logic allows the given load instruction to complete, flushes the younger load instruction, and restarts execution of the processor after the given load instruction in the program order, causing data returned by the given and younger load instructions to be returned consistent with execution according to the program order to satisfy strong load ordering.Type: GrantFiled: January 28, 2022Date of Patent: January 10, 2023Assignee: Marvell Asia Pte, Ltd.Inventors: David A. Carlson, Shubhendu S. Mukherjee, Wilson P. Snyder, II
-
Publication number: 20220156080Abstract: A system and corresponding method enforce strong load ordering in a processor. The system comprises an ordering ring that stores entries corresponding to in-flight memory instructions associated with a program order, scanning logic, and recovery logic. The scanning logic scans the ordering ring in response to execution or completion of a given load instruction of the in-flight memory instructions and detects an ordering violation in an event at least one entry of the entries indicates that a younger load instruction has completed and is associated with an invalidated cache line. In response to the ordering violation, the recovery logic allows the given load instruction to complete, flushes the younger load instruction, and restarts execution of the processor after the given load instruction in the program order, causing data returned by the given and younger load instructions to be returned consistent with execution according to the program order to satisfy strong load ordering.Type: ApplicationFiled: January 28, 2022Publication date: May 19, 2022Inventors: David A. Carlson, Shubhendu S. Mukherjee, Wilson P. Snyder, II
-
Patent number: 11327890Abstract: A network processor includes a memory subsystem serving a plurality of processor cores. The memory subsystem includes a hierarchy of caches. A mid-level instruction cache provides for caching instructions for multiple processor cores. Likewise, a mid-level data cache provides for caching data for multiple cores, and can optionally serve as a point of serialization of the memory subsystem. A low-level cache is partitionable into partitions that are subsets of both ways and sets, and each partition can serve an independent process and/or processor core.Type: GrantFiled: May 29, 2019Date of Patent: May 10, 2022Assignee: MARVELL ASIA PTE, LTD.Inventor: Shubhendu S. Mukherjee
-
Publication number: 20220114101Abstract: A memory request, including an address, is accessed. The memory request also specifies a type of an operation (e.g., a read or write) associated with an instance (e.g., a block) of data. A group of caches is selected using a bit or bits in the address. A first hash of the address is performed to select a cache in the group. A second hash of the address is performed to select a set of cache lines in the cache. Unless the operation results in a cache miss, the memory request is processed at the selected cache. When there is a cache miss, a third hash of the address is performed to select a memory controller, and a fourth hash of the address is performed to select a bank group and a bank in memory.Type: ApplicationFiled: November 18, 2021Publication date: April 14, 2022Inventors: Richard E. KESSLER, David ASHER, Shubhendu S. MUKHERJEE, Wilson P. SNYDER, II, David CARLSON, Jason ZEBCHUK, Isam AKKAWI
-
Patent number: 11269644Abstract: A system and corresponding method enforce strong load ordering in a processor. The system comprises an ordering ring that stores entries corresponding to in-flight memory instructions associated with a program order, scanning logic, and recovery logic. The scanning logic scans the ordering ring in response to execution or completion of a given load instruction of the in-flight memory instructions and detects an ordering violation in an event at least one entry of the entries indicates that a younger load instruction has completed and is associated with an invalidated cache line. In response to the ordering violation, the recovery logic allows the given load instruction to complete, flushes the younger load instruction, and restarts execution of the processor after the given load instruction in the program order, causing data returned by the given and younger load instructions to be returned consistent with execution according to the program order to satisfy strong load ordering.Type: GrantFiled: July 29, 2019Date of Patent: March 8, 2022Assignee: MARVELL ASIA PTE, LTD.Inventors: David A. Carlson, Shubhendu S. Mukherjee, Wilson P. Snyder, II
-
Patent number: 11188466Abstract: A memory request, including an address, is accessed. The memory request also specifies a type of an operation (e.g., a read or write) associated with an instance (e.g., a block) of data. A group of caches is selected using a bit or bits in the address. A first hash of the address is performed to select a cache in the group. A second hash of the address is performed to select a set of cache lines in the cache. Unless the operation results in a cache miss, the memory request is processed at the selected cache. When there is a cache miss, a third hash of the address is performed to select a memory controller, and a fourth hash of the address is performed to select a bank group and a bank in memory.Type: GrantFiled: February 11, 2020Date of Patent: November 30, 2021Assignee: Marvell Asia Pte, Ltd.Inventors: Richard E. Kessler, David Asher, Shubhendu S. Mukherjee, Wilson P. Snyder, II, David Carlson, Jason Zebchuk, Isam Akkawi
-
Patent number: 11093405Abstract: A network processor includes a memory subsystem serving a plurality of processor cores. The memory subsystem includes a hierarchy of caches. A mid-level instruction cache provides for caching instructions for multiple processor cores. Likewise, a mid-level data cache provides for caching data for multiple cores, and can optionally serve as a point of serialization of the memory subsystem. A low-level cache is partitionable into partitions that are subsets of both ways and sets, and each partition can serve an independent process and/or processor core.Type: GrantFiled: May 29, 2019Date of Patent: August 17, 2021Assignee: MARVELL ASIA PTE, LTD.Inventors: Shubhendu S. Mukherjee, David H. Asher, Richard E. Kessler, Srilatha Manne
-
Patent number: 11036643Abstract: A network processor includes a memory subsystem serving a plurality of processor cores. The memory subsystem includes a hierarchy of caches. A mid-level instruction cache provides for caching instructions for multiple processor cores. Likewise, a mid-level data cache provides for caching data for multiple cores, and can optionally serve as a point of serialization of the memory subsystem. A low-level cache is partitionable into partitions that are subsets of both ways and sets, and each partition can serve an independent process and/or processor core.Type: GrantFiled: May 29, 2019Date of Patent: June 15, 2021Assignee: MARVELL ASIA PTE, LTD.Inventors: David H. Asher, Daniel E. Dever, Thomas F. Hummel, Shubhendu S. Mukherjee
-
Patent number: 10977176Abstract: A first memory request including a first virtual address is received. An entry in memory is accessed. The entry is selected using information associated with the first memory request, and includes at least a portion of a second virtual address (first data) and at least a portion of a third virtual address (second data). The difference between the first data and the second data is compared with differences between a corresponding portion of the first virtual address and the first data and the second data respectively. When a result of the comparison is true, then a fourth virtual address is determined by adding the difference between the first data and the second data to the first virtual address, and then data at the fourth virtual address is prefetched into the cache.Type: GrantFiled: November 26, 2018Date of Patent: April 13, 2021Assignee: Marvell Asia Pte, Ltd.Inventors: David Carlson, Shubhendu S. Mukherjee
-
Publication number: 20200183844Abstract: A memory request, including an address, is accessed. The memory request also specifies a type of an operation (e.g., a read or write) associated with an instance (e.g., a block) of data. A group of caches is selected using a bit or bits in the address. A first hash of the address is performed to select a cache in the group. A second hash of the address is performed to select a set of cache lines in the cache. Unless the operation results in a cache miss, the memory request is processed at the selected cache. When there is a cache miss, a third hash of the address is performed to select a memory controller, and a fourth hash of the address is performed to select a bank group and a bank in memory.Type: ApplicationFiled: February 11, 2020Publication date: June 11, 2020Inventors: Richard E. KESSLER, David ASHER, Shubhendu S. MUKHERJEE, Wilson P. SNYDER, II, David CARLSON, Jason ZEBCHUK, Isam AKKAWI
-
Publication number: 20200167285Abstract: A first memory request including a first virtual address is received. An entry in memory is accessed. The entry is selected using information associated with the first memory request, and includes at least a portion of a second virtual address (first data) and at least a portion of a third virtual address (second data). The difference between the first data and the second data is compared with differences between a corresponding portion of the first virtual address and the first data and the second data respectively. When a result of the comparison is true, then a fourth virtual address is determined by adding the difference between the first data and the second data to the first virtual address, and then data at the fourth virtual address is prefetched into the cache.Type: ApplicationFiled: November 26, 2018Publication date: May 28, 2020Inventors: David CARLSON, Shubhendu S. MUKHERJEE
-
Patent number: 10558573Abstract: A memory request, including an address, is accessed. The memory request also specifies a type of an operation (e.g., a read or write) associated with an instance (e.g., a block) of data. A group of caches is selected using a bit or bits in the address. A first hash of the address is performed to select a cache in the group. A second hash of the address is performed to select a set of cache lines in the cache. Unless the operation results in a cache miss, the memory request is processed at the selected cache. When there is a cache miss, a third hash of the address is performed to select a memory controller, and a fourth hash of the address is performed to select a bank group and a bank in memory.Type: GrantFiled: September 11, 2018Date of Patent: February 11, 2020Assignee: Cavium, LLCInventors: Richard E. Kessler, David Asher, Shubhendu S. Mukherjee, Wilson P. Snyder, II, David Carlson, Jason Zebchuk, Isam Akkawi
-
Publication number: 20200034152Abstract: Typical out-of-order machines can be exploited by security vulnerabilities, such as Meltdown and Spectre, that enable data leakage during misspeculation events. A method of preventing such information leakage includes storing information regarding multiple states of an out-of-order machine to a reorder buffer. This information includes the state of instructions, as well as an indication of data moved to a cache in the transition between states. In response to detecting a misspeculation event at a later state, access to at least a portion of the cache storing the data can be prevented.Type: ApplicationFiled: July 30, 2018Publication date: January 30, 2020Inventors: David A. Carlson, Shubhendu S. Mukherjee
-
Patent number: 10042778Abstract: A computer system that supports virtualization may maintain multiple address spaces. Each guest operating system employs guest virtual addresses (GVAs), which are translated to guest physical addresses (GPAs). A hypervisor, which manages one or more guest operating systems, translates GPAs to root physical addresses (RPAs). A merged translation lookaside buffer (MTLB) caches translations between the multiple addressing domains, enabling faster address translation and memory access. The MTLB can be logically addressable as multiple different caches, and can be reconfigured to allot different spaces to each logical cache. Further, a collapsed TLB provides an additional cache storing collapsed translations derived from the MTLB.Type: GrantFiled: March 31, 2017Date of Patent: August 7, 2018Assignee: Cavium, Inc.Inventors: Shubhendu S. Mukherjee, Bryan W. Chin, Wilson P. Snyder, II, Michael Bertone, Richard E. Kessler, Christopher Mikulis
-
Publication number: 20170206171Abstract: A computer system that supports virtualization may maintain multiple address spaces. Each guest operating system employs guest virtual addresses (GVAs), which are translated to guest physical addresses (GPAs). A hypervisor, which manages one or more guest operating systems, translates GPAs to root physical addresses (RPAs). A merged translation lookaside buffer (MTLB) caches translations between the multiple addressing domains, enabling faster address translation and memory access. The MTLB can be logically addressable as multiple different caches, and can be reconfigured to allot different spaces to each logical cache. Further, a collapsed TLB provides an additional cache storing collapsed translations derived from the MTLB.Type: ApplicationFiled: March 31, 2017Publication date: July 20, 2017Inventors: Shubhendu S. Mukherjee, Bryan W. Chin, Wilson P. Snyder, II, Michael Bertone, Richard E. Kessler, Christopher Mikulis