Patents by Inventor Richard F. Rizzolo
Richard F. Rizzolo 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: 10048734Abstract: Adaptive power capping in a chip that includes a plurality of cores in a processing system is provided. An active power demand for the chip is dynamically determined based on observed events of the cores. An average temperature of the chip is computed using one or more on-chip thermal sensors in the cores to estimate leakage power of the chip. A power capping threshold that incorporates the estimate of leakage power is determined based on the average temperature of the chip. Power capping is performed by throttling the cores based on determining that the active power demand for the chip exceeds the power capping threshold.Type: GrantFiled: December 5, 2017Date of Patent: August 14, 2018Assignee: INTERNATIONAL BUSINESS MACHINES CORPORATIONInventors: Charles R. Lefurgy, Preetham M. Lobo, Richard F. Rizzolo, Malcolm S. Allen-Ware, Tobias Webel
-
Publication number: 20180081413Abstract: Adaptive power capping in a chip that includes a plurality of cores in a processing system is provided. An active power demand for the chip is dynamically determined based on observed events of the cores. An average temperature of the chip is computed using one or more on-chip thermal sensors in the cores to estimate leakage power of the chip. A power capping threshold that incorporates the estimate of leakage power is determined based on the average temperature of the chip. Power capping is performed by throttling the cores based on determining that the active power demand for the chip exceeds the power capping threshold.Type: ApplicationFiled: December 5, 2017Publication date: March 22, 2018Inventors: Charles R. Lefurgy, Preetham M. Lobo, Richard F. Rizzolo, Malcolm S. Allen-Ware, Tobias Webel
-
Patent number: 9874917Abstract: Adaptive power capping in a chip that includes a plurality of cores in a processing system is provided. An active power demand for the chip is dynamically determined based on observed events of the cores. An average temperature of the chip is computed using one or more on-chip thermal sensors in the cores to estimate leakage power of the chip. A power capping threshold that incorporates the estimate of leakage power is determined based on the average temperature of the chip. Power capping is performed by throttling the cores based on determining that the active power demand for the chip exceeds the power capping threshold.Type: GrantFiled: January 4, 2016Date of Patent: January 23, 2018Assignee: INTERNATIONAL BUSINESS MACHINES CORPORATIONInventors: Charles R. Lefurgy, Preetham M. Lobo, Richard F. Rizzolo, Malcolm S. Allen-Ware, Tobias Webel
-
Patent number: 9733685Abstract: A method, system, and computer program product for controlling power supplied to a processor is disclosed. A voltage regulator is set to a first voltage regulator set point, wherein the first voltage regulator set point provides a first load line for operation of the processor. A change in an operation of the processor from a first operating condition along the first load line to a second operating condition along the first load line is determined. The voltage regulator is the set to a second voltage regulator set point and the processor is operated at a third operating condition on a second load line corresponding to the second voltage regulator set point.Type: GrantFiled: December 14, 2015Date of Patent: August 15, 2017Assignee: INTERNATIONAL BUSINESS MACHINES CORPORATIONInventors: Charles R. Lefurgy, Karthick Rajamani, Richard F. Rizzolo, Malcolm S. Allen-Ware
-
Publication number: 20170192477Abstract: Adaptive power capping in a chip that includes a plurality of cores in a processing system is provided. An active power demand for the chip is dynamically determined based on observed events of the cores. An average temperature of the chip is computed using one or more on-chip thermal sensors in the cores to estimate leakage power of the chip. A power capping threshold that incorporates the estimate of leakage power is determined based on the average temperature of the chip. Power capping is performed by throttling the cores based on determining that the active power demand for the chip exceeds the power capping threshold.Type: ApplicationFiled: January 4, 2016Publication date: July 6, 2017Inventors: Charles R. Lefurgy, Preetham M. Lobo, Richard F. Rizzolo, Malcolm S. Allen-Ware, Tobias Webel
-
Publication number: 20170168534Abstract: A method, system, and computer program product for controlling power supplied to a processor is disclosed. A voltage regulator is set to a first voltage regulator set point, wherein the first voltage regulator set point provides a first load line for operation of the processor. A change in an operation of the processor from a first operating condition along the first load line to a second operating condition along the first load line is determined. The voltage regulator is the set to a second voltage regulator set point and the processor is operated at a third operating condition on a second load line corresponding to the second voltage regulator set point.Type: ApplicationFiled: December 14, 2015Publication date: June 15, 2017Inventors: Charles R. Lefurgy, Karthick Rajamani, Richard F. Rizzolo, Malcolm S. Allen-Ware
-
Patent number: 9575529Abstract: A processor is provided having a common supply rail, and one or more processor cores, where the one or more processor cores share the common supply rail. Each processor core(s) includes a core dIPC value output and a core throttling signal input, and a chip power management logic, which has at least one input for inputting the core dIPC value, a threshold register for a dIPC threshold value, a chip dIPC register for a current global dIPC value, at least one chip dIPC history register for a historic global dIPC value, a subtractor providing an absolute difference of an average historic global dIPC derived from the historic global dIPC value and the current global dIPC value, a magnitude comparator providing a throttling signal when the absolute difference is above the dIPC threshold value, and at least one output for outputting a core throttling signal to the processor core(s).Type: GrantFiled: September 24, 2015Date of Patent: February 21, 2017Assignee: INTERNATIONAL BUSINESS MACHINES CORPORATIONInventors: Brian W. Curran, Preetham M. Lobo, Richard F. Rizzolo, James D. Warnock, Tobias Webel
-
Publication number: 20160098070Abstract: A processor is provided having a common supply rail, and one or more processor cores, where the one or more processor cores share the common supply rail. Each processor core(s) includes a core dIPC value output and a core throttling signal input, and a chip power management logic, which has at least one input for inputting the core dIPC value, a threshold register for a dIPC threshold value, a chip dIPC register for a current global dIPC value, at least one chip dIPC history register for a historic global dIPC value, a subtractor providing an absolute difference of an average historic global dIPC derived from the historic global dIPC value and the current global dIPC value, a magnitude comparator providing a throttling signal when the absolute difference is above the dIPC threshold value, and at least one output for outputting a core throttling signal to the processor core(s).Type: ApplicationFiled: September 24, 2015Publication date: April 7, 2016Inventors: Brian W. CURRAN, Preetham M. LOBO, Richard F. RIZZOLO, James D. WARNOCK, Tobias WEBEL
-
Patent number: 6971054Abstract: An exemplary embodiment of the present invention is a method for testing an integrated circuit. The method includes generating a test pattern and generating a reference signature corresponding to the test pattern. An integrated circuit test is executed in response to the test pattern and a result signature is generated in response to data output from executing the integrated circuit test. The result signature is compared to the reference signature and a current failing signature is created if the two don't match. The current failing signature is copy of the result signature. Common error analysis is executed in response to creating the current failing signature. Additional embodiments include a system and storage medium for testing an integrated circuit.Type: GrantFiled: May 2, 2002Date of Patent: November 29, 2005Assignee: International Business Machines CorporationInventors: Raymond J. Kurtulik, Franco Motika, Richard F. Rizzolo
-
Patent number: 6751765Abstract: An exemplary embodiment of the invention is a method for LBIST testing integrated circuit. The method includes generating a plurality of multi-bit test patterns and grouping the multi-bit test patterns by a plurality of test pattern partitions including a first test pattern partition having a first number of bits and a second test pattern partition having second number of bits greater than the first number. The first test pattern partition is applied to the integrated circuit to generate a first signature that is compared to a first reference signature to detect a failure. The second test pattern partition is applied to the integrated circuit to generate a second signature that is compared to a second reference signature to detect a failure in the integrated circuit.Type: GrantFiled: November 27, 2000Date of Patent: June 15, 2004Assignee: International Business Machines CorporationInventors: Richard F. Rizzolo, Rocco E. DeStefano, Joseph E. Eckelman, Thomas G. Foote, Steven Michnowski, Franco Motika, Phillip J. Nigh, Bryan J. Robbins
-
Patent number: 6662324Abstract: The present invention, enables complementing the state of either the master (L1) or slave latch (L2) in the shift register latches (SRLs) without changing the state of the other. When this is done after properly loading the LSSD scan chain using a normal scan chain sequence, the next system clock sequence can be used to launch a desired transition from each SRL in the scan chain. The actual mechanism for complementing the state of latches in LSSD scan chains can vary depending on which one of the L1 or L2 latch is being complemented; details of the actual scan chain and Shift Register Latch (SRL) design; and the semiconductor chip circuit technology. The complementing function can be provided as an integral part of the SRL design with minimal impact to system path and performance. An alternate complementing method would be to use a self complementing latch function.Type: GrantFiled: August 21, 2000Date of Patent: December 9, 2003Assignee: International Business Machines CorporationInventors: Franco Motika, Richard F. Rizzolo, Peilin Song, William V. Huott, Ulrich Baur
-
Patent number: 6532571Abstract: A method to improve the testability and analysis of a hierarchical semiconductor chip design formed from a plurality of macros, each macro identifying a particular portion of a semiconductor chip design. This method includes providing a first macro netlist that identifies a logical description of a first portion of the semiconductor chip design and performing RPT analysis on the first macro netlist. The method also includes providing a second macro netlist identifying a logical description of a second portion of the semiconductor chip design and performing an RPT analysis on the second macro netlist. The first macro netlist is combined with the second macro netlist and an RPT analysis is performed on the combination of the first and second macro netlists.Type: GrantFiled: January 21, 2000Date of Patent: March 11, 2003Assignee: International Business Machines CorporationInventors: Richard M. Gabrielson, Kevin W. McCauley, Richard F. Rizzolo, Bryan J. Robbins, Joseph M. Swenton
-
Patent number: 6490702Abstract: A scan chain latch circuit is provided. The scan chain latch circuit includes a first shift register latch and a second shift register latch. The scan chain latch circuit also includes a multiplexor connected between the first and second shift register latches, the multiplexor has a select line for controlling the function of the multiplexor. The multiplexor is configured for implementing an inverting mode such that a logic value may be passed via the multiplexor from the first shift register latch to the second shift register latch in one of a non-inverted state and an inverted state based upon the state of the select line.Type: GrantFiled: December 28, 1999Date of Patent: December 3, 2002Assignee: International Business Machines CorporationInventors: Peilin Song, Richard F. Rizzolo, Franco Motika, Ulrich W. Baur
-
Patent number: 6453436Abstract: A scan chain latch circuit is provided. The scan chain latch circuit includes a first shift register latch, a second shift register latch, and a third shift register latch. A first multiplexor is connected between the first and second shift register latches, and a second multiplexor is connected between the second and third shift register latches. Each multiplexor is configured for implementing a jump mode such that a logic value may be passed via the first multiplexor from the first shift register latch to the third shift register latch.Type: GrantFiled: December 28, 1999Date of Patent: September 17, 2002Assignee: International Business Machines CorporationInventors: Richard F. Rizzolo, Peilin Song
-
Publication number: 20020125907Abstract: An exemplary embodiment of the present invention is a method for testing an integrated circuit. The method includes generating a test pattern and generating a reference signature corresponding to the test pattern. An integrated circuit test is executed in response to the test pattern and a result signature is generated in response to data output from executing the integrated circuit test. The result signature is compared to the reference signature and a current failing signature is created if the two don't match. The current failing signature is copy of the result signature. Common error analysis is executed in response to creating the current failing signature. Additional embodiments include a system and storage medium for testing an integrated circuit.Type: ApplicationFiled: May 2, 2002Publication date: September 12, 2002Applicant: International Business Machines CorporationInventors: Raymond J. Kurtulik, Franco Motika, Richard F. Rizzolo
-
Patent number: 6442720Abstract: The present invention can include a method and system for testing IC chips, including the steps of performing a binary search to a first failing pattern, determining a failing sink latch, performing a back cone trace to determine all source latches, determining source latch logic states, positioning the source latch logic states in a scan chain, exercising a chip scan path by applying logic transitions on the source latches in the absence of a system L1 clock, and observing an exercised failing circuit. The invention can include the use of PICA techniques to observe the exercised failing circuit. In another embodiment, the invention can include using LBIST or a WRP technique to search for the failing pattern. In yet another it includes the step of using an algorithm to exercise the exercised failing circuit.Type: GrantFiled: June 4, 1999Date of Patent: August 27, 2002Assignee: International Business Machines CorporationInventors: Timothy J. Koprowski, Mary P. Kusko, Richard F. Rizzolo, Peilin Song
-
Patent number: 5455931Abstract: A clock tuning system and method for a data processing system with enhanced timing failure diagnostics and unlayering capabilities. Both common and individual phase adjusting capabilities ensure programmable tuning of clock pulses distributed throughout a computer system, thereby facilitating isolation of timing margin failure to specific clock signals or enhancing system performance by shifting timing margin between logic paths. Both single-clock and dual-clock data processing are discussed, as well as clock tuning embodiments for each.Type: GrantFiled: November 19, 1993Date of Patent: October 3, 1995Assignee: International Business Machines CorporationInventors: Peter J. Camporese, Patrick J. Meaney, Brian J. O'Leary, Richard F. Rizzolo
-
Patent number: 5142167Abstract: This invention reduces the Delta I noise on an integrated circuit chip by reducing the changes in current supply required for transitions in logic states of the input/output devices. The invention uses a 3/6 binary code for communicating between integrated circuit chips. This code uses six bits to represent the 16 hex code digits typically used for computer instructions. Three of the six bits are in a high logic state and three of the six bits are in a low logic state for all 16 hex code representations. Therefore, changing from any one logic state to another, does not change the overall current supply required by the six input/output devices. Groups of six input/output devices (corresponding to the 3/6 code) are located relatively close to each other with respect to the power supply pins which supply current to the six input/output devices. As a result, there is a high to low transition for every low to high transition over similar parasitic impedances on the input/output devices.Type: GrantFiled: May 1, 1991Date of Patent: August 25, 1992Assignee: International Business Machines CorporationInventors: Joseph L. Temple, Richard F. Rizzolo, Charles B. Winn
-
Patent number: 4760289Abstract: A masterslice cell wireable to form any of a selected book set of two level differential cascode current switch basic circuits. Twenty percent increased performance is provided as compared with ECL masterslice circuits running at the same power. In spite of increased wire due to differential logic, and potential increased complexity in design software, the invention is actually readily adaptable to existing masterslice design systems.Type: GrantFiled: August 4, 1986Date of Patent: July 26, 1988Assignee: International Business Machines CorporationInventors: Edward B. Eichelberger, Stephen E. Bello, Rolf O. Bergenn, William M. Chu, John A. Ludwig, Richard F. Rizzolo