Abstract: The technology disclosed relates to automatically assigning and optimizing the physical memory layouts of all intermediate dense tensor data in a program. The technology disclosed is an implementation of a compiler analysis and transformation pass which automatically determines required physical layouts in light of kernel operation and performance requirements. The proposed solution also inserts physical layout conversion operations where necessary in cases of unresolvable incompatibilities. The pass takes as input a program acyclic dataflow graph and a set of physical layout constraints for every known operation.

Abstract: A dataflow graph has operation units that are configured to be producer operation units to produce tensors for execution of the application, and to be consumer operation units to consume the tensors for execution of the application. Compile time logic is configured to process the dataflow graph to determine, for the tensors, expected producer memory layouts, expected consumer memory layouts, and current memory layouts. The expected producer memory layouts specify memory layouts required by the producer operation units that produce the tensors. The expected consumer memory layouts specify the memory layouts required by the consumer operation units that consume the tensors. The current memory layouts specify the memory layouts of the tensors. Each of the memory layouts includes a vector dimension and at least one of a vector ordering and a data alignment.

Abstract: A dataflow graph for an application has operation units that are configured to be producers and consumers of tensors. A write access pattern of a particular producer specifies an order in which the particular producer generates elements of a tensor, and a read access pattern of a corresponding consumer specifies an order in which the corresponding consumer processes the elements of the tensor. The technology disclosed detects conflicts between the producers and the corresponding consumers that have mismatches between the write access patterns and the read access patterns. A conflict occurs when the order in which the particular producer generates the elements of the tensor is different from the order in which the corresponding consumer processes the elements of the tensor. The technology disclosed resolves the conflicts by inserting buffers between the producers and the corresponding consumers.

Abstract: A dataflow graph has operation units that are configured to be producer operation units to produce tensors for execution of the application, and to be consumer operation units to consume the tensors for execution of the application. Compile time logic is configured to process the dataflow graph to determine, for the tensors, expected producer memory layouts, expected consumer memory layouts, and current memory layouts. The expected producer memory layouts specify memory layouts required by the producer operation units that produce the tensors. The expected consumer memory layouts specify the memory layouts required by the consumer operation units that consume the tensors. The current memory layouts specify the memory layouts of the tensors. Each of the memory layouts includes a vector dimension and at least one of a vector ordering and a data alignment.

Abstract: A dataflow graph for an application has operation units that are configured to be producers and consumers of tensors. A write access pattern of a particular producer specifies an order in which the particular producer generates elements of a tensor, and a read access pattern of a corresponding consumer specifies an order in which the corresponding consumer processes the elements of the tensor. The technology disclosed detects conflicts between the producers and the corresponding consumers that have mismatches between the write access patterns and the read access patterns. A conflict occurs when the order in which the particular producer generates the elements of the tensor is different from the order in which the corresponding consumer processes the elements of the tensor. The technology disclosed resolves the conflicts by inserting buffers between the producers and the corresponding consumers.

Abstract: Computer implemented method, system and computer usable program code for cache management. A cache is provided, wherein the cache is viewed as a sorted array of data elements, wherein a top position of the array is a most recently used position of the array and a bottom position of the array is a least recently used position of the array. A memory access sequence is provided, and a training operation is performed with respect to a memory access of the memory access sequence to determine a type of memory access operation to be performed with respect to the memory access. Responsive to a result of the training operation, a cache replacement operation is performed using the determined memory access operation with respect to the memory access.

Abstract: A friction plate 22 is formed by pasting a friction member 25 to a core plate 23 of a metal. Numeral 24 designates a spline teeth fitted to a torque transmitting member. The friction member has a nonwoven cloth including aramide fiber and at least one of carbon fiber, carbonized fiber, cellulose fiber and rayon. A thermo-hardening resin includes a friction conditioner, and is impregnated to the friction member. Further, a distribution rate of the friction conditioner is made to be high at a friction surface and is gradually lowered toward a bottom surface. Further, in manufacturing the wet type friction member, a roll, an ultrasonic wave transmitter, a vacuuming apparatus on the like is used.

Abstract: Computer implemented method, system and computer usable program code for cache management. A cache is provided, wherein the cache is viewed as a sorted array of data elements, wherein a top position of the array is a most recently used position of the array and a bottom position of the array is a least recently used position of the array. A memory access sequence is provided, and a training operation is performed with respect to a memory access of the memory access sequence to determine a type of memory access operation to be performed with respect to the memory access. Responsive to a result of the training operation, a cache replacement operation is performed using the determined memory access operation with respect to the memory access.

Abstract: A friction plate 22 is formed by pasting a friction member 25 to a core plate 23 of a metal. Numeral 24 designates a spline teeth fitted to a torque transmitting member. The friction member has a nonwoven cloth including aramide fiber and at least one of carbon fiber, carbonized fiber, cellulose fiber and rayon. A thermo-hardening resin includes a friction conditioner, and is impregnated to the friction member. Further, a distribution rate of the friction conditioner is made to be high at a friction surface and is gradually lowered toward a bottom surface. Further, in manufacturing the wet type friction member, a roll, an ultrasonic wave transmitter, a vacuuming apparatus on the like is used.

Abstract: A friction plate for a wet-type multiplate clutch provided with a core plate and friction linings fixed on opposite sides of the core plate, respectively. Each of the friction linings is formed such that at a section thereof located on a side radially inner than a radial center line thereof, the friction lining has a thickness that gradually increases from an inner circumferential portion of the section toward an outer circumferential portion of the section.

Abstract: In a lockup clutch assembly for use in a torque converter, a friction lining attached to one of a clutch piston and a front cover of the torque converter is divided into an outer-circumference-side friction lining and an inner-circumference-side friction lining. The outer-circumference-side friction lining contains as a filler diatomaceous earth in a higher proportion than the inner-circumference-side friction lining.

Abstract: In a double-wrap band brake assembly, an intermediate band is provided with a first friction material forming a first friction surface from a portion of the intermediate band near an apply bracket to a predetermined intermediate portion thereof, the first friction surface working for forming a lubricant film. The intermediate band is further provided with a second friction material having a thickness that is larger than a thickness of the first friction material and forming a second friction surface from the predetermined intermediate portion to a vicinity of a coupling portion of the intermediate band with a pair of outer bands, the second friction surface working for removing the lubricant film.

Abstract: A double-wrap band brake assembly to be used for braking a rotating member fitted in a double-wrap brake band, which double-wrap brake band comprises an annular middle band with a friction surface formed on the inner periphery thereof, a pair of annular outer bands which are coupled to the middle band in a state that the respective free ends thereof are opposed to the free end of the middle band, and which respectively have friction surfaces on the inner peripheries, an anchor bracket secured to the working end of either one of the middle band and the outer bands and latched to the main body casing, and an apply bracket secured to the working end of the other of the middle band and the outer bands for receiving a force in the direction of constriction of the middle band and the outer bands from an actuator.

Abstract: In a torque converter equipped with a lock-up clutch device, a friction member is put on either one of a front cover and a lock-up piston which are to be brought into contact. On the front cover, coupling members are welded for coupling with an engine. Owing to welding distortion, convex portions are generated on a friction face. Thus, when the friction member is put on either side of the front cover and clutch piston, if relative rotation slippage occurs therebetween, a positive pressure and a negative pressure are produced in front of and behind each convex portion, which leads to the stick/slip. This invention gives grooves 71 each exceeding the range of a welding portion 61 of the coupling member 6 which are formed on the friction face of the friction member 7, which realizes balancing in a change in pressure.

Abstract: In a wet type friction plate having a groove through which lubricating oil passes, the groove is subjected to oil repelling treatment.

Abstract: In a lockup clutch assembly for use in a torque converter, a friction lining attached to one of a clutch piston and a front cover of the torque converter is divided into an outer-circumference-side friction lining and an inner-circumference-side friction lining. The outer-circumference-side friction lining contains as a filler diatomaceous earth in a higher proportion than the inner-circumference-side friction lining.

Abstract: A friction plate for a wet-type multiplate clutch provided with a core plate and friction linings fixed on opposite sides of the core plate, respectively. Each of the friction linings is formed such that at a section thereof located on a side radially inner than a radial center line thereof, the friction lining has a thickness that gradually increases from an inner circumferential portion of the section toward an outer circumferential portion of the section.

Abstract: A wet type friction plate comprises symmetrical segments cut out of a frictional member, and pasted on the metallic core plate so as to form a groove which is gradually narrower from the inner diameter side to the outer diameter side. A method of producing the wet type friction plate, comprising the steps of cutting symmetrical segments out of a frictional member, and pasting the cut-out symmetrical segments on a metallic core plate so as to form a groove which is gradually narrower from the inner diameter side to the outer diameter side.

Abstract: A friction plate 22 is formed by pasting a friction member 25 to a core plate 23 of a metal. Numeral 24 designates a spline teeth fitted to a torque transmitting member. The friction member has a nonwoven cloth including aramide fiber and at least one of carbon fiber, carbonized fiber, cellulose fiber and rayon. A thermo-hardening resin includes a friction conditioner, and is impregnated to the friction member. Further, a distribution rate of the friction conditioner is made to be high at a friction surface and is gradually lowered toward a bottom surface. Further, in manufacturing the wet type friction member, a roll, an ultrasonic wave transmitter, a vacuuming apparatus on the like is used.