# US 8,571,924 B2 · High performance personalized advertisement serving by exploiting thread assignments in a multiple core computing environment > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 8,571,924 B2 - Document ID: US8571924B2 - Status: Granted - Kind: B2 - Application number: 13/176,668 - Filed: 2011-07-05 - Published: 2013-01-10 - Granted: 2013-10-29 - Earliest claimed: 2011-07-05 - Assignees: Yahoo! Inc. - Applicants: Arun Kerjariwal - Inventors: Arun Kerjariwal - Web page: https://patentgrep.com/patent/US8571924B2 ## CPC - G06Q 30/0241: Advertisements - G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR ## Abstract A method, apparatus and computer program product for high performance advertisement serving by exploiting processor thread assignments in a processor having multiple threads. The method commences by receiving digital advertisements, an instance of the digital advertisement being associated with a plurality of decision trees (to relate a user to an advertisement), where the evaluation of a decision tree results in a tree score. After determining a number of cores within the processor, the decisions trees are assigned to processor cores or processor threads by iteratively assigning to a next one of the number of cores, a next decision tree, and iteratively accumulating a plurality of the tree scores to form a tree score subtotal. The tree score subtotals can be compared quantitatively to rank a first digital advertisement against a second digital advertisement, and a ranked advertisement displayed to a user. Some decision trees comprise demographic, behavioral and context quantifiers. ## Claims ### Claim 1 A computer-implemented method for high performance advertisement serving by exploiting processor thread assignments in a processor having multiple threads, the method comprising: receiving, in memory, at least two digital advertisements, an instance of the digital advertisement corresponding to a plurality of decision trees, evaluation of a decision tree resulting in a tree score; determining, from the processor, a number of cores within the processor; iteratively assigning, to a next one of the number of cores, a next decision tree to calculate the tree score; and iteratively accumulating a plurality of the tree scores to form a tree score subtotal, wherein iteratively accumulating a plurality of tree scores to form a tree score subtotal terminates when the tree score subtotal is greater than a tree score subtotal threshold, based in part on an execution rate assessment. ### Claim 2 (depends on 1) The method of claim 1, further comprising: ranking a first of the at least two digital advertisements against a second of the at least two digital advertisements. ### Claim 3 (depends on 1) The method of claim 1, wherein one core selected from the number of cores executes multiple threads. ### Claim 4 (depends on 1) The method of claim 1, wherein the at least two digital advertisements are selected from a candidate ad set. ### Claim 5 (depends on 1) The method of claim 1, wherein the plurality of decision trees are selected in an execution order from a decision tree forest, the execution order determined in part by a tree depth. ### Claim 6 (depends on 1) The method of claim 1, wherein the step of iteratively accumulating a plurality of tree scores is terminated when the tree score subtotal is greater than a graph threshold value. ### Claim 7 (depends on 1) The method of claim 1, wherein the step of iteratively accumulating a score subtotal is terminated based on an execution rate assessment. ### Claim 8 An advertising server network implementing high performance advertisement serving by exploiting processor thread assignments in a processor having multiple threads comprising: a module for receiving, at least two digital advertisements, an instance of the digital advertisement corresponding to a plurality of decision trees, evaluation of a decision tree resulting in a tree score; a module for determining, from the processor, a number of cores within the processor; a module for iteratively assigning, to a next one of the number of cores, a next decision tree to calculate the tree score; and a module for iteratively accumulating a plurality of the tree scores to form a tree score subtotal, wherein iteratively accumulating a plurality of tree scores to form a tree score subtotal terminates when the tree score subtotal is greater than a tree score subtotal threshold, based in part on an execution rate assessment. ### Claim 9 (depends on 8) The advertising server network of claim 8, further comprising: a module for ranking a first of the at least two digital advertisements against a second of the at least two digital advertisements. ### Claim 10 (depends on 8) The advertising server network of claim 8, wherein one core selected from the number of cores executes multiple threads. ### Claim 11 (depends on 8) The advertising server network of claim 8, wherein the at least two digital advertisements are selected from a candidate ad set. ### Claim 12 (depends on 8) The advertising server network of claim 8, wherein the plurality of decision trees are selected in an execution order from a decision tree forest, the execution order determined in part by a tree depth. ### Claim 13 (depends on 8) The advertising server network of claim 8, wherein the step of iteratively accumulating a plurality of tree scores is terminated when the tree score subtotal is greater than a graph threshold value. ### Claim 14 (depends on 8) The advertising server network of claim 8, wherein the step of iteratively accumulating a score subtotal is terminated based on an execution rate assessment. ### Claim 15 A non-transitory computer readable medium comprising a set of instructions which, when executed by a computer, cause the computer to implement high performance advertisement serving by exploiting processor thread assignments in a processor having multiple threads, the set of instructions for: receiving, at least two digital advertisements, an instance of the digital advertisement corresponding to a plurality of decision trees, evaluation of a decision tree resulting in a tree score; determining, from the processor, a number of cores within the processor; iteratively assigning, to a next one of the number of cores, a next decision tree to calculate the tree score; and iteratively accumulating a plurality of the tree scores to form a tree score subtotal, wherein iteratively accumulating a plurality of tree scores to form a tree score subtotal terminates when the tree score subtotal is greater than a tree score subtotal threshold, based in part on an execution rate assessment. ### Claim 16 (depends on 15) The non-transitory computer readable medium of claim 15, further comprising: instructions for ranking a first of the at least two digital advertisements against a second of the at least two digital advertisements. ### Claim 17 (depends on 15) The non-transitory computer readable medium of claim 15, wherein one core selected from the number of cores executes multiple threads. ### Claim 18 (depends on 15) The non-transitory computer readable medium of claim 15, wherein the plurality of decision trees are selected in an execution order from a decision tree forest, the execution order determined in part by a tree depth. ### Claim 19 (depends on 15) The non-transitory computer readable medium of claim 15, wherein the step of iteratively accumulating a plurality of tree scores is terminated when the tree score subtotal is greater than a graph threshold value. ### Claim 20 (depends on 15) The non-transitory computer readable medium of claim 15, wherein the step of iteratively accumulating a score subtotal is terminated based on an execution rate assessment.