# US 2019/0317844 A1 · Methods, systems, articles of manufacture and apparatus to select code data structure types > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2019/0317844 A1 - Document ID: US20190317844A1 - Status: Published application - Kind: A1 - Application number: 16/453,816 - Filed: 2019-06-26 - Published: 2019-10-17 - Earliest claimed: 2019-06-26 - Assignees: Intel Corporation - Applicants: Intel Corporation - Inventors: Justin Gottschlich; Mohammad Mejbah Ul Alam; Shengtian Zhou - Web page: https://patentgrep.com/patent/US20190317844A1 ## CPC - G06F 9/544: {Buffers; Shared memory; Pipes} - G06F 11/302: {where the computing system component is a software system} - G06F 8/30: Creation or generation of source code - G06F 11/3409: {for performance assessment} - G06F 11/3447: {Performance evaluation by modeling} - G06F 11/3476: {Data logging (G06F11/14, G06F11/2205 take precedence)} - G06F 11/3684: {for test design, e.g. generating new test cases} - G06F 16/285: {Clustering or classification} - G06N 3/044: Recurrent networks, e.g. Hopfield networks - G06N 3/0442: characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU] - G06N 3/0455: Auto-encoder networks; Encoder-decoder networks - G06N 3/08: Learning methods - G06N 3/09: Supervised learning - G06N 3/045: Combinations of networks - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS ## Abstract Methods, apparatus, systems and articles of manufacture are disclosed to select code data structure types. An example disclosed apparatus includes an application programming interface (API) engine to generate an abstract data structure (ADS) placeholder in a location of a code sample corresponding to a memory operation, and a data structure selector to select a first candidate data structure having a first candidate data structure type, the first candidate data structure to service the memory operation of the ADS placeholder. The example apparatus also includes a workload engine to select a first candidate workload type to be processed by the selected first candidate data structure, and an execution logger to log first code performance metrics during execution of the code sample during a first iteration corresponding to the first candidate data structure type and the first candidate workload type, and log second code performance metrics during execution of the code sample during a second duration corresponding to a second candidate data structure type and the first candidate workload type. The example apparatus also includes a classification engine to select one of the first candidate data structure type or the second candidate data structure type based on a relative ranking of the first and second code performance metrics. ## Claims ### Claim 1 An apparatus to select a data structure type, the apparatus comprising: an application programming interface (API) engine to generate an abstract data structure (ADS) placeholder in a location of a code sample corresponding to a memory operation; a data structure selector to select a first candidate data structure having a first candidate data structure type, the first candidate data structure to service the memory operation of the ADS placeholder; a workload engine to select a first candidate workload type to be processed by the selected first candidate data structure; an execution logger to: log first code performance metrics during execution of the code sample during a first iteration corresponding to the first candidate data structure type and the first candidate workload type; and log second code performance metrics during execution of the code sample during a second duration corresponding to a second candidate data structure type and the first candidate workload type; and a classification engine to select one of the first candidate data structure type or the second candidate data structure type based on a relative ranking of the first and second code performance metrics. ### Claim 2 (depends on 1) The apparatus as defined in claim 1, further including a data sequence generator to: generate a first sequence of operational data corresponding to the first candidate data structure type; and generate a second sequence of operational data corresponding to the second candidate data structure type. ### Claim 3 (depends on 2) The apparatus as defined in claim 2, further including a sequence normalizer to generate a fixed length representation of respective ones of the first sequence of operational data and the second sequence of operational data. ### Claim 4 (depends on 3) The apparatus as defined in claim 3, wherein the sequence normalizer is to invoke a long short-term memory (LSTM) neural network. ### Claim 5 (depends on 1) The apparatus as defined in claim 1, wherein the workload engine is to select a second candidate workload type for respective ones of the first candidate data structure type and the second candidate data structure type. ### Claim 6 (depends on 5) The apparatus as defined in claim 5, wherein the workload engine is to select different data quantity volumes for respective ones of the first and second candidate workload types. ### Claim 7 (depends on 1) The apparatus as defined in claim 1, further including a code entry detector to identify the memory operation in the code sample. ### Claim 8 A non-transitory computer readable storage medium comprising computer readable instructions that, when executed, cause at least one processor to at least: generate an abstract data structure (ADS) placeholder in a location of a code sample corresponding to a memory operation; select a first candidate data structure having a first candidate data structure type, the first candidate data structure to service the memory operations of the ADS placeholder; select a first candidate workload type to be processed by the selected first candidate data structure; log first code performance metrics during execution of the code sample during a first iteration corresponding to the first candidate data structure type and the first candidate workload type; log second code performance metrics during execution of the code sample during a second duration corresponding to a second candidate data structure type and the first candidate workload type; and select one of the first candidate data structure type or the second candidate data structure type based on a relative ranking of the first and second code performance metrics. ### Claim 9 (depends on 8) The non-transitory computer readable storage medium as defined in claim 8, wherein the instructions, when executed, cause the at least one processor to: generate a first sequence of operational data corresponding to the first candidate data structure type; and generate a second sequence of operational data corresponding to the second candidate data structure type. ### Claim 10 (depends on 9) The non-transitory computer readable storage medium as defined in claim 9, wherein the instructions, when executed, cause the at least one processor to generate a fixed length representation of respective ones of the first sequence of operational data and the second sequence of operational data. ### Claim 11 (depends on 10) The non-transitory computer readable storage medium as defined in claim 10, wherein the instructions, when executed, cause the at least one processor to invoke a long short-term memory (LSTM) neural network. ### Claim 12 (depends on 8) The non-transitory computer readable storage medium as defined in claim 8, wherein the instructions, when executed, cause the at least one processor to select a second candidate workload type for respective ones of the first candidate data structure type and the second candidate data structure type. ### Claim 13 (depends on 12) The non-transitory computer readable storage medium as defined in claim 12, wherein the instructions, when executed, cause the at least one processor to select different data quantity volumes for respective ones of the first and second candidate workload types. ### Claim 14 (depends on 8) The non-transitory computer readable storage medium as defined in claim 8, wherein the instructions, when executed, cause the at least one processor to identify the memory operation in the code sample. ### Claim 15 A computer-implemented method to select a data structure type, the method comprising: generating, by executing an instruction with at least one processor, an abstract data structure (ADS) placeholder in a location of a code sample corresponding to a memory operation; selecting, by executing an instruction with the at least one processor, a first candidate data structure having a first candidate data structure type, the first candidate data structure to service the memory operations of the ADS placeholder; selecting, by executing an instruction with the at least one processor, a first candidate workload type to be processed by the selected first candidate data structure; logging, by executing an instruction with the at least one processor, first code performance metrics during execution of the code sample during a first iteration corresponding to the first candidate data structure type and the first candidate workload type; logging, by executing an instruction with the at least one processor, second code performance metrics during execution of the code sample during a second duration corresponding to a second candidate data structure type and the first candidate workload type; and selecting, by executing an instruction with the at least one processor, one of the first candidate data structure type or the second candidate data structure type based on a relative ranking of the first and second code performance metrics. ### Claim 16 (depends on 15) The method as defined in claim 15, further including: generating a first sequence of operational data corresponding to the first candidate data structure type; and generating a second sequence of operational data corresponding to the second candidate data structure type. ### Claim 17 (depends on 16) The method as defined in claim 16, further including generating a fixed length representation of respective ones of the first sequence of operational data and the second sequence of operational data. ### Claim 18 (depends on 17) The method as defined in claim 17, further including invoking a long short-term memory (LSTM) neural network. ### Claim 19 (depends on 15) The method as defined in claim 15, further including selecting a second candidate workload type for respective ones of the first candidate data structure type and the second candidate data structure type. ### Claim 20 (depends on 19) The method as defined in claim 19, further including selecting different data quantity volumes for respective ones of the first and second candidate workload types. ### Claim 21 (depends on 15) The method as defined in claim 15, further including identifying the memory operation in the code sample.