Accelerating machine learning and profiling over a network

A local computing device comprising:

• one or more processors;
• and
• one or more non-transitory computer-readable storage media to store a configuration of the computing device and to store instructions executable by the one or more processors to perform operations comprising:
• executing an application executing locally on the local computing device, and not as part of a distributed computer system;
• gathering local usage data associated with a usage, by the locally executing application, of multiple computing resources of the local computing device, the local usage data comprising:
• a clock speed of a central processing unit (CPU) of the local computing device;
• a number of threads associated with the locally executing application that are being executed by the CPU;
• a graphics memory utilization of a graphics processing unit (GPU) of the local computing device;
• and
• a temperature of the GPU;
• determining if a profile exists on the local computing device that is associated with the locally executing application;
• and
• then only one of:
• selecting an existing profile for use by the local computing device only if the existing profile on the local computing device is determined to be associated with the locally executing application, or
• performing the following if no existing profile on the local computing device is determined to be associated with the locally executing application:
• generating one or more packets, wherein each packet of the one or more packets includes at least a portion of the local usage data,
• setting a priority of each packet of the one or more packets,
• sending the one or more packets to a remote server or one or more additional remote computing devices,
• receiving a new profile to modify a configuration of the local computing device to increase throughput of the locally executing application being executed by the local computing device,
• modifying an original configuration of the local computing device based at least in part on the new profile to create a second configuration,
• executing the locally executing application based on the second configuration, and
• determining that throughput of the locally executing application has increased using the second configuration as compared to using the original configuration.

One or more non-transitory computer-readable storage media to store instructions executable by one or more processors of a local computing device to perform operations comprising:

• executing an application executing locally on the local computing device, and not as part of a distributed computer system;
• gathering local usage data associated with a usage, by the locally executing application, of multiple computing resources of the local computing device, the local usage data comprising:
• a clock speed of a central processing unit (CPU) of the local computing device;
• a number of threads associated with the locally executing application that are being executed by the CPU;
• a graphics memory utilization of a graphics processing unit (GPU) of the local computing device;
• and
• a temperature of the GPU;
• determining if a profile exists on the local computing device that is associated with the locally executing application;
• and
• then only one of:
• selecting an existing profile for use by the local computing device only if the existing profile on the local computing device is determined to be associated with the locally executing application, or
• performing the following if no existing profile on the local computing device is determined to be associated with the locally executing application:
• generating one or more packets, wherein each packet of the one or more packets includes at least a portion of the local usage data,
• setting a priority of each packet of the one or more packets,
• modifying a priority of each packet of the one or more packets to a high priority,
• sending the one or more packets having the high priority to a remote server or one or more additional remote computing devices,
• receiving a new profile to modify a configuration of the local computing device to increase throughput of the locally executing application being executed by the local computing device,
• modifying an original configuration of the local computing device, based at least in part on the new profile, to create a second configuration,
• executing the locally executing application based on the second configuration, and
• determining that throughput of the locally executing application has increased using the second configuration as compared to using the original configuration.