Modifying computer configuration to improve performance

Abstract

Embodiments of the present disclosure relate to improving computer performance. An action may be issued to a first client and a second client. A first optimization factor list (OFL) may be generated for the first client and a second OFL may be generated for the second client. After each OFL is generated, the first OFL may be compared to the second OFL. A correlation factor (CF) may be generated between the first client and the second client based on the OFL comparison, wherein the CF indicates a level of similarity between the first client and the second client. A bottleneck causing performance issues in the first client may be determined and computer configuration of the first client may be adjusted based on the bottleneck.

What is claimed is: 
     
         1 . A method comprising:
 issuing, by a processor, an action to a first client and a second client through a network interface, wherein the action affects multiple computer components to test multiple computer components simultaneously, wherein the multiple computer components include computer processing units (CPU), random access memory, graphics processing unit, CPU benchmark testing, and Network I/F;   generating a first optimization factor list (OFL) for the first client and a second OFL for the second client, wherein the OFL includes performance data that includes a client's performance at the issued action, wherein performance data includes how fast the first client and the second client completed the action, wherein the performance data is based on a percentage completed of the action, wherein the performance data is only generated until a certain point in time;   comparing the first OFL to the second OFL, wherein the first OFL and the second OFL are a list of all computer components associated with each respective client;   generating a correlation factor (CF) between the first client and second client based on the OFL comparison, wherein the CF indicates a level of similarity between the first client and the second client, wherein generating the CF includes generating a partial CF value for each factor in the OFL, weighting each factor in the OFL, adjusting each partial CF value using a weight of the factor associated with the partial CF, and accumulating the adjusted partial CF values;   determining that there is a bottleneck causing performance issues in the first client;   adjusting, in response to determining that there is a bottleneck causing performance issues in the first client, a client computer configuration of the first client based on the bottleneck;   generating, by the processor, a relevance factor (RF) based on an adjustment made to a factor of the client computer configuration;   altering, based on the RF of the adjustment made to the factor of the client computer configuration, the OFL and CF, wherein the altering the OFL and CF comprises:
 determining that the RF exceeds a threshold RF value, wherein the threshold RF value is defined as a change effectiveness required to shift a client into a higher performance efficiency class, wherein the RF threshold is based in part on a percentage of improvement in performance, wherein the threshold RF value is numerically defined depending on an RF range; and 
 increasing, in response to determining that the RF exceeds the threshold RF value, the weighting of the factor that was adjusted; 
   classifying, by the processor, the first client and the second client into performance efficiency classes based on action performance; and   determining that the first client and the second client are in different performance efficiency classes, wherein the comparing the OFL of the first client to the OFL of the second client is performed in response to determining that the first and second clients are in different performance efficiency classes.