Efficient mart-aided modeling for microarchitecture design space exploration and performance prediction
read more
Citations
Accurate and efficient processor performance prediction via regression tree based modeling
Microarchitectural design space exploration made fast
An adaptive sampling scheme guided by BART—with an application to predict processor performance
Model guided adaptive design and analysis in computer experiment
References
Greedy function approximation: A gradient boosting machine.
Accurate and efficient regression modeling for microarchitectural performance and power prediction
Efficiently exploring architectural design spaces via predictive modeling
Related Papers (5)
Frequently Asked Questions (10)
Q2. What is the title of the paper?
In this paper, the authors propose to use the state-of-the-art tree-based predictive modeling method combining with advanced sampling techniques from statistics and machine learning to explore the microarchitectural design space and predict the processor performance.
Q3. What is the way to measure the performance of the model?
since some of the architectural design parameters are nominal (no intrinsic ordering structure) and the others are discrete (having a small number of values), the authors use the following defined distance before applying the maximin distance criterion.
Q4. What is the main reason for using adaptive sampling?
Adaptive sampling, also known as active learning in machine learning literature, involves sequential sampling schemes that use information gleaned from previous observations to guide the sampling process.
Q5. What is the way to measure the performance of the experiment design?
In experiment design, the distance-based space-filling sampling methods are popular, especially, when the authors believe that interesting features of the true model are just as likely to be in one part of the experimental region as another.
Q6. What is the definition of a huge design space?
A huge design space is composed by the product of the choices of many microarchitectural design parameters such processor frequency, issue width, cache size/latency, branch predictor settings, etc.
Q7. What is the reason to use MART?
The reason to use MART, an ensemble of trees, is the following: (1) trees are inherently nonparametric and can handle mixed-type of input variables naturally, i.e. no assumptions are made regarding the underlying distribution of values of the input variables, as well as categorical predictors with either ordinal or non-ordinal structure; (2) trees are adept at capturing non-additive behavior, i.e. complex interactions among predictors are routinely and automatically handled with relatively little input required from the analyst; (3) MART improves the prediction performance from a single tree by using an ensemble of trees.
Q8. What is the value for the maximin distance design?
jwt( ) ([ ]∑ =≠×= pj jjj xxIwtd 1 2121 , xx )where ( )parameterdesign for levels ofnumber log2 thj jwt = and ( )AI is an indicator function, equal to one when A holds, other-wise zero.
Q9. How many configurations are used for each workload?
The total design space for each workload is about 15 million configurations composed of the cross product of 13 design parameter choices.
Q10. What is the weight for each design parameter?
Note that the weight for each design parameter is equal to its information entropy with uniform probability for each of its possible values.