Approximate Computing is based on the idea that significant improvements in performance, energy and memory consumptions can be achieved when small levels of approximation can be tolerated. Indeed, domains such as Machine Learning, Search Engines, Computer Vision and Digital Signal Processing are approximate in nature and has been shown to …
To avoid going berserker and start flipping tables, these are a few of rants I have written over time when someone goes like ‘it works on my machine’. So here you have it, a few guidelines and bits of advice if you want to slowdown your team, stress the sh**it …
Resulting from a preparation for an interview, I realized that I had never thought about the way I solve problems. I just did it. It’s like if someone ask you `describe how you breath’. This blog post results from my efforts to describe a methodology that paradoxically enough, I never had to …
Approximate Loop Unrolling (A-Roll) is an approximate loop optimization that transforms loops in a similar way Loop Unrolling does. However, unlike its exact counterpart, A-Roll unrolls by adding interpolations of the loop’s body, rather than exact copies. The objective of the optimization is to increase the loop’s speed while reducing …
After struggling all day yesterday with dependencies in CMake, this is what I got so far: When looking to manage dependencies using CMake abandon any hope for quick tricks and rapid fixes. You must understand the system deeply or be prepared to suffer several headaches. That’s why I advise you …
In the C2 compiler, loops are optimized during the Machine Independent Optimizations super-phase . Loops is where most of the execution time is spent, so the compiler gives great deal of attention to them. As mentioned elsewhere, Machine Independent Optimizations are performed modifying the Ideal Graph. Loop optimizations are implemented …
In this post I will write down my notes on the behavior, inputs and outputs of some nodes of the Ideal Graph (IG). This is not an introduction to the IG. I’m assuming you have already some knowledge on the subject and have read some of the related documentation. If …
The Java C2 compiler turns Java bytecode into highly optimized native instructions. The process is decomposed in several phases. Knowing the order in which these phases are applied is important for developers interested in modifying Hostpot. In this post I will talk about the phase ordering of the C2 compiler, …
In this post I will be talking about how to make C2 selectively optimize methods based on a newly created annotation. I will also describe the parts of the code that must be modified to implement such feature. Do you need annotations to optimize parts of your code? Annotations are …
In a previous post I wrote about how to build OpenJDK 9. This is cool, but my objective was to mess around with the sources of Hotspot and modify it, so I must have a way to write my changes and debug them. I’m using as main source of reference …