This is another study solely dedicated to understanding Concurrency and this will be learnt with Go, using the book Concurrency in Go by Katherine Cox-Buday as a guide.
I will be building small programs and making documentation of findings in any way possible.
Concurrency is a property of the code, Parallelism is a property of the running program. We write a concurrent code and hope that it is run in parallel. A concurrent code running on one core is not running in parallel but sequentially in a fast way.
Parallelism is a property of the runtime of our code.
Moore's law ("the number of components on an integrated circuit will double every 2 years") started losing it's impact till around the 2012, companies foresaw slowdown in rate of Moore's law and started looking for different ways in increasing computing power, this led to creating multicore processors, this led to solving problem in a simultaneous way (Parallelism) and this stem from Amdahl's law ("it entails, model the potential performance gains from implementing the solution to a problem in a parellel manner i.e gains are bounded by how much of the program must run in sequential manner").
Solving problems in a parallel manner led to Horizontal Scaling, having multiple instance of a program on different CPUs or machines.
-
Common Issues in Concurrency
-
Building blocks of Go Concurrency
-
Concurrency Patterns
Running the following commands can help
go run -race sample.go // compile and run the program
go test -race sample // test the package
go build -race sample // build the command
go install -race sample // install the package
It is a tool that was created at Google and can display profile data either while a program is running or by consuming saved runtime statistics. More
It helps to answer these questions;
- How your program is performing at runtime
- How many goroutines are running
- If the CPU is being fully utilized
- Memory consumption by the program