I'm a little shocked to see Java compare so closely with C code. Java programs are compiled to virtual machine bytecode and runs on the JVM. I expect this to be much slower than C. Maybe it applies only for this limited set of benchmark tests? Or is it because the slowdown does not include the jvm overhead?
Python's and PHP's performance is similar to expectation.
caiqifang
Yes, I am also surprised to see this benchmark. What are some of the features that help Java reaching the performance as a C++ program? Can you force compile the entire program to bytecode before running? Are you allowed to touch and finely tune the JVM?
yey1
@ferozenaina JVM implementation is much better now. And I think they may not count the JVM launch time.
I'm a little shocked to see Java compare so closely with C code. Java programs are compiled to virtual machine bytecode and runs on the JVM. I expect this to be much slower than C. Maybe it applies only for this limited set of benchmark tests? Or is it because the slowdown does not include the jvm overhead?
Python's and PHP's performance is similar to expectation.
Yes, I am also surprised to see this benchmark. What are some of the features that help Java reaching the performance as a C++ program? Can you force compile the entire program to bytecode before running? Are you allowed to touch and finely tune the JVM?
@ferozenaina JVM implementation is much better now. And I think they may not count the JVM launch time.