Dear Researchers,
I am using CPLEX Optimizer to solve a MIP model, and the computation time is big. When I checked the PC performance, I found that about 0.1% of the CPU and 258 MB are used by the application program of CPLEX (oplide) {Attached}. Whereas, I found another background process called "Java (TM) Platform SE binary" exhausts about 99% of the CPU and 81% of the memory {Attached}.
I have two questions:
1-What the difference between the two actions, namely oplide and Java?
2-Is there any tricky way to share the hard disk memory or any external memory with the RAM?
Any answer is appreciated.
Best Regards,
Mahmoud
That application is written in Java. Therefore, the JVM is what is actually using up your memory and doing the "hard" work as instructed by that program, which does not need to actually have the memory allocated to itself (process).
This is normal for the Java application architecture.
What you can do is try to configure your Java Runtime Environment and allow it to use up more RAM if you wish to make it possible for Java programs to be able to use more RAM.
This does not bean that applications will allocate more memory.
Applications allocate as much memory as they need were programmed to allocate in order to perform a certain task. You cannot force an application to arbitrarily allocate more memory if it does not need it. Maybe the algorithm is written in such a way to work with a certain portion of data at the time due to some limitations of the method of implementation that you might not be aware of that are very technical, but are the reason why the program uses as much as it does and not more.
Still, allowing the JVM and the JRE more memory in the heap and overall might allow the app to work better.
About your question regarding sharing hard disk in place of RAM or as additional RAM. This is possible, but will probably make things worse because hard drives are much slower than RAM, so it will end up slowing down your program.
Dear Milan,
Thank you very much for your answer, I really appreciate that.
Indeed, you can pass the -Xmx parameter to the JVM. For instance, -Xmx4000m would allow 4000 Mo to the heap. Note that you need a 64bit JVM if you want to allocate more than (approximatively) 1500 Mo of RAM to the heap.
The -Xmx option set the maximum heap. Less could be allocated when the app is started, and the JVM will take additional RAM when needed. This operation can be time consuming. Therefore, it could be advisable to also specify the minimum heap to allocate to the JVM, using the -Xmm option. For instance "-Xmx4000m -Xmm2000m" will start the app with 2000Mo of heap, and it could claim up to 4000Mo.
- Badges
- Science topic
- Similar topics
- Linguistics
- Composition Studies
- Publication