RPi 2 performance
With the Raspberry Pi 2 is it better to go raspbian or snappy for Java performance?
https://github.com/tinspin/rupy - A tiny Java async HTTP application server.
-
DougieLawson
- Posts: 42157
- Joined: Sun Jun 16, 2013 11:19 pm
- Location: A small cave in deepest darkest Basingstoke, UK
Re: RPi 2 performance
Raspbian every time, especially if you've never used a RPi before.
Snappy won't give you what you think you're getting in just the same way that the Windows stuff won't.
Snappy won't give you what you think you're getting in just the same way that the Windows stuff won't.
Languages using left-hand whitespace for syntax are ridiculous
DMs sent on https://twitter.com/DougieLawson or LinkedIn will be answered next month.
Fake doctors - are all on my foes list.
The use of crystal balls and mind reading is prohibited.
DMs sent on https://twitter.com/DougieLawson or LinkedIn will be answered next month.
Fake doctors - are all on my foes list.
The use of crystal balls and mind reading is prohibited.
Re: RPi 2 performance
Could you elaborate it a bit more for me please? 
(not questioning it - but just curious why)
(not questioning it - but just curious why)
-
DougieLawson
- Posts: 42157
- Joined: Sun Jun 16, 2013 11:19 pm
- Location: A small cave in deepest darkest Basingstoke, UK
Re: RPi 2 performance
Because Raspbian is the general purpose operating system that includes all of the packages you need and currently (because I'm waiting for the postie to deliver some new SDCards) Ubuntu is an unknown beast.
Raspbian is also runs on A,B,B+ & A+ so everything you do on it is backwards compatible for the folks who don't have (and don't desire) the new hardware.
Raspbian is also runs on A,B,B+ & A+ so everything you do on it is backwards compatible for the folks who don't have (and don't desire) the new hardware.
Languages using left-hand whitespace for syntax are ridiculous
DMs sent on https://twitter.com/DougieLawson or LinkedIn will be answered next month.
Fake doctors - are all on my foes list.
The use of crystal balls and mind reading is prohibited.
DMs sent on https://twitter.com/DougieLawson or LinkedIn will be answered next month.
Fake doctors - are all on my foes list.
The use of crystal balls and mind reading is prohibited.
Re: RPi 2 performance
All valid points. Thanks!
Re: RPi 2 performance
Would a Java VM compiled against the Pi2 and its ARM7-Quadcore Cortex-A7 also bring some advantages compared to Raspbian's current Oracle Java8 which is optimised but compiled against ARM6/7 (if it means it uses ARM6 instructions only) ?
(I mean looking at a single core only. Obviously more cores help Java and its threads generally, and coming from Unix this has been so since the first JVMs.)
Since the JVM is also a kind of emulator, it could benefit from an ARM7 compilation. :-)
(I mean looking at a single core only. Obviously more cores help Java and its threads generally, and coming from Unix this has been so since the first JVMs.)
Since the JVM is also a kind of emulator, it could benefit from an ARM7 compilation. :-)
Re: RPi 2 performance
As completely ignorant in the matter - I though that JVM for ARM (deployed to RPi as well as some other ARM platforms) is already capable of utilising ARM7 instruction set (and its Java extension) when/if available. There is no need for extra 're-compilation' for platform.
Generally speaking JVM (at least Oracle's for sure) is not just 'emulator' - it is far, far smarter than that. It profiles code it executes and in 'background' it compiles most executed paths to the target platform (hence ARM6/ARM7 instruction set explicitly extended for Java being really helpful). So, there's no need for explicit ARM7 version of JVM as 'interpreter' part of the is not the bottle neck - but on-the-fly compiled code benefit extra from ARM7 extended instruction set...
Please guys - do correct me where I was wrong in above statements.
Generally speaking JVM (at least Oracle's for sure) is not just 'emulator' - it is far, far smarter than that. It profiles code it executes and in 'background' it compiles most executed paths to the target platform (hence ARM6/ARM7 instruction set explicitly extended for Java being really helpful). So, there's no need for explicit ARM7 version of JVM as 'interpreter' part of the is not the bottle neck - but on-the-fly compiled code benefit extra from ARM7 extended instruction set...
Please guys - do correct me where I was wrong in above statements.
Re: RPi 2 performance
This was specifically what I was wondering about.clicky wrote:As completely ignorant in the matter - I though that JVM for ARM (deployed to RPi as well as some other ARM platforms) is already capable of utilising ARM7 instruction set (and its Java extension) when/if available. There is no need for extra 're-compilation' for platform.
Someday maybe someone from Oracle will browse by the forum.
https://github.com/tinspin/rupy - A tiny Java async HTTP application server.
Re: RPi 2 performance
From JDK 8 for ARM Downloads (http://www.oracle.com/technetwork/java/ ... 87472.html):clicky wrote:
As completely ignorant in the matter - I though that JVM for ARM (deployed to RPi as well as some other ARM platforms) is already capable of utilising ARM7 instruction set (and its Java extension) when/if available. There is no need for extra 're-compilation' for platform.
JDK 8 for ARM is supported on systems based on 32-bit ARM v6 or ARM v7 running Linux.
Happy coding!
Re: RPi 2 performance
Thanks. Also, I think I've read somewhere that ARM7 has better support for JVM than ARM6 and should offer some performance improvements... And that original JVM is capable of using those enhancements out of the box. Now, what enhancements are those I forgot... O: )