plugin performance on difference machines...

hi all,

i am working on several projects at once, one is my starcraft ai. in terms of project size, i am not sure what to count, so i'll give you the compilation times ;)

AI: ~45 seconds

real world project/at work (consists of multiple sbt projects): about 5 minutes

in my ai project, the plugin is very fast, i never have to wait. for the big project, i *usually* can work without waiting for the ide on my machine at home (4.6ghz 6700k) more than a few seconds even in what i imagine to be a complicated scenario for code completion. at work i only have a notebook (two actually, haswell, 2.7ghz & 3.5ghz) and those are *significantly* slower, the 2.7 one noticeable more than the 3.5 one). the felt speed is 25% of my desktop machine on the slow one.

is there a big difference between mobile and desktop cpus, apart from the clock speed? is skylake just better at compilation/compilation like tasks/what the scala plugin is doing? i use copies of the same VM on all machines, all use windows 10 outside and ubuntu 14 inside, so the software should not be the cause.


1 comment
Comment actions Permalink

"is there a big difference between mobile and desktop cpus, apart from the clock speed?"

Recently it was always the clock speed that was the major one. There may be also differences like more cores in a desktop CPU (e.g. 6 or 8 instead of 4) or more L3 cache. But core architecture, L1, L2 cache sizes are exactly the same. Probably not enough to explain the 4x difference you're seeing.

Some high-end Skylake CPUs have big edram L4 cache. Maybe this is causing the difference?

I'm using an old Sandy Bridge 2.4 GHz 4-core i7 mobile processor and I must admit the performance of Scala editing is quite ok. Incremental compilation does magic.


Please sign in to leave a comment.