“every millisecond lost results in $100m per annum in lost opportunity”. You should also be posting your sources so we can run a diff on them to see what you've actually done to 3.5.4 or 3.6.2/3-rt. Subjectivity doesn't sit well, for very obvious reasons. You should also be posting metrics you go with your numbers, you've been asked by people here, LKML (where you admitted that all your tweaks we based on nothing more than your own subjective observations). and really, you act like your a kernel developer (who THINKS he knows better than anyone else), yet you don't even know how to collect logging information? wtf?īut do feel free to actually make counter arguments (if you are able to do that), so far it just seems like you revert to childish remarks that are nothing more than religious fanaticism or just total rubbish, when asked about things or confronted. 3rdly, i have doubts in your superior knowledge claims, when just looking at your rt-linux-user-list posts it's pretty obvious you are asking for help with some pretty silly mistakes (like the rootdevice thing).

2ndly, as far as i know, you cannot compile CONFIG_PREEMPT_RT_FULL=y without threadedirqs, but feel free to post a citation/link where people are doing that. it just comes across as (you) not really knowing what you are doing.

it's really all thaty matters LOL.Īnd again, you already said RT was pretty much deprecated in your mind because your 3.5.4-low jitter kernel, yet here a week or so later you now are basing your LJ kernel around it - with all of your superior knowledge to everyone else, why would you do such contradictory things.

Really all we should care about is getting an extra 2 FPS in Doom3 - because doom3 is such an important part of computing. but I guess that averager is pointless - since people don't actually use multitasking with many demanding applications on modern computers (ie: not C64/amiga days). (on 2 systems much higher specs than your core2duo - ie: i7 and Phenom II x4). The system will halt badly / unresponsive when suspending VMs for a period of time - without touching that code, the problem does not exist at all. You know what happens when you make your changes to fair.c (ridding the 10ms averager bit) when using VMware? I do. okay Genius, so you claim no one has a clue on anything except (unsurprisingly) you. it's funny how you jump to the conclusion that i did though ~ in the future you might want to stop for a second or two and actually put on your critical thinking hat before you speak, since that has pretty much zero relevance to anything anyway. and yet i never said a god damn thing about sunnism or any of that other non-sense.nor did i associate you with it.

Afterfall insanity wikia