IRC logs for #openrisc Sunday, 2014-11-16

--- Log opened Sun Nov 16 00:00:30 2014
LimbWell glad to see fusesoc still builds my nexys4 adaption without complaining. Now to make the processor work04:24
poke53281olofk: I have read a little bit about asm.js. And it looks like, that calls into and out of asm.js modules must go through a ffi. Then it's clear, why it is slow.06:49
olofkI feel sorry for JavaScript sometimes. It gets forced to do all these things that it was not really supposed to do08:23
olofkIt was probably looking forward to taking it easy and getting retired08:24
stekernolofk: this was the talk I was mentioning during orconf:
stekernyour last 2 sentences pretty much sums up that talk08:29
poke53281Yes, I like the video. He was so close in mentioning jor1k :) .08:34
poke53281The pitty is, that we had already a good solution. Java and Javabytecode.08:36
poke53281But unfortunately, Microsoft never installed it as default. And it was buggy and it still is. So it died.08:37
poke53281Then we had Flash. The same thing. But at least everyone installed it. And it survived, because it was the only way to watch videos.08:37
poke53281During those solutions Javascript and especially HTML evolved.08:39
poke53281And after Apple kicked flash out of their devices Javascript remained as the smallest common denominator.08:40
poke53281There are other solutions like Googles PNaCL.08:41
poke53281But no one except Chrome supports it. And not even Chrome. You have to activate somewhere.08:42
poke53281Everything Microsoft does is not accepted by the community. So forget Typescript.08:43
poke53281The only thing that remains is Javascript. And with asm.js it can be fast. At least in principle. You can see the asm.js code as some sort of bytecode, which is strangely compatible to Javascript. And Microsoft and Google might support it. At least both parties think about it.08:45
poke53281But as I said in my talk. asm.js is in principle just syntax checking if the JIT is good, it will optimize the code anyhow close to "native " speed.08:48
poke53281At least the IE came already close to Firefox.08:49
poke53281olofk: I just want to say. I hope, that Javascript will die very soon and that we will have very good standardized alternative.08:51
poke53281But at the moment I don't see an alternative. But now we have more and more of these compilers with language backend (thanks to LLVM) (see emscripten). So at least there is a solution. A solution, which we wouldn't need if the $ยง"$%& companies would work together on this.08:56
mor1kx[mor1kx] bandvig pushed 1 new commit to withfpu:
mor1kxmor1kx/withfpu 9facfe0 Andrey Bacherov: 1) split potentially asynchronous rst and synchronous flush_i...17:12
--- Log closed Mon Nov 17 00:00:32 2014

Generated by 2.15.2 by Marius Gedminas - find it at!