Chinchilla: Tcsc: was actually talking to someone in golang about how a project where they escaped gc by being very careful with what they created, in a C# proj
Gentleman: Even caching the length on a for loop is sometimes slower in JS.
Elvis: You can improve the determinism by a decent amount, but microoptimizations are usually not productive in js
Doria: That’s a really difficult moving target Gentleman :
Elvis: But honestly if you use the array iteration methods a lot you end up with a ton of noise on the profiler
Gentleman: On which profiler, there’s not a “the”
Elvis: Is there any where that’s not the case?
Elvis: Because i’d be interested. even in probably-not-performance critical areas i tend to avoid them all-together because of this.
Gentleman: I don’t use “profilers” like that, i usually just print timestamps and figure out the slow parts.
Gentleman: Certainly more functions will give you more stack frames to look at
Gentleman: But i ***ume you can just filter out any function that’s running faster than some threshold
Gentleman: Ie, i’d only care about the slowest 10%, say.
Elvis: You can, but it’s not obvious what those are if theres a ton of Array.prototype.forEach in there.
Condict: Has anyonre used this and can recommend it, or a better alternative? https://github.com/dapuck/node-svn-interface
Elvis: Gentleman: especially because the function inside those usually just shows up as closure
Mbamalu: 2 click event functions are targeting something on my site and obviously only one is happening. Is there a way to remove just that one so the other works? Specifically: on a WP site the theme has a function in a custom.js for smooth scrolling to anchor links and it’s overwriting a drag-n-drop Visual Composer builder’s “Tabs” functionality. I can post both functions or a link if it helps.
Elvis: But time stamps are effective and i do those a lot because you can use them to make a chart better though
Elvis: Which you can look at without having to open up the debugger, which is nice since opening the debugger effects the performance a lot.
Elvis: Baxx: that video doesnt seem to work. what are you seeing?
Defusco: Http://sendvid.com/7n0vmp8u really? Weird – it’s all black in firefox
Elvis: Yeah it works in chrome
Elvis: Should those be displaying the same thing?
Defusco: What that video doesn’t work in firefox? haha, ace. Yeah they should – I just zoomed in or something in Firefox and it’s been weird since
Kotzen: Tcsc, i do asm/simd coding too. were i look at inter instruction depedancies and latencies. sometimes timing tiny things in a profiler often gives bad results as the profiler itself gets in the way. timing and charting is the best option there too!
Defusco: I’ve tried clearing hist / ingognito etc but it’s still like it, not too sure why
Elvis: Baxx: i would hit ctrl + 0
Elvis: Or whatever it is to reset zoom
Defusco: Yeah it’s still funky after that
Elvis: Zap: yeah, i havent done any simd programming in a while but have done it in the past.
Defusco: Yeah think C-0 did the job, nice one
Boken: Ham is tastier than hamsters
Defusco: My sisters hamster died and the other one started eating it
Valerius: I’ve no personal experience with that so i cannot comment
Staberg: Why is line 9 of the JS code unable to read enum_text? http://jsfiddle.net/72wqm40a/3/
Boken: This npm issue I have is super weird, I have no idea how it manages to make my network go down
Elvis: It was painful to find out that kronos group named their new shader language IR SPIR-V, despite it being largely unrelated to SPIR. remaking the mistakes of js vs java
Batie: Is it possible to select an html element based on its innerhtml?