Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

24-core CPU and I can’t move my mouse

randomascii.wordpress.com

Post reply on HN
Print view View original

151–160 of 509 posts

Re: 24-core CPU and I can’t move my mouse

#151

Post by JabavuAdams »

mpweiher wrote: post #105
titzer wrote: post #45

Full disclosure: I work for Google on Chrome. A Chrome build is truly a computational load to be reckoned with. Without the distributed build, a from-scratch build of Chrome will take at least 30 minutes on a Macbook Pro--maybe an hour(!). TBH I don't remember toughing out a full build without resorting to goma. Even on a hefty workstation, a full build is a go-for-lunch kind of interruption. It will absolutely own a…

> Damn, are we doing this wrong? Yes, and I mean "we" as in "this industry". I just recently talked to someone whose Swift framework(s) were compiling at roughly 16 lines per second. Spurred on by Jonathan Blow's musings on compile times for Jai[1], I started tinkering with tcc a little. It compiled a generated 200KLOC file (lots of small functions) in around 200 ms. Then there are Smalltalk and Lisp systems that are…

I used to be a C++ guy for 20 years, but won't go back unless absolutely necessary. I mean, it's tempting -- there are some cool new language features. I find that the abstractions are leaky, though, so you still have to understand all the hairy edge-cases. The language has gone insane. I'm 40 -- I want to get stuff done before I die, not play clever games to get around my language / system.

Re: 24-core CPU and I can’t move my mouse

#152

Post by jhasse »

pjc50 wrote: post #66
titzer wrote: post #45

Full disclosure: I work for Google on Chrome. A Chrome build is truly a computational load to be reckoned with. Without the distributed build, a from-scratch build of Chrome will take at least 30 minutes on a Macbook Pro--maybe an hour(!). TBH I don't remember toughing out a full build without resorting to goma. Even on a hefty workstation, a full build is a go-for-lunch kind of interruption. It will absolutely own a…

"Concatenated" builds seem to be the best band-aid for this. Concatenate as many source files as possible before compiling, #include a bunch of cpp files into one big file. It makes tracking down errors slightly harder and macros a bit more risky, but greatly improves the overall build efficiency.

Also called unity build sometimes.

Re: 24-core CPU and I can’t move my mouse

#153

Post by jhasse »

claudiug wrote: post #83
titzer wrote: post #45

Full disclosure: I work for Google on Chrome. A Chrome build is truly a computational load to be reckoned with. Without the distributed build, a from-scratch build of Chrome will take at least 30 minutes on a Macbook Pro--maybe an hour(!). TBH I don't remember toughing out a full build without resorting to goma. Even on a hefty workstation, a full build is a go-for-lunch kind of interruption. It will absolutely own a…

Have you give a look to dlang? or even rust

Rust's compile times are even worse.

Re: 24-core CPU and I can’t move my mouse

#154

Post by rbanffy »

hoodoof wrote: post #53

Saw the headline and thought "must be Windows". I'm not a Windows hater, but one of my long standing gripes about Windows is that it just seems to have terrible multitasking compared to OSX. I'm sure there are reasons but it just seems utterly symbolic of Microsoft that they never managed to get Windows to multitask in a rock solid, smooth and reliable way like OSX.

> my long standing gripes about Windows is that it just seems to have terrible multitasking compared to OSX.

Using both Macs and Linux laptops, I'm sometimes shocked at how the Mac sometimes locks up when the Linux machines degrade much more gracefully under heavy loads. I never dug too deep into it, but it feels like it's something with HFS+ under heavy IO. I hope APFS fixes that.

Re: 24-core CPU and I can’t move my mouse

#155

Post by virtualwhys »

jmartinpetersen wrote: post #68
torrent-of-ions wrote: post #38

Earlier quoted context omitted.

Plus ça change, plus c'est la même chose...

"The more it changes, the more it’s the same thing." [0] [0]: https://en.wiktionary.org/wiki/plus_%C3%A7a_change,_plus_c%2...

For English the equivalent meaning would be, "the more things change, the more they remain the same".

I don't think I've ever heard the literal translation in English.

Re: 24-core CPU and I can’t move my mouse

#156

Post by ordu »

mysterydip wrote: post #131
titzer wrote: post #45

Full disclosure: I work for Google on Chrome. A Chrome build is truly a computational load to be reckoned with. Without the distributed build, a from-scratch build of Chrome will take at least 30 minutes on a Macbook Pro--maybe an hour(!). TBH I don't remember toughing out a full build without resorting to goma. Even on a hefty workstation, a full build is a go-for-lunch kind of interruption. It will absolutely own a…

Makes me wonder about a distributed compilation. If everyone had a monster workstation but not everyone compiles at the same time, a theoretical networked compiler (which may already exist as I haven't really checked) could spread the files out among available workstations and bring things back together near the end. As the largest issue is the throwing away of duplicate work, I'd see it as a kind of reverse binary t…

distcc: https://en.wikipedia.org/wiki/Distcc

Re: 24-core CPU and I can’t move my mouse

#157

Post by ComodoHacker »

I remember every other book on Windows programming saying "Process creation/destruction is expensive, use thread pools (or at least process pools) instead, that's the way to go on Windows". Perhaps this mindset is ingrained for Windows QA team too - they don't have [enough] test cases for such scenarios.

Re: 24-core CPU and I can’t move my mouse

#158

Post by masklinn »

qb45 wrote: post #100
masklinn wrote: post #34

Earlier quoted context omitted.

> NtGdiCloseProcess has a system-wide global lock which is used quite often "Holds" rather than "has", and importantly that system-wide lock must be held by things like SendMessageW (which sends a message and waits for its processing before returning) which is pretty critical for UI updates. This is compounded by parallelised build processes as not only do they spawn lots of processes leading to lots of process destr…

> "Holds" rather than "has", and importantly that system-wide lock must be held by things like SendMessageW (which sends a message and waits for its processing before returning) which is pretty critical for UI updates. Do you happen to know anything about it? I'm scratching my head how it's possible that process termination serializes with GUI... Maybe it hogs some lock on process descriptors which SendMessage also n…

> Do you happen to know anything about it? I'm scratching my head how it's possible that process termination serializes with GUI…

Not anything more than what's in the essay. But possibly some (pair of) utility function calls were added for e.g. a cleanup or notification to the OS between W7 and W10 which was not noticed at the time.

> I hope you didn't mean to say that every SendMessage call is completely serialized with each other.

It's my understanding that at least a subset of SendMessage is serialised in the kernel yes, and from the essay:

> functions like SendMessageW, apparently waiting on a kernel critical region[…], deep in the call stack in win32kbase.sys!EnterCrit (not shown)

[0] https://msdn.microsoft.com/en-us/library/windows/desktop/dd7...

Re: 24-core CPU and I can’t move my mouse

#159

Post by int_19h »

paulio wrote: post #134
SmellyGeekBoy wrote: post #120

Earlier quoted context omitted.

I live in two worlds at the moment - supporting and maintaining a large PHP application (for which I use phpStorm) and developing .NET applications (for which I, obviously, use VS). I've been a VS user for nigh-on 10 years now and always found the experience really fantastic. I think it's certainly more coherent that the IntelliJ-based IDEs (which, to be fair, are also very good). Anyway, I'm genuinely surprised to s…

"what specifically do you dislike about it?" It's not the tooling I dislike it's the performance. No matter what I throw at it I still get the same experience. It feels like 95% of everything it does is blocking the UI.

Improving perf in VS is hard without massive rewrites. The fundamental problem is that this is originally a COM app (as in, heavily using COM to componentize itself) designed back in mid-90s. Consequently, you get all the wonders of things such as STA apartments, and code that insists running thereon.

As it gets rewritten, new managed bits don't care about all that stuff. But so long as there's one bit of legacy code anywhere in the particular flow that needs to run on STA (usually it's UI thread), you get this whole "you have 20 cores and 60 logical threads, but all those threads need to sync on STA, so everything is serialized and slow" thing.

Even for the new code, the problem is that all those old COM APIs that it needs to interact with (not just for VS itself, but also for the sake of backwards compatibility with third party extensions) are usually synchronous. So if you want background processing, you need to spawn a thread - but, of course, threads aren't free, either.

Re: 24-core CPU and I can’t move my mouse

#160

Post by jug »

TeMPOraL wrote: post #73
anigbrowl wrote: post #47

Earlier quoted context omitted.

But why? What good is an operating system on a multi-core device that allows anything to get that close to the performance envelope? This is a fine example of competition driving change for change's sake rather than real innovation and everything ending up worse as a result. I like new features as much as the next person, but not when they compromise core functionality. Not being able to type is inexcusable .

I agree. However at this point, I don't see anything an OS can do to help. I see plenty of typing slowdowns every other day now. But I'm not sure just how many of them are the OS's fault. When your typing seems to lag, there are two places that can be slowing it down - the input side (reacting to hardware events) and the output side (drawing and updating the UI). I suppose keyboard buffers are pretty well isolated, a…

Agreed, abstraction seem to be exploding these days and I’m not even sure we are at the end of the road yet! Linux or Windows never had any trouble with essentially realtime keyboard feedback in their Terminal windows. It’s not the OS.
Post reply on HN