www.digitalmars.com         C & C++   DMDScript  

digitalmars.D.announce - DConf 2013 Day 3 Talk 4: LDC by David Nadlinger

reply Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
You know the drill!

reddit: 
http://www.reddit.com/r/programming/comments/1gie4b/dconf_2013_ldc_the_llvmbased_d_compiler_by_david/

hackernews: https://news.ycombinator.com/item?id=5892652

facebook: https://www.facebook.com/dlang.org/posts/658638807483137

twitter: https://twitter.com/D_Programming/status/346598441230671873

youtube: http://youtube.com/watch?v=ntdKZWSiJdY


Andrei
Jun 17 2013
next sibling parent "Alex =?UTF-8?B?UsO4bm5l?= Petersen" <alex lycus.org> writes:
Great talk!

Regarding the ci.lycus.org fleet, credit should definitely go to 
Adam Wilson (C# to D talk) and Kelly Wilson (same person who was 
present in the pie chart) too for providing many of the machines 
hooked up to the master node.

The fleet doesn't do a whole lot of work most of the time, so if 
you have a project that

1) has a sane build system;
2) you're willing to respond to build failures on;
3) and is 'significant' enough,

feel free to email me and I'll see what I can do.

(By 'significant' I mean "has enough impact to be useful for a 
reasonable amount of D programmers". This is of course pretty 
subjective, but we have to be a bit conservative about how many 
projects we add so that we don't end up having lots of stalled 
builds in the queue.)
Jun 17 2013
prev sibling next sibling parent reply "bearophile" <bearophileHUGS lycos.com> writes:
Andrei Alexandrescu:

 http://www.reddit.com/r/programming/comments/1gie4b/dconf_2013_ldc_the_llvmbased_d_compiler_by_david/
Slide 14: PFFT (SSE) seems slow on LDC2: if you can extract a small test case LLVM devs will appreciate a lot a bug report (they fixed many lacks of optimizations submitted by me). If you have a link to the PFFT code them maybe I can do that myself. Slide 25:
 Implicit invariants often hard to track down
Then maybe it's a good idea to add such invariants to the dmd front-end code, even before its port to D. Bye, bearophile
Jun 17 2013
parent reply "nazriel" <spam dzfl.pl> writes:
On Monday, 17 June 2013 at 13:47:20 UTC, bearophile wrote:
 Andrei Alexandrescu:

 http://www.reddit.com/r/programming/comments/1gie4b/dconf_2013_ldc_the_llvmbased_d_compiler_by_david/
Slide 14: PFFT (SSE) seems slow on LDC2: if you can extract a small test case LLVM devs will appreciate a lot a bug report (they fixed many lacks of optimizations submitted by me). If you have a link to the PFFT code them maybe I can do that myself.
I don't know what PFFT stands for (can't google it either, funny results shows up) but if it related to vectorization then maybe LDC has been slower because it was built against LLVM 3.3 while LLVM 3.4 brings more vector optimizations. Maybe all what has to be done, is rerunning benchmarks against LDC + LLVM 3.4 ?
 Slide 25:
 Implicit invariants often hard to track down
Then maybe it's a good idea to add such invariants to the dmd front-end code, even before its port to D. Bye, bearophile
Jun 17 2013
next sibling parent reply Justin Whear <justin economicmodeling.com> writes:
On Mon, 17 Jun 2013 17:41:22 +0200, nazriel wrote:

 On Monday, 17 June 2013 at 13:47:20 UTC, bearophile wrote:
 Andrei Alexandrescu:

 http://www.reddit.com/r/programming/comments/1gie4b/
dconf_2013_ldc_the_llvmbased_d_compiler_by_david/
 Slide 14:

 PFFT (SSE) seems slow on LDC2: if you can extract a small test case
 LLVM devs will appreciate a lot a bug report (they fixed many lacks of
 optimizations submitted by me). If you have a link to the PFFT code
 them maybe I can do that myself.
I don't know what PFFT stands for (can't google it either, funny results shows up) but if it related to vectorization then maybe LDC has been slower because it was built against LLVM 3.3 while LLVM 3.4 brings more vector optimizations. Maybe all what has to be done, is rerunning benchmarks against LDC + LLVM 3.4 ?
 Slide 25:
 Implicit invariants often hard to track down
Then maybe it's a good idea to add such invariants to the dmd front-end code, even before its port to D. Bye, bearophile
My guess is Parallel Fast Fourier Transform.
Jun 17 2013
parent "jerro" <a a.com> writes:
On Monday, 17 June 2013 at 15:56:21 UTC, Justin Whear wrote:
 On Mon, 17 Jun 2013 17:41:22 +0200, nazriel wrote:

 On Monday, 17 June 2013 at 13:47:20 UTC, bearophile wrote:
 Andrei Alexandrescu:

 http://www.reddit.com/r/programming/comments/1gie4b/
dconf_2013_ldc_the_llvmbased_d_compiler_by_david/
 Slide 14:

 PFFT (SSE) seems slow on LDC2: if you can extract a small 
 test case
 LLVM devs will appreciate a lot a bug report (they fixed many 
 lacks of
 optimizations submitted by me). If you have a link to the 
 PFFT code
 them maybe I can do that myself.
I don't know what PFFT stands for (can't google it either, funny results shows up) but if it related to vectorization then maybe LDC has been slower because it was built against LLVM 3.3 while LLVM 3.4 brings more vector optimizations. Maybe all what has to be done, is rerunning benchmarks against LDC + LLVM 3.4 ?
 Slide 25:
 Implicit invariants often hard to track down
Then maybe it's a good idea to add such invariants to the dmd front-end code, even before its port to D. Bye, bearophile
My guess is Parallel Fast Fourier Transform.
It's Pretty Fast Fourier Transform. The code is at https://github.com/jerro/pfft/tree/experimental (I linked to the experimental branch because master branch is quite outdated)
Jun 17 2013
prev sibling parent "Kai Nacke" <kai redstar.de> writes:
On Monday, 17 June 2013 at 15:41:24 UTC, nazriel wrote:
 On Monday, 17 June 2013 at 13:47:20 UTC, bearophile wrote:
 Andrei Alexandrescu:

 http://www.reddit.com/r/programming/comments/1gie4b/dconf_2013_ldc_the_llvmbased_d_compiler_by_david/
Slide 14: PFFT (SSE) seems slow on LDC2: if you can extract a small test case LLVM devs will appreciate a lot a bug report (they fixed many lacks of optimizations submitted by me). If you have a link to the PFFT code them maybe I can do that myself.
I don't know what PFFT stands for (can't google it either, funny results shows up) but if it related to vectorization then maybe LDC has been slower because it was built against LLVM 3.3 while LLVM 3.4 brings more vector optimizations. Maybe all what has to be done, is rerunning benchmarks against LDC + LLVM 3.4 ?
As far as I followed the LLVM commit list, the change is that the loop vectorizer is now enabled at -O2 (instead of -O3). I don't know the options David has used but I assume something like "ldmd2 -O -inline -release". "-O" from ldmd2 maps to -O3 and therefore enables the loop vectorizer. Regards Kai
 Slide 25:
 Implicit invariants often hard to track down
Then maybe it's a good idea to add such invariants to the dmd front-end code, even before its port to D. Bye, bearophile
Jun 18 2013
prev sibling next sibling parent reply Andrej Mitrovic <andrej.mitrovich gmail.com> writes:
On 6/17/13, Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> wrote:
 youtube: http://youtube.com/watch?v=ntdKZWSiJdY
There seems to be some audio glitching every couple of seconds (at the beginning). I've noticed this in other videos as well. It's mostly minimal though, not much harm done.
Jun 17 2013
next sibling parent reply "Steven Schveighoffer" <schveiguy yahoo.com> writes:
On Mon, 17 Jun 2013 11:19:14 -0400, Andrej Mitrovic  
<andrej.mitrovich gmail.com> wrote:

 On 6/17/13, Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> wrote:
 youtube: http://youtube.com/watch?v=ntdKZWSiJdY
There seems to be some audio glitching every couple of seconds (at the beginning). I've noticed this in other videos as well. It's mostly minimal though, not much harm done.
I noticed that during the conference. There were several talks where the mic was giving the AV guys trouble. Each of the speakers was using a clip-on remote mic. Of course, live, it wasn't as big a deal, as we could hear the person talking :) But it would be more glaring for the recording. I'm sure if there are any glitches that omit an important piece of the talk, point them out and the speaker and/or conference attendees can help discern what was being said. -Steve
Jun 17 2013
parent Walter Bright <newshound2 digitalmars.com> writes:
On 6/17/2013 8:35 AM, Steven Schveighoffer wrote:
 I'm sure if there are any glitches that omit an important piece of the talk,
 point them out and the speaker and/or conference attendees can help discern
what
 was being said.
I was saying "flip that!".
Jun 17 2013
prev sibling next sibling parent reply "Joseph Rushton Wakeling" <joseph.wakeling webdrake.net> writes:
On Monday, 17 June 2013 at 15:19:27 UTC, Andrej Mitrovic wrote:
 There seems to be some audio glitching every couple of seconds 
 (at the
 beginning). I've noticed this in other videos as well. It's 
 mostly
 minimal though, not much harm done.
https://www.youtube.com/watch?v=m6jsXQm5IrM#t=106s :-)
Jun 17 2013
parent Marco Leise <Marco.Leise gmx.de> writes:
Am Mon, 17 Jun 2013 22:03:09 +0200
schrieb "Joseph Rushton Wakeling"
<joseph.wakeling webdrake.net>:

 On Monday, 17 June 2013 at 15:19:27 UTC, Andrej Mitrovic wrote:
 There seems to be some audio glitching every couple of seconds 
 (at the
 beginning). I've noticed this in other videos as well. It's 
 mostly
 minimal though, not much harm done.
https://www.youtube.com/watch?v=m6jsXQm5IrM#t=106s :-)
"What's this wire doing here? It's dangerous!" *pulls* -- Marco
Jun 17 2013
prev sibling parent "David Nadlinger" <code klickverbot.at> writes:
On Monday, 17 June 2013 at 15:19:27 UTC, Andrej Mitrovic wrote:
 On 6/17/13, Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> 
 wrote:
 youtube: http://youtube.com/watch?v=ntdKZWSiJdY
There seems to be some audio glitching every couple of seconds (at the beginning). I've noticed this in other videos as well. It's mostly minimal though, not much harm done.
Yes, there were, and it was rather irritating for me as well. Actually, they even changed my wireless beltpack during the presentation, at the point where there is pretty much silence in the video and they cut to a shot of the bored audience. David
Jun 18 2013
prev sibling next sibling parent Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
On 6/17/13 8:25 AM, Andrei Alexandrescu wrote:
 You know the drill!

 reddit:
 http://www.reddit.com/r/programming/comments/1gie4b/dconf_2013_ldc_the_llvmbased_d_compiler_by_david/


 hackernews: https://news.ycombinator.com/item?id=5892652

 facebook: https://www.facebook.com/dlang.org/posts/658638807483137

 twitter: https://twitter.com/D_Programming/status/346598441230671873

 youtube: http://youtube.com/watch?v=ntdKZWSiJdY


 Andrei
HD video up: https://archive.org/details/dconf2013-day03-talk04 Andrei
Jun 17 2013
prev sibling next sibling parent Nick Sabalausky <SeeWebsiteToContactMe semitwist.com> writes:
On Mon, 17 Jun 2013 08:25:50 -0400
Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> wrote:

 You know the drill!
 
 reddit: 
 http://www.reddit.com/r/programming/comments/1gie4b/dconf_2013_ldc_the_llvmbased_d_compiler_by_david/
 
 hackernews: https://news.ycombinator.com/item?id=5892652
 
 facebook: https://www.facebook.com/dlang.org/posts/658638807483137
 
 twitter: https://twitter.com/D_Programming/status/346598441230671873
 
 youtube: http://youtube.com/watch?v=ntdKZWSiJdY
 
 
 Andrei
Torrents/links up: http://semitwist.com/download/misc/dconf2013/
Jun 17 2013
prev sibling next sibling parent "Joseph Rushton Wakeling" <joseph.wakeling webdrake.net> writes:
... slightly more serious response: really nice talk, David, and 
thanks for the mention of Dregs. :-)
Jun 18 2013
prev sibling next sibling parent Jacob Carlborg <doob me.com> writes:
On 2013-06-17 14:25, Andrei Alexandrescu wrote:
 You know the drill!

 reddit:
 http://www.reddit.com/r/programming/comments/1gie4b/dconf_2013_ldc_the_llvmbased_d_compiler_by_david/


 hackernews: https://news.ycombinator.com/item?id=5892652

 facebook: https://www.facebook.com/dlang.org/posts/658638807483137

 twitter: https://twitter.com/D_Programming/status/346598441230671873

 youtube: http://youtube.com/watch?v=ntdKZWSiJdY
David mentions in the talks that git submodules make it more complicated to do merges. I'm not sure I understand why. git submodules are just regular repositories that are included in other repositories. One can work on these repositories separately, then it shouldn't be much difference. -- /Jacob Carlborg
Jun 19 2013
prev sibling parent Jacob Carlborg <doob me.com> writes:
On 2013-06-17 14:25, Andrei Alexandrescu wrote:
 You know the drill!

 reddit:
 http://www.reddit.com/r/programming/comments/1gie4b/dconf_2013_ldc_the_llvmbased_d_compiler_by_david/


 hackernews: https://news.ycombinator.com/item?id=5892652

 facebook: https://www.facebook.com/dlang.org/posts/658638807483137

 twitter: https://twitter.com/D_Programming/status/346598441230671873

 youtube: http://youtube.com/watch?v=ntdKZWSiJdY
About the Mac OS X support. Is TLS the only problem on Snow Leopard? Have you considered moving the code dealing with TLS from the dynamic linker into the executable? -- /Jacob Carlborg
Jun 19 2013