Re: GHC 2013 in Paris
On Friday, 23 August 2013 at 16:31:17 UTC, Iain Buclaw wrote: I've jsut been told there's no reception on Saturdays, so after 9.00am might be a slight problem, but I think they will leave a number for you to call. I *could* excuse myself to let you in, but that would require relying on me to remember what time you are arriving. :o) Hum... The forum was down yesterday, so I couldn't post this any earlier. I actually only intend to see your talk, so I won't be there before 16:00. I'll be at the gates *before* 15:30, and see if/how I can get it from there. FWI, I'll be the guy who walks with a crutch. I don't if anyone will be seeing this today, but if you don't see me in the building by 15:50, then it means I failed to get in, and am waiting outside...
Re: GHC 2013 in Paris
On 24 August 2013 09:42, monarch_dodra monarchdo...@gmail.com wrote: On Friday, 23 August 2013 at 16:31:17 UTC, Iain Buclaw wrote: I've jsut been told there's no reception on Saturdays, so after 9.00am might be a slight problem, but I think they will leave a number for you to call. I *could* excuse myself to let you in, but that would require relying on me to remember what time you are arriving. :o) Hum... The forum was down yesterday, so I couldn't post this any earlier. I actually only intend to see your talk, so I won't be there before 16:00. I'll be at the gates *before* 15:30, and see if/how I can get it from there. FWI, I'll be the guy who walks with a crutch. I don't if anyone will be seeing this today, but if you don't see me in the building by 15:50, then it means I failed to get in, and am waiting outside... I'm feeling a little rough this morning, so am running late (sorry Eles!) - just having breakfast, and will be down about 10.30... Regards -- Iain Buclaw *(p e ? p++ : p) = (c 0x0f) + '0';
Re: I'm porting some go code to D
On 24 Aug 2013 03:15, Jesse Phillips jesse.k.phillip...@gmail.com wrote: On Friday, 23 August 2013 at 23:54:55 UTC, Rory McGuire wrote: So I'm porting so #golang code to #dlang and there is all these blasted go statements.So I thought I'd give implmenting it in D a shot. What do you guys think? Fire away :). I'd suggest posting long snippets of code to https://gist.github.com/ or http://dpaste.dzfl.pl/ A couple lines is fine but whole implementations, definitely want a more color friendly solution. Good point.
Re: I'm porting some go code to D
Daemonic Threads often end with a segfault, so if your main thread exists, the other threads will probably segfault.
Re: I'm porting some go code to D
On 24 Aug 2013 11:00, David d...@dav1d.de wrote: Daemonic Threads often end with a segfault, so if your main thread exists, the other threads will probably segfault. Thanks, I wonder what they're accessing that they shouldn't.
Re: I'm porting some go code to D
On Friday, 23 August 2013 at 23:54:55 UTC, Rory McGuire wrote: So I'm porting so #golang code to #dlang and there is all these blasted go statements.So I thought I'd give implmenting it in D a shot. What do you guys think? Fire away :). /** * chan allows messaging between threads without having to deal with locks, similar to how chan works in golang */ class chan_(T) { shared Mutex lock; struct Container(T) { T value; Container!T* next; I'm probably missunderstanding somehting about the TLS model, but from what I know, for something like this, shouldn't you make the class instance itself go into shared storage instead instead of all the members?
Re: I'm porting some go code to D
On Friday, 23 August 2013 at 23:54:55 UTC, Rory McGuire wrote: So I'm porting so #golang code to #dlang and there is all these blasted go statements.So I thought I'd give implmenting it in D a shot. What do you guys think? It's a cute idea, but not really practical, I'm afraid – Goroutines are managed by a scheduler in the Go runtime library, whereas D threads directly map to OS threads. Thus, if the Go application you are porting uses many Goroutines (and the Go code I've seen usually does so very liberally), the performance of the D equivalent is going to be horrible. David
Re: I'm porting some go code to D
David Nadlinger: It's a cute idea, but not really practical, I'm afraid – Goroutines are managed by a scheduler in the Go runtime library, whereas D threads directly map to OS threads. Can't Rory McGuire add a scheduler to his code? How much code does it take? Bye, bearophile
Re: I'm porting some go code to D
On 24 Aug 2013 11:25, Moritz Maxeiner mor...@ucworks.org wrote: On Friday, 23 August 2013 at 23:54:55 UTC, Rory McGuire wrote: So I'm porting so #golang code to #dlang and there is all these blasted go statements.So I thought I'd give implmenting it in D a shot. What do you guys think? Fire away :). /** * chan allows messaging between threads without having to deal with locks, similar to how chan works in golang */ class chan_(T) { shared Mutex lock; struct Container(T) { T value; Container!T* next; I'm probably missunderstanding somehting about the TLS model, but from what I know, for something like this, shouldn't you make the class instance itself go into shared storage instead instead of all the members? I have no idea if shared on the class makes all it's parts shared. It was a struct but I had problems with passing it to the spawn func when I was using std.concurrent . I'm trying to port a cassandra cql library from go so it's really just to help with that. I'm having a hard time imagining how to implement select from go. Could probably use std.concurrent.receive because that can handle multiple types at once. Sorry I'm rambling...
Re: SIMD implementation of dot-product. Benchmarks
On Sunday, 18 August 2013 at 05:26:00 UTC, Manu wrote: movups is not good. It'll be a lot faster (and portable) if you use movaps. Process looks something like: * do the first few from a[0] until a's alignment interval as scalar * load the left of b's aligned pair * loop for each aligned vector in a - load a[n..n+4] aligned - load the right of b's pair - combine left~right and shift left to match elements against a - left = right * perform stragglers as scalar Your benchmark is probably misleading too, because I suspect you are passing directly alloc-ed arrays into the function (which are 16 byte aligned). movups will be significantly slower if the pointers supplied are not 16 byte aligned. Also, results vary significantly between chip manufacturers and revisions. I have tried to write fast implementation with aligned loads: 1. I have now idea how to shift (rotate) 32-bytes avx vector without XOP instruction set (XOP available only for AMD). 2. I have tried to use one vmovaps and [one vmovups]/[two vinsertf128] with 16-bytes aligned arrays (previously iterates with a). It works slower then two vmovups (because loop tricks). Now I have 300 lines of slow dotProduct code =) 4. Condition for small arrays works good. I think it is better to use: 1. vmovups if it is available with condition for small arrays 2. version like from phobos if vmovups is not avalible 3. special version for small static size arrays I think version for static size arrays can be easily done for phobos, processors can unroll such code. And dot product optimized for complex numbers can be done too. Best regards Ilya
Re: I'm porting some go code to D
On Saturday, 24 August 2013 at 13:26:32 UTC, bearophile wrote: David Nadlinger: It's a cute idea, but not really practical, I'm afraid – Goroutines are managed by a scheduler in the Go runtime library, whereas D threads directly map to OS threads. Can't Rory McGuire add a scheduler to his code? How much code does it take? It would be very nice to have a builtin scheduler in the runtime. It would make async non-threaded programming so much nicer in D. /Jonas
Re: I'm porting some go code to D
On Sat, Aug 24, 2013 at 3:26 PM, bearophile bearophileh...@lycos.comwrote: David Nadlinger: It's a cute idea, but not really practical, I'm afraid – Goroutines are managed by a scheduler in the Go runtime library, whereas D threads directly map to OS threads. Can't Rory McGuire add a scheduler to his code? How much code does it take? Bye, bearophile I imagine that it will be fine on Linux because threads truly are lightweight on Linux, but its not going to be great on windows. Go's scheduler is basically the same as the way vibe.d works. Fibers. I haven't tried it but I could use fibers to do the same sort of thing, though it would be easiest to use in Vibe.d because even in go you have to do something that will wait/sleep in order to have your program advance unless you set GOPROCS to a value greater than 1. Go's scheduler is not preemptive so there are the equivalent of Fiber.yield() spread throughout the standard library, I think. I'm not sure how threads + fibers would have to interact in such a system.
Re: I'm porting some go code to D
On Saturday, 24 August 2013 at 20:03:58 UTC, Rory McGuire wrote: On Sat, Aug 24, 2013 at 3:26 PM, bearophile bearophileh...@lycos.comwrote: David Nadlinger: It's a cute idea, but not really practical, I'm afraid – Goroutines are managed by a scheduler in the Go runtime library, whereas D threads directly map to OS threads. Can't Rory McGuire add a scheduler to his code? How much code does it take? Bye, bearophile I imagine that it will be fine on Linux because threads truly are lightweight on Linux, but its not going to be great on windows. Funny, I always thought otherwise, because Windows only has threads. Processes are just a means of grouping threads on Windows, as there isn't the distinction between threads and processes that UNIX systems used to make. Then again, I lost track how the performance on Linux systems changed across the whole Processes - LinuxThreads - NPTL - Posix Threads evolution. -- Paulo
Re: GHC 2013 in Paris
On 24 August 2013 09:48, Iain Buclaw ibuc...@ubuntu.com wrote: On 24 August 2013 09:42, monarch_dodra monarchdo...@gmail.com wrote: On Friday, 23 August 2013 at 16:31:17 UTC, Iain Buclaw wrote: I've jsut been told there's no reception on Saturdays, so after 9.00am might be a slight problem, but I think they will leave a number for you to call. I *could* excuse myself to let you in, but that would require relying on me to remember what time you are arriving. :o) Hum... The forum was down yesterday, so I couldn't post this any earlier. I actually only intend to see your talk, so I won't be there before 16:00. I'll be at the gates *before* 15:30, and see if/how I can get it from there. FWI, I'll be the guy who walks with a crutch. I don't if anyone will be seeing this today, but if you don't see me in the building by 15:50, then it means I failed to get in, and am waiting outside... I'm feeling a little rough this morning, so am running late (sorry Eles!) - just having breakfast, and will be down about 10.30... Hi Guys, Thanks for making it down. Are you going to be around tomorrow? Feel free to ping me or message me and we'll arrange something. :o) Will be at the IRIS building at 9.00am (this time I will be present at that time). And will be staying in Paris until around 2.00pm, which is when I have to check in for train back to London. Regards -- Iain Buclaw *(p e ? p++ : p) = (c 0x0f) + '0';