Re: [Mono-dev] ObjectPool
On Wed, 2012-07-25 at 07:17 -0400, Greg Young wrote: Yuriy will reply with more details but by just making Take return Creator() and nooping release many of our issues have gone away (including visible issues with concurrentstack). From what I have seen we also saw massive performance gains. Just for the curiosity: are you using SGEN? I've always thought that ObjectPool in ConcurrentStack was used to reduce the pressure on the garbage collector (Boehm, specifically). It is probably much less important with generational collector like SGEN. -- Regards, Konrad ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-dev] ObjectPool
On Wed, 2012-07-25 at 07:28 -0400, Greg Young wrote: Yes we are using sgen. Good to know the history there as well. Cheers, Greg Well, *actually* this is my guess ;) -- Regards, Konrad ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-dev] ConcurrentStack with value type in 2.10
On Mon, 2012-07-23 at 13:45 +0100, Alan wrote: I cannot reproduce the problem either. What exact version of 2.10 did you test against? It's possible the bug has already been fixed in a newer release of the 2.10 series. I could reproduce it on 32-bit machine, only with sgen. -- Regards, Konrad ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-dev] Organize usings statement in MonoDevelop 3.02
On Tue, 2012-05-29 at 02:18 -0700, siquylee wrote: I'm using MonoDevelop 3.02 and I don't find Organize usings statement from context menu when the cursor is in 'using' sections. Could you please tell me where I could enable this statement? Try Edit - Preferences - Other - Source Analysis - Enable Source Analysis. Then you should notice yellow bulb when cursor is located in unneeded usings. You can click (or press Alt+Enter) and remove unneeded ones. -- Regards, Konrad ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
[Mono-dev] Confusion with throw; and throw e
Hello all, consider this simple C# program: using System; namespace ETest { class MainClass { public static void Main(string[] args) { try { Fun(); } catch(Exception e) { throw; } } public static void Fun() { throw new Exception(Test); } } } Here's the stack trace: System.Exception: Test at ETest.MainClass.Fun () [0x0] in Main.cs:21 at ETest.MainClass.Main (System.String[] args) [0x0] in Main.cs:11 And another simple program: using System; namespace ETest { class MainClass { public static void Main(string[] args) { try { Fun(); } catch(Exception e) { throw e; } } public static void Fun() { throw new Exception(Test); } } } and the stack trace is exactly the same. Shouldn't the second program's exception originate from the Main.cs:15, i.e. throw e line? The generated IL looks good. -- Regards, Konrad ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-dev] Confusion with throw; and throw e
Here's another example: using System; using System.Runtime.CompilerServices; namespace ETest { class MainClass { public static void Main(string[] args) { try { Fun(); } catch(Exception e) { Console.WriteLine(e.StackTrace); } } public static void Fun() { try { Fun2(); } catch(Exception e) { throw; } } [MethodImpl(MethodImplOptions.NoInlining)] public static void Fun2() { throw new Exception(Test); } } } Stack trace is the same if throw is changed to throw e. Is this a known bug? I've found a bug considering throw; not preserving the stack trace, but this one is opposite - throw e does not reset it ;) -- Regards, Konrad ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-dev] Google Summer of Code project ideas
Hello again, On Thu, 2012-04-05 at 10:06 +0700, Alexander Mezin wrote: Yeah, I agree on the general term that it could and should be easier to use, just thought maybe you did not discovered it yet (frankly, because of the hidden nature of this feature not so many people are aware of it). Personally the ReSharper approach is my favourite, however yours is interesting too. I prefer ReSharper's approach too, but I think implementing it will require much more work, I'm not sure that I'll finish such project. If you did not encounter it yet, you should definitely look into new code analysis API, based on NRefactory. It is currently developed on the newresolver branch. Look at the sample context action here: http://mikemdblog.blogspot.com/2012/03/how-to-write-context-action-using.html It seems that I'll have time to look into it after proposals deadline. And the deadline is tomorrow, I think I must start writing proposal. Could you be a mentor for this project? Frankly, I am not very familiar with this API, but if no other person wants to be a mentor, I could. Therefore I'm redirecting this to mailing list too. Can anyone tell me should I contact/fill anything at this point? -- Regards, Konrad ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
[Mono-dev] Count leading zeros (or find first one) in Mono.SIMD
Hi, would it be possible to have implemented functionality like count leading zeros in Mono.SIMD? Instructions like these are available on most architectures. Newer SSE instructions would be welcomed too. -- Regards, Konrad ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-dev] Count leading zeros (or find first one) in Mono.SIMD
On Tue, 2012-02-28 at 11:49 -0300, Rodrigo Kumpera wrote: Mono.SIMD needs update to the latest versions of SSE. Right now we're on SSSE. The problem is the lack of engineering bandwidth, I would love to have the 4.X stuff. I guess I could do some work on it. The questions are: a) whether current API is to be extended or there are some reasons to rebuild it fundamentally? IMO it looks fine, just new (wider) structures and methods need to be added; b) question above applied to the underlying JIT machinery; c) is there any documentation about it or should I just rather look at the source code? -- Regards, Konrad ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-dev] Count leading zeros (or find first one) in Mono.SIMD
Are you talking about AVX? AVX is a very tricky matter and I don't have an answer for it upfront, but a safe guess is that it will take quite some work and API changes. This is for using the wider 256bits registers. Using AVX with 128bits register is fine. Yes, I thought about AVX too, but in that case it may be reserved for future. For adding missing instructions, it's just a matter of adding them. Pretty simple. b) question above applied to the underlying JIT machinery; Same as above. c) is there any documentation about it or should I just rather look at the source code? Just look at the source code, those are some entry level changes if you fancy doing so. Thanks, I'll report if something emerges. -- Konrad ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-dev] LLVM with current Mono
Hi Zoltan, 2011/10/20 Zoltan Varga var...@gmail.com: Hi, You should use the branch in the LLVM repo corresponding to your mono version, i.e. mono-2-10. Zoltan Generally it was what I did. Specifically: LLVM from mono-2-10 branch, last commit ab6947203a665efa0d7676f19cabf2fa70e3af60 configured with: ./configure --enable-optimized --enable-targets=x86 x86_64 --prefix=/usr/ Mono 2.10.5 tarball from mono site. Configured with: ./configure --prefix=/usr/ --enable-llvm=yes LLVM compilation went fine, hovewer Mono's ended with error attached in previous message. -- Regards, Konrad ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list