|
From: | Vincent Caron |
Subject: | [Nel] Trouble with block memory allocator |
Date: | Mon, 24 Feb 2003 20:56:17 +0100 |
User-agent: | Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.3b) Gecko/20030210 |
Hello,I'm having a hard time with the CSTLBlockAllocator and CSTLBlockList classes. I can't make them compile with gcc 3.2, VC.NET (with and without STLport). I can't make the default stub compile either. After wasting a wealth of time on unreadable compiler errors, I looked more closely at the code. I'm suspicious about the real gain about this hack, and I would like to check some facts before going further into the STL nightmare. I'd be pleased to hear rationales and testimonies about the CSTLBlockAlllocator.
I wrote a simple benchmark, but I can't run it since I did not manage to compile the aforementioned classes. I'm curious about the results, if anybody with a compiling NeL tree could try it, that'd be cool :
http://zerodeux.net/misc/nel/memory_test.zipUnpack in code/nel/, it will create a test/memory/ path. VC.NET project files included, it's a basic win32 console project statically linked with the NeL misc library. The result is the average number of alloc+dealloc couples executed per second, the greater the better.
[Prev in Thread] | Current Thread | [Next in Thread] |