Discussions on more advanced topics such as monolithic vs micro-kernels, transactional memory models, and paging vs segmentation should go here. Use this forum to expand and improve the wiki!
Including a cache hierachy and the size of heap management functions, has anyone captured performance statistics of their heap utility -- particularly about cache movement required to execute the heap management utility?
I guess I'd like ratios corresponding to cache levels which measure cacheHit:cacheMiss for the heap management stuff.
Also, it'd be interesting to somehow apprehend the detriment of overhead for various memory-design approaches (e.g. garbage collection). I don't think I'm aware of the best methods to identify this jizz
Very nice, I'll have to experiment with a Linux kernel a lot. If this isn't enough I'm sure there are more tools like it, or I could write my own. I was just checking to see if anyone has already crunched the numbers.
Some of the information seems like it'd be a red herring. I'd better have the source of whatever tool I'm using. I'm thinking I'll just go for writing a little utility program myself; it won't be too hard.
Oh. You're the author? Good! I sent you an MSN message asking for clarification regarding this claim:
"-It can calculate library call speed, and the speed of the operating system's memory manager"
If you also benchmark the malloc/free performance during "Complex algorithmic performance" then I'd trust it much more. Individual scores + aggregated scores would be nice info. I'm really much more interested in malloc/free with lots of **** between than a repeated-case stress test.
It'd be very useful to be able to just run something like this in the background and sample over the duration of the entire system's execution; writing aggregated scores to a binary log file (with some kind of sampling parameters).
The more I think about it, the more it sounds appropriate to just write my own
Oranos wrote:Oh. You're the author? Good! I sent you an MSN message asking for clarification regarding this claim
sorry, i didnt recived anything. msn is discontinued by ms, so please use skype instead
that malloc and free test is just simply mallocing and freeing the same memory area repeatly. its not a serious stresstest for malloc/free. i just put this test in to be aware of the default allocator performances beethwen different operating systems, this test is also not multithreaded. i would not take this malloc result seriously at all, its only gives some info when its compared with other systems.
