On Mon, 04 Dec 2006 00:57:50 -0500 Wendy Cheng <wcheng@xxxxxxxxxx> wrote: > > > > > >>>Did you look at improving that lock-lookup algorithm, btw? Core kernel has > >>>no problem maintaining millions of cached VFS objects - is there any reason > >>>why your lock lookup cannot be similarly efficient? > >>> > >>> > >>> > Yes, just found the new DLM uses "jhash" call (include/linux/jhash.h). > I'm on an older version of DLM that uses FNV hash algorithm > (http://www.isthe.com/chongo/tech/comp/fnv/). Will do some performance > test runs to compare these two methods. I'd be surprised if the choice of hash algorithm itself makes much difference. But we can't say much about it unless we can see the code (ie: your code). Is it a simple old hash-to-find-the-bucket-then-walk-a-list implementation? If so, what does the bucket count distribution look like? What is the average walk length? Does it use a single lock, or hashed locking, or a lock-per-bucket? etc. - To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
