I just released a new version of Hoard that incorporates lots of changes (one of which I described in an earlier post).
This release incorporates a number of fixes and improvements, including:
- Better per-thread allocation, improving speed for Unix platforms that do not support
__thread(a fast way to do thread-specific data) - Added interception of Solaris threads API (
thr_*) to work with older Solaris programs (the subject of the earlier post) - Fixed a rare race condition accidentally introduced in 3.7.1
- Increased checking to protect against heap corruption or other errors
- And Hoard now uses GNU-supported hooks on platforms with glibc (especially Linux), allowing it to work better with legacy programs (even emacs!).
Emery Blogger 
Does the new version of Hoard have any fixes in it to handle more than 64 cores? SAP BO is relating that libhoard is causing their application to not work properly. Seems a little fishy to me.
Thanks for the fantastic mem allocator.
You might find this interesting:
“War of allocators: hoard or hoards?”
http://webkit.sed.hu/node/42
This benchmark shows use of hoard in WebKit. The results aren’t very hoard friendly — some problem with benchmark itself?
Thanks for the pointer. I am in communication with the author to try to find out what the issue might be.