RFR :7088419 : (L) Use x86 Hardware CRC32 Instruction with java.util.zip.CRC32 and java.util.zip.Adler32
david.r.chase at oracle.com
Thu May 16 15:09:05 PDT 2013
On 2013-05-16, at 5:27 PM, Alan Bateman <Alan.Bateman at oracle.com> wrote:
> The current proposal doesn't change the API at this time but I wonder if you have considered adding parallelUpdate methods to complement the serial methods?
I'm not sure I understand. Right now, the update methods run in parallel for sufficiently large inputs, otherwise serial. I'm still unsure whether their appetite for tasks should be throttled somewhat above and beyond the size limit. There's benefit to having CRC run 3x faster if it uses 4 threads to do it, but using 32 threads to run 9x faster is maybe not a win, even though it is nominally "faster". And here I mean "threads", not tasks, though I don't have a lot of visibility how much of a multiprocessor I am eating when I break something into 32 or 64 tasks.
More information about the hotspot-compiler-dev