Hi John, Thanks for the thoughts and advice.
At this point i plan to proceed with the current patch (i will add tests for buffer views), and then follow up more more analysis and recommendations for SPARC. The situation is a little murky after some further performance analysis, non-optimal loop unrolling and address calculations may be dominating. Without eyeballing generated code and shedding some light on this i am fumbling around in the crepuscule. Paul. > On 31 Mar 2016, at 08:33, John Rose <john.r.r...@oracle.com> wrote: > > On Mar 30, 2016, at 2:36 AM, Paul Sandoz <paul.san...@oracle.com > <mailto:paul.san...@oracle.com>> wrote: >> >> When access is performed in loops this can cost, as the alignment checks are >> not hoisted out. Theoretically could for regular 2, 4, 8 strides through the >> buffer contents. For such cases alignment of the base address can be >> checked. Not sure how complicated that would be to support. >> >> I lack knowledge of the SPARC instruction set to know if we could do >> something clever as an intrinsic. > > A couple of partial thoughts: > > If we had bitfield type inference, we would be able to deduce that the low > bits of p and p+8 are the same. Graal has this (because I gave them the > formulae[1]). C2 may be too brittle to add it into TypeInt. Bitfield > inference on expressions of the form p&7 and (p+8)&7 would allow commoning > tests in an unrolled loop, hoisting the alignment logic to the top of the > loop body, and (perhaps) through the phi to the loop head. > > [1]: > http://hg.openjdk.java.net/graal/graal-core/file/ea5cc66ec5f2/graal/com.oracle.graal.compiler.common/src/com/oracle/graal/compiler/common/type/IntegerStamp.java#l460 > > <http://hg.openjdk.java.net/graal/graal-core/file/ea5cc66ec5f2/graal/com.oracle.graal.compiler.common/src/com/oracle/graal/compiler/common/type/IntegerStamp.java#l460> > > An intrinsic that would guide the JIT more explicitly would (I think) need an > extra argument. Something like: getLongUnaligned(p, uo, ao), where uo and > ao are both longs, but only ao is required to be naturally aligned > ((ao&7)==0). Yuck. Maybe this could be pattern-matched in C2; it would be a > kludge. > > — John
