On 05/24/2011 12:37 PM, Fernando Cassia wrote: > Well guess what? 64 bit code is bigger (bigger pointers) and thus > slower, because CPU cache is less effective, with bigger code. > In my previous job I've done quite a bit of benchmarking. First, the OS itself is faster in 64-bit because it can use the additional GP registers, and memory access is linear. I've seen many cases where 32-bit apps run faster than 64-bit apps, but I have also seen cases where 32-bit apps run faster using a 64-bit OS than they do with the same 32-bit OS. I have not tested Java, where the article you reference (http://su.pr/23Ic43) is strictly a Java article. While I was at Digital in the 90s, the Alpha was a pure 64-bit platform, and we came up with some ways to allow 32-bit applications to be built and run under Tru64. But, most of my benchmark testing on both x86 and IA64 points the fact that most 64-bit apps will be faster or not suffer any performance degradation. The bottom line here is you can't make a blanket statement that 64-bit code is slower. And if you have a decent optimizing compiler, your performance could improve.
-- Jerry Feldman <g...@blu.org> Boston Linux and Unix PGP key id: 537C5846 PGP Key fingerprint: 3D1B 8377 A3C0 A5F2 ECBB CA3B 4607 4319 537C 5846
signature.asc
Description: OpenPGP digital signature
-- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
