You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@harmony.apache.org by "Alexey Varlamov (JIRA)" <ji...@apache.org> on 2007/01/30 08:17:33 UTC

[jira] Commented: (HARMONY-3063) [drlvm][jit] ArrayCopyRepMove does not work on EM64T

    [ https://issues.apache.org/jira/browse/HARMONY-3063?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12468520 ] 

Alexey Varlamov commented on HARMONY-3063:
------------------------------------------

Um, George, you disabled the optimization for arrays of floats:
Index: vm/jitrino/src/codegenerator/ia32/Ia32InstCodeSelector.cpp
======================================
@@ -2585,7 +2603,6 @@
         case Type::UIntPtr:
         case Type::UInt32 :
         case Type::Single :
-        case Type::Float  : 
         case Type::Object  : 
         case Type::SystemObject  : 
         case Type::SystemString  : 

I agree that float cannot fit into 4 bytes and it was a hidden bug, so now it will assert instead. But should not we fix this more accurately?

> [drlvm][jit] ArrayCopyRepMove does not work on EM64T
> ----------------------------------------------------
>
>                 Key: HARMONY-3063
>                 URL: https://issues.apache.org/jira/browse/HARMONY-3063
>             Project: Harmony
>          Issue Type: Bug
>          Components: DRLVM
>         Environment: EM64T
>            Reporter: George Timoshenko
>         Assigned To: Alexey Varlamov
>         Attachments: arraycopy_rep_move_em64t_enabling.patch
>
>
> The optimisation crashes on EM64T. In particular SPECjbb2005 fils with enabled genArrayCopyRepMove switch.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.