Details
-
Type:
Improvement
-
Status:
Closed
-
Priority:
Major
-
Resolution: Fixed
-
Affects Version/s: 2.9.3
-
Fix Version/s: None
-
Component/s: Instruction Architecture: Intel
-
Labels:None
-
Number of attachments :
Issue Links
Activity
LiGen
made changes -
| Field | Original Value | New Value |
|---|---|---|
| Attachment | IA32_SSE2.patch [ 31658 ] |
Ian Rogers
made changes -
Ian Rogers
made changes -
| Resolution | Fixed [ 1 ] | |
| Assignee | Ian Rogers [ ianrogers ] | |
| Status | Open [ 1 ] | Resolved [ 5 ] |
Ian Rogers
made changes -
| Status | Resolved [ 5 ] | Closed [ 6 ] |
Ian Rogers
made changes -
Introduce two non-terminals called float_load and double_load. When these are called they save a memory operand on a stack in the BURS_Helpers, this is pulled out again in the corresponding float add/sub and double add by the call to consumeMO again from the BURS_Helpers. This makes the code a little cleaner and smaller. And thanks for Ian's kindly help.