[Openchrome-users] [openChrome] #346: memory speed calculation affected by CPU throttling

[OpenChrome Trac]

#346: memory speed calculation affected by CPU throttling
-------------------+--------------------------------------------------------
Reporter:  dsd     |       Owner:  somebody      
    Type:  defect  |      Status:  new           
Priority:  major   |   Component:  initialization
 Version:          |    Keywords:                
Blocking:          |   Blockedby:                
-------------------+--------------------------------------------------------
 At http://dev.laptop.org/ticket/9694 we found that openchrome reports
 significantly faster memory speeds under Phoenix BIOS compared to
 OpenFirmware.

 We traced this down to openchrome's time_function() being inappropriate
 for systems where the CPU frequency can change dynamically. (the slower
 the CPU runs, the faster the reported bandwidth)

 openchrome should use a timing methodology that is unaffected by CPU
 throttling, for example the ACPI timer register or maybe HPETs.

 (thanks to Mitch Bradley for diagnosis)

-- 
Ticket URL: <http://www.openchrome.org/trac/ticket/346>
openChrome <http://www.openchrome.org/>
The openChrome project