[llvm-testresults] Geryon-X86-64 nightly tester results

[Chris Lattner]

On Dec 15, 2009, at 7:05 PM, Daniel Dunbar wrote:

> 
> On Dec 15, 2009, at 6:01 PM, Chris Lattner wrote:
> 
>> 
>> On Dec 15, 2009, at 2:48 PM, Evan Cheng wrote:
>> 
>>> This is a significant optimizer regression. Any ideas?
>> 
>> I upgraded the machine to 10.5.8 yesterday.  So this is what the future looks like!
> 
> That's strange, did the uname not change?
> 
> And uh, I'm dubious. Upgrading to 10.5.8 made some proper subset of the tests slower? Either way, seems like a bug somewhere. It would be interesting to have some idea which tests end up actually using the OS much (for example, pumping out lots of output).

Upgrade implies a reboot, reboots (or lack thereof) can definitely affect performance.

-Chris