That's what I *thought* happens and while the 'do-regression' script implies that, the evidence suggests otherwise though which confused me.
Consider, i deleted the build/ALPHA_SE/ directory on 'z/m5/regression/zizzer/m5' afternoon yesterday. Then when the regressions ran, that directory wasnt built although the report listed all the regressions being run. So my question is how can it be running locally if the relevant build directories in ".../zizzer/m5" arent being updated? Doesnt quite make sense to me just yet. On Thu, May 14, 2009 at 9:42 AM, Ali Saidi <[email protected]> wrote: > It tries to run on the pool first, if that fails it then runs on > zizzer locally. > > Ali > > On May 14, 2009, at 8:07 AM, Korey Sewell wrote: > >> Although it looks like the regressions are running on: >> /z/m5/regression/zizzer/m5 >> >> I dont think that is the right directory as last night the >> build/ALPHA_SE directory did not get updated after I deleted it >> earlier in the day. >> >> A closer look at the regression output showed it building and >> running out of : >> /n/poolfs/z/home/m5test >> >> So I went to that directory and deleted build/ALPHA_SE just in case. >> >> In the do-regression script, it didnt look like it was operating out >> of m5test ("qsub timeout, retrying locally"), but I guess I need to >> look a little closer at what's going on in those scripts... >> >> >> On Wed, May 13, 2009 at 4:21 PM, Steve Reinhardt <[email protected]> >> wrote: >>> Or if you want, just delete build/ALPHA_SE and wait for tonight's >>> regression >>> to run to rebuild it automatically. Otherwise you have to be >>> careful about >>> permissions (e.g., generating a binary that the regression user >>> doesn't have >>> permission to run, etc.). >>> >>> Steve >>> >>> On Wed, May 13, 2009 at 1:19 PM, Korey Sewell <[email protected]> >>> wrote: >>>> >>>> That makes sense....Just to be clear, I should delete the >>>> build/ALPHA_SE and rebuild it. Then, *hopefully* the regressions >>>> will >>>> run with no hiccups. >>>> >>>> On Wed, May 13, 2009 at 3:05 PM, Steve Reinhardt >>>> <[email protected]> wrote: >>>>> You probably need to delete the build/ALPHA_SE directory in the >>>>> regression >>>>> tree... what you've done is set it so that the default options for >>>>> ALPHA_SE >>>>> now include InOrderCPU, but since the regressions are still >>>>> running off >>>>> the >>>>> options as they were initialized last time it built from scratch, >>>>> that >>>>> doesn't affect it. >>>>> >>>>> Steve >>>>> >>>>> On Wed, May 13, 2009 at 12:03 PM, Korey Sewell <[email protected]> >>>>> wrote: >>>>>> >>>>>> Thanks Nate. >>>>>> >>>>>> The simerr is: >>>>>> ksew...@zizzer:/z/m5/regression/zizzer/m5$ more >>>>>> >>>>>> build/ALPHA_SE/tests/fast/quick/00.hello/alpha/linux/inorder- >>>>>> timing/simerr >>>>>> Traceback (most recent call last): >>>>>> File "<string>", line 1, in <module> >>>>>> File "/z/m5/regression/zizzer/m5/src/python/m5/main.py", line >>>>>> 359, in >>>>>> main >>>>>> exec filecode in scope >>>>>> File "tests/run.py", line 61, in <module> >>>>>> execfile(os.path.join(tests_root, 'configs', config + '.py')) >>>>>> File "tests/configs/inorder-timing.py", line 40, in <module> >>>>>> cpu = InOrderCPU(cpu_id=0) >>>>>> NameError: name 'InOrderCPU' is not defined >>>>>> >>>>>> >>>>>> I'm not sure why InOrderCPU isnt getting built given that I edited >>>>>> ALPHA_SE and in the src/cpu/inorder/SConsopt file I have >>>>>> InOrderCPU as >>>>>> a default CPU build. >>>>>> >>>>>> I have yet to dig into if the regressions build the CPU models >>>>>> from a >>>>>> certain list, but I was sure that it was building as a default. >>>>>> Guess >>>>>> not! >>>>>> >>>>>> Anything else I have to do here to get InOrderCPU to build as >>>>>> default? >>>>>> >>>>>> On Wed, May 13, 2009 at 12:29 PM, nathan binkert >>>>>> <[email protected]> >>>>>> wrote: >>>>>>> He does have sudo access, but that's not the right directory. >>>>>>> >>>>>>> Look in: /z/m5/regression/zizzer/m5/build >>>>>>> >>>>>>> On Wed, May 13, 2009 at 8:07 AM, Korey Sewell <[email protected]> >>>>>>> wrote: >>>>>>>>> On Wed, May 13, 2009 at 9:32 AM, Steve Reinhardt <[email protected] >>>>>>>>> > >>>>>>>>> wrote: >>>>>>>>> zizzer (note the sender's email address) >>>>>>>> >>>>>>>> I'm having a little trouble figuring out this error... The >>>>>>>> regression >>>>>>>> output message lets me know that "M5 exited with nonzero >>>>>>>> status 1". >>>>>>>> >>>>>>>> But, where can I find the "simout"/"simerr" so that I can get >>>>>>>> the >>>>>>>> exact >>>>>>>> error? >>>>>>>> I figured it might be in z/m5/m5test/build/* but it seems I dont >>>>>>>> have >>>>>>>> permissions to that directory: >>>>>>>> ksew...@zizzer:/z/m5$ cd m5test/ >>>>>>>> -bash: cd: m5test/: Permission denied >>>>>>>> >>>>>>>> >>>>>>>> -- >>>>>>>> - Korey >>>>>>>> _______________________________________________ >>>>>>>> m5-dev mailing list >>>>>>>> [email protected] >>>>>>>> http://m5sim.org/mailman/listinfo/m5-dev >>>>>>>> >>>>>>>> >>>>>>> _______________________________________________ >>>>>>> m5-dev mailing list >>>>>>> [email protected] >>>>>>> http://m5sim.org/mailman/listinfo/m5-dev >>>>>>> >>>>>> >>>>>> >>>>>> >>>>>> -- >>>>>> - Korey >>>>>> _______________________________________________ >>>>>> m5-dev mailing list >>>>>> [email protected] >>>>>> http://m5sim.org/mailman/listinfo/m5-dev >>>>> >>>>> >>>>> _______________________________________________ >>>>> m5-dev mailing list >>>>> [email protected] >>>>> http://m5sim.org/mailman/listinfo/m5-dev >>>>> >>>>> >>>> >>>> >>>> >>>> -- >>>> - Korey >>>> _______________________________________________ >>>> m5-dev mailing list >>>> [email protected] >>>> http://m5sim.org/mailman/listinfo/m5-dev >>> >>> >>> _______________________________________________ >>> m5-dev mailing list >>> [email protected] >>> http://m5sim.org/mailman/listinfo/m5-dev >>> >>> >> >> >> >> -- >> - Korey >> _______________________________________________ >> m5-dev mailing list >> [email protected] >> http://m5sim.org/mailman/listinfo/m5-dev >> > > _______________________________________________ > m5-dev mailing list > [email protected] > http://m5sim.org/mailman/listinfo/m5-dev > -- =========== Korey L Sewell PhD Candidate Computer Science & Engineering University of Michigan _______________________________________________ m5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/m5-dev
