[openib-general] Re: Re: Userspace testing results (manykernels, many svn trees)

[Nishanth Aravamudan]

On 24.01.2006 [23:19:52 +0200], Michael S. Tsirkin wrote:
> Quoting r. Nishanth Aravamudan <nacc at us.ibm.com>:
> > Subject: Re: [openib-general] Re: Re: Userspace testing results (manykernels, many svn trees)
> > 
> > On 24.01.2006 [21:39:23 +0200], Michael S. Tsirkin wrote:
> > > Quoting r. Roland Dreier <rdreier at cisco.com>:
> > > > Subject: Re: [openib-general] Re: Re: Userspace testing results (manykernels, many svn trees)
> > > > 
> > > >     Michael> 1 sec = 5.37731e+14 usec
> > > > 
> > > >     Michael> which seems to indicate something's still wrong.
> > > > 
> > > > BTW this number is pretty close to 2^32 times bigger than 1e6, so the
> > > > problem is probably still using long long to return the result of
> > > > mftb (which will result in shifting the result by 32 bits, ie
> > > > multiplying by 2^32).
> > > 
> > > Hmm.
> > > Maybe make clean wasnt run after updating?
> > > Could it be un on rev 5174?
> > 
> > Heh, here's what happens with 5174:
> > 
> > Correlation coefficient r^2: 0.773428 < 0.9
> > 1 sec = inf usec
> > 1 sec = inf usec
> > 1 sec = inf usec
> > 1 sec = inf usec
> > 1 sec = inf usec
> > 1 sec = inf usec
> > 1 sec = inf usec
> > 1 sec = inf usec
> > 1 sec = inf usec
> > 1 sec = inf usec
> > 1 sec = inf usec
> > 1 sec = inf usec
> > 1 sec = inf usec
> > 1 sec = inf usec
> > 1 sec = inf usec
> > 1 sec = inf usec
> > 1 sec = inf usec
> > 
> > And so forth...
> > 
> > Thanks,
> > Nish
> 
> Hmm. Looks like mftb is returning wrong data.
> Could you uncomment lines setting DEBUG and DEBUG_DATA at the top?
> This will print all mftb values out.

Here you go:

x=100 y=21172
x=110 y=21985
x=120 y=23709
x=130 y=26136
x=140 y=27815
x=150 y=29919
x=160 y=32063
x=170 y=33921
x=180 y=35829
x=190 y=37941
x=200 y=40042
x=210 y=42064
x=220 y=43935
x=230 y=45957
x=240 y=47818
x=250 y=50004
x=260 y=51942
x=270 y=54173
x=280 y=56043
x=290 y=57951
x=300 y=59798
x=310 y=62034
x=320 y=63812
x=330 y=65928
x=340 y=67835
x=350 y=69929
x=360 y=71869
x=370 y=73876
x=380 y=75872
x=390 y=77927
x=400 y=79865
x=410 y=81923
x=420 y=84076
x=430 y=85937
x=440 y=87819
x=450 y=89999
x=460 y=91831
x=470 y=93975
x=480 y=95994
x=490 y=97743
x=500 y=99868
x=510 y=101917
x=520 y=103806
x=530 y=105724
x=540 y=107811
x=550 y=110012
x=560 y=111876
x=570 y=113934
x=580 y=115857
x=590 y=117735
x=600 y=119811
x=610 y=121907
x=620 y=124025
x=630 y=125983
x=640 y=127945
x=650 y=129729
x=660 y=131898
x=670 y=133723
x=680 y=135868
x=690 y=137742
x=700 y=139815
x=710 y=141800
x=720 y=143812
x=730 y=145768
x=740 y=147790
x=750 y=149935
x=760 y=151826
x=770 y=154003
x=780 y=155618
x=790 y=157746
x=800 y=159846
x=810 y=161874
x=820 y=163914
x=830 y=165901
x=840 y=167634
x=850 y=169793
x=860 y=171833
x=870 y=173895
x=880 y=175657
x=890 y=177807
x=900 y=179827
x=910 y=181875
x=920 y=183834
x=930 y=185884
x=940 y=187860
x=950 y=189922
x=960 y=191883
x=974 y=194497
x=980 y=195672
x=990 y=197907
x=1000 y=199626
x=1010 y=201735
x=1020 y=203874
x=1030 y=205728
x=1040 y=207861
x=1050 y=209737
x=1060 y=211836
x=1070 y=213749
x=1080 y=215880
x=1090 y=217817
x=1100 y=219729
x=1110 y=221755
x=1120 y=223704
x=1130 y=225783
x=1140 y=227729
x=1150 y=229786
x=1160 y=232030
x=1170 y=233770
x=1180 y=235803
x=1190 y=237791
x=1200 y=239841
x=1210 y=241768
x=1220 y=243862
x=1230 y=245735
x=1240 y=247678
x=1250 y=249795
x=1260 y=251850
x=1270 y=253771
x=1280 y=255840
x=1290 y=257928
x=1300 y=259880
x=1310 y=261752
x=1320 y=263702
x=1330 y=265601
x=1340 y=267671
x=1350 y=269570
x=1360 y=271756
x=1370 y=273750
x=1380 y=275702
x=1390 y=277765
x=1400 y=279695
x=1410 y=281620
x=1420 y=283692
x=1430 y=285587
x=1440 y=287872
x=1450 y=289621
x=1460 y=291676
x=1470 y=293590
x=1480 y=295702
x=1490 y=297790
x=1500 y=299529
x=1510 y=301597
x=1520 y=303741
x=1530 y=305806
x=1540 y=307870
x=1550 y=309780
x=1560 y=311539
x=1570 y=313756
x=1580 y=315725
x=1590 y=317589
x=1600 y=319559
x=1610 y=321651
x=1620 y=323878
x=1630 y=325848
x=1640 y=327742
x=1650 y=329615
x=1660 y=331551
x=1670 y=333784
x=1680 y=335505
x=1690 y=337610
x=1700 y=339742
x=1710 y=341609
x=1720 y=343717
x=1730 y=345641
x=1740 y=347780
x=1750 y=349626
x=1760 y=351749
x=1770 y=353782
x=1780 y=355740
x=1790 y=357413
x=1800 y=359755
x=1810 y=361621
x=1820 y=363584
x=1830 y=365768
x=1840 y=367582
x=1850 y=369810
x=1860 y=371521
x=1870 y=373702
x=1880 y=375905
x=1890 y=377659
x=1900 y=379704
x=1910 y=381626
x=1920 y=383601
x=1930 y=385635
x=1940 y=387715
x=1950 y=389671
x=1960 y=391704
x=1970 y=393599
x=1980 y=395572
x=1990 y=397692
x=2000 y=399776
x=2010 y=401853
x=2020 y=403711
x=2030 y=405478
x=2040 y=407577
x=2050 y=409618
x=2060 y=411603
x=2070 y=413642
x=2080 y=415601
x=2090 y=417823
a = -8.02523
b = 199.818
a / b = -0.0401626
r^2 = 0.999999
Warning: measured timestamp frequency 199.818 differs from nominal 1600 MHz
1 sec = 1.00195e+06 usec
1 sec = 1.00198e+06 usec
1 sec = 1.00207e+06 usec
1 sec = 1.00207e+06 usec
1 sec = 1.00207e+06 usec
1 sec = 1.00207e+06 usec
1 sec = 1.00207e+06 usec
1 sec = 1.00207e+06 usec
1 sec = 1.00207e+06 usec
1 sec = 1.00207e+06 usec

Just an FYI, when I tried redirecting stdout and stderr to a file, the
program never would print the "1 sec = ..." lines. If I just redirect
stderr (which I had to grab all the numbers above), it works fine. Dunno
what's up with that or if it's just a problem on my end.

> > P.S. Is there any way to specify how long to run clock_test from the
> > command line? It's a bit of a pain the grid to kill a process...
> > 
> > Thanks,
> > Nish
> 
> I'll make something up, for now I guess you can just add a hard-coded
> counter in for(;;) in clock_test.c

Great, thanks! That will make debugging with future revs a lot easier
for me.

Thanks,
Nish