Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Test takes a very long time on a 32 cpu machine #9

Open
doug65536 opened this issue Dec 8, 2019 · 1 comment
Open

Test takes a very long time on a 32 cpu machine #9

doug65536 opened this issue Dec 8, 2019 · 1 comment

Comments

@doug65536
Copy link

The test has been running about 24 hours now on my 16-core, 32-thread 3950X with dual channel 64GB DDR4-2400 ECC memory on Linux (Ubuntu x86_64 19.04 LTS). I compiled it from the source tarball on the website. Is the test expected to take days to complete on a machine with many CPUs? I'm starting to doubt that it will ever finish.

Here is some of the latest output:

run time = 1.54818 -> next test with repeat factor=11710012914
RESULT	datetime=2019-12-07 20:56:01	host=doug-dt	version=0.6.2	funcname=PermRead64SimpleLoop	nthreads=24	areasize=65536	threadsize=2816	testsize=67584	repeats=4291965	testvol=290068162560	testaccess=36258520320	time=1.5481844719615764916	bandwidth=187360206624.78201294	rate=4.269850116050120098e-11
Running nthreads=24 factor=11710012914 areasize=98304 thrsize=4224 testsize=101376 repeats=2772257 testvol=281040325632 testaccess=35130040704
Make permutation: filling filling filling filling filling filling filling filling filling filling filling filling filling filling filling permuting filling cycle= permuting permuting cycle=528 cycle=528 permuting permuting permuting cycle=528 permuting cycle=528 cycle=528 permuting permuting filling permuting cycle=528 cycle=528 permuting filling permuting permuting cycle=528 filling cycle=528 filling permuting cycle=528 filling cycle=528 filling permuting cycle= permuting permuting cycle= permuting cycle= permuting permuting528528 permuting528 cycle= filling cycle= permuting528 permuting528 filling cycle=528 cycle=528 cycle= permuting cycle=528528528 cycle= cycle=528528 cycle=528
run time = 1.50383 -> next test with repeat factor=11680189154
RESULT	datetime=2019-12-07 20:56:02	host=doug-dt	version=0.6.2	funcname=PermRead64SimpleLoop	nthreads=24	areasize=98304	threadsize=4224	testsize=101376	repeats=2772257	testvol=281040325632	testaccess=35130040704	time=1.5038301280001178384	bandwidth=186883026479.68893433	rate=4.2807525919800250103e-11
Running nthreads=24 factor=11680189154 areasize=131072 thrsize=5504 testsize=132096 repeats=2122128 testvol=280324620288 testaccess=35040577536
Make permutation: filling filling filling filling filling permuting filling permuting filling permuting filling permuting permuting cycle=688 cycle= filling688 filling permuting filling permuting filling permuting cycle=688 filling permuting filling filling permuting cycle=688 filling filling filling filling permuting permuting permuting filling filling cycle=688 filling filling cycle=688 permuting cycle=688 cycle= cycle=688 permuting permuting cycle=688 cycle=688 permuting permuting cycle= cycle=688 cycle=688 cycle=688 cycle=688 cycle=688 permuting filling688688 cycle=688 permuting permuting cycle= cycle=688 permuting cycle=688 permuting permuting688 cycle=688 cycle=688 cycle=688
run time = 1.48091 -> next test with repeat factor=11830762260
RESULT	datetime=2019-12-07 20:56:04	host=doug-dt	version=0.6.2	funcname=PermRead64SimpleLoop	nthreads=24	areasize=131072	threadsize=5504	testsize=132096	repeats=2122128	testvol=280324620288	testaccess=35040577536	time=1.4809095459640957415	bandwidth=189292196172.2545166	rate=4.2262703702375861194e-11
Running nthreads=24 factor=11830762260 areasize=196608 thrsize=8320 testsize=199680 repeats=1421967 testvol=283938370560 testaccess=35492296320
Make permutation: filling permuting filling filling filling filling filling filling filling filling filling filling permuting filling filling filling permuting filling permuting filling cycle=1040 permuting cycle=1040 cycle=1040 permuting permuting permuting filling permuting filling cycle= permuting filling cycle= permuting permuting cycle= cycle=1040 permuting cycle=1040 permuting permuting cycle= permuting1040 filling permuting filling permuting cycle=1040 permuting cycle=1040 filling permuting1040 permuting cycle=1040 cycle= cycle=1040 cycle=1040 cycle=10401040 cycle=1040 cycle=1040 cycle=1040 cycle=1040 filling permuting1040 cycle=1040 cycle=10401040 filling permuting permuting cycle=1040 cycle=1040
run time = 1.4521 -> next test with repeat factor=12220991348
RESULT	datetime=2019-12-07 20:56:05	host=doug-dt	version=0.6.2	funcname=PermRead64SimpleLoop	nthreads=24	areasize=196608	threadsize=8320	testsize=199680	repeats=1421967	testvol=283938370560	testaccess=35492296320	time=1.4521038149832747877	bandwidth=195535861575.62045288	rate=4.0913211190705924444e-11
Running nthreads=24 factor=12220991348 areasize=262144 thrsize=11008 testsize=264192 repeats=1110192 testvol=293303844864 testaccess=36662980608
Make permutation: filling filling filling filling filling filling permuting filling filling permuting filling permuting filling cycle= permuting filling permuting filling permuting cycle=1376 filling filling permuting permuting cycle= permuting permuting cycle= cycle=1376 filling filling permuting cycle= cycle=1376 filling permuting filling permuting cycle= permuting1376 filling permuting permuting permuting filling cycle=1376 cycle=13761376 cycle=1376 filling filling permuting cycle=13761376 cycle=13761376 cycle=1376 permuting permuting filling permuting filling permuting cycle=1376 cycle= permuting permuting1376 cycle=1376 cycle=1376 cycle=1376 cycle=1376 cycle=13761376 cycle=1376 cycle

Does that look right?

@robobenklein
Copy link

Just to leave some additional info here, I ran this on a 7950X with 128GB DDR5 and had to limit it to 4GB of tested memory (without -S 0 or with -M 4G), for which it finished in over a day.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants