[LTP] [PATCH v1] realtime:matrix_multi: main function to return exit status
Bogdan Lezhepekov
bogdan.lezhepekov@suse.com
Fri Nov 12 14:04:01 CET 2021
Hi Joerg,
No, this is a braindead test. Our results are pretty consistent with ours and that's why we excluded one from the regression loop.
The return value was fixed just because I was fixing it everywhere wherever it caught my eye 🙂
-Bogdan
________________________________
From: Joerg Vehlow <lkml@jv-coder.de>
Sent: Friday, November 12, 2021 14:07
To: Bogdan Lezhepekov <bogdan.lezhepekov@suse.com>; ltp@lists.linux.it <ltp@lists.linux.it>
Subject: Re: [LTP] [PATCH v1] realtime:matrix_multi: main function to return exit status
Hi,
is the test even returning any valid results anywhere?
I don't know if systems are running too fast for the test nowadays,
but in my tests, they always divide by 0, even with very high iteration
counts even in qemu.
e.g:
$ matrix_mult -i10000000
---------------------------------------
Matrix Multiplication (SMP Performance)
---------------------------------------
Running 10000000 iterations
Matrix Dimensions: 100x100
Calculations per iteration: 8
Number of CPUs: 4
Running sequential operations
Min: 0 us
Max: 160 us
Avg: 0.0000 us
StdDev: 0.2082 us
Running concurrent operations
Min: 0 us
Max: 1693 us
Avg: 0.0000 us
StdDev: 0.6652 us
Concurrent Multipliers:
Min: -nan
Max: 0.0945
Avg: -nan
Criteria: 3.00 * average concurrent time < average sequential time
Result: FAIL
If this is not a result of my test system, we should probably remove the
test from the
profile, at least as long as it was not fixed.
Joerg
On 9/13/2021 5:31 PM, Bogdan Lezhepekov via ltp wrote:
> The original version always returned 0.
>
> Signed-off-by: Bogdan Lezhepekov <bogdan.lezhepekov@suse.com>
> ---
> testcases/realtime/func/matrix_mult/matrix_mult.c | 10 +++++-----
> 1 file changed, 5 insertions(+), 5 deletions(-)
>
> diff --git a/testcases/realtime/func/matrix_mult/matrix_mult.c b/testcases/realtime/func/matrix_mult/matrix_mult.c
> index 61ab887fc..e702c0ff9 100644
> --- a/testcases/realtime/func/matrix_mult/matrix_mult.c
> +++ b/testcases/realtime/func/matrix_mult/matrix_mult.c
> @@ -188,7 +188,7 @@ void *concurrent_thread(void *thread)
> return NULL;
> }
>
> -void main_thread(void)
> +int main_thread(void)
> {
> int ret, i, j;
> nsec_t start, end;
> @@ -308,7 +308,7 @@ void main_thread(void)
> criteria);
> printf("Result: %s\n", ret ? "FAIL" : "PASS");
>
> - return;
> + return ret;
> }
>
> int main(int argc, char *argv[])
> @@ -319,7 +319,7 @@ int main(int argc, char *argv[])
> numcpus = sysconf(_SC_NPROCESSORS_ONLN);
> /* the minimum avg concurrent multiplier to pass */
> criteria = pass_criteria * numcpus;
> - int new_iterations;
> + int new_iterations, ret;
>
> if (iterations <= 0) {
> fprintf(stderr, "iterations must be greater than zero\n");
> @@ -348,7 +348,7 @@ int main(int argc, char *argv[])
> printf("Number of CPUs: %u\n", numcpus);
>
> set_priority(PRIO);
> - main_thread();
> + ret = main_thread();
>
> - return 0;
> + return ret;
> }
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linux.it/pipermail/ltp/attachments/20211112/9efd3bcd/attachment.htm>
More information about the ltp
mailing list