*** WARNING: THE ATTACHED DOCUMENT(S) CONTAIN MACROS ***
*** MACROS MAY CONTAIN MALICIOUS CODE ***
*** Open only if you can verify and trust the sender ***
*** Please contact infosec@redhat.com if you have questions or concerns **

Mon, Nov 16, 2015 at 07:41:07PM CET, jtluka@redhat.com wrote:
>Thu, Nov 12, 2015 at 02:47:16PM CET, jtluka@redhat.com wrote:
>>Thu, Nov 12, 2015 at 12:19:53PM CET, olichtne@redhat.com wrote:
>>>On Wed, Nov 11, 2015 at 02:55:11PM +0100, Jiri Prochazka wrote:
>>>> This patch adds support for measuring both local and remote CPU utilization:
>>>>   New option for netperf client have been added:
>>>>     cpu_util - when set to "local", netperf is run with -c argument, enabling measure of local CPU utilization during run,
>>>>                when set to "remote", netperf is run with -C argument, enabling measure of remote CPU utilization during run,
>>>>                when set to "both", netperf is run with both -c and -C arguments
>>>>
>>>>   Parsing of throughput and confidence interval has been made cleaner:
>>>>     Using omni netperf output, all relevant data are now stored in key=val format,
>>>>     which can be parsed more easily and is the same for all netperf test types.
>>>>
>>>> Changes in v2:
>>>>   Changed format of cpu_util argument (described above)
>>>>
>>>>   Renamed keys in res_val dict to from loc_cpu_util and rem_cpu util
>>>>   to LOCAL_CPU_UTIL and REMOTE_CPU_UTIL
>>>>
>>>>   Fixed incorrect real confidence value, now the touple correctly
>>>>   returns (CONFIDENCE_LEVEL, THROUGHPUT_CONFIDENCE_INTERVAL/2)
>>>>
>>>> Changes in v3:
>>>>   Removed redundant error value parsing from _parse_confidence method
>>>>
>>>> Signed-off-by: Jiri Prochazka <jprochaz@redhat.com>
>>>
>>>Looks good, after jtlukas tests finish and there's no new problems i'll
>>>ack and push it.
>>
>>Tests are ok. I've checked whether confidence is same with old and new
>>approach and it's the same.
>>
>>Acked-by: Jan Tluka <jtluka@redhat.com>
>
>I've noticed that the UDP_STREAM performance dropped after this change.
>First I thought it is because of cpu utilization data collection but
>disabling this did not help.
>
>Very likely this is because of OMNI output. I've looked in the netperf
>code and there are few other "THROUGHPUT" values:
>
>src/nettest_omni.c:
>
>  LOCAL_SEND_THROUGHPUT,
>  LOCAL_RECV_THROUGHPUT,
>  REMOTE_SEND_THROUGHPUT,
>  REMOTE_RECV_THROUGHPUT,
>
>I'm wondering if the previous value that we reported is some of these.

>_______________________________________________
>LNST-developers mailing list
>lnst-developers@lists.fedorahosted.org
>https://lists.fedorahosted.org/admin/lists/lnst-developers@lists.fedorahosted.org

Cut from netperf code:

nettest_omni.c send_omni_innner()

                                remote_bytes_received = (uint64_t)omni_result->bytes_received_hi << 32;
                                remote_bytes_received += omni_result->bytes_received_lo;
                                remote_bytes_xferd = (double) remote_bytes_received +
                                    remote_bytes_sent;


                                        if ('x' == libfmt)
                                            /* it was a request/response test */
                                            thruput = calc_thruput((double)trans_completed);
                                        else if (NETPERF_RECV_ONLY(direction))
                                            thruput      = calc_thruput(bytes_xferd);
                                        else
                                            thruput = calc_thruput(remote_bytes_xferd);


                                local_send_thruput = calc_thruput((double)bytes_sent);
                                local_recv_thruput = calc_thruput((double)bytes_received);
                                remote_send_thruput = calc_thruput((double)remote_bytes_sent);
                                remote_recv_thruput = calc_thruput((double)remote_bytes_received);


nettest_unix.c

  local_thruput = calc_thruput(bytes_sent);
  ...
  messages_recvd    = dg_stream_results->messages_recvd;
  bytes_recvd   = send_size * messages_recvd;
  remote_thruput    = calc_thruput(bytes_recvd);
_______________________________________________
LNST-developers mailing list
lnst-developers@lists.fedorahosted.org
https://lists.fedorahosted.org/admin/lists/lnst-developers@lists.fedorahosted.org