On 18/11/2008, Jose Pablo Sarco <[EMAIL PROTECTED]> wrote: > Ok, but, how I can calculate the total time from the start of the first > sample to the end of the last sample with only the jtl file? I can't use the > timestamp because different requests are trigger between the start and the > end request that I'm calculating. >
Why not? The JTL file will have the timestamp (default start-time) and elapsed time of each request. Just keep track of the minimum start-time and maximum end-time (start+elapsed) for each request you are interested in. At the end, take the difference. This is necessary since entries are not necessarily in exact order of start time - for example, if request 1 takes 2 seconds, and request 2 starts shortly after request 1 but only takes 1 second, then the sample for request 2 will appear in the log before request 1. > Regards, > > Jose > > > > > -----Original Message----- > From: sebb [mailto:[EMAIL PROTECTED] > Sent: Martes, 18 de Noviembre de 2008 11:26 a.m. > To: JMeter Users List > Subject: Re: How calculate Throughput? > > On 18/11/2008, Jose Pablo Sarco <[EMAIL PROTECTED]> wrote: > > Hi all, > > > > > > > > I'm using Jmeter 2.3.2 on XP. I'm trying to process the CSV file > > generated by Jmeter listeners, but I have some problems with Throughput > > calculation. I found some definitions like: > > > > > > > > *Throughput = Number of requests / Total time to issue the requests. > > > > *For a single request, the throughput depends only on the time to issue > > a single request. > > > > *For multiple requests, the throughput depends also on the gaps between > > the requests. E.g. if a request is issued every 30 seconds, the > > throughput will be 2/min, and is largely independent of the time each > > request takes. > > > > > > > > I understood the definition for a single request (1/elapsed), but I > > don't understand very well how calculate it when I have multiple > > requests, I tried with: samples number/samples elapsed > > (3/(elapsed1+elapsed2+elapsed3)), but the result is different to the > > Throughput displayed by the aggregate listener. > > > > > > > > Somebody know the formula to calculate the Throughput for multiple > > requests? > > > > *Throughput = Number of requests / Total time to issue the requests. > > The total time is measured from the start of the first sample to the > end of the last sample, which will generally be greater than the sum > of the individual times. > > The throughput measures the number of requests per unit of time, so > you have to measure the overall elapsed time. > > > > > > > > > Regards, > > > > > > > > Jose > > > > > > > > > > > ################################################################################################################# > > Este documento puede incluir información confidencial y propiedad de > Neoris y deberá ser leído solamente por la > > o las personas a quienes está dirigido. Si usted ha recibido este mensaje > por error, por favor avise > > inmediatamente al remitente contestando y eliminando este correo. > Cualquier punto de vista u opiniones expresadas > > en este mensaje son del remitente y no necesariamente coinciden con > aquellas de Neoris. Este documento no deberá > > ser reproducido, copiado, distribuido, publicado, ni modificado por > terceros sin la autorización por escrito de Neoris. > > > > Este mensaje ha sido verificado contra virus. Visítenos en www.neoris.com. > > > ################################################################################################################# > > > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > > > > ################################################################################################################# > Este documento puede incluir información confidencial y propiedad de Neoris > y deberá ser leído solamente por la > o las personas a quienes está dirigido. Si usted ha recibido este mensaje > por error, por favor avise > inmediatamente al remitente contestando y eliminando este correo. Cualquier > punto de vista u opiniones expresadas > en este mensaje son del remitente y no necesariamente coinciden con aquellas > de Neoris. Este documento no deberá > ser reproducido, copiado, distribuido, publicado, ni modificado por terceros > sin la autorización por escrito de Neoris. > > Este mensaje ha sido verificado contra virus. Visítenos en www.neoris.com. > > ################################################################################################################# > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > > --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
