[parsec-users] OS effect in benchmarks

Mahmood Naderan nt_mahmood at yahoo.com
Fri Nov 11 01:01:22 EST 2011

>From Fig. 3 we can conclude that the majority of time is spent in user instruction. If we consider "LCK" and "LAT" as OS behavior then bodytrack and swaptions should be excluded because they experience high OS behavior. 

Do you agree with that?

// Naderan *Mahmood;

From: kishore kumar <kishoreguptaos at gmail.com>
To: Mahmood Naderan <nt_mahmood at yahoo.com>; PARSEC Users <parsec-users at lists.cs.princeton.edu>
Sent: Wednesday, November 9, 2011 7:17 PM
Subject: Re: [parsec-users] OS effect in benchmarks

The following paper might answer your questions..........

K. K. Pusukuri, R. Gupta, L. N. Bhuyan, Thread Reinforcer: Dynamically Determining Number of Threads via OS Level Monitoring. In proceedings of IEEE International Symposium on Workload Characterization (IISWC), Austin, Texas, USA, Nov. 2011. 

2011/11/9 Mahmood Naderan <nt_mahmood at yahoo.com>

>I want to know is there any study about OS activity effect in PARSEC suite? should the effect be considered or system emulation is enough?
>// Naderan *Mahmood;
>parsec-users mailing list
>parsec-users at lists.cs.princeton.edu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.cs.princeton.edu/pipermail/parsec-users/attachments/20111110/500dac23/attachment.htm>

More information about the parsec-users mailing list