[parsec-users] serial version of parsec suite
kishoreguptaos at gmail.com
Thu Feb 2 15:24:06 EST 2012
If you are running four different single threaded programs then you have no
need to care about lock-contention. But you need to consider memory
bandwidth as programs share memory channels.
BTW, my point about lock-contention is that, for example, as you know,
running 8 single threaded ferret instances is different compared running a
ferret instance with 8 threads on a multicore machine. In the later case
you need to consider lock-contention also if you want to understand
scalable performance of these programs on multicore machines.
Regarding using single threaded programs or multithreaded totally depends
on the problem that you are interested in.
Kishore Kumar Pusukuri
On Thu, Feb 2, 2012 at 11:50 AM, Mahmood Naderan <nt_mahmood at yahoo.com>wrote:
> Seems that I asked the question wrongly. The problem is not running one
> program 4 times. I meant 4 different programs on 4 cores. As I said:
> core 0 runs ferret
> core 1 runs fluidanimate
> core 2 runs dedup
> core 3 runs facesim
> Each core runs the serial version of application (not multithreaded).
> I really don't understand why should I care memory bandwidth or lock
> contention in this configuration.
> To biswabandan:
> The paper says this in section 5.1:
> We simulate both 4-core (for sequential workloads) and 16-core (for
> parallel workloads) CMP systems
> So I think it separate the workloads based on parallel or sequential
> // 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: Thursday, February 2, 2012 6:56 PM
> Subject: Re: [parsec-users] serial version of parsec suite
> Apart from memory bandwidth, one more important factor that influence
> scalable performance of multithreaded program is lock-contention. I have
> observed that compared to memory bandwidth, most of the PARSEC programs are
> not scalable on a multicore machine with a large number of cores (e.g. 64
> core machine) is because of lock-contention.
> Kishore Kumar Pusukuri
> On Thu, Feb 2, 2012 at 2:23 AM, Mahmood Naderan <nt_mahmood at yahoo.com>
> >The main characteristic of PARSEC suite is that it is a collection of
> multithreaded applications. However it is possible to run a serial version
> of those applications. Now I want to know is it fine to use serial version
> like SPEC benchmarks? For example, is it fair to run ferret, fluidanimate,
> dedup and facesim on a 4 core CMP (each core runs one thread) and report
> the results?
> >// Naderan *Mahmood;
> >parsec-users mailing list
> >parsec-users at lists.cs.princeton.edu
> parsec-users mailing list
> parsec-users at lists.cs.princeton.edu
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the parsec-users