|
Message
From: "Jim Dempsey" <tapedisk@a...>
Date: Fri, 29 Aug 2003 16:56:33 -0500
Subject: Re: [openrisc] OR1200 Power Consumption
I would think it an apples to oranges comparison to compare the published
consumption of the ARM946E (0.8mW/MHz) to a value attained via synthesis.
There are other factors involved (e.g. OtherBone/WishBone). Brian should be
considering the completed system power consumption.
Jim
----- Original Message -----
From: "Damjan Lampret" <lampret@o...>
To: <openrisc@o...>
Sent: Friday, August 29, 2003 1:56 PM
Subject: Re: [openrisc] OR1200 Power Consumption
> Brian,
>
> there is no precise power consumption analysis, more like an estimate. If
> you want a very precise estimate, and a low power consumption, synthesize
> OR1200 with clock gating, and run dhrystone code and use node toggling to
> calculate power conumption. The result should be in the range of ARM946E.
>
> regards,
> Damjan
>
> ----- Original Message -----
> From: "Brian Adams" <brian.adams@a...>
> To: <openrisc@o...>
> Sent: Friday, August 29, 2003 8:41 PM
> Subject: [openrisc] OR1200 Power Consumption
>
>
> > Damjan,
> >
> > I've been comparing some of the specs for the OR1200 with that of
> > various ARM processors. (specifically the ARM7 and ARM9 varieties) From
> > the info in the OR1200 overview it compares quite favorably in speed and
> > area, but not so well in power. For example, the 0.18u ARM920T is
> > advertised as consuming 0.8mW/MHz, while the OR1200 is advertised at
> > <0.5W at 250 MHz (@50%). Assuming a linear relationship that would mean
> > the OR1200 is ~2mW/Mhz. Does this seem right to you. Do you know what
> > configuration you used to get the above figure? I'm trying to justify
> > (to others) using the OR1200 for many reasons, but 2x power consumption
> > is troubling. I was hoping there was a simple (any possibly obvious)
> > explanation.
> >
> > Thanks,
> > Brian
> >
> >
>
>
|
 |
Browse
