XO battery/performance

[Martin Langhoff]

On Mon, May 28, 2012 at 4:43 PM, Yioryos Asprobounitis
<mavrothal at yahoo.com> wrote:
> I do not think I claimed a sophisticated test.

Sorry, I didn't mean to hurt here. But it's really not testing what
you thought it was testing, as others have pointed out.

>> If you want an idea of low-level performance, I can suggest
>> running LMBench.
>
> Thanks for the advice. If you can point to an lmbench rpm or how to compile it without bitkeeper, by now a commercial software, would be even better.

True, it seems to want bitkeeper. I remember I read the makefile, and
it uses some "bk tellmeversion > .version" thing to get the a revision
or version string. Just put a random string in the generated file, and
if you want remove that part of the makefile.

> The XOs use 100% CPU, if you do anything with them other than looking at the screen.

The CPU is only one part of a complex system. Cranking out audio
consumes power, reading from the camera consumes lots of pwoer,
writing to eMMC or SD consumes power.

> I would think that a test that uses 100% cpu is as close to "normal" as it gets when we are talking XO use.

Hmmmm. Sorry, I disagree here. You have usage cases, and they are very
different from each other. There isn't much "normal".

 - writing a document
 - making a drawing
 - using TurtleArt
 - using one of the TamTams
 - recording or playing back video
 - reading content on websites
 - watching a video online
 - reading an ebook

Some of those are moderate power consumers, some are big power hogs,
some are pretty efficient. A special case needs to be consider for
reading a page on a website, which can be extremely efficient or a
power hog, depending on animations on the page.

cheers,



m
-- 
 martin.langhoff at gmail.com
 martin at laptop.org -- Software Architect - OLPC
 - ask interesting questions
 - don't get distracted with shiny stuff  - working code first
 - http://wiki.laptop.org/go/User:Martinlanghoff