[X4U] XP <> Vista benchmarks [C]

[Stroller]

On 2 Feb 2007, at 19:22, Robert Ameeti wrote:
>
> The problem is the overhead that MS has mandated to ensure copy  
> protection for HD that all hardware mfgs must incur including the  
> graphics cards guys.
>
> See <http://www.cs.auckland.ac.nz/~pgut001/pubs/vista_cost.html>  
> for why Vista is slower than XP.

[CONTINUED]

Mr Gutmann's article would have been FAR more useful in assessing the  
actual "hidden cost of Vista" had he compared the playback  
performance of Vista & XP using unprotected hi-def content against  
playing DRM'd content using the DRM channels he describes. Since the  
AACS content-protection of Blu-ray & HD-DVD has now been  
circumvented, full-quality HD-DVD rips are available via BitTorrent  
with the DRM removed (although I don't know whether this was  
available at the time of the article's original publication). These  
can be played on XP or on Vista without meeting the DRM requirements,  
and playback performance of the the unencrypted rip could easily be  
compared against that of the DRM'd HD-DVD (of the exact same movie at  
the exact same resolution / quality / bit-rate).

This would allow one to usefully say that "your P4 3.2ghz playing  
DRM'd content under Vista gives the same playback performance of a  
2.5ghz P4 under XP playing unprotected content". An overhead of (say)  
50% would be enough to impress consumers a little bit with how  
craptastic DRM actually is, but it probably wouldn't stop many of  
them buying hi-def movies if that's what they want to watch.  
Unfortunately you'll never be able to headline that "an old 486 could  
play back this movie if it wasn't for Vista's DRM" because the  
playback demands of hd-content is so demanding, anyway. The whole  
point of a hi-def movie is that it's at a resolution of 1920 x 1080  
pixels - my back-of-a-fag-packet maths indicates that's about 6 times  
the bits to throw around your system than required for a standard 720  
x 480 NTSC DVD.

The content of Gutmann's article that is actually measurable and  
tangible is restricted to about the first 3 paragraphs of the  
"Unnecessary CPU Resource Consumption" section. And even then he  
doesn't ACTUALLY measure it, but uses phrases like "considerable cost  
to both ends of the connection". I like his comment "finding SSL  
being run inside a PC from one software module to another is just  
weird" - it amuses me, and he's quite right, but this surely might as  
well be applied to "the hidden cost of playing back DRM'd content on  
any licensed Blu-Ray / HD-DVD playback software".

Gutman says "twenty years ago, in their work on the ABYSS security  
module, IBM researchers concluded that the use of encrypted buses as  
a protection mechanism was impractical" and when I look up my  
thesaurus to describe this statement I come up with antonyms for  
insighful such as "obtuse", "vacuous" and "vapid". Well, duh! Of  
course encrypted buses were impractical on am Amiga or a 68000  
Macintosh II. Those things had like 7mhz processors and cost hundreds  
of thousands of pounds - nowadays ISPs give away routers with 200mhz  
processors for free when you subscribe to their service. Vista proves  
that "encrypted buses as a protection mechanism" are now practical  
(whether we like it or not) and this is only a testament to the  
improvements made to computing power over the last 2 decades.

 From reading the MythTV list I understand that you're basically  
advised to get a modern dual-core processor if you want to play back  
any hi-def content - that's playing it unprotected content under  
Linux. Once you've got that sort of processing power kicking around  
it's unclear that the overhead of DRM is actually punishing (despite  
any hyperbole Mr Gutmann or I might wish to use in constructing our  
descriptions of it).

Stroller.