[openbeos] Re: A question/concern re OSS
- From: "Shannon Mackey" <coalesce@xxxxxxxxxxxxxxxxxx>
- To: openbeos@xxxxxxxxxxxxx
- Date: Tue, 15 Jan 2002 23:27:25 -0500
>>Helmar Rudolph <helmar@xxxxxxxxx> wrote:
>>
>>>distros, who initiates and coordinates talks with companies
>>>like AMD (for SMP, etc support) or any other one we would
>>>like to see support the BeOS? BeUnited? And if BU does it,
>>Who ever on this list is the best person for that patricular
>>job. If someone of BeUnited has had contact with AMD
>>in the past, he might announce it here on the mailinglist,
>>and we may elect him to do it. Probably any other person.
>>
>
>I've considered this part for a while now. What does Intel or AMD have
>to gain by helping us? An OS that is so efficent its users don't need
>to buy the Fastest chip currently available every 18 months. I can
see
>why they would want Windows to dominate. I've not upgraded to a
faster
>chip since I found BeOS. I'm still happy with a P2 400 for nearly
>everything. WIN XP is completely out of my relm with my systems. I
would
>HAVE to buy the latest and greatest to use it. THATS what keeps Intel
and AMD
>going. This is an honest question.. What Does AMD have to gain by
helping
>us?
>LewisB
>
An honest answer, from someone who's been there... AMD doesn't have
anything to gain by helping OBOS, BUT OBOS, as was BeOS, is likely to
work with most all AMD motherboards/processors anyway., though not take
advantage of any special features :(
you all might find this interesting
Well, though I'm not with AMD anymore. I was testing motherboard
chipsets and data buses with K6 and K7(athlon). I was even running
BeOS on Dual K7 systems in '98! Why they took 2 years to release; one
can only guess. BeOS ran like a champ on every board/processor that
came through during '97-'98.
I spent most time analyzing the PCI and AGP buses. We used a nifty
tool that would measure the efficiency (actual data : overhead traffic)
of the PCI bus between North and South Bridges. It was quite
revealling to look at various OSs.
Win98 generally sat at around 5% efficiency, only reaching 10% while
viewing a movie fullscreen at 640x480 resolution. NT sat around 12%,
and could peak at 35% with the same movie situation. BeOS, on the
other hand, remained at 50%, whether or not it was idling, or you were
running 6 movies to a rotating cube, while capturing/watching video via
USB cameras, playing music, etc., etc.
While Microsloth did annoyingly affect the development direction of
some aspects of AMDs devices, AMD did also, as a matter of course test
with 35 different OSs with every new spin of the silicon or board sets.
We tested a dozen flavors of unix, a half dozen flavors of DOS, half
dozen flavors of winblows, half dozen RTOSs, even Next, and BeOS.
Though, even if our designs were "to spec," but winblows had a problem
with it, it might just get changed to meet the winblows "standard"
Shannon
"there'll be one corporation, selling one little box, it'll be what you
want, and tell you what you want, and cost what ever you've got" -Greg
Brown
Other related posts:
