Daniel Phillips wrote:
On Friday 06 May 2005 09:16, Timothy Miller wrote:So I had a nice long reply typed up and hit the wrong button... Oh well. Here is my short reply.
On 5/6/05, Daniel Phillips <[EMAIL PROTECTED]> wrote:
My opinion is that if we want to boot Windows, and we want to supportWas it decided whether to implement (a few) text modes only, vs vga graphics modes as well? It would seem sensible to implement only text modes for the time being, in order to leave more room on the FPGA for our own (non-vga) graphics mode(s).
If it is going to be text mode only, does that reduce the number of
registers that need to be fully documented? Or do we want to fully
document the vga graphics-related registers anyway, and just mark
the ones that are not involved in text modes?
safe mode, we need to support some graphics modes.
How important is it that the very first FPGA release boot Windows? How about just booting Linux for the first release, and spend some saved gates on more multipliers and memory control logic?
Maybe this is something we need to add to the survey:
a) Would you buy this card _only_ if it boots Windows (where nobody really needs an open graphics card anyway)
b) Would you want this card even more if it _doesn't_ boot Windows :-)
c) Do you dual boot? Don't you have a working video card already, that already has Windows drivers? Why not keep using that one to boot Windows?
We can always follow up a few weeks later with a logic upgrade to boot Windows. But I for one would not want to delay the Linux release by even a week to accommodate that.
And: will Windows drivers even be ready by then? Is anybody really passionate enough to write those drivers, particularly given the extra level of pain they can expect for low-level Windows code?
/me listens for a thundering herd of volunteers...
We need the prototype first. We need enough of a driver for the prototype to prove that it does something. From that we can get people to buy development cards. We need enough of a base driver for the development card to be able to use it for development in both Windows and Linux. A prototype graphics core for the development card would be good when it's ready so the general community has a starting point, but it's not absolutely necessary.
There will be many updates to the graphics core as development continues. At this point, hardware and driver development is progressing in parallel. The big bonus being that is the driver developers find an issue, we can decide if it should be fixed in hardware or software. If its a hardware fix, the core gets changed and those changes get propagated out to the development community.
At some point in here, a user could actually buy the development card just to use as a "cool" graphics card.
Then we get to the ASIC. I personally don't think the ASIC should be release until the core supports what an average user would expect. We may not be targeting the average user, but the ASIC is going to be the first general public face for the group and needs to be good.
Patrick M _______________________________________________ Open-graphics mailing list [email protected] http://lists.duskglow.com/mailman/listinfo/open-graphics List service provided by Duskglow Consulting, LLC (www.duskglow.com)
