Thread: RE: [GD-Design] Game UI's
Brought to you by:
vexxed72
From: Tom F. <to...@mu...> - 2002-06-26 10:33:21
|
> -----Original Message----- > From: ham...@tm... [mailto:ham...@tm...] > Please note that 3 can have an extra advantage, depending on > precisely how much the implementation fits within the mores > of whichever OS(es) it runs on: thanks especially to nVidia, Ouch - careful with that hype - you'll have someone's eye out. IIRC, nVidia were almost the last major IHV to introduce multi-monitor cards. > an awful lot of "normal" games-playing people now have two or > more monitors on their PC - it's no longer the preserve of > professional graphic artists etc - and number 3 allows them > to actually take advantage of that extra monitor, without the > GUI designer having the headache of designing specifically > for multi-monitor. [snip] > [confession: I tend to get really pi**ed off when a game with > loads of dialogues refuses to play nice with multi monitor - > I've got the hardware, I got it working under > windows/X11/whatever, and its extremely frustrating not to be > bale to use it]. The problem with multimonitor is that it's a huge pain at the moment. If using OpenGL, the moment you enable the second monitor, all hardware rendering support vanishes. Various IHVs have fudged it so it doesn't (by tricking Windows into thinking it's running on a single large monitor) but that causes all sorts of problems of its own, especially the different-sized monitor case. Under DirectX, you can use multimonitor, and have been able to do for some time, but itmeans keeping track of multiple devices, each with its own set of texture, VBs, etc, even though they actually all live on the same card. This is changing for the better in DX9 I believe, so that a single device can have multiple RAMDAC surfaces (which is the way to go), but it does require new driver support, etc - so you need to decide when you're shipping and what the market support will be. So I'm going to express cautious support for multimonitor, especially if using DX9 and shipping in a year or so. But it is cautious. > Adam M Tom Forsyth - purely hypothetical Muckyfoot bloke. This email is the product of your deranged imagination, and does not in any way imply existence of the author. |
From: Tom H. <to...@3d...> - 2002-06-26 12:52:24
|
At 03:25 AM 6/26/2002, Tom Forsyth wrote: >If using OpenGL, the moment you enable the second monitor, all hardware >rendering support vanishes. Various IHVs have fudged it so it doesn't (by >tricking Windows into thinking it's running on a single large monitor) but >that causes all sorts of problems of its own, especially the different-sized >monitor case. Actually .. this isn't accurate any more (and hasn't been for quite some time). Right now my buddy runs a GeForce2 as primary AGP, and PCI TNT card as secondary. Both are hardware accelerated with OpenGL. The place where things get ugly is when a window straddles two monitors ... only the part of the window on the display where the window was created still works properly by default ;) If a window moves from one monitor to another, you can destroy the old RC and get a new one within that monitor and get acceleration again. You _might_ even be able to use the shared texture and display list stuff and only use one copy of textures and display lists for dual-head cards ... but that's probably a stretch ;) I don't know how things work with mutli-mon and going full screen .. specifically switching resolutions and generally dealing with devices, but I suspect there's some form of solution. >Under DirectX, you can use multimonitor, and have been able to do for some >time, but itmeans keeping track of multiple devices, each with its own set >of texture, VBs, etc, even though they actually all live on the same card. >This is changing for the better in DX9 I believe, so that a single device >can have multiple RAMDAC surfaces (which is the way to go), but it does >require new driver support, etc - so you need to decide when you're shipping >and what the market support will be. I'm not personally familiar with things under DX at the moment, but that sounds accurate. >So I'm going to express cautious support for multimonitor, especially if >using DX9 and shipping in a year or so. But it is cautious. I wouldn't make multi-mon a required feature ... nor would I put a lot of effort into it. Maybe a couple of days or a week ... tops. There just aren't enough people that use it yet. Tom |
From: Tom F. <to...@mu...> - 2002-06-26 13:59:37
|
> -----Original Message----- > From: Tom Hubina [mailto:to...@3d...] [snip] > Actually .. this isn't accurate any more (and hasn't been for > quite some > time). Right now my buddy runs a GeForce2 as primary AGP, and > PCI TNT card > as secondary. Both are hardware accelerated with OpenGL. The > place where > things get ugly is when a window straddles two monitors ... > only the part > of the window on the display where the window was created still works > properly by default ;) If a window moves from one monitor to > another, you > can destroy the old RC and get a new one within that monitor and get > acceleration again. You _might_ even be able to use the > shared texture and > display list stuff and only use one copy of textures and > display lists for > dual-head cards ... but that's probably a stretch ;) > > I don't know how things work with mutli-mon and going full screen .. > specifically switching resolutions and generally dealing with > devices, but > I suspect there's some form of solution. Sorry, my bad. I haven't kept up with multimon OGL for a while. So now it's just as klunky as DirectX 8 is, i.e. two devices, two contexts, two copies of everything and all that. Gagh. That's far too much work for most apps to cover such a tiny portion of the market. [snip] > >So I'm going to express cautious support for multimonitor, > especially if > >using DX9 and shipping in a year or so. But it is cautious. > > I wouldn't make multi-mon a required feature ... nor would I > put a lot of > effort into it. Maybe a couple of days or a week ... tops. There just > aren't enough people that use it yet. Agreed. The problem is that "a couple of days" doesn't sound like long enough to port any engine to knowing about multiple contexts, which seems to be what is required under both OGL and DX8. Things should be better in DX9. Obviously it's still in beta right now, but it looks like you can simply create a second display chain of surfaces, SetRenderTarget to them, render your scene, and it all Just Works. Now _that_ is something that would only take a few days, assuming you already had the concept of multiple viewports in your rendering engine (which most do). > Tom Tom Forsyth - purely hypothetical Muckyfoot bloke. This email is the product of your deranged imagination, and does not in any way imply existence of the author. |