WQXGA LCD for OS/2?

[Dariusz Piatkowski]

Has anyone attempted to use an LCD with greater then the 1920x1600 resolution?

I have a Samsung SyncMaster 245T right now which is hooked up to my ATI Radeon
X850 PE card. That display is set up as WUXGA, which is the 1920x1600
resolution.

But...I'm looking at the purchase of Samsung 305T monitor, that beast will do
2560 x 1600, which would be awefully nice!

The monitor does require DVI-D Dual Link hookup, which I think the ATI card
already provides, although I'm checking on this because while the card does have
2 DVI-I connectors I'm not actualy sure if the 2560x1600 resolution is
supported, nor whether the drivers will combine the 2 DVI-I ports into a single
signal feed. My SNAP drivers do allow me to define whatever resolution I might
want to try, just never tried anything that high on this video card.

Thanks!

[Steve Wendt]

On 09/29/12 06:59 am, Dariusz Piatkowski wrote:

> I have a Samsung SyncMaster 245T right now which is hooked up to my ATI Radeon
> X850 PE card. That display is set up as WUXGA, which is the 1920x1600
> resolution.

I guess you mean 1920x1200.

> The monitor does require DVI-D Dual Link hookup, which I think the ATI card
> already provides, although I'm checking on this because while the card does have
> 2 DVI-I connectors I'm not actualy sure if the 2560x1600 resolution is
> supported, nor whether the drivers will combine the 2 DVI-I ports into a single
> signal feed. My SNAP drivers do allow me to define whatever resolution I might
> want to try, just never tried anything that high on this video card.

The odds of it working are low - I can tell you it was never tested.

[Dariusz Piatkowski]

Hi Steve!

On Sat, 29 Sep 2012 21:34:35 UTC, Steve Wendt <spa...@forgetit.org> wrote:

> On 09/29/12 06:59 am, Dariusz Piatkowski wrote:
>
> > I have a Samsung SyncMaster 245T right now which is hooked up to my ATI Radeon
> > X850 PE card. That display is set up as WUXGA, which is the 1920x1600
> > resolution.
>
> I guess you mean 1920x1200.

Oopss..yes, absolutely. In my attempts to get all the correct resolution
acronyms right I missed the most relevant part, the correct resolution itself.
Yes, my SNAP display settings today are set to do 1920x1200.

> > The monitor does require DVI-D Dual Link hookup, which I think the ATI card
> > already provides, although I'm checking on this because while the card does have
> > 2 DVI-I connectors I'm not actualy sure if the 2560x1600 resolution is
> > supported, nor whether the drivers will combine the 2 DVI-I ports into a single
> > signal feed. My SNAP drivers do allow me to define whatever resolution I might
> > want to try, just never tried anything that high on this video card.
>
> The odds of it working are low - I can tell you it was never tested.

OK, so I'm thinking the question therefore is more about the following: if
'gamode add' allows me to add this resolution to the list, and if I can get all
the timings/settings synced up between the SNAP and the LCD itself, is there
anything int he SNAP drivers themselves that may prevent me from successfully
running at that resolution?

So when you say the odds are 'low'...what sort of reservations do you have?

I'm thinking this would be very experimental.

Thanks!

[Steve Wendt]

On 09/30/12 12:54 pm, Dariusz Piatkowski wrote:

>>> The monitor does require DVI-D Dual Link hookup, which I think the ATI card
>>> already provides, although I'm checking on this because while the card does have
>>> 2 DVI-I connectors I'm not actualy sure if the 2560x1600 resolution is
>>> supported, nor whether the drivers will combine the 2 DVI-I ports into a single
>>> signal feed.
>>

>> The odds of it working are low - I can tell you it was never tested.
>
> OK, so I'm thinking the question therefore is more about the following: if
> 'gamode add' allows me to add this resolution to the list, and if I can get all
> the timings/settings synced up between the SNAP and the LCD itself, is there
> anything int he SNAP drivers themselves that may prevent me from successfully
> running at that resolution?
>
> So when you say the odds are 'low'...what sort of reservations do you have?

You expressed it yourself in the quoted text - dual link DVI. As I
recall, 1920x1200 is the upper limit for a single TMDS. Even that
required some extra tricks, since 1600x1200 was the limit with regular
timings:
http://www.anandtech.com/show/1799/2

6-7 years ago, video cards with 2 DVI connectors were rather rare; LCD
panels at the time also weren't all that great, so anyone who cared
about image quality and didn't have a fortune to spend stuck with a CRT
for a few more years.

[tho...@antispam.ham]

Steve Wendt <spa...@forgetit.org> writes:

> Dariusz Piatkowski wrote:

>>>> The monitor does require DVI-D Dual Link hookup, which I think the ATI card
>>>> already provides, although I'm checking on this because while the card does have
>>>> 2 DVI-I connectors I'm not actualy sure if the 2560x1600 resolution is
>>>> supported, nor whether the drivers will combine the 2 DVI-I ports into a single
>>>> signal feed.

>>> The odds of it working are low - I can tell you it was never tested.

>> OK, so I'm thinking the question therefore is more about the following: if
>> 'gamode add' allows me to add this resolution to the list, and if I can get all
>> the timings/settings synced up between the SNAP and the LCD itself, is there
>> anything int he SNAP drivers themselves that may prevent me from successfully
>> running at that resolution?
>>
>> So when you say the odds are 'low'...what sort of reservations do you have?

> You expressed it yourself in the quoted text - dual link DVI. As I
> recall, 1920x1200 is the upper limit for a single TMDS. Even that
> required some extra tricks, since 1600x1200 was the limit with regular
> timings:

Dell sells a 2560x1600 monitor (same resolution as the Apple Cinema display)
with a DisplayPort input, something offered on the newer ThinkPads.
That avoids dual link DVI, but isn't a solution to the problem. Pick your
poison.

I think the eCS folks are going to need to address the video driver issue
at some point. I'm running dual head (effectively 3200x1200) using SNAP
on a Matrox G450, but I'm at Warp 4.52; if "upgrading" to eCS 2.x on new
hardware means downgrading my desktop size, it's not going to happen.

[Dariusz Piatkowski]

Well, there is good news and bad news.

Good news: the ATI X850 XT PE card I have does have 2 DVI connectors

Bad news: both the connectors on this card are single link DVI, which as you
pointed out simply do not have the required bandwidth to support the 2560x1600
resolution

Sooo...I'm thinking it's 'game over' as far as that attempt goes. The LCD panel,
used at that, is much to pricey to risk that kind of a $$$.

Strangely enough, depending on what chip actually implements the TMDS protocol
(most were Sil chips) some of them did support dual link DVI interface when both
single link DVI ports were NOT used. The card auto-detected this and would
switch over to run the display MAX OUT only on that single port. I simply have
no way to tell whether the genuine ATI card I have falls into this though.

For what it's worth I did attempt the 'gamode add' command:

[G:\snap]gamode add 2560 1600 32
Failed to add mode 2560 x 1600 x 32!

So at least the driver is refusing this, but that could also be due to the LCD
capability being detected as NOT supporting that mode (my current 245T LCD that
is).

Thanks again Steve!!!

BTW: I've been meaning to ask you : where the heck in the driver is the max
memory and GPU clock stored? Or better yet, how the heck can I find it?

[Dariusz Piatkowski]

You know...if I can pick up another 245T LCD panel at a good price I may very
well take the same approach you have...a 3200x1200 would sure be nice!!!

Any things one needs to watch out for / be aware of if taking this approach? In
general, how do apps behave? Any problems with stuff popping up in the middle of
the virtual screen?

[Steve Wendt]

On 10/03/12 05:43 pm, Dariusz Piatkowski wrote:

> [G:\snap]gamode add 2560 1600 32
> Failed to add mode 2560 x 1600 x 32!
>
> So at least the driver is refusing this, but that could also be due
> to the LCD capability being detected as NOT supporting that mode

More likely, it exceeds some maximum pixel clock value for the driver,
or maybe even some global maximum. Video modes can be configured
independent of the attached monitor.

> BTW: I've been meaning to ask you : where the heck in the driver is the max
> memory and GPU clock stored? Or better yet, how the heck can I find it?

You mean the gamemclk stuff? I don't think those actually work on ATOM
BIOS cards; it always says 400MHz I think.

[tho...@antispam.ham]

Dariusz Piatkowski wrote:

>> Dell sells a 2560x1600 monitor (same resolution as the Apple Cinema display)
>> with a DisplayPort input, something offered on the newer ThinkPads.
>> That avoids dual link DVI, but isn't a solution to the problem. Pick your
>> poison.
>>
>> I think the eCS folks are going to need to address the video driver issue
>> at some point. I'm running dual head (effectively 3200x1200) using SNAP
>> on a Matrox G450, but I'm at Warp 4.52; if "upgrading" to eCS 2.x on new
>> hardware means downgrading my desktop size, it's not going to happen.

> You know...if I can pick up another 245T LCD panel at a good price I may very
> well take the same approach you have...a 3200x1200 would sure be nice!!!
>
> Any things one needs to watch out for / be aware of if taking this approach? In
> general, how do apps behave? Any problems with stuff popping up in the middle of
> the virtual screen?

Definitely check which video cards SNAP supports for dual head mode;
it's in the SNAP documentation.

DeScribe behaved oddly in dual head mode, but SciTech solved that
problem with the environment variable:

SET SNAPHRES=2

(I have them side-by-side; if you stack the displays, I suspect the
environment variable would be SNAPVRES.)

You can specify which head popups are centered on in the configuration
notebook, though it's not foolproof.

[Dariusz Piatkowski]

Hi Tholen!

So it appears to be a workable solution...at least worth trying as attemptign to
get the 2560x1600 LCD probably just isn't going to happen.

Thanks for the info!

[Dariusz Piatkowski]

Hi Steve!

Yea, the gamemclk indeed.

'show' option gives:

[G:\snap]gamemclk show
Memory clock settings for ATI Radeon X850 Series (device 0):

Default: 400.00 MHz
Current: 400.00 MHz
Maximum: 0.00 MHz

'set' option on the other hand produces:

[G:\snap]gamemclk set 410
Memory clock too high!

[G:\snap]gamemclk eset 500
Device does not support a programmable engine clock.

...strangely enough, the SNAP SDD kit source tree has a version of the gamemclk
(with source code) which accepts the above settings and claims to successfully
implement them as well...I think I will go back to that and try to analyze
whether the clocks are actually being re-set.

The X850 XT PE version I have actually has a standard mem clock = 590 and engine
clock = 540...so the default settings are quite a bit off...not quite sure what
the final OS/2 performance differential that would actually translate to
though...if anything, I'm curious...

[Alex Taylor]

On Tue, 2 Oct 2012 11:55:50 UTC, tho...@antispam.ham wrote:

> I think the eCS folks are going to need to address the video driver issue
> at some point. I'm running dual head (effectively 3200x1200) using SNAP
> on a Matrox G450, but I'm at Warp 4.52; if "upgrading" to eCS 2.x on new
> hardware means downgrading my desktop size, it's not going to happen.

Um, why would it? Same OS, same card (I assume), same video drivers.

--
Alex Taylor
Fukushima, Japan
http://www.altsan.org

Please take off hat when replying.

[tho...@antispam.ham]

"Alex Taylor" <mai...@reply.to.address> writes:

>> I think the eCS folks are going to need to address the video driver issue
>> at some point. I'm running dual head (effectively 3200x1200) using SNAP
>> on a Matrox G450, but I'm at Warp 4.52; if "upgrading" to eCS 2.x on new
>> hardware means downgrading my desktop size, it's not going to happen.

> Um, why would it? Same OS, same card (I assume), same video drivers.

If it's the same card, then it's not "new" hardware.

[Steve Wendt]

The odds of running an OS/2 desktop at 3200x1200 on any hardware
released since 2006 are pretty close to zero...

[tho...@antispam.ham]

Then we're in agreement. "It's not going to happen."

The eCS folks will need to do something about that, sooner or later.

[Dariusz Piatkowski]

Instead of re-starting from scratch, why wouldn't the SNAP model be picked up
and developed further? Did it all come down to cost? Or was there a technical
reason...I had always wondered given the well established base of SNAP drivers
out there.

[tho...@antispam.ham]

"Dariusz Piatkowski" <dariusz@_NO-SPAM_mnsi.net> writes:

> Instead of re-starting from scratch, why wouldn't the SNAP model be picked up
> and developed further? Did it all come down to cost? Or was there a technical
> reason...I had always wondered given the well established base of SNAP drivers
> out there.

My understanding is that the SNAP model has been picked up. The
question is, how much can they do with it in the absence of
information? I thought that one reason SciTech got out of the
business is because video card manufacturers were no longer
willing to provide details about their hardware, fearing the
competition from other card makers.

[Steve Wendt]

On 10/10/12 05:02 pm, tho...@antispam.ham wrote:

> I thought that one reason SciTech got out of the business is because
> video card manufacturers were no longer willing to provide details
> about their hardware, fearing the competition from other card
> makers.

That can be an issue, but mainly for Nvidia.
http://www.phoronix.com/scan.php?page=news_item&px=MTEyNTE

[Mark Dodel]

On Wed, 10 Oct 2012 23:57:59 UTC, "Dariusz Piatkowski"
<dariusz@_NO-SPAM_mnsi.net> wrote:

-> On Wed, 10 Oct 2012 16:50:49 UTC, tho...@antispam.ham wrote:
->
-> > Steve Wendt <spa...@forgetit.org> writes:
-> >
-> > >>>> I think the eCS folks are going to need to address the video driver issue
-> > >>>> at some point. I'm running dual head (effectively 3200x1200) using SNAP
-> > >>>> on a Matrox G450, but I'm at Warp 4.52; if "upgrading" to eCS 2.x on new
-> > >>>> hardware means downgrading my desktop size, it's not going to happen.
-> >
-> > >>> Um, why would it? Same OS, same card (I assume), same video drivers.
-> >
-> > >> If it's the same card, then it's not "new" hardware.
-> >
-> > > The odds of running an OS/2 desktop at 3200x1200 on any hardware
-> > > released since 2006 are pretty close to zero...
-> >
-> > Then we're in agreement. "It's not going to happen."
-> >
-> > The eCS folks will need to do something about that, sooner or later.
->
->
-> Instead of re-starting from scratch, why wouldn't the SNAP model be picked up
-> and developed further? Did it all come down to cost? Or was there a technical
-> reason...I had always wondered given the well established base of SNAP drivers
-> out there.

Mensys has the SNAP source but they don't have anyone qualified who
can work on it or who is interested in working on it.

Mark


--
From the eComStation 2.1 GA desktop of Mark Dodel

Warpstock 2013 - http://www.warpstock.org
Warpstock Europe 2013 - http://www.warpstock.eu

[Dariusz Piatkowski]

Hi Mark,

OK, makese sense. I'm curious if we could try somewhat of a drastically
different approach to this.

Is there a realistic chance that we as a community could, through let's say some
type of a vote maybe, nominate/pick 1-3 video cards/chipsets for development.
Goal here being that we basically focus the job on a very limitted set of
hardware, but put some real effort (support and financial as required) to
complete the driver development.

I for one can't imagine that the typical OS/2 video requirement is that
crazy...what I am guessing is that the SNAP drivers no longer have the support
for the newer versions of the chipset. Quite frankly though, my ATI X850 XT PE
does just about everything I want it to do, the only thing missing at the moment
is higher resolution and maybe more roboust support for the various LCD panel
configurations.

Has anyone considered this? If I could obtain a new version of SNAP for a
non-ATI chipset/card that does give me the stuff I'm looking for, I'd have no
issues moving up to the supported hardware. That in my mind is the price we must
be willing to pay for carrying on with the OS/2 platform.

Am I the only one thinking this?

[Roderick Klein]

We do have a suitable developer but he is working on other stuff now for
eCS 2.2.

Roderick Klein
Mensys