VOGONS

Common searches


Reply 20 of 41, by Nicht Sehr Gut

User metadata
Rank l33t
Rank
l33t

Originally posted by MajorGrubert You are right, fsvga.sys looks like a smoking gun.

So it's part of the problem.

I believe Snover is right and it is used only to display the splash screen at boot time.

So it's not part of the problem. Ok, I'm confused.

BTW, sympathies on your recent crash.

Reply 21 of 41, by MajorGrubert

User metadata
Rank Member
Rank
Member
Nicht Sehr Gut wrote:

Ok, I'm confused.

Ok, I was not very clear. I believe fsvga.sys is not part of the problem, in spite of its name. I say this based on its dependencies and the fact that it is not loaded by Windows 2000 after you are logged in, even if you run full screen DOS programs.

Regards,

Major Grubert

Athlon 64 3200+/Asus K8V-X/1GB DDR400/GeForce FX 5700/SB Live! 5.1

Reply 22 of 41, by Schadenfreude

User metadata
Rank Member
Rank
Member
DosFreak wrote:
Hmm, Everyone that plans on testing list your: […]
Show full quote

Hmm, Everyone that plans on testing list your:

Video Card
OS (and SP Level)
Motherboard

Video Card: Rage Mobility M4
OS (and SP Level): Windows XP SP1
Motherboard: Dell Inspiron 8000 Laptop

VGATEST 0 hangs XP. (Woohoo!)
VGATEST 1 seems to work, but detects the wrong amount of video RAM. 256K?
VGATEST 2 seems to work - provided it does not test anything beyond mode 7 text modes... (screen goes blank for a little, then drops to a prompt, after testing mode 7).
VGATEST 3 will hang XP without fail. After pressing enter several times, it will eventually fill the LCD screen with a ghostly fog pattern in shades of white, grey and black. I can even see faint LCD "burn-in" of the Windows XP Home login screen. Wonder what is up with this. Anyhow, the laptop fails to respond when this happens, no CTRL-ALT-DEL or anything. I have to hold in the power switch.
VGATEST 4 and above will blank the screen then jump back to the prompt.

This could be caused by my using a laptop. I'll test with a desktop next.

Considering what VGATEST 3 actually does... wouldn't profile.exe be a better testing utility?
showthread.php?threadid=517

Reply 23 of 41, by dvwjr

User metadata
Rank Member
Rank
Member

Originally posted by Schandenfreude:

VGATEST 0 hangs XP. (Woohoo!)

This is the only useful mode for the three PCs I have tested with VGATEST.EXE, a Matrox G450 PCI with XP(SP1), an NVidia GeForce 4 MX440 with XP(SP1) and an old Trident S3-864 1MB with NT4(SP6a). In the DOS boxes many of the other VGATEST.EXE command-line options result in no action, but starting with the VGATEST.EXE command-line option of zero will allow as many of the modes to be tested as a WinXP card/driver combination will allow.

VGATEST 1 seems to work, but detects the wrong amount of video RAM. 256K?

It doesn't detect the "wrong" amount of memory, the original IBM VGA had just 256kb of RAM, addressable in two banks of 128kb or four banks of 64kb. This is actually an old test program to test actual VGA or SuperVGA cards of the early to mid 1990s. It is testing what is was designed for, not the future which we now inhabit...

Free VGA project - Hardware Level VGA and SVGA Video Programming Information Page

VGATEST 2 seems to work - provided it does not test anything beyond mode 7 text modes... (screen goes blank for a little, then drops to a prompt, after testing mode 7).

Well VGATEST command-line option 2 is one test mode that will work, it is only supposed to test the standard VGA text modes: 0h, 01h, 02h, 03h and 07h. Those are the only "standard" VGA text modes, so the program is working as designed when nothing beyond mode 07h is tested.

VGATEST 3 will hang XP without fail. After pressing enter several times, it will eventually fill the LCD screen with a ghostly fog pattern in shades of white, grey and black. I can even see faint LCD "burn-in" of the Windows XP Home login screen. Wonder what is up with this. Anyhow, the laptop fails to respond when this happens, no CTRL-ALT-DEL or anything. I have to hold in the power switch.

Must be an unusual reaction by your laptop display adapter/driver under WinXP. This specific series of graphic mode tests: mode 04h(320x200/4 color), mode 05h(320x200/4 color), mode 06h(640x200/BW), mode 0Dh(320x200/16 color), mode 0Eh(640x200/16 color), mode 0Fh(640x350/BW), mode 10h(640x350/16 color), mode 11h(640x200/2 color), mode 12h(640x480/16 color), mode 13h(320x200/256 color). If your laptop chokes on this simple series of VGA graphics modes, then your laptop fail where a 1987 IBM VGA card will succeed under WinXP in fullscreen mode. Hell, most of them are IBM CGA modes from the early 1980s. Something very strange here.

VGATEST 4 and above will blank the screen then jump back to the prompt.

Yep. Any of the SVGA tests will vary by adapter. The Matrox G450 will refuse to do anything but return to the DOS prompt for VGATEST.EXE command-line options 4-8, but the old 1MB Trident S3-864 will attempt and succeed on one SVGA mode under VGATEST.EXE command-lines options 4-5. What can you say... But I have found that if you start with VGATEST.EXE command-line option 0, it will allow you to test every one of the modes tested by VGATEST.EXE...

This could be caused by my using a laptop. I'll test with a desktop next.
Considering what VGATEST 3 actually does... wouldn't profile.exe be a better testing utility?

I think that the reason that this old test might be better is that it actually does not rely on the video adapter/BIOS or WinXP video driver reporting on what video modes it thinks that it can do, it actually TRIES them, letting the human with eyeballs determine whether or not the tests succeeded or failed...

What I hope the VGATEST.EXE results confirm is what is apparent from most of the posts here at VOGONS, that SVGA mode 101h(640x480/256 color) will work on most video adapter under WinXP(SP1). Looks like Microsoft put enough SVGA support in WinXP to help most video adapter/BIOS to support SVGA mode 101h, and since NOLFB.COM makes any DOS graphics app think that only bank-select SVGA video memory is available most 640x480/256 color games should work, provided they incorporated the fall-back bank-select along with the preferred linear frame buffer coding...

Even my old Dell Optiplex XMT5120 tower I use as a multi-boot test server platform for various server OSs with show the SVGA mode 101h(640x480x256 color) fullscreen under Windows NT4(SP6a). With NOLFB, many of the games using SVGA mode 101h work, and WinNT4 is no spring chicken...

Let the tests continue...

dvwjr

Reply 24 of 41, by MajorGrubert

User metadata
Rank Member
Rank
Member
DosFreak wrote:
Hmm, Everyone that plans on testing list your: […]
Show full quote

Hmm, Everyone that plans on testing list your:

Video Card
OS (and SP Level)
Motherboard

Here we go:

Video Card: Abit GeForce 4 MX440 (AGP 8x) with Nvidia driver 44.03
OS: Windows 2000 SP3, vga.sys patched by vgafix.exe
Motherboard: ASUS A7V333 (Via KT333)

VGATEST 1: ok, 256K
VGATEST 2: all text modes ok (0,1,2,3 and 7, including the scroll tests)
VGATEST 3: all modes ok (4,5,6, D, E, F, 10, 11, 12 and 13)
VGATEST 4: all modes ok (100, 101, 103, 105, 107)
VGATEST 5: all modes ok (111, 114, 117, 11A)
VGATEST 6: modes 112, 115, 118 appeared distorted, the graphics would only take three quarters of the screen horizontally, with the rightmost quarter black; mode 11B had the last quarter partially filled with random lines (and it did not look like a monitor adjust problem, because my monitor always reported vertical frequency being at 60Hz)
VGATEST 7: all modes ok (102, 104, 108)
VGATEST 8: a black screen (but monitor did not sleep due to wrong frequencies), sometimes showing the message "Scrolling down 3 lines"
VGATEST 0: obviously the same results as running all the tests in a row, with the same problems at tests 6 and 8

Update: listed video driver version in the specs

Regards,

Last edited by MajorGrubert on 2003-06-09, 15:07. Edited 1 time in total.

Major Grubert

Athlon 64 3200+/Asus K8V-X/1GB DDR400/GeForce FX 5700/SB Live! 5.1

Reply 25 of 41, by DosFreak

User metadata
Rank l33t++
Rank
l33t++

Windows XP Pro SP1
Geforce FX 5200 PCI
VIA VT82C598 Apollo MVP3
Nvidia 43.45 drivers
DirectX 9.0

Usage: VGATEST.EXE <test number>
0 : Run all tests
1 : VGA memory test
2 : VGA text test
3 : Standard VGA Graphics test
4 : VESA Super VGA 8 BPP test
5 : VESA Super VGA 15/16 BPP test
6 : VESA Super VGA 24 BPP test
7 : VESA Super VGA 4 BPP test
8 : VESA Text test

A=Windows XP ME Boot Disk
B=Windows XP SP1 VGA.SYS 5.1.2600.1106
C=Modified VGA.SYS

VGATEST:
G=Good N=Not

1=G

2
1=G
2=G
3=G
4=G
5=G
6=G
7=G
8=G
9=G

3
1=G
2=G
3=G
4=G
5=G
6=G
7=G
8=G
9=G
10=G

4
1=Vesa 100= A=G B=G C=G
2=Vesa 101= A=G B=G C=G
3=Vesa 103= A=G B=N C=G
4=Vesa 105= A=G B=N C=G
5=N

5
1=Vesa 111= A=G B=G C=G
2=Vesa 114= A=G B=N C=G
3=Vesa 117= A=G B=N C=G
4=N

6
1=Vesa 112= A=G B=G C=G
2=Vesa 115= A=G B=N C=G
3=Vesa 118= A=G B=N C=G
4=N

7
1=Vesa 102= A=G B=N C=G
2=Vesa 104= A=G B=N C=G
3=N
4=N

8
1-8=N

If some of these modes are past 1024X768 then they do not work on this computer since it's

monitor only supports 1024X768.

So it looks (at least from this test) like this hack enables the same VESA modes under

Windows XP as I have under DOS!

I noticed the Matrox post a couple of posts back...since we may not have full access to vga.sys would it be possible to code matrox + nvidia into the vga.sys or will different vga.sys have to be provided for different video cards?

How To Ask Questions The Smart Way
Make your games work offline

Reply 26 of 41, by Nicht Sehr Gut

User metadata
Rank l33t
Rank
l33t

Originally posted by DosFreak So it looks (at least from this test) like this hack enables the same VESA modes under Windows XP as I have under DOS!

They key words being "looks like". Do they actually work? Need real-world examples...

Reply 27 of 41, by Schadenfreude

User metadata
Rank Member
Rank
Member
Nicht Sehr Gut wrote:

They key words being "looks like". Do they actually work? Need real-world examples...

Well, VGATEST displays color bars and stuff. It's not animated, though, so you cannot tell if scrolling is all messed up.

Awaiting your testing with VGATEST. VBETEST reads from the BIOS and cannot be trusted. 😉

Reply 28 of 41, by Schadenfreude

User metadata
Rank Member
Rank
Member
dvwjr wrote:

This is the only useful mode for the three PCs I have tested with VGATEST.EXE, a Matrox G450 PCI with XP(SP1), an NVidia GeForce 4 MX440 with XP(SP1) and an old Trident S3-864 1MB with NT4(SP6a). In the DOS boxes many of the other VGATEST.EXE command-line options result in no action, but starting with the VGATEST.EXE command-line option of zero will allow as many of the modes to be tested as a WinXP card/driver combination will allow.

No offense, but I've proved this to be absolute bollocks. In real DOS, there is not a difference between running the tests separately versus together. Running in XP yields the same results - same versus together, there is not a difference for me. It STILL crashes for me in XP though.

It doesn't detect the "wrong" amount of memory, the original IBM VGA had just 256kb of RAM, addressable in two banks of 128kb o […]
Show full quote


It doesn't detect the "wrong" amount of memory, the original IBM VGA had just 256kb of RAM, addressable in two banks of 128kb or four banks of 64kb. This is actually an old test program to test actual VGA or SuperVGA cards of the early to mid 1990s. It is testing what is was designed for, not the future which we now inhabit...

Free VGA project - Hardware Level VGA and SVGA Video Programming Information Page

Then I must ask - does it really test the RAM? If you have 32 MB of video RAM, what is it testing? The first 256kb? That is what does not make sense.


Well VGATEST command-line option 2 is one test mode that will work, it is only supposed to test the standard VGA text modes: 0h, 01h, 02h, 03h and 07h. Those are the only "standard" VGA text modes, so the program is working as designed when nothing beyond mode 07h is tested.

Okay. Well - I have it working now with command-line 2 and 3 even better. By holding in the <FN> key on the laptop and pressing F7 - the "FONT" setting for the LCD - I can get all the VGATEST 2 tests and VGATEST 3 tests displaying something. I don't know much about what this does, but it has two visible modes:
1. centered in the middle of the screen, looking like less than 300 horiz res.
2. full screen. Occasionally the graphics modes in full screen seem to be appearing twice on the LCD screen side by side or overlapping. (VGATEST 3 mode 4 is notorious for this.) but only certain modes - Mode 10 and 11 are not a problem.

Sometimes the VGATEST 3 tests display nothing UNTIL I hit <FN><F7> a couple of times. Fortunately, this does not count as a keypress to jump to the next test. This even happens using a DOS boot disk, so I consider this some sort of LCD/video card/firmware flaw. (unless it is a feature?)

But I CAN get them working in DOS using a DOS boot disk. And in Windows XP.

VGATEST 4 and above is the BIG issue.

Must be an unusual reaction by your laptop display adapter/driver under WinXP. Yep. Any of the SVGA tests will vary by adapte […]
Show full quote


Must be an unusual reaction by your laptop display adapter/driver under WinXP.

Yep. Any of the SVGA tests will vary by adapter. The Matrox G450 will refuse to do anything but return to the DOS prompt for VGATEST.EXE command-line options 4-8, but the old 1MB Trident S3-864 will attempt and succeed on one SVGA mode under VGATEST.EXE command-lines options 4-5. What can you say...

Unusual? You said it. I've learned that the weird "fractal fog"/milky white display that fades in on the black screen and then disappears just as mysteriously when testing VGATEST 4 (and higher?) is what happens when the LCD screen on your laptop loses signal. It's quite fascinating to watch. Also called "white creep", "white creeping death", "white screen of death" OR WSOD and various other phrases, it has as many possible causes as a CRT monitor losing signal. Anything from hardware issues (broken/loose internal cable or wiring, video chip flaws) to software (video firmware, video drivers, system BIOS, etc.)
Additionally, the machine apparently hangs when this happens - I have to hold in the power switch, as no keys can shake it out of this once started. Not even <FN><F7> or <FN><F8> (the one that disables the LCD and enables the CRT VGA connector, and vice versa).

Using MS-DOS 6.22 and Windows ME boot disks and VGATEST.EXE, I've determined that (well) aside from the <FN><F7> pressing needed with VGATEST 2 and 3 sometimes, VGATEST runs fine. Following DosFreak's format:

Windows XP Pro SP1
ATI Rage Mobility M4
Dell Inspiron 8000
Dell/ATI Drivers M6.77-021118a-007010C-Dell (mainly v6.13.10.5023)
ATI BIOS: Bk4.4.3 VR004.095.004.025T.000.000 mfshh4l
(which is weird, as I read that as 4.095. But Dell's Inspiron 8000 A22 BIOS notes refer to it as 4.096? and Dell version A21 BIOS used version 4.100. VERY strange.)
DirectX 9.0

Usage: VGATEST.EXE <test number>
0 : Run all tests
1 : VGA memory test
2 : VGA text test
3 : Standard VGA Graphics test
4 : VESA Super VGA 8 BPP test
5 : VESA Super VGA 15/16 BPP test
6 : VESA Super VGA 24 BPP test
7 : VESA Super VGA 4 BPP test
8 : VESA Text test

A=Windows XP ME Boot Disk
B=Windows XP SP1 VGA.SYS 5.1.2600.1106
C=Modified VGA.SYS

VGATEST:
G=Good M=Maybe N=Not

1=G

2
0=M
1=M
2=M
3=M
7=M

(where is DosFreak getting modes 4-6 and 8-9? What the hell? This contradicts dvwjr!)

3
4=M
5=M
6=M
D=M
E=M
F=M
10=M
11=M
12=M
13=M

4
1=Vesa 100= A=G B=Crash C=Crash
2=Vesa 101= A=G B=Crash C=Crash
3=Vesa 103= A=G B=Crash C=Crash
4=Vesa 105= A=G B=Crash C=Crash
5=Vesa 107= A=G B=Crash C=Crash

5
1=Vesa 110= A=G B=Crash C=Crash
2=Vesa 111= A=G B=Crash C=Crash
3=Vesa 113= A=G B=Crash C=Crash
4=Vesa 114= A=G B=Crash C=Crash
5=Vesa 116= A=G B=Crash C=Crash
6=Vesa 117= A=G B=Crash C=Crash
7=Vesa 119= A=G B=Crash C=Crash
8=Vesa 11A= A=G B=Crash C=Crash

6
1=Vesa 112= A=G B=Crash C=Crash
2=Vesa 115= A=G B=Crash C=Crash
3=Vesa 118= A=G B=Crash C=Crash
4=Vesa 11B= A=G B=Crash C=Crash

7
any=N (no crash, black screen, then drop to prompt)

8
any=N (no crash, black screen, then drop to prompt)

.........................

So you say, AH! Schadenfreude is out of the woods!

NOT!

Stiletto pointed out THIS VGA test utility for me...
http://www.simtel.net/pub/pd/51865.html

During testing using my DOS boot disk, Test 9 invariably snow-crashed the computer after testing various save/restores. In fact, it seemed about to be done - and then it crashed as described above. Loss of signal. IN MS-DOS!!! Geezus. But the rest worked.

Testing almost all the tests in XP snow crashed. But in DOS, only test 9 was the issue. Weird... Again, in XP, both with and without the patch.

...........................

So Schadenfreude says, to himself, "I am very unlucky! Let me test SDD 6.53's PROFILE.EXE as mentioned in the old VESA testing with NOLFB thread...."

Unbelievably, in XP it lists NO modes.
I have never heard of this happening before. This is both with and without the patch.

In DOS, it lists a whole string of them. 😮

..............................

Recently, I had updated the Inspiron 8000 BIOS to A22 from A06 (I think it was A06, i forget what now...)

So far, I believe what I have discovered is a flaw in the firmware. Before installing A22, I used WinMe. Now I have WinXP. So it's all strange.

Dell is going to blame either ATI or the authors of these 15-yr-old testing utilities. ("What? You cannot run such a DOS program in XP?!?! And your laptop came with ME but you installed XP? No more support!") ATI is going to blame either Dell of the authors of these 15-yr-old testing utilities. ("Hey, we don't support our laptop chips - talk to the vendor.")

I SO look forward to taking this up with technical support. 🙄

I also plan on creating a whole separate thread to tell people what my status is on this. I'll be surprised if this is ever fixed - Dell last updated the BIOS in November as with the drivers.

I ALSO plan on attaching a standalone CRT and seeing what happens versus the LCD.

...

Don't worry, I'll test on a desktop machine at some point. But I think I have proved that video on laptops can STILL be a headache and a half. 😁

Reply 29 of 41, by DosFreak

User metadata
Rank l33t++
Rank
l33t++
Originally posted by Schadenfreude […]
Show full quote

Originally posted by Schadenfreude

2
0=M
1=M
2=M
3=M
7=M

(where is DosFreak getting modes 4-6 and 8-9? What the hell? This contradicts dvwjr!)

How so? Do you mean my results or the actual modes? In my results I've taken to listing the # of consecutive modes and then the video modes that I'd bothered to note are the only ones where I can actually see what mode it is in when the prog identifies it.

Oh, I see. I included the scroll test as a seperate # since it is a different test and requires a key press like all of the other tests.

This is kind of funny. I opened up a prompt in XP in a window and ran VGATEST 2. Weird seeing the prompt change dimensions so fast.

How To Ask Questions The Smart Way
Make your games work offline

Reply 30 of 41, by pati

User metadata
Rank Newbie
Rank
Newbie

GREAT GREAT GREAT! I got URW working with this at last 😁 But now the only problem that exists is that you have to use 60hz display mode when playing it 🙁 Someone of you geniuses should come up with some solution to this so i woudn't have to ruin my eyesight with 60hz 😀
Otherwise superb, just superb work!

Reply 32 of 41, by Kahenraz

User metadata
Rank l33t
Rank
l33t

Quadro 4 750XGL
Win2k Pro sp4
Dell Dimension8100 intel850 chipset

After installing this app my floppy drive disapeared from My Computer.. It didn't come back when I fixed the files, but it did reapear after I did an uninstall of vdmsound 2.0.4. Very odd..

Because my floppy drive was missing I wasn't overly concerned with seeing what failed and what did not, but I can tell you that with the original files I had a lot of monitor frequency errors and with the patched files I had none.

Reply 33 of 41, by Kahenraz

User metadata
Rank l33t
Rank
l33t

I tried running an old dos program today and *BAM* my monitor messed up on me. Lucky for me I found this patch yesterday!! I re-patched, rebooted (my floppy was still present), and it worked like a charm 😀

Reply 34 of 41, by eL_PuSHeR

User metadata
Rank l33t++
Rank
l33t++

This thread is quite interesting but i've got a question. If this is a common problem with GF4 cards (like mine - GF4 Ti4200 under WinXP - i cannot access VESA modes beyond 640x480). The question is: Wouldn't it be a bios related problem (GF4 bios) rather than a driver problem? If it isn't, then how is that GF3s do not exhibit this issue? I'm saying all this because after reading this thread i've come to the conclusion no one has gotten satisfactory results with this patch. I'm at work now, so i will test it myself when i arrive home:

My specs:

+ Windows XP Prof + sp1
+ Creative GeForce4 Titanium 4200 64MB (really manufactured by MSI)

Reply 35 of 41, by MajorGrubert

User metadata
Rank Member
Rank
Member
eL_PuSHeR wrote:

This thread is quite interesting but i've got a question. If this is a common problem with GF4 cards (like mine - GF4 Ti4200 under WinXP - i cannot access VESA modes beyond 640x480). The question is: Wouldn't it be a bios related problem (GF4 bios) rather than a driver problem?

This is not a BIOS problem. In order to change video modes the BIOS will actually set a bunch of parameters in the card using some I/O ports. The vga.sys driver from Windows 2000/XP does not allow access to some ports that are used by the GF4, therefore the BIOS is unable to correctly switch the card into the desired video mode.

If it isn't, then how is that GF3s do not exhibit this issue?

Because the GF3, with a different hardware project, uses a different set of ports, and also because vga.sys incorrectly assumes that the GF4 does not need access to all the ports it uses.

I'm saying all this because after reading this thread i've come to the conclusion no one has gotten satisfactory results with this patch.

I've tested the patch under Windows XP Pro SP1 and under Windows 2000 SP4, in both cases the patched version allowed me to use video modes that were not available before, although not all modes available for the card. By the way, my card is a GF4 MX440 (manufactured by ABit), so YMMV.

Regards,

Major Grubert

Athlon 64 3200+/Asus K8V-X/1GB DDR400/GeForce FX 5700/SB Live! 5.1

Reply 37 of 41, by Teller

User metadata
Rank Newbie
Rank
Newbie

Fantastic I now got blood full speed and res. xp Gforce 4

Thanks

😁

Vista 32 Bit (SP 1)
Intel Quad 2.4 Gig
4 Gig RAM
Dual SLi NVidia G-Force 9600GS
Forceware 174.90 (latest available)
Video Bios 64.94.21.00.26

Reply 38 of 41, by AvalonH

User metadata
Rank Member
Rank
Member

I just tried vgafix.exe and I am surprised as it works excellent.
Card: Leadteck GF4 TI4200.
XP PRO SP1.
Before I run vgafix.exe no resolution above 640x480 would work (out of sync screen) and the refresh rate could not be changed in any vesa mode.
After the patch all the modes I use in DOS worked in XP.

640x480 @100hz
800x600 @100hz
1024x768 @100hz
1280x1024 @100hz
1600x1200 @100hz

Reply 39 of 41, by DosFreak

User metadata
Rank l33t++
Rank
l33t++

Hey, was searching the net tonight and looky what I found!

http://www.nomissoft.com/newsflash.html
http://www.nomissoft.com/cgi-bin/ax.cgi?http: … es/winxpfix.zip

Looks like he updated it for XP SP2...which I'm thankful for since I was just trying an updated videoport.sys (which doesn't work for me), so I tried testing vgafix again but it wasn't compatible with XP SP2.

Of to test this updated ver. Post results peoples!

How To Ask Questions The Smart Way
Make your games work offline