SCSI card causing Quake 2 lockups on G400?

[Boom]

It's the weirdest thing. Ever since I installed this Tekram DC-315U SCSI card and Plextor 8/2/20 CDRW, Quake 2 has been giving me lockups. Immediately when I start the game, my computer locks up hard -- the reset button is the only exit. Also, Half-Life never used to lock up, but now it does after about 1 hour of continuous play.

Disabling the SCSI card fixes the lockups. So does unplugging the CDRW (the only device on the SCSI card). And running Quake 2 in software mode with the SCSI and CDRW on works OK too. The really strange thing is that with the 5.15 drivers, Quake 2 runs fine in OpenGL on the first run, but if I exit, and then run the game again, it locks up. With the 5.25 drivers, it locks up on the first run.

Note that 2D works perfectly fine regardless (I haven't run any D3D stuff for more than 10 minutes straight so I don't know about any D3D lockups). So what do you think is the problem here? The SCSI card? The CDRW? The drivers?

[Maggi]

Give us some details on your machine and config ... Mobo, CPU, clockspeeds, OS, peripherals etc.

------------------
Cheerio,
Maggi
________________________
Asus P2B-DS @ 103MHz FSB
2x P3-450 @ 464MHz
512MB CAS2 SDRAM
Millenium G400 32MB DH

[SCompRacer]

Your Plextor terminated and SCSI ID set properly? Check IRQ assignments in device manager for sharing.

I upgraded from a Plexwriter 4/12 to the 8/2/20. It is a great drive.

[xortam]

I think Matrox is going to soon acknowledge that the G400 does have a problem coexisting with SCSI controllers. I too use SCSI and have had lockups since I got my Max two weeks ago. I'm working with a Matrox TSE on the problem but he's only had me pursuing interrupt sharing so far. I stripped my NIC and sound card out of the system and only have a Buslogic (Mylex) Raid+ SCSI controller and G400 on the busses. I've disabled the on-board SCSI BIOS (ASUS P2B-S) and USB IRQ, but it still hangs. I did notice that the system was more stable after I removed the loading of a few real-mode ASPI drivers. The Ziff-Davis 3D benchmarks would hang almost immediately in 1280x1024x32 with them and would many times make it half way through the bank of 3D tests without them (most commonly on test 9, but has locked up many places, even when just using the desktop, also at 1024x768x32). I'll let the forum know if we come to any conclusions but I'm hoping I simply have a defective card and Matrox will ship me out a new one. To be continued ...

------------------

• ASUS P2B-S, PIII 450MHz, Award ACPI BIOS v1010, 128 MB RAM
• MYLEX FlashPoint RAID+ (BIOS v2.02N) running RAID 0 on two 9 GB IBM DDRS 39130D Disks
• Diamond MX300 sound card
• Matrox Millennium G400 Max Dual Head - English
• NEC 5FG monitor
• YAMAHA CRW4416S and NEC Multispin 3x CDs
• 3Com Fast EtherLink XL 10/100Mb TX NIC
• US Robotics 56K Voice FaxModem Pro
• Trying out Pioneer DVD-303S SCSI (but no S/PDIF ... waaahhh!!!)
• Note--All SCSI devices (except disk drives on RAID) are connected to onboard AIC7890 U2W SCSI
• Mainly running Win98 v4.10.1998

[Boom]

I have checked the termination thoroughly, everything looks fine. I left the Plextor at default, which is ID4, parity on, terminated. I tried different IDs and removing the termination (since the SCSI card has active termination), but there is no improvement in the lockups. SCompRacer, what SCSI card are you using with your Plextor?

There is also no IRQ sharing in my system:

Celeron 433 + Slotket
Abit BH6 v1.1 Bios NK
128MB Micron PC133 SDRAM (2 64MB sticks)
G400DH32 (IRQ11)
Tekram DC-315U SCSI controller (IRQ10)
Linksys EtherPCI II net card (IRQ12)
SBLive!Value (IRQ 5 + 9 for SB emulation)
USR 33.6 internal FAX/Modem (COM2 IRQ3)
Logitech Firstmouse+ PS2->Serial (COM1 IRQ4)
Plextor 8/2/20 CDRW, Pioneer 103S DVDROM
WD Expert AC18000, Quantum Fireball CR 8.4GB
USB Disabled

Well, I'll try running some D3D stuff and see if that locks up too.

[Ali]

I dont think SCSI is a problem with the G400

I have a Adaptec 2940UW and 6 devices hanging off it and everything is fine. It also worked great when sharing an IRQ with the G400 (before I got rid of my last IDE device).

I had a Zip insider controller card at one stage (cant remember its nmber, but it was an adaptec) and that worked fine too.

System
P3 600@618
Abit BX6 Ver 2.02
128Mb GH Ram
12Mb SLI rig
D-Link NIC
Soundblaster 64 ISA
G400 32Mb DH @ 220 mem
Adaptec 2940UW:
0 - Cheetah HDD
1 - Medalist Pro HDD
2 - Plextor 36X slot CD rom
3 - Internal Zip 100
4 - Richo 7060S 6X4X24 CD_RW
5 - Toshiba W1101 DVD-RAM

[xortam]

Ali, I hope you're right about SCSI not being a problem with the G400. My Matrox TSE just gave me the go-ahead to call customer service for an exchange. I'll have to revisit this issue if I still have a problem with the replacement card. What worries me is that the lockup problem seemed to be exasperated by the loading of some real-mode ASPI drivers (that and the many users on this board who have had problems with SCSI environments). I've wondered that my card had been previously returned by another user because Matrox had reused the anti-stat bag (old sealing sticker under new) and the serial number was older than someone else on this forum who also got their board through Matrox at about the same time. Probably take another week to get the exchange done so I'll let you know if its O.K.
Boom, the only issued with SCSI id's is to not have more than one device (or host) on the bus with the same ID. Otherwise, which ID you pick is unimportant. Bad termination will very likely manifest itself as a lockup. Remember, there's two ends of the SCSI bus that need to be terminated and non-device active termination is the most assured (and I believe required for differential SCSI). I've got three SCSI busses on my system, so I've been through some termination problems. Good luck with you're system

[SCompRacer]

Boom, please check sig for system specs. No card, but onboard 7890 chipset, equivalent to Adaptec 2940U2W.

xortam, I wonder if it is the RAID controller, not the SCSI that is the problem. In RAID level 0, data is split across the drives resulting in higher data throughput. What about the ASPI layer? Is it Adaptec or Mylex drivers?

Even if one uses the latest version 3.02a drivers from Adaptec, you get later versions of the ASPI layer drivers if you have Adaptec EZ CD Creator 3.5c core update or latest version 4. DL ASPICHK.EXE from Adaptec to see what you got.

Another utility that has caused background problems is Adaptec EZ SCSI 5.0. Many times some don't even mention it as they think it's transparent and does not affect the system. I have had problems with Drive Preparer and Quick Scan, and disabled them in MSCONFIG/Startup.


------------------
P III 500 @ 560 via 112 FSB, ASUS P2B-LS w/1010 BIOS, 256 MB 8ns ECC RAM ECC Enabled since 5.13 driver release, current 5.25 drivers, Seagate 9 GB LVD Cheetah & 9 GB UW Barracuda, 10 GB IDE DiamondMax, Plextor Ultraplex 32 X/Plexwriter 8/2/20 RW, Hitachi GD 2000 IDE DVD CD ROM, SCSI ZIP Insider, G400 MAX, SB Live, Sigma Hollywood + DVD decoder, US Robotics 56K Sportster, Sony GDM400PS, UMAX 2400S SCSI scanner, Soundworks 5.1, HP 6P Laser & 890C, LAN & USB enabled. SB Live shares IRQ w/MAX, bus mastering enabled, Win98 SE and NT 4.0 SP4.



[This message has been edited by SCompRacer (edited 10-05-1999).]

[Brent]

Good tips everyone. I strongly confirm the checking of your ASPI layer. Also, see if there is a setting in your BIOS for PCI latency and try lowering it from 64 to 32 or lower. I wonder whether it is not a Quake problem as having installed Quake from 4 different SCSI and IDE CD Roms/ DVD Roms, all 20 speed and higher, I always get a message about not having a drive faster than 2x speed or something to that effect. If I use Virtual Drive to copy Quake to a IDE Raid array, it hangs terribly and eventually will not launch at all but if I copy it just to a single IDE drive, no problems at all.

[xortam]

SCompRacer , My RAID controller is SCSI. Not sure what you're point about the RAID level is but I implemented RAID for its performance improvement, not fault tolerance (hey, backups are good enough for this box). I've disabled the loading of the real-mode ASPI drivers during this G400 acceptance testing so they're no longer an issue. I also use ZD's startup manager and disabled all of my startup so I'm not running any extraneous software. I stripped my machine down to the bare bones (removed all but the RAID controller and Max, disabled on-board SCSI BIOS and USB IRQ and also in device manager) but still got the lock ups. I can't get away from the RAID controller because it hosts my boot-up drive. I suppose you boot off of you're IDE drive. I considered enabling my IDE and going for the slightly faster Pioneer 10x IDE slot DVD (rather than the 6x SCSI) and possibly booting off of a "slower" IDE drive thus freeing me from SCSI for debugging like this and easing disaster recovery procedures. The big thing that's stopping me (besides being a SCSI bigot) is that I don't want to spend the extra IRQ(s). Got a better idea how to test w/o the RAID controller?

[SCompRacer]

xortam, I know your RAID is SCSI, not IDE, but it is independant of the onboard SCSI, and is not redundant but uses drive striping, thats what I was guessing at being the problem.

I never had a RAID setup, but I do boot off of two different SCSI drives with different op systems, as well as an IDE drive with no trouble.

The IDE drive is a gaming drive as well as a "Frankendrive" for testing before implementing any changes on my Cheetah. Even if you tested like this and it worked, you would still want a solution for your setup.


------------------
P III 500 @ 560 via 112 FSB, ASUS P2B-LS w/1010 BIOS, 256 MB 8ns ECC RAM ECC Enabled since 5.13 driver release, current 5.25 drivers, Seagate 9 GB LVD Cheetah & 9 GB UW Barracuda, 10 GB IDE DiamondMax, Plextor Ultraplex 32 X/Plexwriter 8/2/20 RW, Hitachi GD 2000 IDE DVD CD ROM, SCSI ZIP Insider, G400 MAX, SB Live, Sigma Hollywood + DVD decoder, US Robotics 56K Sportster, Sony GDM400PS, UMAX 2400S SCSI scanner, Soundworks 5.1, HP 6P Laser & 890C, LAN & USB enabled. SB Live shares IRQ w/MAX, bus mastering enabled, Win98 SE and NT 4.0 SP4.

[Boom]

After some more extensive testing today on this matter, I am now even more perplexed as to the cause of the lockups. Some additional info: D3D will not lock up. If busmastering is off, running Quake2 in either OGL or software mode kicks me back to Windows - no lock up. AGP 1x and 2x make no difference. Uninstalling/reinstalling Quake2 makes no difference. Uninstalling/reinstalling G400 drivers makes no difference. Clean reinstall of 98SE makes no difference.

My ASPI info via ASPICHK.EXE from Adaptec:
WNASPI32.DLL v4.60 (1019) 5/10/99
WINASPI.DLL v4.60 (1019) 5/10/99
APIX.VXD v4.60 (1019) 5/10/99
ASPIENUM.VXD v4.60 (1019) 5/10/99

SCompRacer and Ali, since you guys have no issues with your SCSI, if they're different from mine, can you post your ASPI info?

[xortam]

I know you'd rather hear from somebody who isn't experiencing lockups, but I thought I'd add my observation about ASPI versions. I was running 1014 and just updated the latest drivers from Adaptec which put me up to 1017. I'm running Win98 (4.10.1998) while you're running Win98SE. I assume you've got you're 1019 drivers from the OS install. I wonder if there is any difference from the 1017 drivers and why Adaptec doesn't have them available for download.

Just to make sure you have you're termination right (you're earlier post didn't quite convince me that you have it right) ... again there are two ends to the bus. I take it you have the SCSI controller terminated with active termination; that's one end. You say you removed the termination from you're writer. The last device on the bus must also be terminated or you must have a terminator chained off of that device. Active term -- Host Adp -- CDRW -- Active term.

[SCompRacer]

Boom, my ASPI layer is identical to yours. You have Adaptec EZ CD Creator with 3.5c core update? That is only way I have found you get those versions.

I found after installing Win98 SE, you find ASPIENUM.VXD is missing from the ASPI layer, and aspicheck reports it incomplete. Versions also report differently, like version 1.0, 4/24/99.

[xortam]

SCompRacer, I can't seem to get up to the 1019 ASPI version. I installed Adaptec EZ CD Creator with 3.5c core update (Cdc-35c.exe) last night and it didn't touch my ASPI files in the system directory (verified by ASPICHK). Also installed the 103 patch, but it didn't upgrade them either. I had to run ASPI32 to get from 1014 to 1017. After reading you're response this morning, I decided to uninstall CD Creator and reinstall it(and patch 103). Still running at 1014. Are you certain this is how you updated those files? I've never used EZ-SCSI. I wonder if that might help?

Brent, My BIOS has always been running a PCI latency of 32 (default). Haven't played with the setting yet, but I suppose I can give a try sometime after everything else settles down. Very worrisome comment about you not being able to run Quake from you're IDE RAID array. RAID should be very transparent and is handled at the firmware level. The only thing that comes to mind is that they're could be some problem with ASPI compatibility in the RAID miniport driver for the RAID controller. You can still access you're RAID array without this driver, but at reduced performance (so I read). That's something I'll have to try as well.