Athlon chipsets that support PCI bus parking(?)

[BlueEyedBeezlebub]

Found this online a while ago, my audigy makes my entire system unstable

A little birdie told me about some very interesting events that took place here around or about Computex yesterday. Apparently, a VIA engineer was speaking directly to members of the press about the VIA PCI bus problem that has caused performance and compatibility troubles for some users, and he let slip the precise nature of the problem. It seems VIA's south bridge chips didn't include a PCI extension called bus parking that Intel implemented in its post-BX series of chipsets. Many PCI card makers simply assumed bus parking would be available to them, and when it wasn't, all heck broke loose?snap, crackle, pop on your SoundBlaster.

This must be why when the audigys in my system and i try to transfer large amounts of files it crashes the system.. or even if i try to leave the system on for more than an hour with the audigy in it reboots.

What athlon chipsets do support PCI bus parking? SIS745? Nforce2?

I feel at a standstill, I like my system but wish the audigy worked in it, I dont really want to 'waste' money on a Santa Cruz and get a lesser card than what I already have, and dont want to settle for onboard sound as it sucks on my Z560s, but like I said dont want to buy a new motherboard (nforce2) for the onboard sound of the nforce or to be able to use my audigy...

 

[Peter]

I'd like to emphasize that when a certain PCI card causes problems because the chipset doesn't do PCI bus parking, THEN THAT'S THE CARD'S FAULT. Bus parking is an optional not a mandatory feature of PCI. A card that relies on it being there is broken by design.

That said, VIA have been so kind to add that feature in the 8235 south bridge, Intel's chipsets use to have it. As for SiS and NVidia, I don't know.

But anyway, given Creative's constant struggle with driver quality, I'd rather investigate in that direction than toward hardware.

Soundblaster Live! series cards (not the 5.1, just the original ones) have another PCI problem, but that's a different story altogether.

regards, Peter

 

[tcsenter]

I'd like to emphasize that when a certain PCI card causes problems because the chipset doesn't do PCI bus parking, THEN THAT'S THE CARD'S FAULT. Bus parking is an optional not a mandatory feature of PCI. A card that relies on it being there is broken by design.

Creative still up to their old tricks, eh? I received a few years ago from Creative Support this revealing answer to a problem I was having between one of their products and a VIA chipset (paraphrased):

"We develop our peripherals around Intel standard, VIA is an insignificant market. To solve your problem, please buy Intel motherboard."

For a few years, VIA was going to heroic lengths to 'fix' incompatibilities that, by any objective standard, was not VIA's responsibility to fix, it was the peripheral manufacturer's responsibility. But there's only so much VIA can do on their end after the silicon is finalized.

There are only three or four chipset makers, and several DOZEN peripheral makers. Its is a whole lot easier for each peripheral maker to develop products with consideration for a couple small differences between three or four platforms, than it is for each chipset maker to solve the problems of several DOZEN peripheral makers.

 

[Peter]

... and of course there's a PCI standard to adhere to on both ends ...

 

[BlueEyedBeezlebub]

Hey Peter, I have a question you seem very knowledgable.

I dont want to upgrade or spend any more money on computers for a while.. maybe a video card sometime.
Heres the story and the question, I've been trying to get my KT266A working nicely with my audigy (which I do think is a good sounding card) but have had no success.. I had the system stable as could be, I tested for 2 days uptime with no errors without the audigy, newest hyperions, the latest nvidia drivers from windows update, the 40.72s, because the latest on nvidias website will not work with the latest hyperions for one reason or another on my system.. this is on XP SP1 btw.

Then I thought after two days of IMing and gaming that the system proved itself so I put the audigy in, did a driver install (newest from creative) only and the system started showing cant write to memory location XXXXXX errors and rebooting every 2 hours.

So I did another format, put no hyperion/4in1s on, installed the latest nvidia drivers from Nvidia's site (41.09, they took like a dream), the audigy with latest drivers from creatives site only, and the system seems stable so far.. been about 14 hours uptime.

edit update- but once i tried playing UT2K3 with EAX the system performed a squeal of death, yet in windows was competely stable.. EAX must just use more PCI bandwidth that killed it..

Now my question is, assuming the setup I am running now is stable.. would my computer be faster installing the latest Hyperions and not using the audigy, or not installing the latest hyperions and using the audigy?

I always thought the 4in1s gave quite a performance increase but in your opinion is it worth it to not use my audigy?

Also remember I can't use the latest Nvidia drivers with the hyperions for one reason or another.

Thank you and others are welcome to reply to this also.

 

[tcsenter]

Now my question is, assuming the setup I am running now is stable.. would my computer be faster installing the latest Hyperions and not using the audigy, or not installing the latest hyperions and using the audigy?

At least according to VIA support, always install the 4-in-1 driver AFTER the Creative drivers. Its also ok to install them before and after. But the 4-in-1 is written to overwrite/modify settings installed by the Creative drivers that are known to cause problems between VIA chipsets and the Sound Blaster Live series. May apply to the Audigy, too.

You have a KT266, which is mature enough that it has all the native support it needs in XPSP1. I doubt the latest Hyperion drivers offer anything in the way of improvement for the KT266A over what is also found in versions 4.42 or 4.43 of the 4-in-1. The last couple revisions of the 4-in-1 have been updating support for newest chipsets like AGP 8x KT400 and P4X400, changing little or nothing for older chipsets like KT266A.

Even the "gains" found with the Hyperion drivers have been so small as to only be measurable by benchmarking, you'll never "see" the difference in real world terms.

 

[BlueEyedBeezlebub]

ok thank you, now i still dont understand why I can't use the latest nvidia drivers (41.09) with the latest hyperions?
i would go ahead and install the hyperions since i have the audigy installed correctly but dont want my video messing up.. I would assume nvidia can pull my games more performance than via..

 

[BlueEyedBeezlebub]

ok, i tried unreal tournament 2k3 with eax3 and got a lockup.. the squeal of death.

so i installed the hyperions, figuring couldnt hurt at that point and now instead of lockup the game errors out. but the newest nvidia drivers do work if you install them before the hyperions.

im downloading the latest game patch but found the same error with the demo when hyperions were installed.

i dont know if i like either creative or via.. the only reliable candidate here seems to be nvidia (Nforce2 with good audio and a decent chipset comes to mind?)

 

[Peter]

... if they fixed the PCI throughput problem of nForce-1, well, then yes.

VIA and SiS make fine chipsets with little compatibility issues - apart from Creative's stuff that only seems to run right for everyone when it's on an Intel chipset. With other's chipsets, results vary wildly. See above for Creative's opinion on that.

There are other vendors of really nice sound cards ... and they have to try harder than Creative in engineering, since they can't match Creative's marketing budget.

If you're aiming at the higher end of consumer PC audio, you can e.g. look around for a sound card that uses the Envy24 chip series (now owned by VIA).

 

[BlueEyedBeezlebub]

now the system reboots every 2 hours again, i have the hyperions installed over the sb drivers

i'll probably just order an nforce2.. this system doesnt like being nice even without an audigy, i need 4point sound and if im giong to spend money on a soundcard might as well get the mobo fixed at the same time

 

[Peter]

Then your prime suspects should be the RAM and the CPU cooling. Check the mounting and attachment of the latter, and grill the former in a memtest86 session - www.memtest86.com

 

[tcsenter]

now the system reboots every 2 hours again, i have the hyperions installed over the sb drivers

In addition to Peter's recommendations...

When I said the Hyperion probably wouldn't offer you anything in terms of a benefit that isn't in version 4.43 or 4.42 of the 4-in-1, what I was saying is that, since you are having problems with the Hyperions, why not just use an earlier version of the 4-in-1? Also, you should have all the native support needed provided by XPSP1 (i.e. you don't need a 4-in-1 at all).

Go back to the last stable configuration you attained with the Audigy installed, which according to your description:

So I did another format, put no hyperion/4in1s on, installed the latest nvidia drivers from Nvidia's site (41.09, they took like a dream), the audigy with latest drivers from creatives site only, and the system seems stable so far.. been about 14 hours uptime.

That sounds pretty stable to me. What was wrong with that?

As some wise person once said, the enemy of good is better.

A couple more notes based on personal experience...

ACPI has been known to wreak this kind of havoc in some configurations. If you're up to it, you could try disabling it, or alternatively, if you already had it disabled, try enabling it.

I specifically recall one time I experienced a conflict between a PCI sound card and PCI NIC. The audio was terrible, choppy, scratchy, and it would hang the computer on a few ocassions. This began only after installing the NIC. Pull it out, all was fine again. There were no conflicts that Windows could discern, everything in Device Manager indicated a good configuration, and the two were not even sharing the same IRQ.

I was pulling my hair out trying to troubleshoot this problem. Format C: x 2, install Windows, trying different driver versions, blah blah blah.

I had pulled the cards out and accidentally installed them in reverse, the sound card in the PCI slot formerly occupied by the NIC, and the NIC in the PCI slot formerly occupied by the sound card. I didn't realize it until I booted the computer and magically all the problems were gone, everything was getting along fine. I thought, "Alright...what the hell did I do?" So I began looking around and noticed that I installed the cards in different slots. DOH!

 

[BlueEyedBeezlebub]

So I did another format, put no hyperion/4in1s on, installed the latest nvidia drivers from Nvidia's site (41.09, they took like a dream), the audigy with latest drivers from creatives site only, and the system seems stable so far.. been about 14 hours uptime.

That sounds pretty stable to me. What was wrong with that?
[/quote]

It was great, but if you read my post after that one, when I tried using EAX3 in UT2K3 it performed a squeal of death.

Then I installed the 4in1s and the game doesnt crash the computer but it exits out with an error, and its the latest version also.

 

[BlueEyedBeezlebub]

i ran memtest and came up with a bunch of errors
test 3, 2 errors
test 4, 19 errors
test 5, 23566 errors
test 6, 63 errors

i suppose i'd better get to pulling out sticks and seeing which one is bad

update: ok i tried both sticks by themselves in every slot and they all produce errors in memtest so im assuming its the board
looks like im goign to have to order an nforce2...

moral of the story is, without 4in1s you get squeal of death, with them games exit unexpectedly (with or without audigy) but i did get windows stability with and without the 4in1s as long as i didnt put in the sound card..
interesting but i couldnt get what i wanted!

 

[Peter]

Don't assume. You'd be surprised at how many broken DIMMs are actually being out there. First of all, make sure you're using default speeds in BIOS - and if these DIMMs then still don't work, try another. Only if you then still get errors it's time to suspect the mainboard.

Particularly if you have errors in the "easy" tests like #3. Also, watch whether the errors there reoccur in the same spot. That's another pointer to bad RAM.

 

[KF]

>i ran memtest and came up with a bunch of errors
>test 3, 2 errors
>test 4, 19 errors
>test 5, 23566 errors
>test 6, 63 errors

Another case where a specific problem in an odd circumstance turned out to be a case of "why did it work at all?"

 

[BlueEyedBeezlebub]

hey if you guys are subscribed to this thread, great.. but heres an update if you are!

i bought a maxtor 80gb 8mb buffer drive, because i was trying to format the system and it kept coming up sometimes recognizing the HD and sometimes not.. well my old drive had been dropped before and had put its years in of service so i got the new maxtor.. which is fast btw!
and it let me install windows agian but i still have the same problems as before, but some of my random reboots have gone away.

i notice now when im doing high transfer on the ide bus (cdrom and hard drive), that is when it reboots. seems to be the ide bus to me.. i have the latest 4in1s and seems the only fix for something like this is a new mobo.. so im going to try for a while longer to fix these problems but not going to put a whole lot more work into this, an nforce2 would be great with my new maxtor 8mb cache drive!

 

[Peter]

HellooOOooooo ... throw that fscking broken RAM out, for the sake of whatever you believe in!

 

[BlueEyedBeezlebub]

hehe.. i didnt really update you with this but im using the via ide miniport drivers and the system is stable.. when i remove them i get the constant reboots again.. but i have at least a days uptime with the ide miniport drivers from 10/7/01 off of viatech

 

[Peter]

Originally posted by: Peter
HellooOOooooo ... throw that fscking broken RAM out, for the sake of whatever you believe in!

You also listening, or just writing?

 

[BlueEyedBeezlebub]

ok ok ok, im listening! dont get mad i need your help!

is there a way to individually test the ram sticks? I tested each one by themselves and they both came up with errors in memtest.. both are bad???

Also is there a way to 'work around' bad ram like that linux program or would another board better handle ram errors? ive read that intel boards work well with bad ram compared to others, or at least isnt as picky.
i wouldntget an intel but curious if i should throw the ram in the trash or not.

thanks

 

[Peter]

Well if you tested them individually, then it's time to get new ones. Bad RAM that exhibits unworkable cells won't work right ANYWHERE. You might be lucky on slower boards when using RAM that is bad in the sense of not being as fast as its specification claims, but for RAM with actual broken bit cells, no way.

Get new, quality RAM. memtest86 it. Meanwhile, stuff the DIMMs you have elsewhere and test them there too. Then you'll see.

 

[BlueEyedBeezlebub]

my current ram is two sticks of Crucial PC2100 cl2.5, one was for a ecs k7s5a and the other for my gigabyte7vtxe.

 

[mechBgon]

Incidentally, if you don't already have your hard drive at the end connector of the IDE cable, it's advisable to make sure that end connector is occupied by a drive.

 

[BlueEyedBeezlebub]

yes i use the last end connector but always keep my hard drive on its own ide channel