Hi all.
I mostly use gigabyte motherboards and I have build many
systems with them. Very few had problems related to motherboard.
Up until the nforce4 chipset.
I have all kind of troubles
slow boots
No booting at all
corrupted data
Raid problems
you name it, I have it ....

Two things to say
To Nvidia people: Nvidia should try better....
To gigabyte people: Gigabyte maybe concider another chipset ...
To both: some word and support on their web pages would be appreciated ...

To all of you wich read this post: avoid nforce4 like hell

The Activearmor with the 6.79, x86-64 drivers is working flawlessly for me (EpoX 9NPA+Ultra) and handles bit-torrent flawlessly. Bit-Torrent used to cause an instant BSOD with the older set of network drivers. The best part is that the latest drivers resolve some of the CPU and memory leak issues with the Activearmor.

I never encountered any HDD problems with the Nforce 4 IDE and SATA drivers. The only problem that I had was my with PX-716SA back when I was forced to use beta drivers for XP X64.

I have found that it is better to avoid the Nvidia versions of IDE and Active armor drivers, the performance gains on the IDE drivers is not worth the corruption and hassles, and this last weekend we had a LAN party with a few people who kept getting disconnected from the game server, turns out it had to do with hardware-level TCP offloading acceleration, if you disabled this in the ethernet propeties the problem went away.

I still prefer AMD chipsets to Intel though, for one intel performs like a brick, and for 2, the Nforce chipsets are generally stable and well performing, as long as you avoid their IDE and Ethernet stuff...

I have an intel 845 chipset that i use for work and it constantly has problems with performance and stability. yuck

As unpopular as this opinion is here and on Arstechnica these funky chipset issues are why I STILL only build Intel based systems for myself and for my clients. Intel CPUs are more expensive and AMD beats them in gaming benchmarks but their mobo chipsets are unbeatable. I've never seen any problems relating to the mobo from the BX chipset (surely the GOLD standard of ALL time?), the 845, 865, 925...and those are just the ones I have experience with.

This contrasts sharply with the tons of issues I've seen with numerous VIA, SIS & sadly nForce based mobos that I've wortked on or heard about first hand.

(let the flaming begin I'm ready)

;)

Jetway 939GT2-Ultra-G (nForce4 Ultra).

Buggy, BUGGY IDE drivers (latest nvidia drivers as of last month - now one version behind).

Performance under Windows x64 was abysmal. I could not download anything from the Internet without the file being corrupted (I was not even able to view html pages properly). Strangely enough, file transfers over my Linksys router on the local network were fine (I was paranoid enough to CRC the files I'd transferred).

Reverting to 32-bit windows fixed my Internet connectivity. However, I'd periodically get an "encryption error" message in Remote Desktop that would dump me back to my own desktop.

I have an MSI Neo2 and have problems with the NVidia ethernet port if I have the checksum offloading on and am doing some "heavy" internet use.

I put together a system for my brother with a Chaintech board with an Nforce 4 Ultra chipset and that board will not even load Windows with AA activated.

Got the Asus SLI deluxe and after the first few initial days with BSODs and lockups i dumped the Active Armor and NVFirewall and all has been sweet ever since....

no sound problems, no corruption etc.

Sometimes, it and other times with newer drivers, it does.

looks like you messed up your tags or something.

I am testing for 8 months with the Asus A8N-SLI , with graphic cards , from Low end to High-end ( 7800 GTX ) . Stabiltiy is very important for me.

When using the right Nforce 4 standalone package driver, nothing is ever happened to me.

No ethernet issues, or IDE drivers with booting.

AMD 64 3500 2200@2500
1 GB Corsair TwinX PC3200
Windows XP SP2 / Vista Beta 2 / Windows 2000 / 2003
All possible drivers

Epox 9npaj sli, Ive noticed the active armour corruption when downloading FTP files via HTTP. Turning off active armour's cpu offload fixes the problem, but obviously at a cost.

I have very infrequent sound stutters, that I attribute to the sound codec IRQ and SATA HD controller IRQ sharing the same number.

#51 and #79, Thanks a bunch for the information!

I'm really quite happy with my new ULi M1695 based board (of course, the Asrock 939Dual-SATA2). This has to be one of the easiest and most issue-free builds I've ever made. Full speed AGP (for the X800XT AIW that I invested a lot of money in) and PCI-E x16 for upgrades is nice, the future CPU socket might be nice, but the board has just impressed me so far with its stability. I'm running at a 20% overclock (Athlon 64 X2 3800+ 2.0Ghz -> 2.4Ghz) without even raising the voltage, and keeping my memory down at CAS2-3-3-8. My previous nForce2 board was plagued with problems, involving multiple replaces on a Epox 8RDA+ board that ended up just being thrown in the trash. The Abit NF7S2 I replaced it with is okay for stability, but I still don't trust the system for even so much as folding now. It just holds the disks (the system has a total of 1.5TB of storage) for my... completely legal file collection.

Edit: I should also mention I'm using stock cooling on that 3800+, and that it never peaks above 52C, unlike the XP 2500+ Barton which peaks at 65C at barely overclocked levels.

I have an Asus A8N-SLI-Deluxe. The Nvidia NIC runs fine without AA or Nvidia firewall. I stopped caring about getting them working, and uninstalled those drivers.
Other than having to disable the two main features that made me buy the board, it runs fine.

So, shall we initiate a petition to AMD for it to start manufacturing chipsets?

The most popular boards always show up with the most popular problems - we have some spectacualr problems with dell machines at work - guess what chip they use. There are a few lemons, but I've never seen problems with the nforce 4 boards i've worked with.

Is anyone else considering these issues significant enough to not consider the NForce 4 as the chipset for a new system? I like building systems, but I don't like dealing with issues like the ones Damage described, or the VIA PFD issues (even though I had a KT133A chipset for a while, and it ran great for 2.5 years as my primary gaming rig). I'm willing to consider a different A64 chipset maker if they have a more solid track record; every company has a bad product somewhere in their roadmap, and the torch of "the board to get" isn't restricted to one company.

My other issue is regarding the soundcard; right now I have an NForce 2 with SoundStorm, but obviously I won't be able to migrate that, and it doesn't exist in the newer NVidia lines (it's replaced by the sound offering of the boardmaker's choosing, which is sometimes cheap and terrible). My guidelines are good DACs (therefore no crabs!), and stable drivers (therefore no Creative!).

Isn't that cute, HardOCP posted a negative review of an ATI chipset and The Tech Report in turn has to trash an nVidia chipset. How convenient... :-/

I'm using BeOS/Zeta here together with W2K, and have no issues whatsover under Zeta, but use an older version of nForce drivers under Windows since it has the least issues. File transfers seem faster under Zeta than under Windows on identical hardware. That probably means that:

a) windows is the bottleneck
or
b) the drivers are the bottleneck.

Interesting.

I thought it was common knowledge that you should NEVER EVER UNDER ANY CIRCUMSTANCES install nVidia's IDE drivers. I've learned the hard way since nForce1 to nForce2. I didn't even bother installing them with nForce4.

On a VNF4 Ultra board, the network device works fine under Windows XP (SP2, current drivers, patches, yadda yadd), so long as you use the Windows firewall. If I try the NVIDIA firewall, any time I get the "program such-and-such is trying to access the internet" bubble and click 'allow' - instant BSOD and reboot.

The reverse-engineered(!) Linux driver (called forcedeth) works fine in 100 Mbps mode (don't have any gigabit hardware to test with).

I have an Asus A7N8X-X (nForce2 400). No ActiveArmor, but I have had problems with the new-style IDE drivers, in that Windows XP will take several times longer to boot and then will be sluggish when accessing the disk. Using the old-style IDE drivers works fine, though.

I just got a Nforce4 board and I have found that it is best to let windows search for and install the drivers versus using the Nvidia setup to install the drivers. System appears to be stable.

I repeat -use the Pro2200 chipset not the ultra.
Tyan k8E-SLI

I'm wondering if... as we move to FB-Dimm for memory in a few years, and PCI-E for main I/O (with HT or similar for interconnects between cpus)

will AMD just make the chipset less and less relevant as time goes on by integrating the majority of it onto the cpu?

I mention FB-Dimm and PCI-E because pincount is already quite insane

in superficially comparing the nf3 5.10 and 5.11 driver revisions it seems that that only changes have been made to the ethernet driver along with the addition of "Network management tools version 4.88". My problems don't seem to lie in network performance or corruption, but rather, random inconsistent BSOD's. Has anyone noticed any other differences between the drivers? How about end user result...any stability or performance differences?

I'm running an nf3 and out of the all of the systems I have built for myself, friends, or family over the years using platforms of all various sorts, I must say that this is the only one to have given me any grief! The thing that really "bakes my noodle" is that the grief can't really seem to be explained. I can't really identify the causation because it is a SFF board so it could be just as equally likely derived from that as a result but either way; frustrating to say the least.

I've got a couple friends who are running nf2 ulta boards and although they still work, they seemed to have had their fair share of troubles and i can't remember the amount of times they've asked me for help troubleshooting a problem where research showed it to be an explicit problem with their chipsets.

As a result of this I must say I am interested to see what kind of competition & penetration ATI can offer with their rx480 chipsets....especially once/if they address the southbridge performance. It does seem that even with such a new product they've had their mind in the right places and are serious about the mobo business. I'm not saying it won't have its own fair share of problems but, haha, if this thread is any indication who knows, it might provide a worthy alternative for those nfocre camels whose back has been broken.

I can now confirmed that large file downloads do not happen with 6.79s on my system. I'm starting to think the issue resides on chipset implementation. It seems most of problems are associated with Maxtor HDDs with defective firmware and motherboards that have two NIC chips.

I've got an Asus A8N-SLI Deluxe. I'm switching out my CPU and putting in a 4200 dual core.

To do so, I have to upgrade my BIOS and my nForce drivers. I've been through this hell before and I'm very careful about how I do this. The windows based BIOS software hasn't worked since I upgraded to BIOS 1.08. Reinstalling it did no good, so I have to do a manual BIOS upgrade.

First thing I did was upgrade the drivers. From earlier experience, I know not to install the firewall software, so I don't upgrade that. Reboot, everything works just fine. Cool.

So I upgrade the BIOS. Nightmare. Can't get the damn set up screen to come up. Can't get it to boot. Can't get it to read from the floppy drive to get it to reflash. Pulled out an old keyboard because this newfangled one has one of those "function key lock" features and I figured that might be causing the problem. Finally able to "alt-f2" into the BIOS utility. Set up the BIOS, reboot the machine. Keep in mind, I have not replaced the CPU yet.

BSOD. reboot. BSOD. reboot. BSOD. reboot. Go back into the BIOS to see if there's anything noticeably wrong. No problems there. Windows starts to load, but something is going horribly, horribly wrong. I'm willing to lay money on the IDE drivers but I can't get to Windows to remove them. Even safe mode BSODs.

Sigh.

To top it off, I started to hear this loud whine....looks like my northbridge fan has bit the dust. So, on top of having to reload Windows because I can't figure out a way of backing down the IDE drivers, I'm gonna have to pull the freaking motherboard out to replace the fan.

DAMMIT.

Again....NOT perfect, just MUCH closer.

god... ya know i wonder where people get these problems?
I have an NF3 250Gb based mobo and a Maxtor drive and i havent even had a hiccup out of my system!

Maby some people just need to stop screwing with their System settings to get that extra 1/1,000,000 of a Frame per Second out of it. then maby they wold have a stable system!

Great, I just ordered an nForce4 Ultra-based motherboard... :( Well, it seems it's still a good idea to stay away from the nVidia SW IDE drivers. I currently run an nForce 2-based machine, and while I did not have any lockup or corruption-related issues when I installed the IDE drivers, I did lose access to my optical drives from Exact Audio Copy - my ripping weapon of choice -, so I promptly removed them. Does anyone know if this issue still persists with nForce 4?

It's absolutely revolting that even after one year from release ActiveArmor is still as bug-ridden as it is. nVidia's engineers really need to get their act together, or somebody should start handing out pink slips. This is simply unacceptable. How hard can it be to track down an issue like this? We're talking about a very common problem, not some elusive bug that shows up only when the stars are properly aligned!!!

EDIT: Could this be the reason why AMD is considering getting back into the chipset business? Right now, their good name is being dragged through the mud!

On a side note, is it possible that the problems lie with particular motherboard implementations of nForce chipsets? Even with the original nForce and nForce 2, motherboards were incredibly sensitive to the slightest variations in current and voltage levels, and I think that's why nForce 2 boards were very picky with RAM. Could it be that we're suffering due to cost-cutting measures of mobo makers? If so, nVidia should still kick some butts and make their customers follow their specifications to the T.