nForce PCI writes issue @ Tech Report

[NFS4]

Maybe it's not just VIA:Q And I have an MSI K7NPro coming in today to replace my Shuttle AK31 Rev 3.1, actually it should be here any second...

Since I purchased the MSI K7N420 Pro motherboard late last year, I've
participated in one form or another in the MSI nForce forum. During that
time, there has been a complaint or two about low write performance from
hard drives, in particular with IDE RAID. The nForce boards' peak write
speeds on drive benchmarks using PCI disk controllers would never crest over
20MB/s, while read scores have hit as high as 113MB/s. Due to a relative
lack of complaints and the fact that most users were using Windows XP, which
has its own hard drive writes performance problems, the issue was largely
ignored.

In light of some recent tests from SCSI and IDE RAID users from Windows
2000, it's begining to seem that there is a problem with nForce, not just
with the OS. In fact, Asus apparently has taken steps to address the
problem. Among the fixes in Asus' latest beta BIOS for its nForce board is a
correction for poor PCI write performance.

http://www.tech-report.com/news_reply.x/3675/

 

[Adul]

I was just going to Msg you about this.

Guess you'll find out how it is.

 

[Pabster]

How the hell do PCI latency and performance issues "slip by" chipset validation? There's just no excuse for it. Even Intel has had some problems in various chipsets in this arena. For the life of me, I can't figure it out. Do they just assume because the majority of users are using cheap, low-performing (or mainstream) IDE devices and 90% of them use a Creative sound card that stuff like this can slip by?

 

[Athlon4all]

Uh oh.

 

[TheCollective]

It appears to be an issue with the MSI board itself and not the nForce chipset in general. drruhe's post shows this as he is using an Abit board.

 

[oldfart]

This is one of my pet peeves about motherboard reviews. IDE and IDE RAID read/write performance should be a standard part of performance testing. The tests are too heavily based on CPU/mem system performance, and very light on other important areas. HD access is the slowest part of a PC. More attention should be paid to this area by the reviewers. A few less 3DMark or Q3 graphs would be a welcome change.

 

[NFS4]

Originally posted by: oldfart
This is one of my pet peeves about motherboard reviews. IDE and IDE RAID read/write performance should be a standard part of performance testing. The tests are too heavily based on CPU/mem system performance, and very light on other important areas. HD access is the slowest part of a PC. More attention should be paid to this area by the reviewers. A few less 3DMark or Q3 graphs would be a welcome change.

I agree 100%.

 

[Adul]

nfs4, get your mb yet?

 

[Pabster]

oldfart wrote:

"This is one of my pet peeves about motherboard reviews. IDE and IDE RAID read/write performance should be a standard part of performance testing."

Exactly! (You took the words right out of my mouth!)

You'd think, particularly with the VIA PCI fiascoes, that reviews would take a better look at overall PCI subsystem performance, in addition to IDE / RAID performance. It seems these areas are vastly overshadowed by a few MB/s of RAM bandwidth or an extra USB port, or what have you. I've seen on-board sh!t audio get better coverage.

 

[Nocturnal]

Originally posted by: Adul
nfs4, get your mb yet?

 

[First]

Originally posted by: oldfart
This is one of my pet peeves about motherboard reviews. IDE and IDE RAID read/write performance should be a standard part of performance testing. The tests are too heavily based on CPU/mem system performance, and very light on other important areas. HD access is the slowest part of a PC. More attention should be paid to this area by the reviewers. A few less 3DMark or Q3 graphs would be a welcome change.

I agree. Thanks for the input.

 

[Adul]

yeah

Come to think of it. might has well test it with IDE, SCSI(under win2k), external USB 1.1, 2.0, firewire HDs, etc

stress testing

 

[NFS4]

Originally posted by: Adul
nfs4, get your mb yet?

I called home from work and yep, both my motherboard and wireless optical logitech mouse arrived. Tonight should be fun

 

[joohang]

Reason why I like Anand's reviews is because he tends to benchmark and comment beyond Sandra and Q3 Arena FPS. There is still good room for improvement, though.

 

[TheCollective]

Bump. Ryu still needs some help.

 

[NFS4]

Damn MSI board I got was screwed!!:|:|

Kept locking up at post. I reset the CMOS, took out all the cards, reseated memory, stuck in just one memory chip, etc etc. Just got an RMA to NewEgg.

Guess I'll have to order a NEW motherboard now...I'm not dealing with refurbished stuff again.

 

[Adul]

Refurbish crap is like that.

 

[oldfart]

Evan Lieb, it's nice see to you guys look at the forums for input. I'm looking forward to some more comprehensive motherboard reviews.

 

[foofoo]

hi nsf4,
i just changed 2 days ago from my shuttle ak31a to an abit nv7-133r (cant beat it for $89 from newegg) and i'm very happy. the shuttle developed an error with the atapi ide controller.
did you consider the abit?

 

[TheCollective]

Bump for the nForce users out there.

 

[Gstanfor]

I can't be a lot of help in this thread, because I don't use RAID, a PCI IDE card, or a PCI SCSI card. I'm not likely to start using any of those cards any time soon either, like a good 90% of the general computing population.

I do however use a PCI TV Tuner card and that gives excellent results in the K7N420, whereas it was basically non functional in VIA KT133 and VIA KT133A boards.

Due to the way things work on the nforce (PCI data is sent to and from the southbridge on an 800mb/s hypertransport link and PCI data only accounts for 133mb/s) I doubt it is a hardware issue, probably just a driver/BIOS issue. As others have suggested elsewhere, if it were hardware based it would affect any PCI card with high bandwidth requirements (fast ethernet, TV Tuners, etc), which it does not appear to do.

Greg

 

[NFS4]

Originally posted by: foofoo
hi nsf4,
i just changed 2 days ago from my shuttle ak31a to an abit nv7-133r (cant beat it for $89 from newegg) and i'm very happy. the shuttle developed an error with the atapi ide controller.
did you consider the abit?

Actually, I ordered the NV7-133R and a Western Digital 800JB last night from NewEgg

 

[Sunner]

Seems like it's damn hard to get the PCI stuff right.
Heck everyone seems to have problems with it, Intel, VIA, now nVidia.

 

[jaeger66]

Originally posted by: Gstanfor

Due to the way things work on the nforce (PCI data is sent to and from the southbridge on an 800mb/s hypertransport link and PCI data only accounts for 133mb/s) I doubt it is a hardware issue, probably just a driver/BIOS issue. As others have suggested elsewhere, if it were hardware based it would affect any PCI card with high bandwidth requirements (fast ethernet, TV Tuners, etc), which it does not appear to do.

Greg

Not quite right, the Hypertransport link is between the north and south bridges. The PCI bus is still a 133MB/s bus running to the south bridge.

 

[jaeger66]

Originally posted by: NFS4
Actually, I ordered the NV7-133R and a Western Digital 800JB last night from NewEgg

Should have stuck with the MSI, it's the only true nForce board out there.