Gigabytes Ga-7N400/7NNXP - Ruled Out as Workstation Boards ? - READ IN

[PR3ACH3R]

Hello All,

As A Long Time Anandtech Visitor & a Pro DAW Builder,
I Wanted To Share with you The Following Findings:
Any Input you might Have Will Be Appreciated.

i Came across a VERY SERIOUS BUG in GA-7N400 Pro.
It Concerns the ITE Raid Chip ,

When Used With Any Pro Audio Card ,For recording ,
The ITE CHIP will send Horrible Repeating Clicks every second or so
When WRITING TO THE DRIVE !

The clicks are sent to ALL THE DEVICES ON THE PCI BUS,
& Repeat every second or so WHILE THE DRIVE IS WRITING.

The Clicks DO NOT APPEAR when used with the
STANDARD IDE Channels of the Nforce2 !

NO One Noticed it ....YET



This was the ONE board That COULD HAVE MADE it
To Being a Great Workstation Nforce2 based MB,
As All others dont have PATA raid .

Im Waiting For Gigabytes Reply on this one...(3 days passes already)
Mass RMA's sure looks like a Possibility ,as i dont believe This issue will be easy to fix..



Please Feel Free to comment & Verify (Flame?!) My Findings ..

**
TO PRO AUDIO GUYS / MUSICAINS:

(pro Video cards should be tested as its very likely a problem with video as well)

If you own a Pro Audio card (Motu ,Rme ,Aardvark ,or any other Pro Card With ASIO Support)
HERES HOW TO TEST:

You Will Need To Record at least 2 Tracks-(A Dual mono Or STEREO Track is OK) & ONLY in WAV Format.

PREFERABLY 24Bit WAV Or Even 32 Bit WAV Files ,if you have ANY
PRO(ish) Audio Software.If you dont Just Dl a Demo of one.

If You Only Own An Audigy/Audigy2
you Can Try ,& Might even get very little or no clicks,Because Im Not sure Your Audigy can Saturate The PCI Bus Like a Pro Card.

Also Make Sure The Software uses The path to a Drive on the iTE
For all its storage options ,Both Audio & Temp Files.

I know The Audigy Has ASIO Drivers (Lower Latency & higher Performing Drivers Pioneered By a Company Called Steinberg-Steinberg.Net)
So you'll need to try it with Something like Steinbergs Cubase SX Software ,& The ASIO Drivers in a Latency of about 5-10 Milliseconds
To get Some Action going on the PCI Bus.

Without a Big Enough Hit on The PCI Bus ,your mileage may vary.

@ AnandTECH
Id just LOVE to see a Full Review With these issues Checked On AnandTech ,
It would be a first Review of a NFORCE2 based System as a PRO WorkStation.

 

[mechBgon]

Welcome to the Forums PR3ACH3R Thanks for the firsthand info there. Which PCI slot was the audio card in? Were all of the board's integrated audio/MIDI/Gameport features switched off? Also, what operating system (and if it was WinXP, were the SW Performance IDE drivers installed? No = good).

 

[PR3ACH3R]

Thanks For The Warm Welcome ! :beer:

PCI Slot used was PCI 5 - But All Were tried out.PCi 5 (weird i know..)Was the Best one,
sharing an irq with the Nvidia PCI Management System.

all of the board's integrated audio/MIDI/Gameport features were indeed switched off,
& back on again - as the setting had no bearing on this issue & no effect either way.

OS was WinXPSP1(a) on All tests.
Nvidias SW IDE drivers were tried a few times - but gave higher cpu loads for some reason,
But did not affect this issue

 

[mechBgon]

I heard some good remarks about the onboard SATA RAID on the NF7-S, if you wanted to try something else. Or are you trying to avoid SATA?

Personally, I lean in the SCSI direction

 

[PR3ACH3R]

well ...

I Had Some Experience with Silicon Images previous Attempts at PATA Controllers..
Add to that the fact that none of my clients want to pay the price premium for serial ata drives (& rightfully so i think)
& The Nice Specs of this Controller ,on paper at least,I Opted for this board as a true candidate for Athlon Workstatons.

& check out the specs of this controller ,Sure looks nice on paper

RAID Function

Compatible with the ATA/
ATAPI-6 specification and supports two IDE channels with 4 drives
Supports ANSI ATA proposal PIO Modes 0, 1, 2, 3, 4 with flow control, DMA Modes 0, 1, 2 and Ultra DMA modes 0, 1, 2, 3, 4, 5, 6
Programmable active pulses and recovery time for data port access timing
512 bytes FIFO for data transfer per IDE channel
Supports RAID 0/1/0+1 function
Supports JBOD function
Supports Scatter/Gather function for DMA/UDMA transfer
Supports pre-fetch and post-write function for PIO mode per IDE channel

>>>>>>
Includes one embedded CPU and firmware on our chip to handle the RAID function. It can reduce the driver's loading and improve the system's stability


PCI Interface

Host interface compiles with PCI local bus specification revision 2.2
Supports PCI Power Management v1.1 capability
Supports one Flash/ROM interface for expansion ROM of the PCI card
Miscellany

Supports the drivers for Windows 95/98/Me/XP, Windows NT 4.0, Windows 2000 and Linux
Package

128-pin QFP

 

[mechBgon]

I wonder if this same effect would occur if you used an add-in PCI controller of some kind.

 

[PR3ACH3R]

If Any of you guys ,have contacts with GigaByte in any way,
Or have a better option to reach their tech support/bug reports departments,
please submith this bug report ,as its been a few days & even though i did mail all tech support departments,
None were answered or even acknowledged.

The Taiwan Form processing ASP page just fails on submittal (DOH...)
So i had to Mail..

Thanks again

 

[mechBgon]

If you feel like experimenting a bit, set the BIOS for PIC mode instead of APIC and let us know how that goes.

 

[Insane3D]

Just on a side note...S-ATA drives are not really any more money now...and more and more models are coming out. I just instaled a Raptor on my NF7-S, and I am likely going to add one of the new Maxtor 120GB, 8mb S-ATA drives in the near future. A 120GB WD "SE" IDE drive is about ~$110, and a 120GB S-ATA drive from Seagate or Maxtor is about $130..so you are talking maybe $20 more.


Also, after several crap boards from Gigabyte, I not only stopped using them, but hence forth I refer to them as "Gigasuck".

 

[PR3ACH3R]

Hehe Insane3D ...
thanks for the reply

About the price difference ..Well the diference is a lot more then 20$ in a lot of parts of the world,

& on a sidenote - even Paying 5$ more for running a drive on an Abit NF7-S Serial ATA controller ...
is money for nothing performance wise ,as this controller sits on your PCI bus like any other Addon PATA controller ..

So any extra BANDWIDTH youll get by any addon controller or SupaDriVe Turbo XRi whatever will have a hard time
Showing Any of its potential muscle..

& Any Speed that it can get off your pci (33mbps tops ..) thats what youll get ....not a kb more ..
So as you can see ,paying more for sATA drives on boards like these are ..Far From a Logical decision ,Unless you
like the cabling that much better.

BTW

No Respone FROM GIGABYTE YET.

I Wonder If Someone From The ANANDTECH ,will Find interest in this.
As I This is a Really Nice board ,except the above mentioned issue.

& a TRUE CANDIDATE for NFORCE2 based WORKSTATION - providing the issue with the iTE will be solved..

 

[Insane3D]

Actually no on the S-ATA drives not being any faster. I just ordered myself a 120GB Maxtor S-ATA drive with the 8mb buffer, and from the tests I've seen, it is the fastest desktop, non-SCSI HD out there right now, other than the Raptor. The S-ATA controller on the PCI bus never gets anywhere near saturating it's theoretical limit, and to my knowledge, there is no extra performance from native S-ATA solutions like on the Intel 865/875 chipsets. You do realize no current desktop drive gets any where near the 150mb/s that S-ATA allows right? Even running two S-ATA drives in RAID would never saturate the PCI bus.

 

[mechBgon]

Any Speed that it can get off your pci (33mbps tops ..) thats what youll get ....not a kb more ..

To add to what Insane3D said... a 32-bit 33MHz PCI bus has a theoretical bandwidth of 132 megabytes per second, and on an nForce2 board it is going to have a practical ceiling of about 120 megabytes/sec (confirmed on mine, and this is the case with the original nForce as well ). The ITE controller is on the PCI bus just like with any of the competing SATA RAID solutions, so its only benefit to you is that you can use those PATA drives.

It almost looks like you define a "workstation" board by the presence of a PATA RAID controller. Anything the matter with an Asus A7N8X-Deluxe as an alternative, other than that? Mine is holding up well at work, pulling double duty as a workstation and occasional fileserver. ~:heart:

Incidentally, have you reported this bug to the manufacturers of the audio cards yet? They might be interested in finding out what the issue is, and either updating their drivers or talking to ITE about it.

 

[PR3ACH3R]

mechBgon ,pro Audio manufucturers are actually Tired of Controllers such as this one implemented in such a way that they hog The bus
like the iTE does..

The Reason I posted this thread is Because i LIKE the product .

Both The Asus & The Abit are not as good as this board Is ..
We Tried Them Both a few times ,& even with the iTE Bug ,The 7N400 pro is still the better choice all around.

I Realize that my findings have very little to do with Most Potential clients ,
& Actually ,i assume a lot of people will never be able to test it ,due to lack of need/hardware/knowldge etc.

I Just Read The Anandtech 7NNXP Review ,& I am very disappointed that the iTE wasnt even tested ,so I Guess Im alone here.

Btw Gigabyte Still havent Replied ,Even though i mailed them a week ago..
The Anandtech review claims Gigabyte replied in 33 hours..

I Wonder who in Gigabyte Should I contact ,Or what e-mail address i should mail to,
To get a Reply in ANY time frame.