Gigabyte Boards and DPC Latency

[Fallen Kell]

DarkRogue, yes, I think laezyre is saying that. He isn't the only one, which is why Colin mentioned earlier that this one isn't on Gigabyte's website yet because it doesn't fix the problem for everyone, which means there is still something going on that they are looking at.

 

[LOUISSSSS]

MB: GA-P35-DS3P
Bios: Award F10 2/27/2008

DPC Latency Checker:
maxed out at about 5000 and got a "red" every 10 interval units or so. got a good amount of "yellows" tho..

real-world:
no music stuttering in iTunes... don't know what other symptoms there would be. system is 100% stable under XP Sp3.

 

[cactusdog]

Originally posted by: laezyre


I was wondering if anyone has been able to get gigabyte to replace their motherboard for this reason. I bought mine over a month ago and can not return it to the reseller at this point.

Thats a good question, I have heard a few people returning their boards and getting a replacement but the problem is, all their boards have the problem to some degree.

I dont think Gigabyte will offer it officially, or they will go broke but you could try.

Probably better off to cut your losses and get another brand like i'm gonna do if they dont fix it in a week or two.

Luckily for you your board isnt that expensive, i paid $350 for my X38DQ6. I also spent several more hundreds on new RAM/PSU/HD before i knew it was a motherboard problem.

@Louissssss, i bet your system will feel a whole lot faster with the bios fix.

 

[Mir96TA]

I was noticing that too
I thought it was some thing to do CODEC.
My problem is more sensitive with Live meetings
Sure I enough I have this issue
Here is the link for my result
http://s64.photobucket.com/alb...n=view&current=DPC.jpg
MB GA P35C DS3R Rev1.1
with Bios Ver F10
CPU E3100
Mem: 4 Gig PC PC2 6400
Video : GTX 8800 640 Meg from EVGA
Sound Card : Audgy 2 Value
PCI WiFi Card (Disable)
PCI 1394 Card
OS XP Pro SP3
my email address is mirmhussain@gmail.com

 

[chriskwarren]

@Louis

I am in the same boat as yourself. Mostly yellows and I have not seen the problem in real world apps yet, however I have not had the time to try any real audio/video work as of yet.

 

[chriskwarren]

Yeah I think we have to give Gigabyte a chance to fix this via bios first.

What are our options though if this doesn't get fixed? Another model won't fix it as the problem, as you mentioned, is across their entire product line.

I would think that this should be fixable, or there will be a lot of people who will be very upset!

 

[Blazer7]

That's not 100% accurate. It seems that many boards are affected by this but this is in no way the case with their entire line. Hell, my old 680 is constantly in the green.

 

[chriskwarren]

Lol is that the 680i board from Mr. TheBeagle's epic thread?

The beta bios is helping some people so I think/hope a fix is on the way. Is anyone else able to report no issues with this?

 

[Syrup]

It would be great if we could get Colin to check in every week or so just for peace of mind.

In addition lets all do our little part to research this thing out and keep this thread alive until its resolved. If I come across any updates on other sites I'll be sure to bring them back here.

No change for me personally, E8400, DS3L, XPSP3 still runs very stable but continues to have the latency issue. It didnt appear to change when I updated to SP3 on XP. So long as mine runs stable and doesnt affect my experience im going to be patient and give Gigabyte a few weeks to fix it. If theres no fix or at least something close I will likely move to the Abit IP35E. Yay for inexpensive motherboards \o/.

Syrup

 

[DarkRogue]

Originally posted by: Syrup
In addition lets all do our little part to research this thing out and keep this thread alive until its resolved. If I come across any updates on other sites I'll be sure to bring them back here.

Agreed.

I'm not sure how reliable this information is, but I stumbled across this thread on the TomsHardware forums while Googling aound for this issue. It seems fairly widespread with some people noticing it more than others. Anyway, according to that thread, Gigabyte apparently changed/updated something in one of the BIOS's which began this DPC latency problem. One user in that thread tried various BIOS's and found out (for the DS3R at least) R4 BIOS was the last one to run without this problem, and R5 onward has it. Another user posted that when they asked Gigabyte how they fixed it, they responded by saying they re-used the old code.

Again, I don't know the legitimacy of that information, but it's something, and I guess Gigabyte has an idea of what the problem is.

As for me, my DPC is still about the same. No audio/video skipping thankfully, but still regular spiking. Although, my spiking seems to differ from other people's who seem to get a sharp high spike regularly, I get a big spike, then a second of normalcy, then a medium spike, another second or two of normalcy, followed by a small spike... and repeat.

 

[laezyre]

I have tried the bios fix from Colin and the bios fix from the gigabyte support team that they sent me. They were both supposed to help. Neither one helped at all.

I have constant yellow spikes and frequent red spikes. I just had a 23,000 us spike with my system idle. If I run cpu-z and run the latency program I have constant red spikes every few seconds with yellow in between. I ran cyrsis and had a lot of red spikes while loading, but moslty yellow and green while playing. I never have more than a rare spike below 500 us.

There are no more patches available at the moment. I feel like I have a defective board.
I'm not holding my breath on a new board from gigabyte.


Any suggestions on a replacement?

 

[dajeepster]

Originally posted by: laezyre

There are no more patches available at the moment. I feel like I have a defective board.
I'm not holding my breath on a new board from gigabyte.


Any suggestions on a replacement?

yeah ... good luck trying to get help from gigabyte :disgust:

 

[laezyre]

Thanks, I feel much better now.....

 

[cactusdog]

Gigabyte are lot better at responding to problems than most companies in my experience. At least they answer you in a day or two if you contact support.

 

[Winterpool]

How many non-Gigabyte users have tried the latency tool?

I've run the tool on my Windows Vista Ultimate system (Abit IP35-E, Xeon E3110) and observe regular 1000-microsecond spikes, though I only get a yellow spike after several minutes. My pattern is mostly 500 microseconds or less, but I do get 1000-microsecond spikes every three or four measurements. This pattern seems consistent whether I run media or not (have quite a few Firefox windows and tabs open though).

 

[sempur]

I tried it on my old Asus P4PE with P4 northwood and WinXP. No problems. Highest was 147 after a few minutes.

 

[SnakeChomp]

I'm running the GA-EP35-DS3P running F3n, and have occasionally been seeing spikes of this nature. I don't mean "I occasionally get a spike of 1000us". I mean that I occassionally get a spike of 999559us, according to the tool linked in this thread. Whenever this happens, it manifests itself as the computer "skipping" for periods of a second, ignoring mouse input so that the cursor jumps around once a second. It doesn't happen so often as to render the machine unuseable, but I would hate to have this manifest during a game or something of that nature.

To the above poster, I also get 1000us measurements every 5th or so measurement, but that is causing no issues for me. Its the super huge spikes that produce a noticeable problem.

 

[VirtualLarry]

I have two machines in front of me. One has a Gigabyte P35-DS3R and an E2140@3.2Ghz, with F9 released BIOS. I ran the DPC latency checker, and I get occasional "spikes" of 400us. Nothing to be concerned about. Everything is smooth on this rig.
The other machine is a Foxconn board with a 6150 NV chipset and an AMD 3800+ 939 CPU. It was getting spikes every four bars or so of 260,000us! The mouse would freeze every few split-seconds. I unplugged my USB printer and restarted the PC, and now things are down to normal. I will see if the spikes return, with the USB printer unplugged. The system never used to have that problem when I built it.
Both rigs are running XP SP2.

Edit: The problem mysteriously re-appears, even without the USB printer attached. It seems to just spontainiously happen, and it requires a full power-down to fix. Well, until it happens again. It never used to have this problem either.

 

[FellowTraveller]

I think it's important to keep in mind that some of the spikes may be due to kernel-level drivers or processes. I run Vista x64, but have also tested DPC checker in Vista 32 and XP. The spikes are gone in Vista 64 and XP, but I still get them in Vista 32. I have yet to narrow it down, but that tells me it's a device driver or system process causing it in that case.

Before the beta bios, I experienced 1000us+ spikes in all three operating systems.

 

[SnakeChomp]

Doh. I may have only had 1000us spikes last night, but upon booting today (without having changed anything since last night) the 1000us spikes are now either 10000us or 300000us. It seems to vary each time I boot with no pattern seen when disabling devices. Its giving me audio drop outs now, so its becoming a real problem. =( Every fourth sample in this tool is a huge red spike.

 

[cactusdog]

Originally posted by: SnakeChomp
It seems to vary each time I boot with no pattern seen when disabling devices. .

I found the same thing. It seems totally random with each reboot. Sometimes its good and sometimes its bad.

 

[SnakeChomp]

An update on my issue. I happened to move my USB keyboard from this new machine to my old laptop that I am replacing to use it for a bit, and noticed that plugging it back into the new machine using a different usb port made my spiking return back down to the 1000us level and no higher. I had previously had the keyboard plugged into my monitor (which acts as a USB hub). Doing this was the cause of my major spiking. Maybe this information will be useful for other people, as this machine no longer has any ill effects that I can notice from the dpc spikes (aside from the previously mentioned 1000000us spikes that happen when opening some programs but only last a few seconds). I was considering returning this motherboard after waiting a week for gigabyte to get me the beta bios, but now that I know what causes my issue I can just work around it and have a useable machine.

More specifically, it seems that as long as I boot the new machine with the usb keyboard plugged directly into the motherboard (instead of through a usb hub) the spiking is OK. After booting I can move the keyboard back into the USB hub without incurring the spikes.

 

[SnakeChomp]

An additional update on my issue. It may not be the keyboard thing as I originally thought. At least, I now have a reliable way to induce the DPC latency issue at will. I have the ESATA bracket installed (GA-EP35-DS3P motherboard) and an esata enclosure with 500gb western digital sata drive installed. If I plug the e-sata cable into the enclosure while the enclosure is powered off, it creates dpc latency issues (spikes of 200000us). If I turn the enclosure on, the spikes go away. I turn the enclosure back off, and the spikes return. I unplug the e-sata cable, and the spikes go away. I do this all while windows is running without rebooting. I have the esata bracket plugged into the intel sata controller, which is in AHCI mode to support hot-plugging. My spikes are now only 1000us, which has no performance impacts that I can percieve.

Hope this helps someone out there.

 

[chriskwarren]

Originally posted by: SnakeChomp
An additional update on my issue. It may not be the keyboard thing as I originally thought. At least, I now have a reliable way to induce the DPC latency issue at will. I have the ESATA bracket installed (GA-EP35-DS3P motherboard) and an esata enclosure with 500gb western digital sata drive installed. If I plug the e-sata cable into the enclosure while the enclosure is powered off, it creates dpc latency issues (spikes of 200000us). If I turn the enclosure on, the spikes go away. I turn the enclosure back off, and the spikes return. I unplug the e-sata cable, and the spikes go away. I do this all while windows is running without rebooting. I have the esata bracket plugged into the intel sata controller, which is in AHCI mode to support hot-plugging. My spikes are now only 1000us, which has no performance impacts that I can percieve.

Hope this helps someone out there.

This post in the gigabyte forum also confirms your post:
http://forums.tweaktown.com/f6...index2.html#post250338

Guy unplugged the esata cable that was not plugged into anything and spikes practically vanished.

Not my particular issue but may help some.

 

[Green Gigs]

I have a GA-EP35-DS3R rev. 2.1 that I thought was running fine until I came across the DPC latency thread and used the checker. I spike up and down from 30 to 600 to 1000 repeatedly. Opening and running the Dynamic Energy Saving utility causes spikes in the 40,000 range and does create an audible disruption when listening music.