Gigabyte P35 motherboards and DPC

Has anyone else using a Gigabyte P35 motherboard experienced spikes in DPC Latency Checker? A web search has shown that this seems to be a widespread problem, and Gigabyte doesn't seem to want to address it. It doesn't affect Asus P35 boards, so it must not be a problem with the chipset. Here is what I'm referring to:

I have a Gigabyte G33 based board and I'm seeing the same thing. It doesn't seem to affect my low latency operation anymore after switching to a MOTU PCI424 based setup!

Rico

You're not by chance running Vista are you?

ORIGINAL: losguy

You're not by chance running Vista are you?

No. I'm using XP Pro 32 bit.

Hi Dan

I use Gigabyte! I remember several years ago I noticed that too but it didn't effect latency and/or PC performance at all. Hmm!!

Hi SongCraft,

I just put together a new system, so I'm trying to see what it's capable of. Compared to my old system I can get very low latencies, but it would seem that the spikes must be limiting it from reaching it's full potential. Even my notebook (single core Pentium M 715) doesn't show these spikes.

What soundcard are you using and what latencies are you able to set?

I got exactly the same problems with my Gigabyte GA P35-S3g with an INTEL Q6600 quad core and XP SP 2.
Spikes in idle don't go over 850 but i remember my old MSI K8n Neo 2 with AMD X2 4200+ that remains a lot lower and
stable with no spikes at all!
I've try everything,turn off almost everything to see if spikes go lower (networking cards,cd rom,unused usbs,unused ide channels) nothing change!

ORIGINAL: Dan DAmico

Hi SongCraft,

I just put together a new system, so I'm trying to see what it's capable of. Compared to my old system I can get very low latencies, but it would seem that the spikes must be limiting it from reaching it's full potential. Even my notebook (single core Pentium M 715) doesn't show these spikes.

What soundcard are you using and what latencies are you able to set?

Back then I was using an M-Audio 24/96 pci card on a P3 running SONAR 4, I can't remember precisely what latency I had but there was no problem at all with latency although I was limited to the amount of tracks (18 to 20 audio tracks) due to my system being a P3 / 800 with 512 Ram. I did turn off all nonessential services and other apps whilst using SONAR but the spikes remained.

Maybe with some Gigabyte boards there's an indication of activity in regards to the OS, systems files, interrupts and whatever as these could be what's causing the spikes, I'm not sure! Also I think that particular Gigabyte board may have some sort of security features built-in (could be that) but I can't be sure.

Interesting issue. If the Asus P35 doesn't exhibit the DPC latency spikes then it must be something about the Gigabyte that's doing it. Question is, what? I'm curious...

I did a clean install of Windows XP without any chipset, audio or video drivers to see if the spikes remain. To my dismay, they are still there.
I had an Acronis backup of my full install so I just restored it.

I've read that the GA-P35-DS4 rev 1 early (before F6) BIOS did not have this problem, but the later ones do. My motherboard has only 3 available BIOS, and I've tried them all to no avail. Gigabyte must have changed something. I've sent them an email about it, but I'm told they have isolated it to the southbridge, but haven't had enough demand for a fix to do anything. I'm hoping that if enough people contact them, they will release a BIOS that fixes the problem.

I'm sure it has been looked into already but are you using the latest BIOS? Updating mine improved PCI performance drastically. (I'm using an ASUS P5k fwiw)

Im getting some spikes as well, biggest one just past 500.
Rico had me check this a while back. Nice to have a good friend sonar user just down the street!

I must say this gigabyte rig runs awesome, OC'd from 1.8 to 3.0ghz

Still, I agree, ya wonder wut causes those spikes and do the effect something

ORIGINAL: losguy

Interesting issue. If the Asus P35 doesn't exhibit the DPC latency spikes then it must be something about the Gigabyte that's doing it. Question is, what? I'm curious...

I'm getting a flat line and no spikes with my Opteron based K8 board, so it seems to be isolated to Intel based boards, or a particular chipset.

ORIGINAL: Lanceindastudio

Im getting some spikes as well, biggest one just past 500.

Lance,

Could you send an email to Gigabyte tech support requesting that they address the problem? I'm hoping to get a few people besides me to contact them and get them to look for a fix.

I agree the performance is really good on these boards, but I believe it could be even better if they fix the problem.

ORIGINAL: Alndln


I'm getting a flat line and no spikes with my Opteron based K8 board, so it seems to be isolated to Intel based boards, or a particular chipset.

My Opteron, and even my notebook computer are all flat, also. I don't think it's the chipset because the Asus P35 boards don't have the problem. I think it's a BIOS issue in the way Gigabyte is using the chipset. Earlier BIOS don't seem to have the problem.

...well Dan, yours is better than mine, which occurs like clockwork every 63 seconds.

[image][/URL][/image]

This is a relatively recent occurance on a HP ZD7000 series laptop. I have done everything that I know and have recently learned to track this without success.



Laptop Specs:
Processor Pentium 4 3.06 GHz
Installed Memory 1024 MB (DDR SDRAM)
Hard Drive 80 GB EIDE, 4200 RPM
Display 17 in. WXGA+ TFT Active Matrix
Operating System Microsoft Windows XP Home Edition
Mobile Technology Wi-Fi Certified
Processor
Processor Manufacturer Intel
Processor Type Pentium 4
Processor Speed 3.06 GHz
Motherboard
Bus Speed 533 MHz
Memory
Installed RAM 1024 MB
RAM Technology DDR SDRAM
Max Supported RAM 2 GB
RAM Speed 333 MHz
Installed Cache Memory 512 KB
Hard Drive
Hard Drive Rotation Speed 4,200 RPM
Storage Controller Type EIDE
CD / DVD
CD / DVD Type CD-RW/DVD-ROM
Display
Display Tech WXGA+ TFT Active Matrix
Display Size 17 in.
Display Max. Resolution 1440 x 900
Video
Graphic Processor NVIDIA GeForce FX Go5600
Installed Video Memory 64 MB
Audio
Audio Output Type Sound card • Speaker(s)
Operating System
Operating System Microsoft Windows XP Home Edition
Platform PC
Technical Features
Expandability 1 x SD Card Slot, 1 x Memory Stick Card Slot, Docking Station/Port Replicator, 1 x MMC Card Slot, 1 x Type III PC Card Slot, 1 x Memory Stick PRO, 1 x SmartMedia Card
Expansion Ports 4 x USB 2.0, 1 x IEEE 1394a (FireWire), VGA out, Parallel Port, RJ-11 (modem), RJ-45 (NIC), S-Video out
Security Features Kensington Lock Slot, Power-On Password
Input Method Keyboard • Numeric Keypad • Scroll Button • Touchpad
Networking
Networking Type Integrated 10/100 Network Card • Infrared Port • Integrated Wireless LAN
Data Link Protocol Ethernet • Fast Ethernet • IEEE 802.11b • IEEE 802.11g
Modem
Modem Type Fax / Modem
Analog Modulation Protocol ITU V.90, ITU V.92
Battery
Battery Technology Lithium ion

I have tried to locate/fix the "spiking" issue with:

Doubledawg
DPC Spike Checker v1.03 (TcElectronics)
DPClat
KRview
MotherboardMonitor
Process Explorer
Ulysses
RATTV3

I have done the "turn off everything tilll it crashes" routine including all drivers and controllers.

This spiking issue continues.

The next step is the big "F".

F@KKER

/format, that is ;)

[edit] I have also (after the issue manifested) updated to SP2 and SP3, checked the HP site for and installed all the latest drivers, tried tweaking the NVIDIA GeForce card with Riva Tuner, ran XP w/o the video card drivers- still no success.

I have a P35 series Gigee-byter and your DPC screen looks like mine.
But. Mine runs like a deer so I don't care what that screen says.

ORIGINAL: F@KKER

This is a relatively recent occurance on a HP ZD7000 series laptop. I have done everything that I know and have recently learned to track this without success.

That is very discouraging. Are you getting dropouts on your audio?

ORIGINAL: bitman

I have a P35 series Gigee-byter and your DPC screen looks like mine.
But. Mine runs like a deer so I don't care what that screen says.

Mine runs good, too. Maybe I just need to ignore it. Which model is your motherboard?

ORIGINAL: F@KKER
...well Dan, yours is better than mine, which occurs like clockwork every 63 seconds.

Ouch. Looks like you were pretty thorough, so you probably already tried disabling the WiFi? That's been culprit #1 for laptops. (But you probably knew that already...)

ORIGINAL: Dan DAmico

ORIGINAL: bitman

I have a P35 series Gigee-byter and your DPC screen looks like mine.
But. Mine runs like a deer so I don't care what that screen says.

Mine runs good, too. Maybe I just need to ignore it. Which model is your motherboard?

I don't remember and I'm all compfy on the couch and all the stuff and manual is downstairs.

P35 somthin. Is has 3 PCI slots and (1) 1x pcie and (1) 16x pcie if that helps.

Hey Dan, after reading my own post- I must apologize to you as my post appears as a hijack- sorry.

But in answer to your question, yes I am getting dropouts. I have spent a good 4 weekends on top of hours at a time weeknights and still no success. There is a rather large (45 pages or so) thread at the sos forum on this.

Additionally, I just tried installing the recovery disk drivers, there is no change. I just read that someone did a full format and the spikes are still there, so it must be either very low level -perhaps at the bios level or maybe a hardware issue.

This sucks because the laptop was my portable DAW and I spent a good amount of $ on specific gear just for this application.

losguy, yes, I disabled wi-fi -I even tried the remove hardware route to no avail. I did the same with usb and network adapters.



F@KKER

[edit] I failed to mention that I was running the laptop as it came from HP with 512 MB of Ram and it worked with no issues (other than FW-1884 issues, in which case I would use the Echo I/O sound card) and so I just installed 2 gb ram, still no change.

Dan,

Note the scale at the left hand side of the table. This tool reports in micro-seconds. The longest wait on a deferred call is still under 1 millisecond. You'll never notice a thing when the dpc queue is showing maximums at this level.

I can't get my ASIO buffers to adjust any lower than 2ms (128 samples) -- the driver doesn't go any lower. Generally, I run at 256 samples, corresponding to a 3ms latency which is virtually instantaneous.

Relax.

Yes and no. Even if it gets no higher than 1 msec, that's 1 msec eating into the latency that you would have otherwise had. If you're wanting to run at low latency, or loading your CPU up, or both, this is mission-critical time that's getting eaten up with system whatnot.

I've flashed latest beta bios from gigabyte site (F3b) and spikes increased over 1000 micro seconds (over the yellow line)!!!!!
Ro back to latest non beta (F2) and spikes go back under the yellow line (max 900 micro seconds).
That's it....

Seems that a lot of people have the same problem :

http://www.tomshardware.co.uk/forum/247210-12-audio-latency-problem-ds3r

ORIGINAL: KWILD

I've flashed latest beta bios from gigabyte site (F3b) and spikes increased over 1000 micro seconds (over the yellow line)!!!!!
Ro back to latest non beta (F2) and spikes go back under the yellow line (max 900 micro seconds).
That's it....

What model is your motherboard? Also, have you or would you be willing to email Gigabyte tech support to ask them to address the problem?

Like i've written some posts ago my board is a GA P35-S3G , yes im think to contact the Gigabyte tech support,but i think is better to do that all togheter.
Let me know if u are thinking to write a single comulative mail asking to fix that problem.

Thanx.

http://myspace.com/_kwild

Hi Gigabyte P35 MB owners - see here for my post / solution:-

http://forum.cakewalk.com/tm.asp?m=1353232

Thanks,
Tim

ORIGINAL: timboe
Hi Gigabyte P35 MB owners - see here for my post / solution:-
http://forum.cakewalk.com/tm.asp?m=1353232

Hey Tim, great work digging this out.

Over on the Tom's Hardware forum, someone posted this:

On my P35-DS3R Rev. 1, the problem was solved by going from F11 back to F5 (original BIOS that came with the board). F11 added a High Precision Event Timer in the system devices, so maybe that's the source of this latency issue...

Maybe someone could try disabling the High Precision Event Timer to see if it fixes the problem. I can upload F5 later if anyone would like to try it. BTW, I also had higher NB temps with F11 for some reason.

Seems a reasonable thing to ask. Maybe someone here could try disabling this in their BIOS, before rolling it back, just to see if it works? It might be a way forward instead of backward...

Unfortunately, all the BIOS for my motherboard (GA-EP35-DS4) suffer from this problem.. I have an email into the Gigabyte tech support to see if I could use an early BIOS from a Ga-P35-DS4 rev 1, but so far haven't heard back. I'm not planning to switch to Vista anytime soon.

Just to update who ever may want to know, I reformatted my laptop, re-installed XP Home w/sp1 and DPClat shows the same latency spikes.

F@KKER

ORIGINAL: F@KKER

Just to update who ever may want to know, I reformatted my laptop, re-installed XP Home w/sp1 and DPClat shows the same latency spikes.

F@KKER

Bummer. Sorry to hear that.

Hi Losguy

Yep, I tried disabling HPET before "downgrading" the BIOS

It does reduce spiking BUT it results in a much higher [although more steady] DPC reading.

From memory, with HPET disabled, I got a steady 35us <-> 40us reading [with a few peaks].

If you can roll back your BIOS, dont hesitate - I was running BIOS F10 on my rig and went back to F1 and F2 - both were just as good.

Apart from fixing my Vista DAW DPC issue, I noticed zero difference in my DAW performance from going from F10 to F1 or F2.

Hope this helps,
Timboe

timboe,

So, just to clarify, you went down to BIOS F1 or F2 and were able to reduce the spiking you were experiencing? And this was on Windows Vista only?

Curious to see if any XP ( SP2) users were able to reduce this problem by droping down to a lower BIOS.
We are using the P35-DS3R as well as Win XP Pro and experiencing the same spiking problem here as well. Currently using F9 BIOS.

Steve H.

ORIGINAL: Dan DAmico

ORIGINAL: Lanceindastudio

Im getting some spikes as well, biggest one just past 500.

Lance,

Could you send an email to Gigabyte tech support requesting that they address the problem? I'm hoping to get a few people besides me to contact them and get them to look for a fix.

I agree the performance is really good on these boards, but I believe it could be even better if they fix the problem.

Yes I will email gigabyte about this tomorrow as Im very tired! Thanx for all the efforts here.

My board runs amazing, but I would love to help those that may not have the same situation, and as well, maybe improve my system!

Cheers, Lance

" ....... So, just to clarify, you went down to BIOS F1 or F2 and were able to reduce the spiking you were experiencing? And this was on Windows Vista only?...."

Yes - either F1 or F2 - both were identical in terms of elimination the DPC spikes - also, keep HPET "Enabled" in the BIOS

" ...... We are using the P35-DS3R as well as Win XP Pro and experiencing the same spiking problem here as well. Currently using F9 BIOS ..... "

My MB is a P35C-DS3R Revision 1.x.

My MB is identical to your non-C model, provided you MB is a Revision 1.0 or 2.0.

Go here for the Revision 1.0 for your MB and grab F2 [there is no F1]

http://www.gigabyte.com.tw/Products/Motherboard/Products_Overview.aspx?ProductID=2543

and go here for Revision 2.0 of your MB and grab F2 [there is no F1]

http://www.gigabyte.com.tw/Products/Motherboard/Products_Overview.aspx?ProductID=2626

If you have a Revsion 2.1 of your MB, you cannot flash back - either stick with XP X32 or find a Revsion 1.0 or 2.0 and swap them over.

Hope this helps,
Tim
PS: the Sticker on the side of the MB box will tell you which revision your MB is

timboe,

We have ver 2.1 and with XP 32, the spikes are peeking into the 700's.
Odd thing is...the board is a ver 2.1 ( indicated on the circuit board ) but the mobo shipped with BIOS f9. Updated to F11 with the same results. Dissapointing. Will send off and e-mail to Gigabyte in hopes of a resolution.

Steve H.

Who cares? I get the same spikes on my gigabyte mobo and run at 2ms latency with no dropouts whatsoever.

sessionthree,

Thats great!
On some systems however, the more complicated the projects gets and the more crouded the data streams become, the over all throughput and efficiency of the motherboard becomes paramount. Several of our projects that have in excess of 20+ tracks recording simultaneously while streaming 100+ from disk with effects are showing signs of problems that the Asus boards did not have. Add to that ANY type of external sync ( ADAT, AES, SPDIF ) to the projects as well as hardcard DSP processors and things get bogged down fast. Low latency operation simply exposes the problem faster with the real problem being at the mix end of the project when everything in the system has to play nice with each other and hand back all those remote calls to the CPU in a timely fashion.

Steve H.

Hi All,

I'm in the same boat as Featherlight, recently had my dead ga-p35-ds3p replaced with a rev 2.1 ga-ep35-ds3p and now dpc latency shows near exactly the same screen shots as on the first page :(
I was wondering if it's at all possible to flash the bios with an early (F4 I think is the earliest) version of the ga-p35-ds3p rev 2.0 bios? Would that even work, and if I tried and it didn't would I end up with a dead m/b?

Thanks for your help!

Glenn

Any news from Gigabyte?Someone get replys?

ORIGINAL: Featherlight

sessionthree,

Thats great!
On some systems however, the more complicated the projects gets and the more crouded the data streams become, the over all throughput and efficiency of the motherboard becomes paramount. Several of our projects that have in excess of 20+ tracks recording simultaneously while streaming 100+ from disk with effects are showing signs of problems that the Asus boards did not have. Add to that ANY type of external sync ( ADAT, AES, SPDIF ) to the projects as well as hardcard DSP processors and things get bogged down fast. Low latency operation simply exposes the problem faster with the real problem being at the mix end of the project when everything in the system has to play nice with each other and hand back all those remote calls to the CPU in a timely fashion.

Steve H.

QFT. Excellent answer.

ORIGINAL: KWILD

Any news from Gigabyte?Someone get replys?

It's been a week since I emailed Gigabyte, and still no reply. My tech support page says my emails are being processed.

I have run many tests and found some surprising results. I get much better performance with the WDM drivers in Sonar than the ASIO. I'm using a Delta 66 and a Delta 2496. I can go down to 64 samples in the Delta control panel with the WDM drivers, but only 128 samples with the ASIO drivers. I ran the Sonar benchmark test and wasn't able to run any multiband compressors at 64 samples with ASIO without pops and clicks, even though CPU usage was below 10%. But I was able to run 55 multiband compressors at 64 samples with WDM.

ORIGINAL: KWILD

Any news from Gigabyte?Someone get replys?

There's been a few new posts at http://www.tomshardware.com/forum/247210-30-audio-latency-problem-ds3r
If I understood correctly one of their users got a beta bios for a rev 2.1 ep35-dsr board which fixed the problem, things are looking promising!

Glenn

Interesting. One poster there reported that one fix he found suddenly introduced memory errors in *good* memory modules.

When he changed BIOS version the memory problems were solved, but his audio latency problem returned.

This would seem to indicate there is some BIOS problem with *another* latency factor, namely cas latency (CL 2 etc) in the memchips.

Apparently there's some internal timing conflict trying to satisfy both customers at the same time.

Dan,

Where exactly did you email Gigabyte support?

My P35-DS4 Rev 2.1 acts pretty much like yours (all bars ~20% higher).

Hi Ed,

I was wondering if you had the same issue with your board.

You can email them here: Gigabyte Tech Support

This is somewhat disappointing thread to read since I just built a machine using the GA-P35-DS4(rev2.0). I don't have it setup for audio work yet, but when I do I will try a lot of tests with it and watch the DPC. I hope Gigabyte resolves it. the Main reason I selected this mobo is because its one of the few modern mobo's that still supports an LPT port, which I need for my trustee old school MOTU midi interface. Ironically, something tells me that very midi interface is going to make my DPC results even worse. ha ha.. Yikes.

Hey Dewdman,

If you are so inclined, I think it might help goose Gigabyte a little if they got a few emails asking them to fix it. I posted a link to tech support a few entries up.

Dan

Just sent an email to them.

I might just do it every week.

BTW, this has not affected my performance yet. FWIW, I hav not pushed this machine either.

I was wondering if you had the same issue with your board.

I'll post a sceenshot of my DPC Latency Checker tonight.

When I perform this test my read out it almost exactly the same as this one. I have a GIGABYTE GA-P35-DS3L.

Just a note for all of you the software that came with the motherboar EasyTuner5, if you have the running my Latency was solid over 500us. When I close that program back to what you see below.

The cd that came with my board says version 1.05 how do I know what F# I have to change to an earlier version to fix this problem.

I must say though that when i turned off the EasyTuner I was able to drop all my latencies to achieve below 4.4 ms and a 60 track audio song.
Much much better than I could do before.

ORIGINAL: Dan DAmico

Has anyone else using a Gigabyte P35 motherboard experienced spikes in DPC Latency Checker? A web search has shown that this seems to be a widespread problem, and Gigabyte doesn't seem to want to address it. It doesn't affect Asus P35 boards, so it must not be a problem with the chipset. Here is what I'm referring to:

The cd that came with my board says version 1.05 how do I know what F# I have to change to an earlier version to fix this problem.

F# is a BIOS version number.

How do I get an image directly into a reply?

You have to upload it to a website or file server and link to it via the image button.

ORIGINAL: Dan DAmico

Has anyone else using a Gigabyte P35 motherboard experienced spikes in DPC Latency Checker? A web search has shown that this seems to be a widespread problem, and Gigabyte doesn't seem to want to address it. It doesn't affect Asus P35 boards, so it must not be a problem with the chipset. Here is what I'm referring to:

I'm seeing the same exact thing, same pattern and everything! (see my sig below for system specs).

However, this doesn't seem to impact my performance at all. Maybe the DPC latency checker is not accurate on this config?

-mr moon

I emailed them about it, but for me as well no apparent performance hit

That thought has crossed my mind, too. But I don't really know if it's impacting the performance. I mean, I get good performance and low latency, but it seems that maybe even better performance can be achieved if the spikes didn't occur.

ORIGINAL: Dan DAmico

That thought has crossed my mind, too. But I don't really know if it's impacting the performance. I mean, I get good performance and low latency, but it seems that maybe even better performance can be achieved if the spikes didn't occur.

My system is overclocked; is everyone else's as well?

...Probably not.

Anything else common? I'm using F2 BIOS. All SATA II drives, Nvidia-based video.

Any other ideas?

-mr moon