Testing / comparing : Intel D975XBX2 / Asus P5B DX ***56K WARNING***

[STaRGaZeR]

Nice sticks. Are yours the 667: 3-3-3-8, or 800: 4.4.4.10?

667 3-3-3-8 exactly, but yours will do the same i guess

Sticks are in slots 2&4, and I´m observing some annoying behavior. Stick in slot 4 is doing something like degrading itself. It will no longer boot at the default 1,84v, as it did before. The strange thing is that two weeks ago, when i got this board, both boot perfectly at that voltage. Same with my previous iCFX3200.

It looks like BX2 is really kiling progressively the sticks put in slots 3&4

[GamingDaemon]

Anandtech did a great article about a year ago that investigated the "speed improvements" of running RAID 0 vs. running a single drive. They looked at the high speed WD Raptor 15K discs and put them against slower drives like the 7200RPM Seagate Barricudas.

The result? Virtually ZERO speed improvements with RAID 0 in games or regular windows applications.

So unless your box is running a web server or database, I would recommend that if anything you go to RAID 1 so that you have a little extra insurance in the event a drive goes out on you.

Cool. I will stick with what I have and maybe do RAID 1 down the line. Thanks jmpage2.

[GamingDaemon]

I think it really depends on the game. I've seen benchmarks for some games that take as much as a 50% performance hit when running under Vista.

I run BF2142 under Vista with the release version Nvidia drivers on my 8800 and enjoy a nice solid 99 frame rate in the game. It runs buttery smooth in single player, but I get a lot of lags/hiccups when playing multi-player even though the framerate counter never drops below about 50fps. It definitely feels like driver/OS interaction issues going on. Also if I turn up all of the details in BF2142 I've had some weird application crashes, etc, that I suspect are driver related (seeing as the rig is stable in testing for up to 12 hours with no errors).

Yeah, the interaction between Vista/nVidia or Vista/ATi is still buggy, so it comes down to the actual system/game. The only good news is that Vista seems to be working hard on the problem. I will be installing Obvlivion tonight...I'll let you know how that works out...

[ArcticOC]

Got replay from Intel today on the C1E issue.

Hello Jorn,

Thank you for contacting Intel(R) Technical Support.

In regards to this issue with the C1E technology on your Intel(R) Desktop Board D975XBX2, please provide us with the following information:

-The model of the processor you are using (please provide us with the product code from the box as well).

-How did you determine that C1E was not being fully disabled?

-Have you disabled Enhanced Intel SpeedStep(R) Technology as well?

We will be waiting for your reply.

Please do not hesitate to contact us again if you need further assistance.

Sincerely,

Paul C.
Intel(R) Technical Support

My reply:

Thank you for your quick reply!

Q1: The model of the processor you are using (please provide us with
the product code from the box as well?
A1: I have this problem with an E4300, product code: BX80557E4300SL9TB

Q2: How did you determine that C1E was not being fully disabled?
A2: CPU-Z always monitor the CPU-speed in realtime, and when my system
is in IDLE state, the multiplier instantly drops form 9x to 6x. I`m
using version 1.39 of CPU-Z. When i stress my system, the multiplier
goes back to the "normal" 9x, giving me the correct totalt frequency.

Q3: Have you disabled Enhanced Intel SpeedStep(R) Technology as well?
A3: Yes, it has been disabled in the BIOS (EIST)

The "Badaxe 2" is targeted at power/entusiast users, so an
error/problem of this kind is very damaging for the boards reputation.
This is also becoming a quite known issue in many tech-forums, as many
other also have the same problem. I really hope this "bug" can be
corrected in an upcoming BIOS, as I and many others like this board
alot.

Best regards
Jorn Ove Edvardsen

[finance_coder]

Hi guys, first post here.

I've built many computers, but this is the first I've really tried to overclock. I've got an XBX2, E6600 and 2x1GB G.Skill "HZ" DDR2-800 memory.

Right now, I'm at strap (ref freq)=266, mem speed=800, fsb=333 (2.99 GHz). All voltages are stock. System is stable.

Not a particularly interesting OC (definitely not xtreme!), but I'm waiting until I receive my Thermalright Ultra-120 before I push it any higher (using the retail heatsink/fan right now; load temps are in the low 60s).

Anyway, here's what's interesting: while trying to achieve this OC, I first pushed my mem speed down to 667. The system was stable, except for my Promise SATA II TX4 150 PCI SATA card. This card has four drives attached to it; with the memory speed set at 667, the card would only detect three of the four drives. I rebooted many times, each time shutting down completely and double-checking that my cables were properly seated. No luck.

But when I moved the memory back to 800, the SATA card correctly detected all drives!

Anyone have any idea what might cause this? I plan to try and push the CPU higher, but will surely need the ability to change my memory speed.

As a side note: my goal is to have the max stable OC without increasing any voltages (for heat, silence and longevity of my hardware). I'm not going for any records, just want to get the most value out of my hardware.

One more note: you may be asking, why use the Promise SATA card when there's a perfectly good Marvell SATA chip built in to the XBX2? Well, this system runs Linux exclusively, and (at this time), the Promise drivers are more stable/mature than the Marvell ones.

Thanks!

[chaotic646]

Got replay from Intel today on the C1E issue.



My reply:

Thank you for your quick reply!

Q1: The model of the processor you are using (please provide us with
the product code from the box as well?
A1: I have this problem with an E4300, product code: BX80557E4300SL9TB

Q2: How did you determine that C1E was not being fully disabled?
A2: CPU-Z always monitor the CPU-speed in realtime, and when my system
is in IDLE state, the multiplier instantly drops form 9x to 6x. I`m
using version 1.39 of CPU-Z. When i stress my system, the multiplier
goes back to the "normal" 9x, giving me the correct totalt frequency.

Q3: Have you disabled Enhanced Intel SpeedStep(R) Technology as well?
A3: Yes, it has been disabled in the BIOS (EIST)

The "Badaxe 2" is targeted at power/entusiast users, so an
error/problem of this kind is very damaging for the boards reputation.
This is also becoming a quite known issue in many tech-forums, as many
other also have the same problem. I really hope this "bug" can be
corrected in an upcoming BIOS, as I and many others like this board
alot.

Best regards
Jorn Ove Edvardsen

I have a feeling they are going to blame CPU-Z. I emailed them yesterday but they havent replied yet.

gtj - Is there any other way to prove C1E isnt disabled besides the CPU-Z answer Arctic gave? Just wondering because they will probably ask me the same thing very soon.

[GamingDaemon]

Hi guys, first post here.

I've built many computers, but this is the first I've really tried to overclock. I've got an XBX2, E6600 and 2x1GB G.Skill "HZ" DDR2-800 memory.

Right now, I'm at strap (ref freq)=266, mem speed=800, fsb=333 (2.99 GHz). All voltages are stock. System is stable.

Not a particularly interesting OC (definitely not xtreme!), but I'm waiting until I receive my Thermalright Ultra-120 before I push it any higher (using the retail heatsink/fan right now; load temps are in the low 60s).

Anyway, here's what's interesting: while trying to achieve this OC, I first pushed my mem speed down to 667. The system was stable, except for my Promise SATA II TX4 150 PCI SATA card. This card has four drives attached to it; with the memory speed set at 667, the card would only detect three of the four drives. I rebooted many times, each time shutting down completely and double-checking that my cables were properly seated. No luck.

But when I moved the memory back to 800, the SATA card correctly detected all drives!

Anyone have any idea what might cause this? I plan to try and push the CPU higher, but will surely need the ability to change my memory speed.

As a side note: my goal is to have the max stable OC without increasing any voltages (for heat, silence and longevity of my hardware). I'm not going for any records, just want to get the most value out of my hardware.

One more note: you may be asking, why use the Promise SATA card when there's a perfectly good Marvell SATA chip built in to the XBX2? Well, this system runs Linux exclusively, and (at this time), the Promise drivers are more stable/mature than the Marvell ones.

Thanks!

I have no idea why it would have problems recognizing the SATA drives, but obviously something was unstable. When you say you are stable now, have you been running Orthos? What is your memory divider? 3:2? I would think you would want to try a memory speed of 533 to get a 1:1 divider, but then i am still new to this.

P.S. I have come to the same conclusion about OC'ing...a little goes a long way...not trying for any records either...just a good balance of speed, power and temps

[Maglin]

667 3-3-3-8 exactly, but yours will do the same i guess

Sticks are in slots 2&4, and I´m observing some annoying behavior. Stick in slot 4 is doing something like degrading itself. It will no longer boot at the default 1,84v, as it did before. The strange thing is that two weeks ago, when i got this board, both boot perfectly at that voltage. Same with my previous iCFX3200.

It looks like BX2 is really kiling progressively the sticks put in slots 3&4

Hey Stargazer. Are you mem sticks Micron D9's? And where you running them at 2.32v? You sig looks to confirm the 2. My D9 died in slot 4 and was running 2.32v @ 4-4-4-10 @ 928 mhz. When my new stick arrives I'll be tring to shoot for 2.0 vDimm and see how far I can push them at that voltage. Crucial has a lifetime warrenty and offers crossshipping RMA. So if this happens again I'll be working with them to get me some better ram that isn't D9's and see if it isn't just a IC issue with the XBX2.

[JamesAvery22]

Sorry for the non-oc related post but does anyone have 5.1 audio working with the onboard sound, Vista, and Red Orchestra? The RO forums have been less than helpfull...

[STaRGaZeR]

Hey Stargazer. Are you mem sticks Micron D9's? And where you running them at 2.32v? You sig looks to confirm the 2. My D9 died in slot 4 and was running 2.32v @ 4-4-4-10 @ 928 mhz. When my new stick arrives I'll be tring to shoot for 2.0 vDimm and see how far I can push them at that voltage. Crucial has a lifetime warrenty and offers crossshipping RMA. So if this happens again I'll be working with them to get me some better ram that isn't D9's and see if it isn't just a IC issue with the XBX2.

Well, since i got the board to 4 days ago I was running them at 2,40v 822 3-3-3-1 (and optimized with memset at windows startup), but they work just fine with 2,36v, so I dropped the voltage. My chips are Micron D9GMH. I´m now thinking that in a few weeks my second stick will be dead. Hope not.

People here with OCZ RAM seems to have no memory related issues.

[SiGfever]

Got replay from Intel today on the C1E issue.



My reply:

Thank you for your quick reply!

Q1: The model of the processor you are using (please provide us with
the product code from the box as well?
A1: I have this problem with an E4300, product code: BX80557E4300SL9TB

Q2: How did you determine that C1E was not being fully disabled?
A2: CPU-Z always monitor the CPU-speed in realtime, and when my system
is in IDLE state, the multiplier instantly drops form 9x to 6x. I`m
using version 1.39 of CPU-Z. When i stress my system, the multiplier
goes back to the "normal" 9x, giving me the correct totalt frequency.

Q3: Have you disabled Enhanced Intel SpeedStep(R) Technology as well?
A3: Yes, it has been disabled in the BIOS (EIST)

The "Badaxe 2" is targeted at power/entusiast users, so an
error/problem of this kind is very damaging for the boards reputation.
This is also becoming a quite known issue in many tech-forums, as many
other also have the same problem. I really hope this "bug" can be
corrected in an upcoming BIOS, as I and many others like this board
alot.

Best regards
Jorn Ove Edvardsen

This is from Intel on the C1E issue...The first quote is the one that I sent at first followed by their answer. Then I sent the second quote from gtj and the last response is the one that I received today...

Quote:

"Hello again, John,

Thank you once more for contacting Intel(R) Technical Support.

We have done some further research in this situation, and found out that the situation you are having is related to a feature called the Intel(R) Thermal Monitor 2, which lowers frequency and voltage (Dynamic VID) to help reduce power consumption.

Unfortunately, this feature is hardware-enabled in the processor, and cannot be disabled. We apologize for the inconvenience.

Please do not hesitate to contact us again if you need further assistance.

Sincerely,

Paul C.
Intel(R) Technical Support"

Quote:
Paul,

After further research...

"According to the document "Intel® Core™2 Duo Desktop Processor E6000∆ and E4000∆ Sequence Thermal and Mechanical Design Guidelines", TM2 is disabled by default and when it is enabled, it only becomes active when the processor reaches a critical temperature. Further, it appears that the 975XBX2 BIOS does NOT enable TM2 although the same document states that it is required.

The document is at
http://www.intel.com/design/core2duo/documentation.htm"

John

Quote:

"Hello again, John,

Thank you once more for contacting Intel(R) Technical Support.

Regarding this situation, after further researching the information, we have concluded that the option is enabled by default when a processor that supports it is detected by the BIOS.

This board features a thermal protection circuit in the processor voltage regulator area. This circuit protects the processor voltage regulator from overheating and damaging the board. The thermal protection circuit in the processor voltage regulator sensor is triggered at approximately 120C. This trigger will cause the processor to enter a throttling mode (slowing down the processor if it exceeds its maximum operating
temperature) and allow the processor voltage regulator to cool down.


Please do not hesitate to contact us again if you need further assistance.

Sincerely,

Paul C.
Intel(R) Technical Support "

[theonlybabyface]

Maybe you should clarify that in no way did your processor voltage regulator reach 120C at stock. Can't they figure that out??? Also, let then know to stop copying and pasting their answers from some book....

PS - I also submitted this bug last week and never did get back a reply.

[J@mmer]

That's promising. Anyone have thoughts on full surround for gaming and movies? Clear, clean and good?

I have an X-Fi Xtreme Music. Listen to the on-board and then listen to the x-fi and I guarantee you will NEVER use the on-board again.

Specially in games. I play a lot of FEAR Combat and after using the X-Fi it's a completely different game (mainly due to EAX). I have noticed extra sounds that the on-board cannot output also the sound seems crisper and clearer somewhat.

The on-board is more that good enuff for stereo and HT duties but if you're serious about sound and in particular gaming, get an X-fi (or other gaming SC) over the on-board.

[JamesAvery22]

Has anyone done any OC'ing and when you boot into Vista Explorer is hosed? I.E., Control Panel shows "Empty" And stuff like Admin Tools is missing... Stable though...

[J@mmer]

hehehe i would hardly call that stable. It's usually a corrupted OS due to failed OC.

[JamesAvery22]

hehehe i would hardly call that stable. It's usually a corrupted OS due to failed OC.

Orthos runs for hours... And backing the OC down the menus come right back. So hardly a corrupt os...

[J@mmer]

Ummm... missing icons is what I would call a corrupt OS. Backing the OC down fixed it, which means failed OC.

Soooooo... YOU HAD A CORRUPT OS DUE TO FAILED OC.

[NastyDirty]

Has anyone done any OC'ing and when you boot into Vista Explorer is hosed? I.E., Control Panel shows "Empty" And stuff like Admin Tools is missing... Stable though...

And I thought I was the only one. Happens to me all the time, when overclocking. I use RAID 0. The only thing is do is a disk check, and it fixes the problem, or go back to the previous know stable settings, and everything is okay.

[JamesAvery22]

Ummm... missing icons is what I would call a corrupt OS. Backing the OC down fixed it, which means failed OC.

Soooooo... YOU HAD A CORRUPT OS DUE TO FAILED OC.

That would make sense if anything else didn't work... Like if Sisoft failed or Orthos caused a crash or failed... Anything. If Windows gets hosed to a point where menus are just gone they dont tend to magically come back... Thanks though

[JamesAvery22]

And I thought I was the only one. Happens to me all the time, when overclocking. I use RAID 0. The only thing is do is a disk check, and it fixes the problem, or go back to the previous know stable settings, and everything is okay.

Thanks ND, scandisk fixed it. So either OC is killing something and Vista is doing some auto repair(even though my scandisk didn't make any changes and found no errors)? Or there is some weird quirk where maybe the OS is detecting clock speed... Mine was always past a certain speed, 3.50ghz. Even a mhz higher on the bus would cause it to happen.

[jmpage2]

That would make sense if anything else didn't work... Like if Sisoft failed or Orthos caused a crash or failed... Anything. If Windows gets hosed to a point where menus are just gone they dont tend to magically come back... Thanks though

I have to disagree, I've seen all sorts of weird/strange behavior that was directly attributable to overclocking or loading software while overclocked.

[gtj]

No one answered my earlier question about 30 pages back.

How is the on-board sound compared to, say, a Creative X-Fi? This is for watching movies and gaming.

OK, I'll answer as well. Beats Me. Don't game and have digital speakers.

Got replay from Intel today on the C1E issue.

My reply:

Thank you for your quick reply!

Q1: The model of the processor you are using (please provide us with
the product code from the box as well?
A1: I have this problem with an E4300, product code: BX80557E4300SL9TB

Q2: How did you determine that C1E was not being fully disabled?
A2: CPU-Z always monitor the CPU-speed in realtime, and when my system
is in IDLE state, the multiplier instantly drops form 9x to 6x. I`m
using version 1.39 of CPU-Z. When i stress my system, the multiplier
goes back to the "normal" 9x, giving me the correct totalt frequency.

Q3: Have you disabled Enhanced Intel SpeedStep(R) Technology as well?
A3: Yes, it has been disabled in the BIOS (EIST)

The "Badaxe 2" is targeted at power/entusiast users, so an
error/problem of this kind is very damaging for the boards reputation.
This is also becoming a quite known issue in many tech-forums, as many
other also have the same problem. I really hope this "bug" can be
corrected in an upcoming BIOS, as I and many others like this board
alot.

Best regards
Jorn Ove Edvardsen

Excellent response.

I have a feeling they are going to blame CPU-Z. I emailed them yesterday but they havent replied yet.

gtj - Is there any other way to prove C1E isnt disabled besides the CPU-Z answer Arctic gave? Just wondering because they will probably ask me the same thing very soon.

Run Everest, right click on the bottom status bar and take the CPU Debug option then CPUID and MSR Dump. It's MSR 1A0 bit 25:

C1E ON:
MSR 000001A0: 0000-0040-6286-2481
C1E OFF:
MSR 000001A0: 0000-0040-6086-2481

You can also look at CPU Tweakings tool and see and change it.

This is from Intel on the C1E issue...The first quote is the one that I sent at first followed by their answer. Then I sent the second quote from gtj and the last response is the one that I received today...

Quote:

"Hello again, John,

Thank you once more for contacting Intel(R) Technical Support.

We have done some further research in this situation, and found out that the situation you are having is related to a feature called the Intel(R) Thermal Monitor 2, which lowers frequency and voltage (Dynamic VID) to help reduce power consumption.

Unfortunately, this feature is hardware-enabled in the processor, and cannot be disabled. We apologize for the inconvenience.

Please do not hesitate to contact us again if you need further assistance.

Sincerely,

Paul C.
Intel(R) Technical Support"

Quote:
Paul,

After further research...

"According to the document "Intel® Core™2 Duo Desktop Processor E6000∆ and E4000∆ Sequence Thermal and Mechanical Design Guidelines", TM2 is disabled by default and when it is enabled, it only becomes active when the processor reaches a critical temperature. Further, it appears that the 975XBX2 BIOS does NOT enable TM2 although the same document states that it is required.

The document is at
http://www.intel.com/design/core2duo/documentation.htm"

John

Quote:

"Hello again, John,

Thank you once more for contacting Intel(R) Technical Support.

Regarding this situation, after further researching the information, we have concluded that the option is enabled by default when a processor that supports it is detected by the BIOS.

This board features a thermal protection circuit in the processor voltage regulator area. This circuit protects the processor voltage regulator from overheating and damaging the board. The thermal protection circuit in the processor voltage regulator sensor is triggered at approximately 120C. This trigger will cause the processor to enter a throttling mode (slowing down the processor if it exceeds its maximum operating
temperature) and allow the processor voltage regulator to cool down.


Please do not hesitate to contact us again if you need further assistance.

Sincerely,

Paul C.
Intel(R) Technical Support "

Yeah same response I got. I'm going to do some more research on this.

[NastyDirty]

Thanks ND, scandisk fixed it. So either OC is killing something and Vista is doing some auto repair(even though my scandisk didn't make any changes and found no errors)? Or there is some weird quirk where maybe the OS is detecting clock speed... Mine was always past a certain speed, 3.50ghz. Even a mhz higher on the bus would cause it to happen.

What happens is it decides to put certain files in the "unindexed" status, which the OS can't find for use. It usually shows a total of 5 that need to be fixed, when doing a disk check. But, it is a sign of something being unstable, even if orthos is running okay. I wonder if XP users are experiencing the same. I wouldn't consider going back to XP, if it wasn't happening with that OS.

[JamesAvery22]

I have to disagree, I've seen all sorts of weird/strange behavior that was directly attributable to overclocking or loading software while overclocked.

Yeah Im probably completely wrong Just struck me weird because of its repeating behavior and orthos passing. But yeah still could be just a bad oc

[Forceman]

This board features a thermal protection circuit in the processor voltage regulator area. This circuit protects the processor voltage regulator from overheating and damaging the board. The thermal protection circuit in the processor voltage regulator sensor is triggered at approximately 120C. This trigger will cause the processor to enter a throttling mode (slowing down the processor if it exceeds its maximum operating
temperature) and allow the processor voltage regulator to cool down.

That's certainly not right, because the only time it does it is when the CPU is idling, and the voltage regulator would be having an easy time of it.