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

[theonlybabyface]

C1E bug still not fixed...

[ChaosMinionX]

C1E bug still not fixed...

QFT.....

EDIT: I was all over that Intel site when it hit

[ArcticOC]

C1E bug still not fixed...

Well.. Beliving Intel would fix that bug is like beliving in Santa Claus

Guess we have to wait on the next version, but I`m getting to old to belive in Santa anyway

[Poodle]

Getting the feel that they can but won't fix it.

[TWY]

Wonder if anyone has managed to get a Rev. 506 yet?

[argylesocks]

No reason other than I didn't have time to confirm that 4-4-4-15 would be stable enough. I need to make a pass through all 3 machines and re-confirm my settings.

no time? huh? i thought you worked for XS & answering all our questions was your full time job?

[J@mmer]

Yeh just flashed to 2674... nothing new that i noticed. Anyone tried pushing their OC further with this new BIOS?

[STaRGaZeR]

That's what I'm using here. Vmch is at 1.88v real, nb temp is at 26C, room temp is 76f

How did you measure vMCH and NB temp?

[GamingDaemon]

How did you measure vMCH and NB temp?

Yeah, how did you do that?

Also, what exactly was the reason for the BIOS release? I don't see any major improvements in the release notes.

[gtj]

How did you measure vMCH and NB temp?

Back in the early pages of this thread are pics that show where on the MB the points to measure vcore vmch and vdimm are.

Also, what's commonly reported as V1.5 in sensor reading programs is the vmch.

And just to make sure... You can't programmatically measure the NB temp. There's no sensor on or near the chip. You have to stick a thermometer on it. The "Motherboard" sensor is actually in the upper left of the processor area.

[gtj]

no time? huh? i thought you worked for XS & answering all our questions was your full time job?

If I could maintain my current standard of living answering all your questions, I'd do it in a heartbeat. Unfortunately, reality beckons...

And that's not what I was thinking.

[shimmishim]

any signs of unlocked mults?

[GamingDaemon]

Thats what i like about Bx2 always constantly improving their bios. Strange that even though this bios update works for all the OSs but if u select Vista 32 the bios driver is not there?

It's there now...

[STaRGaZeR]

Back in the early pages of this thread are pics that show where on the MB the points to measure vcore vmch and vdimm are.

Also, what's commonly reported as V1.5 in sensor reading programs is the vmch.

And just to make sure... You can't programmatically measure the NB temp. There's no sensor on or near the chip. You have to stick a thermometer on it. The "Motherboard" sensor is actually in the upper left of the processor area.

Yeah I know, thats why i´m asking him, if that measurements are accurate

[aicjofs]

I don't see any major improvements in the release notes.

I was glad to see the floppy thing addressed without media. Yes I still use a floppy even though it needs to go away. In the old BIOS if you used advanced boot it would sit accessing the floppy during boot adding 15-20 sec to your boot time even if nothing in the drive. Work around was to disable boot from removable media but not the best solution.

I personally enable C1E and like the feature when I'm runnning 24/7 but I don't like the fact that when I want to disable it for benching and testing I can't. I wrote them awhile back about C1E and I got back an email back from Intel requesting more information so I thought I'd help the team here by sending the info they wanted back from my perspective just to add one more voice to what others have said before me.

Processor model: QX6700 (I'm at work can't see box, but its an 06 week 47 if that helps)
EIST is disabled in BIOS, windows power management is set to always on.

The C1E effect can be observed by Intel utilities that report clock speed and 3rd party utilities such as CPU-z, CrystalCPUID, RMClock etc. I am aware Intel does not base it's conclusions off 3rd party software however it can be observed with Intel utilities as well.

In layman terms my understanding accroding to documents, C1E will decrease multiplier and voltage during low workload to cutdown power and heat. When the OS is idle(and therefore invoking halt command), I can't stress enough that the whole issue occurs when the OS is idle meaning ~ 0% CPU usage. It can be observed no matter if C1E is disabled or enabled in the BX2 BIOS, the multiplier decreasing and corresponding CPU speed decrease when the OS is IDLE. If the same CPU is placed in another manufactures board and C1E is disabled in that boards BIOS then the CPU Speed/multiplier will NOT decrease when the OS is IDLE. EIST is disabled in BIOS and OS on both board setups. To me this points to 2 examples of the BX2 not actually disabling C1E. 1) The visual evidence of decreased speed follows the design criteria of C1E and therefore must be the mechanism causing the decrease as EIST is removed from the equation. 2) Disabling C1E in another manufactures board causes the CPU not to decrease it's speed when the OS is idle.

Now I heard Intel has thought this issue may be TM2 related or related to the motherboard trying to protect overheating of the voltage regulators. This just seems to be a misunderstanding of end user and maufacture when the problem is being discussed. The issue with C1E comes up when the system is idle. Every case I have seen and including my own is far from thermal margin when the system is idle. No throttling should be in effect for a CPU that is idle and has a functioning cooling system keeping the CPU from it's thermal margin, and the voltage regulators are also under light load and far from overheat protection. In fact the easy way to discount all the other throttling mechanisms built into Core2 and BX2 being the issue we are discussing, is that when at full load(100% CPU usage) no throttling takes place at all. CPU is at full speed, multiplier is at normal value. If no throttling is occuring at 100% CPU load for Core 2 or t he BX2 voltage regulators then it certainly should not be throttling for thermal margins at ~ 0% CPU which is the issue we are currently talking about.

2 examples of the fact C1E can be disabled but not on XBX2 via BIOS. 1) Another motherboard from a different manufacture will not reduce multiplier at OS IDLE with C1E disabled in BIOS. 2.) Now I know Intel is not interested in 3rd party software but RM Clock utility can disable C1E on my QX6700 in my XBX2 in Windows XP after I set C1E to disabled in that software. **After I disable it I no longer see this issue we are discussing on my system.**

The last sentence is very important. It goes to show that C1E is what is causing the multiplier to decrease at OS IDLE, and that on the XBX2 with a QX6700 it is possible to disable C1E and have CPU run at a constant speed.

All this said I use the system with C1E enabled because I like the CPU slowing down when it's idle. However there are times when I like to have C1E off, and the XBX2 BIOS C1E disable is not functioning correctly. You guys are doing great with these BIOS's and I thank you for taking the time to respond to this issue, I'm sure we can get this figured out. Please read what I have said and reread if necessary as it is quite a bit to digest when I'm describing things over email, I tried to lay it out the best I could with supporting evidence that clearly points to C1E.

Thanks again,

I led a whole bunch of people on a BIOS bug on the MSI Neo4 a few years back. Tech support wasn't much help on that, they kept giving canned answers kind of like what we are getting right now. In the end I was talking to the North America motherboard manager and direct contact with BIOS engineer. The language gap was probably the biggest issue. That is avoided in our current situation so I think we'll get it resolved but I can tell you from past experience that it takes time, numbers of people reporting the issue, clearly defined problem, and then more time. It'll work out.

[chaotic646]

My latest reply from Intel:



Thank you for contacting Intel(R) Technical Support.

The specifications your processor supports are: Dual Core, Enhanced Halt State (C1E), Enhanced Intel Speedstep(R) Technology, Execute Disable Bit, Intel(R) EM64T 2, Intel(R) Thermal Monitor 2 and Intel(R) Virtualization Technology. You can look at this information at the following links:
http://processorfinder.intel.com/det...px?sSpec=SL9S8
http://processorfinder.intel.com/det...px?sSpec=SL9ZL

The change on the multiplier of the processor was due to the stress test that at the same time overheats the processor. When this happens, the TM2 (Thermal monitor 2) tries to lower down the voltage at that point if the temperature still remains high the C1E (feature that you can not disable in the processor) would activate so the Halt instruction is executed by the processor allowing the system to lower frequency and the DVID (Dynamic Voltage Identification) to reduce power consumption, lower the voltage and the temperature.

Then again, when running applications that demand less processor power, the system will slow the processor clock speed down. Enabling of EIST can lead to power efficient systems that can run quieter and cooler.

If the Intel(R) Processor Identification Utility showed that the processor does not support the Intel(R) Halt State feature might be because the feature is disable at a BIOS level or the BIOS revision is out of date.

Intel(R) does not rely on third party software information therefore you should contact the software developers of those programs you are running.

Sincerely,

Jonathan V.
Intel(R) Technical Support


And my reply:


Thanks for the quick reply.

"The change on the multiplier of the processor was due to the stress test that at the same time overheats the processor."

My cpu is water-cooled and never gets above 38C under a full load. From what Ive read on the Intel site, cpu overheating is anywhere between 70C and 120C depending on the CPU. As you can see I'm well below that.

"When this happens, the TM2 (Thermal monitor 2) tries to lower down the voltage at that point if the temperature still remains high the C1E (feature that you can not disable in the processor) would activate so the Halt instruction is executed by the processor allowing the system to lower frequency and the DVID (Dynamic Voltage Identification) to reduce power consumption, lower the voltage and the temperature."

Thank you for acknowledging that you can not disable C1E. Let me ask you something - Why would an overclocker care about power consumption? Limited power consumption prevents a processor from reaching its full potential. That's the whole point of me contacting Intel Support - there is an option in bios to turn off C1E but it doesnt really turn it off. Intel released an enthusiast board designed for overclocking but the people who buy it are limited in how much they can overclock because of C1E limiting the power consumption.

[chaotic646]

Okay Im done with Intel customer support, its totally worthless. Heres my final response from them:


Thank you for contacting Intel(R) Technical Support.

I understand your point about disabling the C1E feature so you can overclock the system.

As stated before, unfortunately this feature is not possible to be disabled in a processor level because it is part of its electronic design.

Bare in mind that overclocking is neither recommended nor supported by Intel(R).
Information on this subject can be found at:
http://www.intel.com/support/process.../cs-007627.htm and
http://www.intel.com/support/process.../cs-001614.htm

As well it is important to mention that we do not recommend using a different fan and heatsink than the one that comes with the boxed processor. Using a different fan and head sink than the one that was provided with the processor, as well as overclocking will void the processor warranty.

Sincerely,

Jonathan V.
Intel(R) Technical Support

[hotchoc]

As well it is important to mention that we do not recommend using a different fan and heatsink than the one that comes with the boxed processor. Using a different fan and head sink than the one that was provided with the processor, as well as overclocking will void the processor warranty

Right..... Try explaining that to everyone

[GamingDaemon]

Essentially, Intel has not committed themselves to the Gaming/Overclocking market. They didn't like getting beat by AMD, so they built the Core 2 Duo line of processors which are easy to overclock, and quickly won back the market share in this niche.

But, since their procs are also used in business servers, business desktops, etc., they want to be sure they appeal to that market as well with reliable service. To do that, they need to protect themselves (read: cover their assets) to prevent a landslide of ridiculuous lawsuits. Hence, they have not committed 100% to making an overclocking BIOS.

It is a problem of conflicting interests, where they are trying to appeal to everyone. They should decide what they want to produce: procs or motherboards. Then, they wouldn't seem so schizophrenic.

However, the motherboard industry is simply to appealing to leave it alone. If I was CEO of Intel, I would truly separate these lines of business. I would continue sharing information between them, but they would be wholly seaprate entities so they can pursue their intended audiences without sending mixed messages.

Just my $0.02

[Zucker2k]

Okay Im done with Intel customer support, its totally worthless. Heres my final response from them:


Thank you for contacting Intel(R) Technical Support.

I understand your point about disabling the C1E feature so you can overclock the system.

As stated before, unfortunately this feature is not possible to be disabled in a processor level because it is part of its electronic design.

Bare in mind that overclocking is neither recommended nor supported by Intel(R).
Information on this subject can be found at:
http://www.intel.com/support/process.../cs-007627.htm and
http://www.intel.com/support/process.../cs-001614.htm

As well it is important to mention that we do not recommend using a different fan and heatsink than the one that comes with the boxed processor. Using a different fan and head sink than the one that was provided with the processor, as well as overclocking will void the processor warranty.

Sincerely,

Jonathan V.
Intel(R) Technical Support

Judging from the above, it seems Intel considers disabling CIE a hack? Might explain why they had that feature hidden in configuration mode. I think we're f*@ked. Remember, third party mobo manufacturers like Asus have been known to find ways to circumvent Intel's specifications.

[Xilikon]

Okay Im done with Intel customer support, its totally worthless. Heres my final response from them:


Thank you for contacting Intel(R) Technical Support.

I understand your point about disabling the C1E feature so you can overclock the system.

As stated before, unfortunately this feature is not possible to be disabled in a processor level because it is part of its electronic design.

Bare in mind that overclocking is neither recommended nor supported by Intel(R).
Information on this subject can be found at:
http://www.intel.com/support/process.../cs-007627.htm and
http://www.intel.com/support/process.../cs-001614.htm

As well it is important to mention that we do not recommend using a different fan and heatsink than the one that comes with the boxed processor. Using a different fan and head sink than the one that was provided with the processor, as well as overclocking will void the processor warranty.

Sincerely,

Jonathan V.
Intel(R) Technical Support

LOL, what's the point of selling a motherboard with overclocking if immediatly after selling you one, you cannot do anything with it.

I suggest you try to find a way to contact the BIOS team since the sig on those replies made me believe it's a entirely different departement, team and surely a poorer IQ level. Those guys has no clue how things work !!

[Zucker2k]

I'm beginning to doubt if we'll see a downward multiplier in a future bios release. It makes no sense, the feature has always been present in all the released bioses but only for the extreme edition processors. If they intended to make it available for non-extreme processors they would have done so.

Now moving specifically to the lock-down feature in the bios itself, anyone who has examined the itk files included in the itk bios versions will notice that the default bios multi is set to 13*. This makes sense to you? Of course! Only extreme edition processors are capable of upward multi up to 13 and beyond. So what this means is that the bios initially tries to run with the 13 divider, if your non-extreme processor refuses to comply, the multiplier remains locked. Until we find a way to fool the bios to accept a default multi of 6 (the lowest core 2 duo divider) we won't be able to unlock the multiplier. Note that I have an ES cpu which should be able to run unlocked? Not with the BX2. Why? You guessed right, my E6700 ES has a max multi of 10. I hope some talented hacker out there comes to our aid. I may just revert to my AW9D-Max. Been waiting too long for that golden bios from intel.

PS: I have tried several times to override the multiplier manually without success. I'm sure there is a fool-proof feature built into the non-writable portions of the bios or something.

[ArcticOC]

Well, just did my final reply to Intel. I put in some funny points for them to answer to

Right now I`m thinking that we from now on rather should adress this issue to the PR-division of Intel. Cause what is happening right now is that the tech support team is destroying Intel`s reputation among the power users and as we all know, this group is essential to have on it`s side becuase the mainstream tends to listen to "experts"

Personally my recomandation is worth at least 5-6 sales for Intel the coming months as I as an power user very often get asked what`s hot and not. And these people often tend to bring the message to other and so on.. If you start multiplying these numbers, how many sales will Intel lose on my "no go" on the brand and board?! Multiply that number again with frustrated users on this and other forums and I`m sure the number get quite ugly

Is I see it, this has been a "striptease" or "flirt" with the OC crowd from Intel`s side, but now the bar is closed and no more drinking for us and we must all head on home (Abit, DFI, Asus..) again

[aicjofs]

I suggest you try to find a way to contact the BIOS team since the sig on those replies made me believe it's a entirely different departement, team

Right now I`m thinking that we from now on rather should adress this issue to the PR-division of Intel.

Now we are talking. That's what I was telling everybody in my last post you have to get the information directly to the BIOS team if the Tech Support is blocking you. Or at least the above ideas are the way things had to be addressed with MSI. It's not that the Tech Support dept doesn't want to help but their understanding level only goes so far. i.e. We aren't talking directly to the engineers who create C2D or BX2 we are talking to tech support which primary function is to maintain a product and secondary role is to feedback to the enginners MAJOR issues.

Blah blah about the C1E hardware locked. BIOS writing may be the most difficult coding on the planet. A BIOS engineer didn't put in a C1E "enable" "disable" in the BIOS on a whim or by accident if it was hardware locked. Just silly to even think that. You think the RM Clock coders work for Intel, no they found how to disable it by writing to a MSR in a white paper somewhere, it's probaly not even that big of deal to do, but it's not tech supports job to be writing CPU registers.

Another thing is a lot of people are mentioning overclocking to the tech support. To me this is like saying you had a few drinks to the cops, you are pretty much guilty of what ever happened to you. The C1E issue is there even and stock settings. IMO approaching it that way sends up more flags. "What our board is functioning weird at stock?!" That would get more attention from me then some guy who is overclocking and says he has an issue.

I'm not a know it all, not at all, but I spent at least 3 months battling MSI before we got a BIOS that helped and it wasn't by going the tech support route, and it started the same way this issue is bothering you guys right now when tech support is blocking.

[ArcticOC]

Now we are talking. That's what I was telling everybody in my last post you have to get the information directly to the BIOS team if the Tech Support is blocking you. Or at least the above ideas are the way things had to be addressed with MSI. It's not that the Tech Support dept doesn't want to help but their understanding level only goes so far. i.e. We aren't talking directly to the engineers who create C2D or BX2 we are talking to tech support which primary function is to maintain a product and secondary role is to feedback to the enginners MAJOR issues.

Blah blah about the C1E hardware locked. BIOS writing may be the most difficult coding on the planet. A BIOS engineer didn't put in a C1E "enable" "disable" in the BIOS on a whim or by accident if it was hardware locked. Just silly to even think that. You think the RM Clock coders work for Intel, no they found how to disable it by writing to a MSR in a white paper somewhere, it's probaly not even that big of deal to do, but it's not tech supports job to be writing CPU registers.

Another thing is a lot of people are mentioning overclocking to the tech support. To me this is like saying you had a few drinks to the cops, you are pretty much guilty of what ever happened to you. The C1E issue is there even and stock settings. IMO approaching it that way sends up more flags. "What our board is functioning weird at stock?!" That would get more attention from me then some guy who is overclocking and says he has an issue.

Never ever mention overclocking in any mail to Intel support staff! I haven`t done this in any of my conversation w/intel regarding the issue. Because I know what answer I get back Just mention that the C1E option in BIOS = non working

I will try to dig up some mail adress to the PR division, because like you say I also belive that these "middel men" that have been answering us the last weeks are not pshysicaly writing the BIOS code. They are just the first row in the support chain. The PR division may change this policy when they realize that this could turn in to very negative headlines on hardware site`s all over the www!

What if "AnusTek" did the same?! Headlines would pop up all over (tom`s, anand...) "Overclockers fooled.. the truth about Anus rippoff!"