C2D only goes to 1.120 idle...anyone get .85ish???

[crow_smiling]

AGAIN, that is under load, look at the multiplier, it is at 9(4ghz), not 6 (1.6ghz), so speedstep is not applied at that moment (it is in the bios, just that system load is so high that is hasn't dropped down yet.) My comp goes to, according to CPUZ, to 1.280V uderload (2 instances of Prime95) and RMclock says the normal 1.35V. I will look around for non ES but that is not the point.

You are missing my point. Do you think that it is likely that the chip that Anandtech used was actually running at 4GHz at 1.15V? If not, then CPU-Z is inaccurate when reporting VCore with this ES chip. If CPU-Z was inaccurate with this chip, could it not also have been inaccurate with the ES chip that TOMs had?

I’m speaking from experience as someone that has just sold an ES Core Duo on eBay. I had the same problem with CPU-Z reporting false voltages. But don’t take my word for it, go over to Xtreme forums and see all the screenshots with CPU-Z showing erroneous VCore values and ask around.

Hey, maybe you are right; if so please come back here and tell us the good news.

 

[Black_Knight_MC]

There is a god, thank you for checing into it.

My main arguement is the VID, RMclock only reports 1.162 as the lowest option, in other words, its not even TRYING to lower the CPU voltage even close to .85. Think of that at 6x the CPU is just getting an unnecessay amount of voltage (in no way harming it, much more than it needs to be stable) and just wasting electricity and making extra heat.

 

[odk0037]

Nope, lowest i got is 1.141v. That's with 6x multiplier at 1.6ghz. There is an option in bios that allows me to disable one core, so may be i'll get .85v with only 1 core enable? what else could possibly influence the voltage for the cpu that we might not be aware of?

To crow_smiling, i think what black knight is saying is that on spec it claim .85v~1.35 or something like it at high end, and why aren't we able to achieve the .85v while they can on the ES version. The .85v is from intel, not the ES reviewer with false readings. and btw, i confirmed my reading with the supplied intel desktop utility, so it can't be cpu-z's false reading.

 

[Black_Knight_MC]

Tomshardware review was the ONLY time I have seen a conroe (even ES) working like Intel says. I have been to 10 (and I am still looking) reviews sites and those that had ES's did not get the .85ish voltage. Is there a way to contact TH? I'll take a look but I never checked because I was hoping there was a simple answer or fix, but so far no one has offered any proof other than others not getting it to work like Intel claims either.

 

[crow_smiling]

To crow_smiling, i think what black knight is saying is that on spec it claim .85v~1.35 or something like it at high end, and why aren't we able to achieve the .85v while they can on the ES version. The .85v is from intel, not the ES reviewer with false readings.

I know he’s getting excited by the Intel spec sheet, but there’s something you need to know about the way Intel presents data these days. It tends to use the same data for every chip in a particular range. So you will see that all the E6xxx chips are shown as having a TDP of 65W, from 1.86GHz to 2.67W. Obviously, this is not a true reflection on the individual chips within the range, as they all vary.
Also, the figures that they release at the time they launch a new chips series, can also cover processors that they intend releasing in 6 months time, which may not even have been announced yet. The reason they do this is so that manufacture’s have a set of data that they can use to build systems that covers more than just the chips released at launch. This doesn’t mean that the data they release at launch doesn’t get amended at times, but I’m sure you get my point.

If you look at the data for the Core Duo T2300E you will see a range of 1.1625V - 1.30V. Yet, I am able to reduce the VCore to 0.95V with this CPU using RMClock, in my laptop and also when I tested it in a desktop board.
Things aren’t always what they seem!

 

[Black_Knight_MC]

I am not getting "excited". I had seen what Intel's specs showed and how earlier processors BARELY varied, less than .06 volts at least. If you look at the electrical section of the PDF I linked to you will see how often the .85 minimum is shown and so on. Go look at other PDFs of older p4's and you will see that they also show volts that they ACTUALLY run at, almost exact. If anything, I am trying to find out if Intel is saying they can but cant or if mobos arent supporting these ultra low voltages.

Just cause a CPU is mobile doesn't mean it must be able to run at a lower volt than a desktop one. I don't know how the mobile power scheme works but it probably that it uses less amps than desktops so the volt x amp = watt is actually lower than a desktop. The thing is, why would those mobiles run so close to spec and desktop CPU's be off by around .25? Thats a huge difference from the spec, very large.

Oh and the reason the can all run at the same voltage but be faster as you move up the series is because those chips are higher picks. It is just like forcing your voltage to say, 1.3, and then overclocking until you crash. A 6600 will run at 2.4ghz with 1.35 volts and be stable, thats how 'good' the chip is; a 6700 will run at 2.66ghz with 1.35 and be stable, thats how 'good' that chip is.

 

[crow_smiling]

I am not getting "excited".

Sorry, I was being flippant.

Just cause a CPU is mobile doesn't mean it must be able to run at a lower volt than a desktop one. I don't know how the mobile power scheme works but it probably that it uses less amps than desktops so the volt x amp = watt is actually lower than a desktop.

Good point and worth investigating if I can face another Intel PDF. They publish so much information that you could reverse engineer all their stuff.

The thing is, why would those mobiles run so close to spec and desktop CPU's be off by around .25? Thats a huge difference from the spec, very large.

I find it bizarre and it makes me treat their data with caution.
My T2300E is rated at 1.1625V but I can manage 0.95V; difference = 0.2125V.
My E6400 is rated at 0.85V but I can manage 1.1136V; difference = 0.2636V.
So there’s not that much difference between the figures, except for the fact that one is below their threshold and the other is above, which makes it doubly confusing.

 

[Black_Knight_MC]

I just want to find out what is going on. Going from an 805 to this is still much better but if I go by ASUS my idle temps are the same as the 805, 44-46C. CoreTemp and intels TAT both say per core 39-40C, still better but nothing like toms 25C. Their article goes along with Intel's specs which is why I wish I could ask TH directly and see if they know the answer. I doubt they got some super ES. I think that mobos are more set for voltages for most the intel family, not this new core 2 duo that can drop several multipliers and sub 1 volt.

 

[crow_smiling]

You’re idle temps do seem too high, but I think that’s more to do with your cooling. I’m assuming that you’ve tried re-sitting the heatsink?
If I still had a Gigabyte DS3 I would set the VCore to 0.85 in the BIOIS and see how that affects idle temps. I’d be curious to see if it was stable at 1.6GHz at that voltage also.

 

[crow_smiling]

save me tom cruise!!

Why, what do you want him for?

 

[crow_smiling]

he is a scientology guy
there is a movie line that says the same thing
save me tom cruise!!

Oh, I thought that now that he’s been dropped by his studio that you wanted to offer him a job as your pool boy or something.

 

[Black_Knight_MC]

I just want to find out what is going on. Going from an 805 to this is still much better but if I go by ASUS my idle temps are the same as the 805, 44-46C. CoreTemp and intels TAT both say per core 39-40C, still better but nothing like toms 25C. Their article goes along with Intel's specs which is why I wish I could ask TH directly and see if they know the answer. I doubt they got some super ES. I think that mobos are more set for voltages for most the intel family, not this new core 2 duo that can drop several multipliers and sub 1 volt.

i think what you have to understand is everyboard vendor makes their own version of an intel board. the intel boards are intel engineered and are made to the intel spec. the other vendors may decide to cut corners to save a few bucks. if you want to see proof of the numbers intel puts out you would have to use the boards they used to get the data that was collected

Even the Deluxe models were used, I don't think it is a hardware limitation unless it is impossible because Intel hasn't been fully honest. I think, however, it is simply the motherboard BIOS not allowing so little voltage to goto the CPU.

Oh and to crow_smiling, I WOULD set my Vcore to .85 in the BIOS but my WHOLE argument here is that my mobo and others LOWEST settings to select is 1.20 or 1.22. Obviously I ran lower than that with speedstep (CPUZ says 1.120 and RMclock says 1.162). The thing is, the BIOS needs to allow even lower voltages for speedstep to get to what the CPU is rated at.

 

[Black_Knight_MC]

I understand how that works but this would seem like a feature that it is removed from too many models (ALL as far as everyone knows). Why make a mobo only support half the supported voltages when thats what the new conroes are all about?

 

[Black_Knight_MC]

Thats all I have been asking, is for people who DO run at that voltage. So I need to hope an Intel mobo user sees this and has a C2D and lets us know.

 

[pausert20]

A little history on Intel Board design.

Intel comes up with a new chipset. They build what is called a Reference Validation Platform (RVP). This gets tested in their labs. They can have many flavors of the RVP’s. One is called an Electrical Validation Board (EV). Another is called a Customer Reference Board (CRB). The CRB is designed to whatever the current industry specifications are currently at. These specifications change incrementally over time as the manufacturing process matures.

The CRB design is given to any and all motherboard vendors to use as is or to modify as they want. Some manufactures stay close to the CRB design and some don’t even bother with it. Intel does this to enable all of the other motherboard vendors because they want those motherboard vendors that are Intel competitors to be able to successfully support each and every new processor that Intel releases. Intel bases their own product motherboards off of the CRB design from the RVP design team.

As for what your Core Voltage is at. Just probe your CPU Vreg on the output from the output with a digital multimeter and remove any doubt from your mind about what the software tools are showing you. You should be able to probe either one of the downstream capacitors or inductor and get an accurate reading.

 

[Black_Knight_MC]

I am glad you do. This is what I have been trying to bring up but seems alot of people want to put it out of mind or just cast it off as normal. Either we have some AMD fan boys or Intel fan boy haters that think I am trying to make the CPU seem even better. I am just stating the facts of what has been said it should do, and how Tomshardware got it to run like Intel claims. I would think this would be a worthy topic. Why have that knock out article if none of us that get the C2D's will actually run that low and save that kind of power.

BTW, someone asked about my heatsink and so on. I have cleaned, then smudged on then wipe off, then apply small amount of AS5 THREE times and each yielded the same temps. I have a themaltake case, 2 120mm fans, one in front sucking in at 2,000 RPM and another blowing out (is 2 inchs from freeze 7 pro heatsink exit) at 1400 RPM and my PSU dual fan at 1300+ (increases as temp increases). My ambient air is fine.

 

[Black_Knight_MC]

Well I don't own one and I wouldn't know how to do so without some tutorial. But all that explaination seems odd. All of the other CPUs intel has made run at what the website specs and what the box specs are, they run much more accurately than .25V off from spec, more like less than .1V. But now, with C2D they end up with vendors half ass supporting their CPU voltages, especially a feature that is suppose to make it a better choice than AMD? That is REALLY hard to swallow.

 

[Action_Man]

I have speedstep enabled but I only idle 38-42C. So I ran CPUZ and yes, it is at 6x but the voltage is still up at 1.120. In the comparison of the new C2D's they go down to .85ish (.888 from the article). Is this a mobo limitation? Thanks.

They are producing new steppings for errata fixes and improvements so maybe you got an older one.

Wow you just keep pumping out the stupid posts.

 

[Black_Knight_MC]

Sorry, I don't need critical thinking 101 or shopping 101. We have already shown that the mobo tomshardware was the ASUS P5W DH Deluxe, not even an intel mobo. And this has never been something people have to look for when buying a mobo. They WOULD have to look if their mobo supported the CPU they intended to use (CPU support lists) but they have never had to look to see if the mobo is half ass supporting the CPU's voltages, hell, the information is not even available on mobo sites.

 

[djkrypplephite]

I have speedstep enabled but I only idle 38-42C. So I ran CPUZ and yes, it is at 6x but the voltage is still up at 1.120. In the comparison of the new C2D's they go down to .85ish (.888 from the article). Is this a mobo limitation? Thanks.

They are producing new steppings for errata fixes and improvements so maybe you got an older one.

God I love the new BaronMatrix. You're actually saying things instead of just being like OMG YOU SHOULD HAVE GOTTEN AMD FOOL!!!

But anyway, I would suspect Baron's right. Hell, maybe it's actually going down but a bug in the chip says it's really not. Spooky.

 

[Black_Knight_MC]

It is not all about performance, it is also about power consumption and performance/watt ratios. The knock out review cannot be considered if none of us can actually idle like that and get the eletricity and heat savings.

 

[Black_Knight_MC]

I have already shown in this thread the mobo TH used, page 11 i believe in the article. And are you asking for evidence that proves people have not had to look at mobos to see if they support the voltages the CPU can run at? The fact that there is a lack of complaints or previous issues of motherboards providing incomplete voltage support ranges for CPUs is the proof.

 

[Black_Knight_MC]

Please read the thread, I am not going to re link all of the pages and pictures that the article showed howw their CPU was running at .888V when idle and temp was 25C when idle.

 

[Black_Knight_MC]

That is ridiculous that such a feature would be so specific as to force buyers to email or contact mobo makers to make sure it can do what the CPU can do. Such a specification should be listed on the specifications of the motherboard if it can vary even on the same model. From what you are saying, you claim that some P5WDH Deluxe could do it while some other P5WDH Deluxe could not. There is no place on the sites that lists the type of voltages that the mobo supports for the Core 2 Duo's EIST.

How about you offer proof that intel boards allow the CPU to drop to .85. Thats what I have been LOOKING for, is for an intel mobo user to post about that. I have PMed 2 people so far that I saw in their sigs own one and one of them could not post yet. It seems to me you are not reading the thread, it is still only 3 pages long, I have already gone over what the goal is; to find out if Intel is not honest about the specs, if mobo makers are not supporting it physically, if mobo makers are not supporting it in the BIOS yet, or any other explaination. What has been said so far has NOT ruled out these possibilities.

 

[crow_smiling]

yes i explained this already they make different sku's and hope you are ignorant.
you should really try just using the intel boards they are more expensive but they are engineered correctly

A colleague of mine recently bought an Intel P965 board as it was inexpensive and seemed to match his needs; he didn’t need over-clocking and it had Firewire so he was happy. It’s the DP965LT and you can get one for ~£75 in the UK.

It turns out that the BIOS doesn’t support the SM Bus so it seems to have no hardware monitoring at all! I know this because I asked him to run RMClock on the board to see how low he could set the VCore and he informed me of the situation. There are details of his contact with Intel tech support half way down this THREAD.

Not that I’m expecting it to run at 0.85V. Do you really think there is enough tolerance in the spec to allow such a wide range of voltages for one CPU model running at the same speed (1.6GHz)? For different models running at different speeds, possibly, but not for the scenario being described here.
If this were true, what we are seeing is the majority of C2D’s being significantly over-volted at idle. That’s not realistic.

I think it’s possibly a documentation error, unless the 0.85V figure is meant to cover the 45nm replacements for the current chips. They will hopefully also idle at less than 1.6GHz, which seems unnecessarily high. A 45nm C3D idling at 1GHz might handle 0.85V!
I’ve contacted Intel tech support about this issue and shall post back here when I get a reply.