I7-3720QM(ES) with unlocked multiplier and 8MB L3. Worth getting

I ordered my 3 pipe from that seller, it arrived today in perfect condition with the TIM on it as new.
I can also confirm that the machine came with a 2 pipe heatsink, not a 1 or a 3. I know some folks were saying they've never seen one that didn't come with a triple pipe sink but they're out there.

 

Spoiler

Spoiler

 

The HS is arrived

I´ll post my impressions later...

 

I have ordered 1 too. In fact my current temperature with 3610QM and 8 month old stock thermal grease makes my CPU hit over 100C on Prime 95 but I hope that with experimental heat spring (should have come already), copper plates taken from broken AC adapter and notebook cooler I will be able to hit at least 3.9Ghz on all 4 cores. without melting the laptop. My AC adapter is 120W though, have no idea if gives enough power for that.

BTW people, should I use usual thermal grease to glue those copper plates to heat pipes? because I will not buy thermal adhesive for sure... at least in closest future.

P.S. I bought from ebay Member shirleyfu1117 and because of the stupidest thing of PayPal EVER (which is automatic security card blocking) I could not pay for the item which I have already committed to buy! That stupid block could not be taken away by paypal's reps and they refused even tell me the reason of blocking! They offered me to try later, change browser, try different card (put another over 200$ on different card without being sure that it gonna help) and used stupid copypasting emailing with me without even proper reading my letters! Anyway, that saved me couple bucks actually (at least I hope so). Seller agreed to cancel my order due to paypal's fault and sent me invoice with lower price because of reduced eBay fees!
Don't get me wrong, I do not recommend to make deals outside of eBay, I am just saying that I got almost 15% lower price in the invoice. Anyway he also offers free DHL shipping and perhaps if you decide to ask for lower cost shipping instead of DHL you may try to ask for some little discount too Just don't be too impudent.

 

is there even a thermal glue TIM? i would just use the thermal tape. That is what i used on my G51J

 

@HopelesslyFaithful, yeah it does. I found this no so long time ago from either tijo or another mod here. Here is the first google link on AS Arctic Silver Thermal Adhesive

 

see what the ratings are and use the better one. Just make sure it will hold.

EDIT
Interesting it is in my bookmarks but i dont remember it.

 

Mines arrived and it's installed seems fine, still waiting for my heatsink though.

Upgrademonkey have changed the listing now so it doesn't include "unlocked multiplier" in the title. I asked them why and the replay was - "we do not encourage buyer to overclock ,so removed it"

 

BTW i don't know if that is the one but i remember someone said their was a adhesive TIM but if you use it, it would be permanent. The tap you cna remove freely. Thought that was worth noting so keep that in mind if it is the AS one that is permanent.

 

@thedinks, could you tell me what is the max turbofrequency of iGPU? I am just interested to know because it is another difference between OEMs 3920XM and 3720QM.
@Hopelesslyfaithful I a not using adhesive right now, I will look if your TIM + Cooler will help or I will have to do a copper job. Although thank about tape. Will look closer.

P.S. About your bookmark... that was real funny

 

Sure tell me how

I'm a James D aswell

 

Okay - the first thing I want to say is thx to Fox, John and especially Unclewebb 4 inspiration and information

The second thing: The HS rocks the house - 4 example 3.9ghz vid=0 in 3dk13 fire strike with the stock-cooler 100°C and with the 6-pipe HS 68°C - I think there´s nothing more to say - a really fantastic, outstanding performance.

The CPU is [email protected] and performs very well. max. temperatures 85°C now i´ve go to bed... good n8

a benchmark session: http://forum.notebookreview.com/ali...nware-m18x-benchmark-thread-part-2-a-373.html

 

A lot of that was from a bad TIM job. No way it dropped that much but congrats

 

What are you basing that on? A poor job of pasting can certainly cause problems, but the best paste job on the planet won't help anywhere nearly enough if the heat sink is too small. Seriously, there is a gigantic difference between the wimpy little 1-pipe QM heat sink and 3-pipe XM unit. So, I can easily see it making that much difference all by itself. Now, the difference between the M18x R1 2-pipe XM and the M18x R2 3-pipe Ivy Bridge XM would not be anywhere near that much difference.

 

What about the 2 pipe? I don't think I've ever even seen anyone mention that before but I had one in my M18.

 

Before the M18x R2 was released, the Sandy Bridge XM heat sink was 2-pipe. I've never owned one of them. I went from a 1-pipe QM heat sink to a prototype 3-pipe XM heat sink.

 

So then the 2 pipe that came in my R2 was not the original heatsink? I'm curious as the warranty still hasn't transferred (was listed with a very nice extended warranty) and when I run the service tag it comes up as having completely different parts.
I trust the seller I bought it from but I'm seriously starting to have some doubts given all of the quirks I'm finding here.

 

sorry i just don't see how a 40 C difference will happen between the 1 pipe to the 3 pipe heatsink. 100C vs 68* C plus at 100C be is already throttling and he has another 400mhz* more freq on top of that. That is at least 40+ C when you count clock difference and throttling into the picture. I find it hard to believe that 40+C difference will happen with a heatsink. I had a 20C drop from stock TIM to ICD and i doubt ICD really magically made it drop 20C....a bad paste job probably added a good 10 or so C to that. No way 40+C drop is just from heatsink. Anyways until you show me someone can replicate that i am calling hogwash.

* denotes correction in numbers

EDIT: i would believe in a 20C or even a 30 C drop but 40C give or take...probably give....no way

 

I don't know Slim... Maybe some M18x R2 have a 2-pipe heat sink. I have no way of knowing that, since I have never purchased an M18x R2 with a non-XM CPU. Maybe someone with a 3720QM or 3820QM can chime in and tell us if their system has a 1-pipe or 2-pipe. I know that a number of folks in our community have upgraded to the 3-pipe heat sink with QM processors in their M18x R2. I think Gerd Weller mentioned he had the 1-pipe heat sink.

@HopelesslyFaithful - believe it... I guess I am puzzled... if you saw a 20°C drop going from stock TIM to ICD the difference in heat sinks should be much easier to believe. In case you where not paying attention to the photo, there is about twice as much copper between those to heat sinks. Gerd posted his results... good enough for me. Nobody needs to prove anything here. It's a take or leave deal on face value here. Screen shots are always nice, but we don't have that. But that's OK. If you choose to continue in doubt, that's perfectly fine, too... doesn't matter a bit either way. "Bad paste job" is kind of a poor choice of words. I have had excellent results from the stock Dell TIM with extreme CPU and GPU overclocking. A really excellent paste job can get messed up by something as simple as rough handling in shipping or an accidental bump or drop with a minor jolt to the chassis. It takes no more than a little flex or movement between the die and heat sink to ruin the bond and turn a cool running machine into an overheating mess, especially while overclocked. Also, FYI, if you use the right settings, 100°C does not automatically result in a throttle. The 3920XM shuts down at 105°C and the way I have my settings it runs full speed on all cores until the system turns off. If you leave the CPU thermal settings alone, it will throttle... but it's only an assumption if you don't know what settings Gerd is using.

 

Thanks Mr. Fox. As I said, I'm starting to get a little concerned as it was listed and sold with a 3 year Complete Care warranty (transferrable) but it's been roughly 4 weeks and the warranty still hasn't transferred. When I contacted Dell I was told there was no record of a transfer being requested and the service tag indicates the computer included a 3840 and 2 7970's but when I bought it it had a 3720 and 2 680's in it. The seller explained that it was an error on the part of Dell when it was shipped to him. I have no reason to doubt the seller as I've bought from him in the past but..............in this case someone somewhere is in error and the end result is that I paid extra for something I'm not receiving. I'm trying to be patient and give Dell and the seller time to push things through but my patience is wearing thin.

 

It would not be the first time I have seen a system configuration not match up with the service tag information. I have heard of several people receiving better hardware than what they paid for or what their service tag reflects they should have. The system you purchased might be an example of a CPU downgrade and GPU upgrade, LOL. But, that was a really good mistake. I'd gladly take the 3720QM/680M SLI over 3820QM/7970M CF any day of the week.

 

Without a doubt you are correct, my concern is the warranty. I really want that transferred and I'm hitting a brickwall getting it done.

 

Have you asked Luis Porras or Mark Obee for assistance? Sometimes the warranty transfers take longer than I think they ought to. I have seen it take well over a month even for a system exchange. Luis or Mark might be able to assist with that, or talk to the person/department that are being unresponsive to your inquiry.

 

Cool, it got here. Nice.

Very good job. See, told ya so.

Not really.

You would have to go back and read where he had his settings wrong and Brother Fox and I were telling him so. Adding 15 flex to a 3.9 ghz over clock was like 40 percent of the problem already. The heatsink another 40 percent. Room temps 18 percent. And Tim about 2 percent of the problem. And it was 32C difference and not 40C. (68 from 100 is 32) So i guess you can be a believer now.

 

i was talking about actual difference if you consider the clock difference and throttling from 100C so i am more than right on there would be a good 40 C difference if you counted the throttling and added freq. Also i did not know he added vid to it so that is a different issue, which again proves that a 1 pipe to 3 pipe still does not give you a 40C or even remotely close to that drop in temps so again thanks for proving me right that the heatsink was not the cause of a drop that big. Again my conclusion was right give in the information i knew from the top of my head. If i knew vid was set higher in that test compared to older test i would not have come to that conclusion

 

No, but I spoke with someone at Dell and they told me no transfer request has ever been initiated. I am not sure I buy that though as the seller assures me he did so and I have no reason to doubt him.

 

i would tell the seller it is his problem to fix this and he better have it fixed in the next week. This is honestly not your problem but the sellers problem. There are courses of action you can take if the seller gives you the cold shoulder if you demand him to fix this problem.

 

Hey, I have a question... What system are you using? M18x R2? Do you have this ES CPU or something else? I looked at your profile and read a some of your posts in other threads and could not figure out what system you are running. If you are running a system other than an M18x R2, that might have some influence on your point of view and that is the reason I am asking. If you have some other system, you're not going to get the same degree of control and flexibility from the CPU as you would with an M18x because other systems tend to be crippled by their manufacturers. If that turned out to be the case then some of the dialogue that took place in this thread might be skewed based on limitations that your system has that M18x owners do not encounter.

As I read my way through this thread it seemed like you tended to always be right, offered rebuttals and disputed facts others attempted to share even though they were getting better performance results from their CPUs than what it appears that you are achieving. As an example, even though it was pointed out that the occurrence of throttling was assumed on your part, despite the fact that it can be overcome with the appropriate BIOS settings, you mentioned throttling once again as if it were some sort of an established fact. In your rebuttal to John, you attempted to point out why your statements and disbelief about Gerd's results with the vastly superior heat sink were still accurate even though your math was not. All the minutia was not important. What really mattered is that Gerd went from 100°C to 68°C at the same settings. To what extent the TIM played a part in that seems like a distraction. The main point is that the single pipe heat sink is unsuitable for significant overclocking. The best TIM in the world is not going to fix that problem. Had the temps gone from 100°C to 78°C, or even 88°C, it would have still been an excellent end result and equally meaningful because the CPU does not need to max out at 68°C to achieve exactly the same performance outcome. Effective thermal management is necessary with any processor, and TIM is only one of many equally important elements that have a tremendous effect on temperatures and performance.

I also saw where you posted that you are using the newest version of XTU and found some settings were not working, which could be part of the dilemma. If you have an M18x R2, you need to use XTU 3.1.201.5 (which is not the newest version) to get the best result. An unlocked BIOS with access to disable the MSR lock would also help if you don't already have that. If you have an M18x R2 with this mystery early prototype CPU that software incorrectly identifies as a 3720QM ES CPU, you should be able to run 4.3GHz on 4 cores without throttling. You should be able to adjust the Long and Short Duration PWR Limits, turbo time limit, voltage (flex) and watts (pri plane) with XTU 3.1.201.5 and not encounter that 28 second turbo limit that you mentioned numerous times.

 

It was already quoted so need to go "from your head" He flat out said so. Not sure how this part is missed.

And let me re add it.
If he didnt change any voltages. it would be 80 percent heat sink 18 percent room temp and 2 percent tim.

I have to seriously ask...Is this your first time doing this sort of thing? I can do better temps on silicone heat transfer compound that your icdiamond. It's more about how the heat sink sits than you spending 20 minutes trying to apply paste.

 

I suspect the seller is telling you the truth and the request probably got lost or has not been processed correctly. I may still be in process waiting for someone to take action and the person that said no transfer had been initiated could be mistaken. I would reach out to Luis or Mark with the service tag info and provide all of the seller's contact info along with your own. They may need to have the seller initiate the contact and confirm the transaction again. If that is done through the forum PM system or private email with everyone (seller, you and Alienware) copied on the written communication, it might help move things along. I will also give you a traceable form of documentation about the request being initiated.

 

I have an R4. I currently need to reinstall windows but i am waiting to get an SSD so i don't feel like reinstalling windows twice. I was the one that got all sorts of magical issues with my R4 and XTU. I was getting 47 43 43 39 with my 3720QM OEM.

Wow you are the worse listener ever.

100c vs 68 c equals 32 C difference but if you include that the 68c had also a 400MHz increase (which i see i am mistaken now. I thought he said he had 4.3 and 68. I see now looking back that those were two separate statements) So still he was throttling at 100C which means he wasn't even getting full performance so those numbers were already skewed. So let me break it down this way since you can't seem to follow

100c
3.9GHz
throttling
single pipe heatsink



68c
3.9GHz
no throttling
3 pipe heatsink.


That 32C is much greater because if he is hitting 100C back then. So there is no way he is actually getting 3.9GHz so if he could magically go above 100C to run at full speed there would easily be a 40C plus difference. Again i was right with the information i had at the time.

Fox said a 32C (realistically the difference is higher because of throttling as i pointed out above) said that is common and the heat sink makes that much of a difference which i quickly objected to because that is pure bullocks. I was simply saying with the information i had at that time. There was no way the heatsink as according to Mr fox was creating that large of a drop in temps. At the time i did not read or see or remember he had the vid all messed up so my first guess was that for there to be that great of a drop in temps assuming all figures were the same (as in no voltage change) it would have been a bad TIM job, which is plausible with the information i had and i have seen. Temps can change an easy 20-50C with a bad TIM job. When not enough TIM is used and the heatsink is not properly tighten it can greatly affect temps. As i pointed out earlier i dropped an easy 20 C with ICD from ASUS stock TIM that is because ASUS pasted bad and didn't even tighten the heatsink down right. I quickly learned from my experience that ICD is not that effective and that ASUS just didn't tighten it enough.

Also further evidence to show how pressure and contact affect it greatly. (please note this is a bit different due to we are talking about different TIMs but it gets the point across) I know i still haven't finished my bs on Heat-Spring but HS has shown that pressure and contact affect it a great deal. I have had worse temps with HS so far due to contact issues even though it is 20 times more thermally conductive. Good contact is paramount with temps so if the person who first applied the paste didn't tighten the screws down than nearly no contact could be happening and air bubbles would quickly form and remove any contact it actually had.

So again i was right with the information i had and i still stand by saying fox is full of it when he said 32C is normal because even you are saying it isn't because he had vid all botched. So yes i was wrong on the bad TIM job because i didn't realize he throw in there a crap ton of extra voltage. If he didn't throw in all that extra voltage than a bad TIM job could have been the issue. Again when i say bad TIM job i am referring more so to not applying enough pressure, which is part of the process to applying a TIM.

Also how do you take 20 mins to apply paste? I can apply it in 60s? (the squeezing it out part. Screwing it all back together and cleaning is what takes time.)



EDIT: also to refer to the guy in the past that said I was throttling i was only hitting 85-95C on my 3720qm so i don't know why it would be throttling since it wasn't at TJmax. Also the guy before i stopped listening to because he was telling me i needed more turbo voltage....which is moronic due too that would only make my temps worse and cause more throttling (if it had to do with heat, which it shouldn't. Plus getting another 400mhz out of a CPU shouldn't require more voltage. I tried it anywheres and did nothing if not even hurt it). Also as i stated before XTU was glitching and i have no idea why and what was causing it. I was getting again on an OEM 3720qm 47 43 43 39 if i remember correctly, which is obviously wrong. Also we were debating about the short and long turbo because none of them affected my CPU and allowing it to turbo longer than the 28s window. Even though he was able to which makes no sense what the difference was (I was not at TJmax so it shouldn't be a thermal issue). Also he was arguing with me that turbo boost actually works. I have never seen a single threaded program actually use the higher speed. I ahev only has a 920xm and a 3720qm and i have never seen a single thread program every go higher than the last speed minus maybe 100mhz higher. I could set my 920xm to 3.2GHz on 1 and 2 cores and leave 3 and 4 at 2GHz and it'll run at 2ghz for a single threaded benchmark and for a single threaded game, which makes me wonder why even bother having this feature if it never works.

 

Dude, your ridiculous.
With all that back pedaling and not paying attention. ROTF!
Quit while your seriously way behind.

And the paste thing was an exaggerated joke.

Wow, the forum take everything serious police!

So on to something else now, because paying attention ain't one of them. No matter how you try to long type it.

 

All I can say is, wow. It's incredible how little everyone else knows, including folks that are getting world class performance from their systems... we're all full of it. Amazing... and, here all this time I thought results are what matter the most.

 

whatever fox you are just being difficult. I am done trying to reason with you.

 

No, not being difficult, only trying to be helpful and share meaningful information that has already been put to the test. I had to ask a few questions to try to understand where you are coming from with some of those comments that are not making sense to me. But, that didn't work out... still scratching my head at this point. What you are able to achieve with the 3720QM OEM CPU in your R4 along with whatever settings you are using probably is not all that relevant to what works best with the incorrectly identified/mislabeled "3720QM ES" QBC1 CPU, since it seems to have specs more closely related to the 3920XM than the 3720QM.

God gave me two eyes, two ears, and only one mouth. I tend to learn more by observing and listening. We learn from each other here at NBR by having two way discussions. So, I think the idea of not trying to convince everyone (or reason with me) about how right you are on everything is a good place to start. I'm frequently wrong or mistaken about a lot of things, and when find out that I am I just admit it and learn something new. ( Here's an example of that.)

Now, about that 100°C throttle thing you keep referring to as though it is an immutable fact. Throttling is mostly a problem when one does not know what power settings need to be used use for a proper overclock. Talking is not getting us very far, so let's try a round of show and tell. Have a look and you'll see what I'm talking about...

Spoiler

Kids, don't try this at home...

<embed src="http://www.youtube.com/v/1BkSISNjaEc?hl=en_US&version=3&rel=0" type="application/x-shockwave-flash" width='1280' height="720" allowscriptaccess="always" allowfullscreen="true">

 

Ivy throttels from 105°C and not below!

 

OMG, you people wrote over 2 pages arguing with phrase "TIM job did a lot of work"??
Well it certainly did. Nobody told that it did most of the job, right? Neither anyonуone will be able to tell for sure how much degrees did it cut. Here is nothing to argue about.

 

Depends allot on how it comes out. People like to say what's on their mind then when the stuff hits the fan, they call it a joke. Trust me, I know where the line is with 13.6k post.

As to paste. I could use toothpaste and do the same job. Been done on desktops and can be done here.

 

the fact you even referenced this is pathetic. That is the worse test ever with the most flawed thinking i have ever seen. It is like saying copper and iron both melt at 2000C so they must have the same melting point.


Thermal Compound Roundup - July 2011 | Hardware Secrets

 

I think it's time to take a step away from this thread. Temporarily closing for now.

 

HELP!!!! PLEASE!!!!
Hi people. I was kinda busy last week but I got this CPU yesterday and today finally have free time. I decided that even though my cooling is not very good I still will be able to reach 3.4-3.5 on all 4 cores (as I have a notebook cooler after all). So I put this QBC1 and I am not able to overclock it at all! I was going to use ThrottleStop or Intel XTU because even if I had unlocked BIOS it still wouldn't let me to change multipliers but software had to work! I made a video and will upload it to youtube soon and I really, really want a help. Because why would I need it then?

ALSO I found other bugs of this ES CPU and it made my situation way way worse. This CPU is WORSE THAN 3610QM at stock!!! It has 2.5 or 2.6 GHz without turboboost (my BIOS shows only 2500MHz for some reason) but its TURBO is only up to 3.0GHz for 4 cores! For 1 core it is 3.2GHz while my OEM 3610qm brings me 3.3GHz! Heck in real life it shows even less! Less than 3.1GHz for 1 core and I assume lower stepping makes it work even lower than 3.2.
I would be fine if I could overclock it but I can't!

Please, someone? Can someone check default Turbo speed of your 3720QM? Is it only my ES such or everybody's QBC1 is the same and you didn't notice because nobody cares about default clock when they can overclock it to 4GHz on all 4 cores?

Anyway I must say that it is not 3920XM for sure because at stock it is slower than my 3610QM. And also that I am screwed because I don't kow if it is this CPU faulty or just a rare chance that my laptop has locked BIOS from Ocing in any way or smth. 1 Guy said smth about possible Intel ME firmware Lock.

no way. you are wrong. 105C is Tjunction when CPU starts dying. It throttles between 90 and 100 and this value is set in BIOS.

 

When you are unable to set the Multipliers in Bios, you shouldn´t expect that you can do this with xtu. i would contact msi (or a compareable community) and ask them for a bios, which fully supports Intel xtu.

afaik my stock Multi is 27x and with Turbo 32x on all cores - but that doesn´t care with 43x on all cores.

BTW. and sorry 4 ot: I flashed the unlocked A10 successfully

 

@Gerd Weller, can you printscreen the right corner of Intel XTU>Manual Timing where default and Proposed settings are shown? Are you sure that stock/default turbo 3.2Ghz is for 4 cores, not the one?

 

James he is right on 105... Reason I said 95 or 100 was I guess my r4 has some throttling built in bios I can't edit.. blanking on guys name but we were pming each other about it

 

Hmm... well I think every manufacturer put this value in BIOS, not just r4. When I had 102C CPU was throttling and went to lower than 100C, down to 95 I would say or so... I will look at those logs again.

Spoiler

 

Ok - you are in right. With all 4 cores the cpu clocks just only with 30x 100mhz. But that don´t cares - all you´ll Need is a better/unlocked bios!

 

Hi James. Your CPU behavior could be largely dependent on the BIOS settings and thermal configuration. Without access to make adjustments, you might see very different behavior than someone else with the same CPU and a different motherboard and BIOS. Throttling can be caused by temperatures, based on the BIOS settings. It can also have a power throttle, where the CPU slows down due to lack of power. Having pri plane settings (watts) that are too low will induce power throttling in the absence of high temperatures.

 

3720QM ES QBC1 Doesn't want to Overclock!!! - YouTube

 

I looked at your video. What version of XTU are you using? Click the Intel logo in the upper left corner. The grayed out controls suggests that either XTU will not work with your motherboard/chipset or you have the wrong version. If I use the wrong version with my M18x R2 motherboard, it looks just like that. The only version that works correctly is 3.1.201.5. There are older and newer version available, and only this one works.



Can you take photos of the BIOS screens for CPU setting on your MSI and post them? On Alienware you cannot see the multiplier controls unless you change a couple of other settings first. Does the GE70 have InsydeH20 BIOS?

 

interesting mine is 4.1_____ for both