LIVE!!! Calling all XM chip, 3930K/3960X and ultrabook owners! Heat-Spring TIMs for testing!

It only uses temperature to make sure it does not exceed TJmax. Turbo is based on the short and long duration power limits, and specified TDP's and a few other variables like that but temperature doesnt affect it unless it hits throttling temp. You can observe this pretty well with a system that lets you modify the values and you can extend the long and short durations to be very long, raise the max tdp, and get a cpu to hold indefinitely at full turbo even when it is just barely under TJmax (ie you are 95c+).

 

And what about 1-st gen Cpu-s? I saw such settings in BIOS in laptops with Sandy Bridge but in Clarksfield there are only Core Power limit and CPU power limit.

 

they all use this. Even Core 2s

 

HopelesslyFaithful, do you want me to test with my 2012 Macbook Air? I have bootcamp.

 

Come on already. I want to see the results of this TIM

 

sorry running behind. i am in class and got my grandmas funeral tuesday and homework and a NEW JOB!!! woot. working on this as fast as i can. Does anyone want to figure out google docs? I'll update the HWinfo tutorial later tonight

 

What do you need to figure out about google docs? I can help. But graphs prolly wouldn't be possible there. We can only create separate spreadsheets in one Excel googledoc, one for each tester's log or something.

 

so graphs are not possible in good docs?

Than can we create a separate spread sheet for each person to log there data once there done. Also cna the spread sheet be locked to only the one person? Than I or extide can copy paste to excel right? What form or how do they layout the data for it to be an easy copy paste? I haven't had time to fiddle with excel yet and its been a while since i have messed to graphs.

 

I think it should perform something like indigo xtreme, which according to their faq is made out of indium and bismuth. Here's a review of it. Unfortunately they don't cut pieces to mobile processor sizes though.

 

still has 4x more thermal conductivity but this will be interesting. i think that is made though for only desktops

 

Never used google docs, but graphs seem possible:
Create a chart or graph - Google Docs Help

 

I stand corrected.

HF, with regards to data, we could make a uniform list of which parameters to monitor. I.e., temps, clocks, voltages and such. Save it in CSV format, and then we can copy and paste it. For the graph, however, I'm not sure how much values you'd want to save to convert into a graph--I tried to graph an overnight idle log and it just looked like a big inkblot. )

Right now here's what I've been monitoring:

Core #0 VID
Core #1 VID
Core #2 VID
Core #3 VID

Core #0 Clock
Core #1 Clock
Core #2 Clock
Core #3 Clock

Total CPU Usage


Core #0 Temps
Core #1 Temps
Core #2 Temps
Core #3 Temps
Core Max Temps

That's about it.

 

we want all core temps, GPU temp, MOBO temp and i think that is it. We don't need hardrives. Oh and we need freqs recorded to monitor throttling


which is better?

USB Powered Thermometer + Hygrometer (-40-120'C) - Free Shipping - DealExtreme

USB Powered Thermometer + Hygrometer with Probe (-40~150'C) - Free Shipping - DealExtreme

 

have you made any choices on whos going to be do ing the testing?

 

After I finsih added the how to's and the updated instructions on what we all agree on is the standard I'll ask again who wants to try and pick everyone out. If you meet the criteria and agree to do all the testing that is needed than you'll get one. I got plenty of samples since Jim sent me 5 extra. Now if you don't have the XM chips but have an interesting situation and I warrant it a good candidate than I'll add you. Some guy's on here have bad cooling systems like the HP guy or a couple weird netbooks that run how. Those i want to try but only after i get XM chips and I feel I can trust that person. Also if your international you got to pay for shipping because I am not doing to pay for that out of my pocket. It is +-$20 bucks for international....add 5 of those and i am dropping 100 bucks lol on 5 people...than another 40-50 to still ship lol. I got enough to test 2-4 desktops and like 40-60 laptops i think. Depends I still need the 29xxXM die size


Also we need fan speeds as a standard sorry forgot to include that one in the above post.

 

I don't think you'll be needing a probe or anything. I guess it's just your choice.

Also, I don't think all laptops can monitor fan speed. Mine can't. But I think motherboard temps are checkable in HWInfo under PCH Temperature. So you don't need VID's anymore? Just CPU Temps, GPU Temps, CPU Frequency?

 

i7-2nd gen rPGA988 specs:
download.intel.com/design/intarch/designgd/325676.pdf#page=17

i7-3rd gen rPGA988 specs:
http://www.intel.com/content/dam/www/public/us/en/documents/datasheets/3rd-gen-core-family-mobile-vol-1-datasheet.pdf#page=160

Maybe keep a few mm margin. i7-2960XM vs Ivy Bridge die size:
[Video Added + Watercooling Results] Replacing the internal IHS TIM of an i7 3770K

e.g. cut TIM pieces at sizes of at least:
i7-2nd gen rPGA988: 12 x 24mm
i7-3rd gen rPGA988: 10 x 22mm

Maybe package power would also be a good idea.

 

mine has 4 screw right at each corner of the core.

 

Yeah, I got older since I already figured that
I have already wrote why I missed them

 

updating a few things before i go to bed. I'll ask you guys probably saturday for you to respond once more to terms and than i'll get address/money from international guys and start in 2 weeks after you post 2 weeks of current TIM. by time you get Heat-Spring you will end up with 3-4 weeks of data on current TIM and you can switch them out. I'll include my cell phone in the package I mail to you guys so if needed you can contact me directly. I wont leave you hanging.

for international guys you can use freecall.com and I think you get like 300 mins free a month i think. crappy client and meh service but supercheap

@James commenting on your sig. You get a sophomore in that situation. AKA a wise fool.

lol can't rep you arg sorry have to spread the wealth. remind me in a few days to see if i can rep you again for those finds on die sizes!

EDIT:
For die sizes I am seeing 8.2mmx19.7mm for the 39xxXM

looking at C1 and C2 on page 160

the webpage where this guy stated the sizes seems a little closer from what i saw

Sandy Bridge I7-2960xm = 10.54mm by 21.88mm = 230.61mm squared, Ivy Bridge Mobile Quad Core: 8.52×20.26= 172.6152


EDIT: Am I missing anything? I thought I added all the stuff you guy's posted but not sure. I still need to finish terms and instructions but all your guy's statements and info should be added. Let me know.

 

What do you mean fool? You need to read "Basic information" under "Current Notebook" to know about main laptop. It's VPC-F126FM with upgraded CPU to 840QM and RAM to 8GB (haven't added changes in personal information yet).

 

I have an Alienware M15X running a 920XM OC'ed to 27-27-25-25 turbo multis, (3.6GHz 1&2 core operation, 3.3GHz 3&4 core operation) I might be interested in this. I am currently using an experimental Indigo-Xtreme TIM on both CPU and GPU. How does it compare to that? I see that they have a Low Melt Alloy in their graph, but that is usually Coollaboratory's LiquidMetal Pad, which is supposed to be nowhere near as good as Indigo-Xtreme.

 

indium has a higher thermal conductivity but does not metal like indigo extreme so i am very interested in doing a comparision of the too. Can you get recordings of idle temp and start a 4 week test (whenever you are not using computer) of F@H to see how your temps work out. Follow the guide in the OP for logging data. I have to update but leave all data to be recorded. If you turn off monitoring it still records so don't waste your time messing with it just record all and let me know once you are 2 weeks into test with indigo extreme and i'll mail you a smaple. I'll update later tonight with everything finished. So feel free and start gather your system specs and pictures and post them into a single thread and i'll save all your info and email a sample once you send me the intial testing with indigo extreme. How has that worked btw compared to what you used before?

 

eKretz
I have same noteboock and cpu but using Cool Laboratory Liquid Pro.

 

I'll have everyone start tomorrow testing their current TIMs. I'll finish this when i come home from class.

 

I am not going fold on my 3930k for weeks on end during the summer, the electricity is too expensive. You don't need to spend weeks to get conclusive data. With proper testing procedures you can get results 99% as good in a few hours. It's also impractical for me to fold for such long times on my laptop, but I can still get you results.

BTW: During the winter months I fold on the 3930k 24/7, and I also fold on lots of other systems too, so I don't have anything against folding itself Check my stats

 

that was more so to get average temps and stuff...some people don't have consistent AC. I live in a farm house right now with window AC :/ so I tend to fluctuate in temp a lot especially since my is heavy air cooled.

Anywho 2-4 weeks is what I would personally like. If you absolutely disapprove i would want at least a single week.

 

I am sort of in agreement with the other poster saying that 2-4 weeks is overkill. If the tester's home/computer room is air conditioned, the extra time won't tell you much of anything, as long as the machine has time to heat soak while under load. If no variables change, the temps should be in the same range every time. Important variables will include ambient temperature for sure though. My home is kept at 75°F 24/7 for instance. Another variable that will perhaps have an effect is humidity, but there isn't much we can do about that one, not being in a laboratory setting. I would be alright with giving you a week of spare time logged, if that is ok with you.

My experiment with the Indigo Xtreme has been pretty successful so far. My idle temps dropped about 5°-10°C and my load temps I never really bothered to test before because I had the stock gunk on the machine still. My idles range from the low to mid 40's usually with Indigo, and load maxed out is right around high 80's-low 90's with all 8 threads @ 100%. This is with stock fan settings.

 

eh was going to explain more for my reasoning for 2-4 weeks but i am lazy....updating thread

 

I'm interested, go ahead...I will probably dial back the OC a notch or two in the interests of rock solid stability/results before starting to log, then even if the new stuff doesn't perform as well, I won't get shutdowns and we can directly compare temps.

 

I normally don't like to break 80C high for me is 85C for F@H. I have TS set to throttle hardcore when ~87C is hit. It is annoying because it will jump back and forth between the two values. I wish it would be smarter or you could set up better setting so when that threshold is hit it will drastically drop and reset at a low value from the first until you come and reset it. I would love that. Like 26 26 24 22 but if it hits 87 C it turn to 26 26 24 18 until it is safe level of heat and ramps back up to 26 26 24 20 until i come and reset it or leave it. This way it doesn't cycle like a mad man. I wonder if it does from 2 multiplier settings every second and changed 10C every second does it wear out the CPU?

 

Yeah I have faith in the old XM; and the Alienware cooling system. I am running F@H now as a quick torture test/trial @24x and hovering between 87-90C. Not sure if the cycling you are talking about would wear anything out. Also, I read back a ways where you were wondering about how to get good contact...they make pressure sensitive contact paper so you can see if you get even contact, might be a good idea to have people submit a test print so we can see what kind of contact there is on their die. That could majorly skew results.

STIHial, how do you like the Coolaboratory you are using? I was thinking about trying that stuff until I decided to experiment with the Indigo Xtreme. I actually emailed back and forth with a guy from Enerdyne (who make Indigo) and they were kind of not recommending I try it, but I wanted the best. If this stuff doesn't outperform it I will. E switching back after testing. It's kind of an involved process to install it but it works pretty well.

 

yea check out my cooling mod to my laptop it is a monster now lol. So not meant to work with a 65-76W CPU lol. I hope this works well I want to push another 10-20W than try to mod a water cooling block to it or jerry rig something. After this project is done I'll revamp my cooling mod thread because it is a disaster.

Where do you get that paper? Got a link? I might hold off and buy some if it isn't a lot and include it with everyones' shipment

EDIT: BTW everything is updated as far as I can think of. Am I missing anything?

 

So if you don't have an XM chip you can't get it? /:

 

if you got an interesting situation like an ultrabook or netbook that runs hot or the clevo/sager 11 in netbook with 3610.

I got a good chunk of samples but want to keep it to systems that run near or can run near TJmax. so like 80-100C but if you got an interesting situation let me know and i might send you one.

 

It isn't cheap but if you cut it up to pieces sized for the individual dies it's not bad. About $55 for a sheet the size of a piece of standard printer paper (12"x10.6"). Pretty cool stuff - it can show you the evenness of your clamp pressure as well as measure it in PSI. Take note you want the Ultra Low variety to get a range of 28-85PSI. Here's a link:

Fujifilm Prescale by the Foot Ultra Low:Amazon:Everything Else

You may be able to find it cheaper, not sure. Maybe you could ask if Indium can supply some, I'd bet that they have probably used it in their testing.

 

I have a 2450M that throttles on 90c and above, so from 2.9Ghz to around 2.6-2.7GHz. So to keep it at 2.9GHz I have to keep it at 89c and below

I read on ark.intel that the TJmax is 100c, but it throttles so early /:

 

Möbius are you using Throttlestop? If you tick the boxes for modulation and have power settings correct it shouldn't throttle (hence the name - Throttlestop).

 

92c and it's not 2.9Ghz

 

it could be his TJmax is different. I thought i remember all chips have different TJmaxs and the reading in C is not necessarily right. I set my limit to 15 DTS and it throttles at 79-84C one core is 79 other is 80 one is 83 and other is 84.

At least i thought that was what uncle webb was saying a long time ago that each core has a different tjmax that is based off of a limit of 100 but does not literally mean 100C. Maybe someone else here knows about this and can explain it better or tell me i am flat out mistake ^-^

mobius PM the info I ask in the first thread and i'll let you know

Also can anyone verify the info on the 29xxXM and 39xxXM die size the info posted earlier has conflicting numbers. Don't want to send to big or small of piece

 

anyone wants to be awesome and buy this for me and mail it to my house one sheet would be more than enough! PM me if you are a totally awesome guy!...or gal and i'll give you my address to ship it too!

http://www.amazon.com/Fujifilm-Pres...words=Fujifilm+Prescale+by+the+Foot+Ultra+Low

DAMN!!! stupid win7 and its force restarts!!!! I miss XP and my 35 days of no reboots!!!....only time i rebooted was when it crashed due to my 25% overclock on CPU lol, which was 1-4 weeks on average lol

 

PM sent

 

The C0% is at 100%, doesn't that mean his chip is not throttling? Mobius your multiplier was set at 29 in that test?

And Faithful, seriously, contact Indium first...they may well have some just lying around, and I'm sure they'd be willing to help out in the interests of having an accurate portrayal of their product. I would be happy to buy it, but I am as broke as a joke right now, and living off of handouts. (I have bad back issues, been through several surgeries).

 

sorry to hear man. maybe someone else can. I'll ask Jim tomorrow

EDIT: i could buy it...just got like 2 grand of stuff to buy for this new job lol

EDIT: btw mobius have you applied ICD before? make sure you use a straight line Also crank down with pressure...obviously no retard strength

 

OK looks like I'm in to give this stuff a try. Happy to give you 1 week of pre heat-spring and 1 week of post heat-spring logging results. Hoping this stuff will live up to the hype, but I'm pretty happy with my Indigo Xtreme...we'll see soon.

Machine Specs:

Alienware M15X, 920XM OC'ed to 26,26,24,24 for purposes of this test and stability in longer term folding. Currently idling at low to mid 40'sC and at 100% load while folding she is holding steady in the high 80'sC to low 90'sC. Currently using Indigo Xtreme TIM applied in an involved process from a desktop kit I bought for the express purpose of experimentation.


Also, Hopelessly Faithful, do you know which alloy we have to test? The website lists several:

Indalloy #4 - 99.99% Indium
Indalloy #1E - 52%In/48%Sn
Indalloy #3 - 90%In/10%Ag
Indalloy #290 - 97%In/3%Ag

 

Multiplier set to turbo (29), it throttles to 28 if it reaches 90-91 and so on

 

Well, maybe your temp sensor is off after all. When mine throttled it was Dell's throttling kicking in, and it cut the duty cycle down, so my C0% went to like 50%, but my multiplier held. However Throttlestop took care of that. If you are hitting Intel's throttling maybe they do it differently. I've never been able to hit that one, I get bsod before I ever get there. Highest I've ever been able to hold consistently was ~96C. Pushing any harder (higher multiplier) than that gets me an instant bsod.

 

The Core i5-2450M has a default multiplier of 25. Any multiplier higher than 25 depends on Turbo Boost which can be limited by power consumption, the amount of current flowing through the CPU as well as the core temperature. The 2nd and 3rd Generation CPUs also have a time limit where the amount of Turbo Boost will automatically decrease after a fixed amount of time. Intel recommends and most manufacturers have this limit set to 28 seconds. The ThrottleStop TPL window should show this value.

TJ Max for your CPU is 100C but Intel probably starts to reduce the amount of Turbo Boost available at 90C. If you keep running at full load, the multiplier will continue to decrease so the core temperature does not exceed 90C. The multiplier should not go any lower than 25 until the CPU hits 100C. That's the official thermal throttling point. I think what you are seeing is just how Intel designed Turbo Boost to work. This can not be over ridden with any software because Intel built this safe guard in at the hardware level.

Manufacturers can also choose to set a lower thermal throttling temperature at the bios level but I don't think that is the case here. When that happens, your CPU multiplier will drop like a rock, well under 25 and will sit there for blocks of time like 5 or 10 seconds or far worse. Go find a meat locker so you can do some more testing.

 

It is Indalloy #4 - 99.99% Indium hence why it had 87w m/k because that is the thermal conductivity of indium

Working on my netbook sample

 

I agree to all terms and conditions as stated in the original post to complete all testing data and submit to HopelesslyFaithful immediately after all data has been recorded.