xtremeoverlocking - pushing the 3930K/3960X to 4.5 GHz

[alpha754293]

I just built a 3930K system using the Asus P9X79 WS mobo cooled by the Noctua NH-D14. No problem overclocking it to 4.5Ghz. In fact I run it 4.6ghz 1.34v 24/7 crunching WCG. It is LinX stable and the Noctua keeps it under 70° while crunching.

I used the corsair 800D case which has lots of room.

For a video card I just picked up a 8400GS for $ 15.

Personally i do not like the all in one liquid cooling kits. Expensive, complicated, and fail to beat the best air coolers.


Oh and after reading this entire thread I must say, as a fellow professional in engineering/technology, I am insulted. You come here asking for help and then rip everyone a new one when you feel it threaten's your intelligence. All that does is prove how not smart you are.

Please, get off your high horse and join us back in reality.

It starts off with two very simple questions (actually...one...really) and then people go off on all these stupid tangents about directional airflows and b--ching about counterproposals or that they would mention one component, and I'd mention why I couldn't (or more specifically) didn't want to go with that component that they mentioned. (UD5 - 8 DIMMs; -> lower cost per memory module. UD3 - 4 DIMMs, but preferred; but higher cost per memory module).

Bun-Bun - the question is really simple. It's not that complicated. And while I agree that the direction and the general flow of air is important, as I've mentioned, I've stated why I'm not a big fan of side fans (and actually have the data, both physical and virtual) as to why. Not that I really should or need to justify myself to you guys. But at least I'm working with actual data, not just on gut feel.

They're not a threat. They would be if they could actually argue on a premise than mere "gut feel".

Can I go with an open kit? Yes. But by the same token, I could have gone with a custom waterblock too. *rolls eyes* Or direct die evaporative cooling too! And clearly, if I'm making references to H80/H100, yes, the open kits are an alternative; yes, I don't think that I would be spending the time to actually do what BeepBeep2 suggested in terms of getting the kit, cutting the tubing to size, leak testing it, and then running it. It'll be a fairly quick deployment.

Some of the other points of argument from this thread is quite simply...just dumb. The whole "simulation vs. real world" part - yea. Dumb. If the real world was so important, then why is it that pretty much EVERYBODY is moving to simulations? Why WOULD we run simulations if guys like you think that I'm stuck in the stratosphere somewhere??? To me, that just reeks of people's ignorance about what simulation engineers do (regardless of whether it's mechanics, fluids, or both. Or multiphysics.) Its fine for you guys because you are able to physically pick up a part that's already been tooled, shot and shipped, with your hands, look at it and screw it in the case.

But what do you do when tooling cost half-a-million dollars (actually, the end of line measurement system is a half-a-million dollars), and you DON'T have parts made yet because you haven't crossed the marker for a design freeze for your tooling guys to tool to? Are you able to tell me what the M and M' are??? I mean, I can ALWAYS make the problem more difficult for you guys. MUCH more difficult. But the point of some of those "back-of-the-envelop" type of calculations is to be able to very quickly show why something likely will or won't work. So, if you can get the blade parameters of the fans in question (the parameters that I'll need can be found in just about any sort of turbomachinery handbook), I would then be able to model explicitly the hub and the entire fan performance curve (minus statistical deviations, electrical to mechanical transform inefficiencies; unless I'm fed the line information as well.) As dumb as I am and as intelligent as the rest of this community is - how someone misses that point is beyond me.

This isn't a pissing match to see who's smarter. I came here to ask because I figured that I would be able to get the best, most professional advice and so far, it's been greeted with mostly nothing but disdain for the original question. Are your heads so far up your own arses that you can't see daylight anymore? And one of the major differences is that I know what my limitations are and won't claim to know more than I do. Heck, even if I've done the research, I won't necessarily claim that it's fully complete and accurate either because I haven't actually taken the painstaking task of actually modelling the entire computer in tremendous detail.

Course, I reckon that most of you here probably don't care for that either (since obviously, we're not into facts here). You want to talk about insulting -- try asking someone for help and all they can tell you is "you've seem to have made up your mind already, so good luck" or "what's the point of those calculations? That's not even accurate. You have to take into the account the hub..." (why haven't you mentioned the blade pitch and angle, and pretty much all of the other turbomachinery parameters and characteristics???)

That would be like if I were to ask you how I can make the engine of my car produce more torque, you'd be telling me that I need to increase the clamping force on the forge press by 2 tons. Or that I need to closely monitor the metal porosity or the carbon penetration. While yes, those would be all true, it completely fails and misses at answering the big picture question.

Is it direction of airflow important? Yes. But I'm surprised for such directionally happy people, no one's mentioned going to a 3U rackmount yet. I mean - come on. That would be a no brainer given the direction that this thread is going?

I remember that Asetek (I think is the name of the company) used to come out with some coolers and I think Koolance was another one. Now, I'm not in the OCing world. Never really have been. So I figured that maybe someone else would have mentioned some of the new stuff that they might have that are closed kits that will do better than the Corsair.

And in my initial research, I also did come across Thermalright's Silver Arrow air cooled HSF, but they were testing I think on a quad core 2500K (OC'd to 4.8 GHz). So I would think that with two extra cores, that the thermal load would be even higher.

I didn't know that Noctua had a fan that fit Socket R. (Good to know though).

Now I'm really curious as to why SnowyOwl went with a liquid cooling solution (while Bun-Bun went with an air solution - which he's stated why).

Do you guys recommend leaving TurboBoost on or off when you're OCing?

Why is it that whenever I ask an OCing question, it seems like all the d1ck$ pop out of the wood works? Seriously?

It isn't about a who's smarter or not thing. If I were really on the high horse, I wouldn't even need to ask you guys (cuz I'd be a know it all, already, wouldn't I?) It's no effing wonder why I never go into the whole OCing world. It is my experience from not only this forum, but a whole bunch others that HPCers seem to be more courtesy and professional than OCers and yet I can never understand why that is.

After interacting with you guys in the past day or two, I'm REALLY thinking of going with the $22k, quad-blade, eight-socket system instead because I know that if I EVER have a question, it will be met with professionalism and courtesy rather than by a bunch of discordant and immature pr1ck$. Grow up.

You wanna talk about real world/reality - talk to me when your EAU volume is 140,000 units a year, where the lead time for tooling is a MINIMUM 12 weeks from China, a change (to said tooling) is 4 grand, the end-of-line inspection system is half a million, a gauge change is half-a-million, and you have NO physical parts to work off of/test but the VP of Engineering AND your customer is asking you to answer the question: "How will it perform?" Let's see how well you guys would fair on with that.

[alpha754293]

Hi !
To help you with your first question... This is the major component of my Catia V5/CFD analysis workstation. I use a corsair H100 on low setting @4.6ghz it's fast stable and quiet. Professional cheap graphic card always a must with these software. High power PSU for low noise & added safety when running 24/7 @100%.

Attachment 124846

Have a nice day
David

How do you like the FirePro? I haven't use an ATI/AMD Fire* series card since I think it was a FireGL 8-something maybe...

My workstation at work has a Quadro 4000 and it's alright, mostly. When I have full truck cab models (CAD and mesh), it REALLY starts to struggle. Even if I have a full truck frame model (CAD and mesh), it REALLY struggles with it.

I've thought about possibly running it NOT headless and it taking over as my primary workstation as well, but I dunno. Was thinking that I might splurge and either go with a Tesla C2075 or the Quadro 6000. My current card is a 3D Labs Wildcat Realizm 800 and that's starting to show it's age.

[Bun-Bun]

In answer to your one question, turbo boost is automatically disabled past a certain multiplier. turbo boost only goes so high.

Good luck getting professional service on a blade server. Unless you have a big contact with the company selling it you they wont give you any more help then average retail computer store.

the issue at hand here is not real world vs simulation. Simulation has it's place but is no replacement to real world.

The issue at hand is you are consulting with people who know the real world of the topic of this thread and have the answers to your questions. You are choosing to not accept their advice because you apparently already know better. If you do then why are you here?


In regards to workstation graphics, depending on the application workstation cards offer no benefit over gaming/consumer cards. Worth looking into. I'm not familiar with your software. In my experience the firegl cards suffer major compatibility issues.

[alpha754293]

Is that just for the K-series processors or X-series as well? (Did not know that. Hmmm....interesting.)

Well...simulations ARE limited by what it CAN simulate. In the ideal world, everything would be solved at the molecular, atomic, or even sub atomic (quantum) levels. But of course, for practical reasons, it can't, so even Coulomb friction is really a simplified, bulk representation. (As is pretty much all of the classical mechanics and fluid mechanics equations that we have at our disposal.) But, by in large - it gets the job done. And is pretty good at it, despite it's shortcomings.

As stated before (and I don't know WHYYY people keep missing this) - I started with my own initial research. And since coming here, it has been superceded by the advice given. So...I don't know how or what more I can do. "You are choosing not to accept their advice..." If that were true, then why did I make those changes?

I don't know what more I can do. I don't know if I even SHOULD be doing anything in the first place since y'all seem CONVINCED that I'm not listening anyways. Why do I have to feel like I have to JUSTIFY myself and defend myself when, as demonstrated, the advice has already been taken, adapted, and IMPLEMENTED?

$hit man. It's true what they say - people focus so much on all the negatives.

If you still want to think that, by all means, go ahead. You've probably done so/am doing so anyways. Nothing I can, should, or will do will EVER change that. Once a person is convicted upon their beliefs, it is IMPOSSIBLE to change them. No matter how much evidence is presented before them, they'll STILL think that the Sun revolves around the Earth and that the world is flat. And that's okay. Everybody's entitled to their opinions.

[BeepBeep2]

Is that just for the K-series processors or X-series as well? (Did not know that. Hmmm....interesting.)

Well...simulations ARE limited by what it CAN simulate. In the ideal world, everything would be solved at the molecular, atomic, or even sub atomic (quantum) levels. But of course, for practical reasons, it can't, so even Coulomb friction is really a simplified, bulk representation. (As is pretty much all of the classical mechanics and fluid mechanics equations that we have at our disposal.) But, by in large - it gets the job done. And is pretty good at it, despite it's shortcomings.

As stated before (and I don't know WHYYY people keep missing this) - I started with my own initial research. And since coming here, it has been superceded by the advice given. So...I don't know how or what more I can do. "You are choosing not to accept their advice..." If that were true, then why did I make those changes?

I don't know what more I can do. I don't know if I even SHOULD be doing anything in the first place since y'all seem CONVINCED that I'm not listening anyways. Why do I have to feel like I have to JUSTIFY myself and defend myself when, as demonstrated, the advice has already been taken, adapted, and IMPLEMENTED?

$hit man. It's true what they say - people focus so much on all the negatives.

If you still want to think that, by all means, go ahead. You've probably done so/am doing so anyways. Nothing I can, should, or will do will EVER change that. Once a person is convicted upon their beliefs, it is IMPOSSIBLE to change them. No matter how much evidence is presented before them, they'll STILL think that the Sun revolves around the Earth and that the world is flat. And that's okay. Everybody's entitled to their opinions.

Please, what exactly did you deduct from my last post in this thread? I couldn't help but un-hide your posts and read your nonsense.
You need to read what I said again. A lot of it had nothing to do with "real-world everything, gut-feeling, I'm so retarded that I am someone that thinks things akin to the sun revolving around the earth" vs simulations. Simulations are wonderful things.

Comments like this really tick me off, and you've already made plenty of them:

I couldn't go with the UD3 because it only has 4 DIMM slots. Well...I could, but then the cost of RAM becomes another $300-500.

Will the system be able to run at full load, at 4.5 GHz for a week (or more) straight?

And it would get decent temps and not sound like a jet taking off?

*edit*
"still way below your budget mate of 2000"

add RAM and GPU, and it'll start getting very close/closer to it.

You ask a question such as "and would it get decent temps and not sound like a jet taking off?", because you don't even know that, then turn around and insult the guys' intelligence as if the guy helping you, which left out a GPU and RAM for a reason (because of the ability your software may or not have to use compute abilities of a GPU, and the fact that you were unsure if you wanted 16GB, 32GB, or 64GB of RAM).

In that same response, you said that going with a 4-slot motherboard would mean that you would need to pay $300-500 more for ram. You are obviously very lost in this area. I did link you a kit that costed only $230, did I not? 4 sticks are always preferred because it is easier for the memory controller to drive them.

You're extremely arrogant in your responses, no matter what we're talking about here. This is not completely about "gut feeling", but measured and calculated results and real-world variables too, I even listed several of them! If you want to know why people don't think you accept their advice, it is because you make comments that continually put you in a superior place than the people that are teaching you. Bun-Bun already said it:

Oh and after reading this entire thread I must say, as a fellow professional in engineering/technology, I am insulted. You come here asking for help and then rip everyone a new one when you feel it threaten's your intelligence. All that does is prove how not smart you are.

Please, get off your high horse and join us back in reality.

I'd like to see you explain why this simple waterblock base:


can perform within 10% of this base in overall temperature performance when combined with the exact same components.

You can design any part in a system to be as efficient as you want, but you won't know at all how it performs until you use it and test it in the real world.

We don't want to focus on the negatives here. We all want to help you pick the best parts for your system and help you understand how overclocking works, and understand the variables in real-world parts. Like I said, each CPU part does not have the same tolerances. You could order 5 3830Ks, and one would do 4.5 GHz at 1.45v with a load temp of 80c, another 4.6 at 1.4v and 72c, two 4.8 with 1.4v at 70c, and the other one be dead upon arrival. Then, you could simply swap motherboards and gain 100 MHz on all CPUs. That explains why asking questions about the ability of your system to run stable for X amount of time completely irrelevant.

[Bun-Bun]

@beepbeep

That's easy. There are many uncontrollable variables in a water loop system that can not be possibly taken into account with a single component design/simulation/comparison.

On top of that we are dealing with low heat system (relatively) that are already close to ambient temperature. The point of diminishing returns was met a long time ago.

Is that just for the K-series processors or X-series as well? (Did not know that. Hmmm....interesting.)

Well...simulations ARE limited by what it CAN simulate. In the ideal world, everything would be solved at the molecular, atomic, or even sub atomic (quantum) levels. But of course, for practical reasons, it can't, so even Coulomb friction is really a simplified, bulk representation. (As is pretty much all of the classical mechanics and fluid mechanics equations that we have at our disposal.) But, by in large - it gets the job done. And is pretty good at it, despite it's shortcomings.

As stated before (and I don't know WHYYY people keep missing this) - I started with my own initial research. And since coming here, it has been superceded by the advice given. So...I don't know how or what more I can do. "You are choosing not to accept their advice..." If that were true, then why did I make those changes?

I don't know what more I can do. I don't know if I even SHOULD be doing anything in the first place since y'all seem CONVINCED that I'm not listening anyways. Why do I have to feel like I have to JUSTIFY myself and defend myself when, as demonstrated, the advice has already been taken, adapted, and IMPLEMENTED?

$hit man. It's true what they say - people focus so much on all the negatives.

If you still want to think that, by all means, go ahead. You've probably done so/am doing so anyways. Nothing I can, should, or will do will EVER change that. Once a person is convicted upon their beliefs, it is IMPOSSIBLE to change them. No matter how much evidence is presented before them, they'll STILL think that the Sun revolves around the Earth and that the world is flat. And that's okay. Everybody's entitled to their opinions.

People are taking offense and thinking you are not listening because you go off on rants about theoretical mumbo jumbo (that honestly no one here cares about...to a point) and usually when someone is doing that it's to act superior and comes off condescending.

[BeepBeep2]

@beepbeep

That's easy. There are many uncontrollable variables in a water loop system that can not be possibly taken into account with a single component design/simulation/comparison.

On top of that we are dealing with low heat system (relatively) that are already close to ambient temperature. The point of diminishing returns was met a long time ago.

Hmm...it doesn't take a rocket scientist to figure that much out. However some people are oblivious...

My first comment about water cooling:

Custom water cooling components would be best for you in this situation but I really doubt you want to take the time to cut your own tubing, test for leaks, perform maintenance, research the right parts either. That is partially why you are here, to get help with those types of things, no?

His response:

5) Yes, I actually started off the research on the Corsair H80, but I was reading some reviews...

Then a question:

2. Now I'm curious, are there other better performing, but probably (slightly) more expensive closed-kit liquid cooling systems out there?

I responded:

2. No. However there are open kits that contain all (or most of, excluding some useful accessories), the parts. For these though, you'd need to mount parts, then measure and cut appropriate lengths of tubing and attach them in the loop using barbs and clamps or compression fittings. After that, you would fill the loop with water as you ran the pump (carefully, because it is self lubricated with the water, it is easier with three hands) by jump-starting your power supply in small bursts with no components attached but your water cooling pump. After it is filled, you can leave the power supply on with the only component attached being the liquid cooling pump. That way you can leak test without worry that your components will be ruined with water. You can simply dry them if you have leaks. This type of loop would need to be drained, taken apart, and cleaned about once a year. If you would like to learn more about this, people here can teach you.

...

I've actually ran the CFD simulations for some of the other people here and elsewhere as well. I think one guy - Jake - goes by the handle Cpt. Planet - when he started with his dry ice cooling and he was machining his tubes, he actually had me simulate the thermal performance and we actually spent a little bit of time optimizing his design with CFD before he ran. I think at one point, he might have held a record or two. I forget.

I've also CFD'd waterblocks too. So, CAN I engineer a custom system? I probably can. One of my friends actually builds CNC kits too, for like $2k, so I can CNC my own blocks as well. Course, that would be time spent on getting the system up and running instead of actually having it up and running and doing something useful with it.

...

3. Did I say a damned thing about running CFD simulations on waterblocks? I don't care if you can design your own blocks, OR manufacture them. What I said has absolutely NOTHING to do with that... ...In fact, you completely avoided the practical aspect, and only talked about the engineering aspect, before a part(s) become(s) a product in use.

...Also, I am sorry, but there is much more than fluid flow and restriction to take into account with a water block. There are waterblocks currently on the market with a low pressure drop that perform exactly the same in a loop as a waterblock with 3x the pressure drop when combined with the same radiators and pumps.

[BeepBeep2]

It starts off with two very simple questions (actually...one...really) and then people go off on all these stupid tangents about directional airflows and b--ching about counterproposals or that they would mention one component, and I'd mention why I couldn't (or more specifically) didn't want to go with that component that they mentioned.

Bun-Bun - the question is really simple. It's not that complicated. But at least I'm working with actual data, not just on gut feel.

They're not a threat. (Who is "they"? Everyone that isn't an engineer here?)They would be if they could actually argue on a premise (Which I am?) than mere "gut feel".

Can I go with an open kit? Yes. But by the same token, I could have gone with a custom waterblock too. *rolls eyes* Or direct die evaporative cooling too! (Way to say smart-ass things, does he not realize a "custom waterblock" would gain him measureably nothing over WB's on the market?) And clearly, if I'm making references to H80/H100, yes, the open kits are an alternative; yes, I don't think that I would be spending the time to actually do what BeepBeep2 suggested in terms of getting the kit, cutting the tubing to size, leak testing it, and then running it. (He doesn't know what I meant, I was talking about Swiftech/XSPC/EK kits)...

Some of the other points of argument from this thread is quite simply...just dumb. The whole "simulation vs. real world" part - yea. Dumb. (Bun-Bun just proved you wrong...)

If the real world was so important, then why is it that pretty much EVERYBODY is moving to simulations? (Because simulations are a GREAT tool to maximize efficiency of a single component, or course!)

Why WOULD we run simulations if guys like you think that I'm stuck in the stratosphere somewhere??? To me, that just reeks of people's ignorance (I reek of ignorance?) about what simulation engineers do (regardless of whether it's mechanics, fluids, or both. Or multiphysics.)

Its fine for you guys because you are able to physically pick up a part that's already been tooled, shot and shipped, with your hands, look at it and screw it in the case. (That isn't exactly what we do here. Maybe that is what average joe does, and when it works he is happy.)

But what do you do when tooling cost half-a-million dollars (actually, the end of line measurement system is a half-a-million dollars), and you DON'T have parts made yet because you haven't crossed the marker for a design freeze for your tooling guys to tool to? Are you able to tell me what the M and M' are??? I mean, I can ALWAYS make the problem more difficult for you guys. MUCH more difficult. (Who cares in this situation where efficiency/design improvements would not be worth the cost/time?)

But the point of some of those "back-of-the-envelop" type of calculations is to be able to very quickly show why something likely will or won't work. So, if you can get the blade parameters of the fans in question (the parameters that I'll need can be found in just about any sort of turbomachinery handbook), I would then be able to model explicitly the hub and the entire fan performance curve (minus statistical deviations, electrical to mechanical transform inefficiencies; unless I'm fed the line information as well.) (...well, what about the noise curves? People also prefer lower hums than high pitched whines, you can't plot that data without the real-world)

As dumb as I am and as intelligent as the rest of this community is - how someone misses that point is beyond me. (Nobody does...but then again we weren't even talking about that.)

This isn't a pissing match to see who's smarter. (You are trying to make it that.)

I came here to ask because I figured that I would be able to get the best, most professional advice and so far, it's been greeted with mostly nothing but disdain for the original question. (There was nothing wrong with the original question.)

Are your heads so far up your own arses that you can't see daylight anymore? (No, yours must be.)

And one of the major differences is that I know what my limitations are and won't claim to know more than I do. (Actually, you don't even know how the parts work in harmony and what weight each part carries in achieving the goal, you only care about a part by itself)

Heck, even if I've done the research, I won't necessarily claim that it's fully complete and accurate either because I haven't actually taken the painstaking task of actually modelling the entire computer in tremendous detail. (This has nothing to do with modelling the computer. You can't predict the temperature of your CPU (which I said was a real-world variable) by simulation. You have to take into account the flow rate of the pump, efficiency of all the parts, thermal paste, thermal contact, there are too many things here for you to be able to calculate without the real-world. Also, I'm not talking about gut-feeling. I'm talking about taking measurements.)

Course, I reckon that most of you here probably don't care for that either (since obviously, we're not into facts here). (We are into facts here. We're into very simple facts you can't even comprehend because you won't open your eyes.)

You want to talk about insulting -- try asking someone for help and all they can tell you is "you've seem to have made up your mind already, so good luck" or "what's the point of those calculations? That's not even accurate. You have to take into the account the hub..." (why haven't you mentioned the blade pitch and angle, and pretty much all of the other turbomachinery parameters and characteristics???) (We haven't mentioned the blade pitch and angle, and the other turbomachinery parameters and characteristics because we are consumers. That doesn't mean we can't comprehend roughly what blade pitch, angle, etc is optimal because we've tested a plethora of products!)

That would be like if I were to ask you how I can make the engine of my car produce more torque, you'd be telling me that I need to increase the clamping force on the forge press by 2 tons. Or that I need to closely monitor the metal porosity or the carbon penetration. While yes, those would be all true, it completely fails and misses at answering the big picture question. (...of course, we're not retarded.)

Is it direction of airflow important? Yes. But I'm surprised for such directionally happy people, no one's mentioned going to a 3U rackmount yet. I mean - come on. That would be a no brainer given the direction that this thread is going? (We didn't mention a 3U rackmount because it is irrelevant to what you wanted?)


I remember that Asetek (I think is the name of the company) used to come out with some coolers and I think Koolance was another one. Now, I'm not in the OCing world. Never really have been. So I figured that maybe someone else would have mentioned some of the new stuff that they might have that are closed kits that will do better than the Corsair. (Asetek makes the Corsair H50 and H70, CoolIT makes the H80 and H100. Forget Koolance, you don't know what you're talking about.)

And in my initial research, I also did come across Thermalright's Silver Arrow air cooled HSF, but they were testing I think on a quad core 2500K (OC'd to 4.8 GHz). So I would think that with two extra cores, that the thermal load would be even higher. (Yes. That is a good air cooler to look at though.)

I didn't know that Noctua had a fan that fit Socket R. (Good to know though). (I already told you that it is more commonly called Socket LGA 2011, or more specifically FCLGA 2011)

Now I'm really curious as to why SnowyOwl went with a liquid cooling solution (while Bun-Bun went with an air solution - which he's stated why). (Because liquid cooling is much better. I'm not talking about the all-in-one solution.)

Why is it that whenever I ask an OCing question, it seems like all the d1ck$ pop out of the wood works? Seriously? (They don't, you just come back with arrogance and spite, as if we can't add prices, for example)

It isn't about a who's smarter or not thing. (You already said that.)

If I were really on the high horse, I wouldn't even need to ask you guys (cuz I'd be a know it all, already, wouldn't I?) You kind of act like it!

It's no effing wonder why I never go into the whole OCing world. It is my experience from not only this forum, but a whole bunch others that HPCers seem to be more courtesy and professional than OCers and yet I can never understand why that is. We don't go into an HPC forum and talk like we know everything in response to suggestions.

After interacting with you guys in the past day or two, I'm REALLY thinking of going with the $22k, quad-blade, eight-socket system instead because I know that if I EVER have a question, it will be met with professionalism and courtesy rather than by a bunch of discordant and immature pr1ck$. Grow up. (Yeah, okay.)

You wanna talk about real world/reality - talk to me when your EAU volume is 140,000 units a year, where the lead time for tooling is a MINIMUM 12 weeks from China, a change (to said tooling) is 4 grand, the end-of-line inspection system is half a million, a gauge change is half-a-million, and you have NO physical parts to work off of/test but the VP of Engineering AND your customer is asking you to answer the question: "How will it perform?" Let's see how well you guys would fair on with that. Go step outside and talk to some normal people. The kind that make up majority.

Here you go.

In case you missed the link in quote, http://skinneelabs.com/ .

[alpha754293]

Please, what exactly did you deduct from my last post in this thread? I couldn't help but un-hide your posts and read your nonsense.
You need to read what I said again. A lot of it had nothing to do with "real-world everything, gut-feeling, I'm so retarded that I am someone that thinks things akin to the sun revolving around the earth" vs simulations. Simulations are wonderful things.

You really need to look at the real world instead of a simulation to grasp this...

All evidence the contrary.

Comments like this really tick me off, and you've already made plenty of them:

You ask a question such as "and would it get decent temps and not sound like a jet taking off?", because you don't even know that

"that". what is "that"? I can ALWAYS decrease the temperature (without a phase-change system) by going to a higher airflow, and oft, louder, fan.

Now, to your credit, I did say this:

Budget is set between $2000-2500 (between board, CPU, case, motherboard, power supply, cooling, video card)

So I did leave out RAM in that. My bad. Course, that also was meant to leave the options for the motherboard open. (Although the amount of memory that I might have gone with was listed above that line, and that also as stated, I wasn't exactly sure how much memory I would want (again, goes back to the whole motherboard thing too).

BUT, the original person that replied left out the GPU, whereas, the allocated budget included the cost of a GPU. That would be like if I were to give you ask you for a quote for a trip to France, and you tell me, oh...MacDo is only 8 EUR, hotel is blah, transporation is blah, and then completely leave out the cost of the plane ticket. *rolls eyes*

You know what we call that? An incomplete analysis.

That's like my work: "Oh $hit. We forgot to include something in the MSC.Adams model again. Sorry about that. Now your performance target requirement is TRIPLED. Have a nice day." (Partially true, actually. We've been in negotiations with our customer for the past two weeks for them to take concessions on their required performance target because they effed up. What's funnier is that our customer actually trusts our work MORE than they do from their in-house analysis department. That's good for us, but also bad (ridiculous amount of work, constantly). Point being..."oops..I forgot the GPU. Tack on another $400, so now you're at $2100. haha. so much for it being WELL below your budget mate. Ah well...")

Or in your case, "sorry about the choice of board. Now you'll have to tack on another $1200 for the RAM if you want 64 GB. haha."

Will someone please tell me why I bother writing it into the SOR if people are just going to ignore them? And then b--ch about it when I remind them that it's in the SOR? *rolls eyes* Sighhhh.....

P.S. I don't recall seeing that he ever actually asked if my programs were GPU-compute capable as the reason for why he didn't pick a GPU. Hmm, I must've missed that part. Whoops. My bad.

...then turn around and insult the guys' intelligence as if the guy helping you, which left out a GPU and RAM for a reason (because of the ability your software may or not have to use compute abilities of a GPU

How is whether the program can use the compute abilities of a GPU even REMOTELY relevant here? Does it matter? When someone asks you for advice on picking out a video card, is your first question back to them (in order to assist you with the selection process) "Can your program use the compute abilities of the GPU?"

...and the fact that you were unsure if you wanted 16GB, 32GB, or 64GB of RAM).

In that same response, you said that going with a 4-slot motherboard would mean that you would need to pay $300-500 more for ram. You are obviously very lost in this area. I did link you a kit that costed only $230, did I not? 4 sticks are always preferred because it is easier for the memory controller to drive them.

4x4 GB is as low as $90 (for DDR3-1600, CL9): http://www.newegg.com/Product/Produc...82E16820231545
4x8 GB is as low as $280 (for DDR3-1600, CL10): http://www.newegg.com/Product/Produc...82E16820231490
4x16 GB isn't even available under Newegg's Desktop memory listings, and for a Kingston 16 GB DDR3-1333 ECC Reg stick, it's $425 a pop, which means for four of them, it would be $1700. http://www.newegg.com/Product/Produc...82E16820139932

Conversely, 8x8 GB (under Newegg's Desktop memory listings) is $600 for DDR3-1600, CL10. http://www.newegg.com/Product/Produc...82E16820231508

I absolutely agree with you that 4 DIMMS is easier for the mem controller to manage than 8 DIMMS. But to restrict myself to a 4-DIMM board - a) leaves no room for expansion when needed, and b) even if I went with an 8-DIMM board; that does NOT necessarily precipitate that I MUST get an 8-DIMM kit. Needless and suffice it to say that by getting an 8-DIMM board, I have expandability options that is not possible with a 4-DIMM board; so there are reasons why I would want to keep my options open. The biggest, most memory intensive run that I've EVER done was 90 GB out of 128 GB, and so depending on how accurate I want the solution to be, I can size the problem based on the RAM limitations.

So, CAN I have 64 GB AND stayed with a 4-DIMM board? Yes. The cost for doing so: about $1200.

You're extremely arrogant in your responses, no matter what we're talking about here. This is not completely about "gut feeling", but measured and calculated results and real-world variables too

Funny that you did the calculations for the 80 mm fan and the 120 mm fan, but not the 200+ mm fans. Why is that? What other "real world" variable did you account for that I didn't?

So you're trying to tell me that you can tell me what the air velocity is going to be when it hits the socket?

Oh and in that video that you linked, why did the guy point the smoke machine nozzle at an angle when he was doing the "test"?

What's next? You're going to say that my numbers are BS because I didn't take into account the R/Z components of the blade, or the nominal angle? Or that my pressure differential across it is wrong? What? What? Tell me. Oh wait...no...lemme guess - you're going to tell me that I haven't modeled the surface roughness of the blades properly too, and that I don't have enough elements/cells to accurately resolve the boundary layer flow and that the separate point is also incorrect/inaccurate and that this whole time, the whole point of doing those calculations are useless because you can't use Bernoulli's equation for rotating frames of reference.

But yet, you seem to be able to tell me things that I don't know. And that you are the ONLY person who knows the answer. Who's the arrogant one now?

I'd like to see you tell me what my heat flux rate is going to be across the heat exchanger since my non-real-world simulations won't be able to give me but meaningless, unrealistic, virtual-world only results. *rolls eyes*

I even listed several of them! If you want to know why people don't think you accept their advice, it is because you make comments that continually put you in a superior place than the people that are teaching you. Bun-Bun already said it:


I'd like to see you explain why this simple waterblock base:


can perform within 10% of this base in overall temperature performance when combined with the exact same components.

You can design any part in a system to be as efficient as you want, but you won't know at all how it performs until you use it and test it in the real world.

Give me the prints or the CAD models for the entire upper level system assembly and I'll tell you why.

We don't want to focus on the negatives here.

But you do.

We all want to help you pick the best parts for your system and help you understand how overclocking works and understand the variables in real-world parts.

And what makes you think that I don't already understand about the variability in your so-called "real world" parts. (You know, for someone who's says "simulations are wonderful things", you spend and awful lot of time trying to "explain" to me the difference between simulation and real-world and you spend an even greater amount of time, distinguishing and separating the two. I'm going to guess that you are familiar with the concepts of error bars, and bounding box dimensions, and bounding box methods, and response surface models/methods, and various optimization methods as well; which all work together and says if you vary it by this much, this is what your lower bound of performance is going to be and if you vary it by this much, this is what your upper bound of performance is going to be. And so long as you're between those tolerances, you're going to end up somewhere in between. Like I said, if you can tell me the injecting molding pressure variation as a function of time, I'm certainly that the simulation is only going to be more accurate. Will I be able to tell you why you keep hanging to the left and shooting to the left? No, probably not. (Not at the bulk level anyways). But I probably can't tell you any better than you can tell me why. And I find it oh so amusing that you seem to somehow think that I will be able to give you a very definitive, precise, concrete answer when the inputs going into it have just as much or even greater tolerance span. I can't be more accurate than the data that's being fed into it.

So, I have NO idea why you keep yammering on about "the real world" "the real world" "the real world"....the real world only exists if you're dealing with real parts. If you had the components in your hands, THEN I can ask you how is it going to perform (temperature/voltage). But I'm not asking that now, am I? I'm asking that even in spite of you NOT having those "real world" parts. And so far, your answer has been "well...I don't know" or "I can't tell you what that is going to be". So what good is your rant about "the real world" when you don't even have the real parts, sitting in front of you.

That's the difference. From my simulations perspective, I'm USED to dealing with stuff that does not yet exist. Within a certain confidence interval, I can say..."here's what the expected level of performance is going to be, based on these initial assumptions, and based on this method or process of determining what the outcome is going to be like; following the rules of math and science." When the real world has the same question posed to them, they are completely stuck - you're at a standstill. According to you, you can't answer that question. However, even with the simpliest, and most grotesque of models, I can still start simulating what can happen, without ever cutting steel.

And in my field, being in crash safety - you want to make damn sure that the thing is going to perform before you start tooling. Because the penalty for effing up is death. Maybe not for you. But when I'm being fed wrong, incorrect, inadequate, incomplete, or misinformation - I have a GREAT potential to kill a LOT of people, REALLY, REALLY quickly.

Like I said, each CPU part does not have the same tolerances. You could order 5 3830Ks, and one would do 4.5 GHz at 1.45v with a load temp of 80c, another 4.6 at 1.4v and 72c, two 4.8 with 1.4v at 70c, and the other one be dead upon arrival. Then, you could simply swap motherboards and gain 100 MHz on all CPUs. That explains why asking questions about the ability of your system to run stable for X amount of time completely irrelevant.

I'm going to go out on a limb here and guess that you brought up a completely irrelevant CPU (3830K) in order to help illustrate your point about the variability that can exist within the production of a CPU, right?

How would that be irrelevant?

If I came to you and said, okay BeepBeep2; for $1000 (in labour only + whatever the cost of the components are), build me a system that can run at 4.5 GHz, performing this task for a month - straight; you seriously would have told me that you're going to decline that offer because "...question about the ability of your system to run stable for x amount of time is irrelevant?" You'd pass up on making $1000 bucks to do the build? Hmmm....good to know. I'll keep that in mind.

[alpha754293]

Interesting, it doesn't to the embedded quote properly now. Hmmm....

It starts off with two very simple questions (actually...one...really) and then people go off on all these stupid tangents about directional airflows and b--ching about counterproposals or that they would mention one component, and I'd mention why I couldn't (or more specifically) didn't want to go with that component that they mentioned.

Bun-Bun - the question is really simple. It's not that complicated. But at least I'm working with actual data, not just on gut feel.

They're not a threat. (Who is "they"? Everyone that isn't an engineer here?)They would be if they could actually argue on a premise (Which I am?) than mere "gut feel".

Can I go with an open kit? Yes. But by the same token, I could have gone with a custom waterblock too. *rolls eyes* Or direct die evaporative cooling too! (Way to say smart-ass things, does he not realize a "custom waterblock" would gain him measureably nothing over WB's on the market?) And clearly, if I'm making references to H80/H100, yes, the open kits are an alternative; yes, I don't think that I would be spending the time to actually do what BeepBeep2 suggested in terms of getting the kit, cutting the tubing to size, leak testing it, and then running it. (He doesn't know what I meant, I was talking about Swiftech/XSPC/EK kits)...

Some of the other points of argument from this thread is quite simply...just dumb. The whole "simulation vs. real world" part - yea. Dumb. (Bun-Bun just proved you wrong...)

If the real world was so important, then why is it that pretty much EVERYBODY is moving to simulations? (Because simulations are a GREAT tool to maximize efficiency of a single component, or course!)

Why WOULD we run simulations if guys like you think that I'm stuck in the stratosphere somewhere??? To me, that just reeks of people's ignorance (I reek of ignorance?) about what simulation engineers do (regardless of whether it's mechanics, fluids, or both. Or multiphysics.)

Its fine for you guys because you are able to physically pick up a part that's already been tooled, shot and shipped, with your hands, look at it and screw it in the case. (That isn't exactly what we do here. Maybe that is what average joe does, and when it works he is happy.)

But what do you do when tooling cost half-a-million dollars (actually, the end of line measurement system is a half-a-million dollars), and you DON'T have parts made yet because you haven't crossed the marker for a design freeze for your tooling guys to tool to? Are you able to tell me what the M and M' are??? I mean, I can ALWAYS make the problem more difficult for you guys. MUCH more difficult. (Who cares in this situation where efficiency/design improvements would not be worth the cost/time?)

But the point of some of those "back-of-the-envelop" type of calculations is to be able to very quickly show why something likely will or won't work. So, if you can get the blade parameters of the fans in question (the parameters that I'll need can be found in just about any sort of turbomachinery handbook), I would then be able to model explicitly the hub and the entire fan performance curve (minus statistical deviations, electrical to mechanical transform inefficiencies; unless I'm fed the line information as well.) (...well, what about the noise curves? People also prefer lower hums than high pitched whines, you can't plot that data without the real-world)

As dumb as I am and as intelligent as the rest of this community is - how someone misses that point is beyond me. (Nobody does...but then again we weren't even talking about that.)

This isn't a pissing match to see who's smarter. (You are trying to make it that.)

I came here to ask because I figured that I would be able to get the best, most professional advice and so far, it's been greeted with mostly nothing but disdain for the original question. (There was nothing wrong with the original question.)

Are your heads so far up your own arses that you can't see daylight anymore? (No, yours must be.)

And one of the major differences is that I know what my limitations are and won't claim to know more than I do. (Actually, you don't even know how the parts work in harmony and what weight each part carries in achieving the goal, you only care about a part by itself)

Heck, even if I've done the research, I won't necessarily claim that it's fully complete and accurate either because I haven't actually taken the painstaking task of actually modelling the entire computer in tremendous detail. (This has nothing to do with modelling the computer. You can't predict the temperature of your CPU (which I said was a real-world variable) by simulation. You have to take into account the flow rate of the pump, efficiency of all the parts, thermal paste, thermal contact, there are too many things here for you to be able to calculate without the real-world. Also, I'm not talking about gut-feeling. I'm talking about taking measurements.)

Course, I reckon that most of you here probably don't care for that either (since obviously, we're not into facts here). (We are into facts here. We're into very simple facts you can't even comprehend because you won't open your eyes.)

You want to talk about insulting -- try asking someone for help and all they can tell you is "you've seem to have made up your mind already, so good luck" or "what's the point of those calculations? That's not even accurate. You have to take into the account the hub..." (why haven't you mentioned the blade pitch and angle, and pretty much all of the other turbomachinery parameters and characteristics???) (We haven't mentioned the blade pitch and angle, and the other turbomachinery parameters and characteristics because we are consumers. That doesn't mean we can't comprehend roughly what blade pitch, angle, etc is optimal because we've tested a plethora of products!)

That would be like if I were to ask you how I can make the engine of my car produce more torque, you'd be telling me that I need to increase the clamping force on the forge press by 2 tons. Or that I need to closely monitor the metal porosity or the carbon penetration. While yes, those would be all true, it completely fails and misses at answering the big picture question. (...of course, we're not retarded.)

Is it direction of airflow important? Yes. But I'm surprised for such directionally happy people, no one's mentioned going to a 3U rackmount yet. I mean - come on. That would be a no brainer given the direction that this thread is going? (We didn't mention a 3U rackmount because it is irrelevant to what you wanted?)


I remember that Asetek (I think is the name of the company) used to come out with some coolers and I think Koolance was another one. Now, I'm not in the OCing world. Never really have been. So I figured that maybe someone else would have mentioned some of the new stuff that they might have that are closed kits that will do better than the Corsair. (Asetek makes the Corsair H50 and H70, CoolIT makes the H80 and H100. Forget Koolance, you don't know what you're talking about.)

And in my initial research, I also did come across Thermalright's Silver Arrow air cooled HSF, but they were testing I think on a quad core 2500K (OC'd to 4.8 GHz). So I would think that with two extra cores, that the thermal load would be even higher. (Yes. That is a good air cooler to look at though.) It was initially ruled out because the review with that processor was already hitting close to 50 C. So I figured that add two more cores to that, and I could be looking at much higher temps (80+ C) - not really sure because no data was available for their thermal resistance.

I didn't know that Noctua had a fan that fit Socket R. (Good to know though). (I already told you that it is more commonly called Socket LGA 2011, or more specifically FCLGA 2011) Really? You're going to get your panties in a wad over semantics? Literally?

Now I'm really curious as to why SnowyOwl went with a liquid cooling solution (while Bun-Bun went with an air solution - which he's stated why). (Because liquid cooling is much better. I'm not talking about the all-in-one solution.) And yet

I use a corsair H100...

he uses an all in one solution. So....whyyyy are you talking about non-all-in-one solutions in a question that specifically ASKS why one engineer when with an all-in-one solution while another said that they're absolute crap?

Why is it that whenever I ask an OCing question, it seems like all the d1ck$ pop out of the wood works? Seriously? (They don't, you just come back with arrogance and spite, as if we can't add prices, for example) See my example above. That'd be like two people are arguing the merits of NAFTA between Canada and the U.S. and you come in arguing about Antarctica. Is there realllly nothing on TV on a late Sunday afternoon? While possibly all valid points, but if I'm not inclining on going with that - then it's totally pointless and irrelevant. (which is why I also used the example of how equally senseless and irrelevant it would have been for you to ask me "why aren't you CNCing your own waterblocks?" or "why aren't you going with direct die evaporative cooling?") That'd also be akin to if someone came to you asking for advice for a sports car under $25k US, and you're asking them "why don't you buy a Veyron?" Are so desperate for validation that you need to do that? Really? That you need someone to open wide and go "yes, BeepBeep2 you are correct. An open kit would perform better. And a custom waterblock would be even better than that." Like really? Seriously??? Why do you repeatedly bring it up when you've already guessed it that it doesn't look like I'm going to be going with an open kit??? Is your brain TOTALLY incapable of processing "ah. he's not going there. So let's limit the discussion to just closed systems. What about Asetek or Koolance? What do they have now? Is there any that doesn't SUCK (as much)?" I mean, is your brain incapable of such a thought pattern?

It isn't about a who's smarter or not thing. (You already said that.)

If I were really on the high horse, I wouldn't even need to ask you guys (cuz I'd be a know it all, already, wouldn't I?) You kind of act like it! Think what you will.

It's no effing wonder why I never go into the whole OCing world. It is my experience from not only this forum, but a whole bunch others that HPCers seem to be more courtesy and professional than OCers and yet I can never understand why that is. We don't go into an HPC forum and talk like we know everything in response to suggestions.

After interacting with you guys in the past day or two, I'm REALLY thinking of going with the $22k, quad-blade, eight-socket system instead because I know that if I EVER have a question, it will be met with professionalism and courtesy rather than by a bunch of discordant and immature pr1ck$. Grow up. (Yeah, okay.)

You wanna talk about real world/reality - talk to me when your EAU volume is 140,000 units a year, where the lead time for tooling is a MINIMUM 12 weeks from China, a change (to said tooling) is 4 grand, the end-of-line inspection system is half a million, a gauge change is half-a-million, and you have NO physical parts to work off of/test but the VP of Engineering AND your customer is asking you to answer the question: "How will it perform?" Let's see how well you guys would fair on with that. Go step outside and talk to some normal people. The kind that make up majority. [i]Surprisingly, the people outside have more common sense than you do. In fact, even my bank teller has seen a video of the crash simulations that I've ran (using a public model, as an example, a public demonstration piece); and they're like "that's soo cool." And I'm like "yea...some of the simulations have gotten so good that BMW may only crash ONE vehicle, pass the test and then that's it. They're done." And they're like "really?" and I'm like "yup. And at upwards of quarter-of-a-mil to half-a-million per crash test, that's a VERY good thing.")

[alpha754293]

(Because simulations are a GREAT tool to maximize efficiency of a single component, or course!)

Please tell me you're kidding.

Way to say smart-ass things, does he not realize a "custom waterblock" would gain him measureably nothing over WB's on the market?

Depends. Ever heard of microchannel coolers?

(He doesn't know what I meant, I was talking about Swiftech/XSPC/EK kits)

Actually, you did send me the link to the H2O-X20. They were out of stock on their website. Also said that they are working on shipping the mounting plates for the LGA2011.

(I reek of ignorance?)

Don't act so surprised.

(...well, what about the noise curves? People also prefer lower hums than high pitched whines, you can't plot that data without the real-world)

Not entirely true. Google "Windsor Zug Island". A high amplitude, low frequency noise could be just as annoying as a low amplitude, high frequency noise. Also if that were the case, there'd be a HECK of a lot more double-bass concertos and yet most melodies are in the higher registers (compared to a double-bass). Hmmmm.....I wonder why that is.

(You are trying to make it that.)

It wouldn't be like this if you got your $hit straight!

(No, yours must be.)

Really? That's all you can come up with for a retort? *sigh* pep boys don't know too many words.

(We didn't mention a 3U rackmount because it is irrelevant to what you wanted?)

There were no limitations on the case. I believe the exact question was:

Any recommendations for the case?

(Course, someone could have came back and asked "rackmount or tower?" to which I probably would have replied "either or".)

(Actually, you don't even know how the parts work in harmony and what weight each part carries in achieving the goal, you only care about a part by itself)

Actually...there's this magical thingy called a weighted function. ;o) Check it out. We use that in artificial neural networks a LOT. And also for sensitivity analysis (trying to identify which component(s) out of 500, has the greatest impact on a particular goal/target.)

(This has nothing to do with modelling the computer. You can't predict the temperature of your CPU (which I said was a real-world variable) by simulation

Actually, you can. See IBM J. Res. Dev. 2007, Vol. 51, No. 6 I think it was.

You have to take into account the flow rate of the pump, efficiency of all the parts, thermal paste, thermal contact, there are too many things here for you to be able to calculate without the real-world. Also, I'm not talking about gut-feeling. I'm talking about taking measurements.)

All of which can be measured, and to some degree, simulated (but of course, not perfectly). But you can get pretty darn close.

So you've been able to measure thermal contact? Did you drill into the heatsink for that or the casing? I'm curious now. How did you manage to do that? You would think that a TekScan sensor would be too much of an insulator for it to be able to work properly and wouldn't the case temperature be too high for it?

Again, the real world is only good IF you have real parts to test. What do you do when you don't have ANY parts, but you're being asked the same question? ;o)

Re: component simulation
yes, we do develop stuff on at a component level first. Some of the OEMs try to force us to run the full model every time, but we have demonstrated that using DMIG and superelements, we can represent their full vehicle model in a more simplistic way, and still get a good correlation of the results from it. Periodically, we actually have to send one of our guys to the OEMs because they don't let the full vehicle models out, so we have to bring our optimized-component-level design up to their facility for integration into the full vehicle model, and then run the whole vehicle as a complete system. (And then back-correlate if need be.)

I would LOVE it the day when I will be able to run the full vehicle model as-is, as you see it out on the road. Unfortuantely, computers still, generally, aren't quite nearly powerful enough to make that practical. But we're slowly getting there.

[Bun-Bun]

You are comparing apples and oranges grasping at trying to prove your point when you have none.

For example you took the example of people preferring lower pitch hum fans to higher pitched fans out of context. The point was you can't simulate the results of personal preference.

And again with estimating temperatures. Your reference to some book is meaningless as everyone here knows that cpu's vary between batches and even within batches. That is real world result that can't be predicted until you have the chip in hand.

Given your field you must have read some QA/QC books. Think about it.

finally you are comparing the design of complex assemblies to computers builds using already known and built parts that can be tested in real.

Apples and oranges.

[OC Nub]

We don't go into an HPC forum and talk like we know everything in response to suggestions.

Noone asked for your advice or suggestions, this was your thread. You came here asking, and this is exactly what you have done.

After interacting with you guys in the past day or two, I'm REALLY thinking of going with the $22k, quad-blade, eight-socket system instead because I know that if I EVER have a question, it will be met with professionalism and courtesy rather than by a bunch of discordant and immature pr1ck$. Grow up. (Yeah, okay.)

You are free to leave at any time, noones asking you to stay.

[l0ud_sil3nc3]

not sure it trolling or serious...

[OC Nub]

not sure it trolling or serious...

Im wondering the same thing.

My suggestion, before making any hardware purchases, or posting in anymore over clocking forums, would be to first buy a box of these



Uploaded with ImageShack.us

"What is Xanax?

Xanax (alprazolam) belongs to a group of drugs called benzodiazepines. It works by slowing down the movement of chemicals in the brain that may become unbalanced."


We could all use a little more balance.

[BeepBeep2]

Please tell me you're kidding.
...please, you're going to tell me that you can explain why a waterblock like the first photo I posted can perform almost as well as the microchannel waterblock common to almost every waterblock now on the market?
It's easy. The first block is much less restrictive, and allows water to flow throughout the loop faster. I may not be a thermal/fluids science engineer, but I know that radiators perform better with faster water flow because I and others have tested it in person.

Depends. Ever heard of microchannel coolers?
Yes, of course, and you mean like the second waterblock I posted? That is a microchannel base for a waterblock, used to cool a CPU.


Actually, you did send me the link to the H2O-X20. They were out of stock on their website. Also said that they are working on shipping the mounting plates for the LGA2011.
Try this: http://www.frozencpu.com/products/15..._Hot_item.html


Don't act so surprised.
Like Bun-Bun said, you are comparing apples to oranges, grasping at straws to try and prove everyone, most specifically myself, because I am a "stupid person" according to your signature. Who cares if I'm arguing with a thermal/fluids science engineer about an environment he is unfamiliar with?

Not entirely true. Google "Windsor Zug Island". A high amplitude, low frequency noise could be just as annoying as a low amplitude, high frequency noise. Also if that were the case, there'd be a HECK of a lot more double-bass concertos and yet most melodies are in the higher registers (compared to a double-bass). Hmmmm.....I wonder why that is.
Okay, so Canadians (and Americans) complaining about an electronic sounding hum that pulses and shouldn't be there?...I'm talking about the differences in pitch emitted by the DC motors in common case fans. I am talking about similarly rated fans, even fans that produce exactly the same amplitude of noise. I don't know what you're trying to prove here, because that isn't what I was talking about...(?)


It wouldn't be like this if you got your $hit straight!
Okay, but my $hit is pretty straight. You've been making what I say look a lot worse than it is...


Really? That's all you can come up with for a retort? *sigh* pep boys don't know too many words.
Yeah, since you talk about everyone else having their heads up their asses seemingly from the inside of your anal cavity.


There were no limitations on the case. I believe the exact question was:



(Course, someone could have came back and asked "rackmount or tower?" to which I probably would have replied "either or".)
You're right, but you came to a forum full of people that use ATX specification parts, why would we suggest something we aren't familiar with? I mean, it is a lot easier to cool an overclocked processor in something bigger than a rackmount anyway. You opened your thread with the title "xtremeoverclocking - pushing the 3930K/3960X to 4.5 GHz", and you don't understand why nobody asked if you wanted a rackmount?


Actually...there's this magical thingy called a weighted function. ;o) Check it out. We use that in artificial neural networks a LOT. And also for sensitivity analysis (trying to identify which component(s) out of 500, has the greatest impact on a particular goal/target.)
Please read what Bun-Bun said above about the actual products...there is a lot of variation in parts and batches. AMD even makes every CPU in their line-up with handicapped silicon, they disable bad cores or cache in lower-end parts, chip harvesting to make money on what would be a worthless die.

Actually, you can. See IBM J. Res. Dev. 2007, Vol. 51, No. 6 I think it was.
Alrighty.

All of which can be measured, and to some degree, simulated (but of course, not perfectly). But you can get pretty darn close.
Exactly. ...to some extent, I would call it a long-stretch to say "pretty darn close" yet for a system (like a vehicle!)...but to some degree, simulated

So you've been able to measure thermal contact? Did you drill into the heatsink for that or the casing? I'm curious now. How did you manage to do that? You would think that a TekScan sensor would be too much of an insulator for it to be able to work properly and wouldn't the case temperature be too high for it?
Stupid man's way of measuring contact:
1. Apply thermal paste in desired fashion. Either spread it, or put a blob in center of CPU.
2. Mount waterblock or heatsink.
3. Remove waterblock or heatsink, check spread.
4. If spread did not cover/touch at the very least, most of the IHS, rinse and repeat.

This is a really good read/review of thermal paste as you will want to buy some. I apologize if the tests are not scientific enough for you, but sometimes people do learn from trial and error (how we have learned) instead of from an engineering book.


Again, the real world is only good IF you have real parts to test. What do you do when you don't have ANY parts, but you're being asked the same question? ;o)
Sir, we have all the parts we are discussing in the real world here, and we have been discussing real world performance as noted by myself and others. I don't know why you keep pulling in other **** to our conversation. At skinneelabs.com and the liquid cooling section (also do a google search for "Martin's Liquid Lab")...we do not just do things by "gut feeling" but measurement in a controlled environment. Martin has tested well over 40 fans in a controlled environment, with an SPL meter and anemometer, here, here, and here. He also went through and perceptually rated the fans noise level by his tolerance. I may be wrong to assume that most people will like a low hum instead of a high whine when at the same amplitude, but that is the general consensus on this forum. The google search about "Windsor Zug Island" is completely irrelevant to our subject...

I really don't like to focus on negatives, and I'm not as stupid as you think I am, neither are the other members here...but we just know different things.

I'm sure any of the members here would be open and receptive to learning more about the engineering...however you've tried to prove everything we've seen, touched, and measured over multiple "experiments" and people, wrong. You also read up about the subject, but we've delved a lot deeper as far as knowing the way the hardware acts and behaves when it is used. There's nothing wrong with that, but it seems we (most specifically I,) have damaged your ego a bit.

[Leeghoofd]

Most users jumped to the H80/100 cooling solutions as it as the easiest to install with quad channel rams (as the dimm slots are located left and right of the CPU socket) Some jus tuse them coz they look cool :p If you have taller dimms you can have compatibility issues when installing big air coolers. Hence why you need to make sure IF you opt to go for an air cooler that you best opt straight for low profile dimms ( eg Vengeance LP or G.SKill Ares rams )

Guys have a beer :p It's the way of the internet. Some things get out of proportion or completely misunderstood. However quoting individual phrases comes over totally wrong... so let's focus on the build...

FYI I ran prime 95 4500mhz yesterday on the 3960X ES with the H100 at medium fan setting. Reaching 75-80°C average ( hot ES sample, as the 3950K ES is way cooler ) BUT the PWM area got around 90-94°C on the Sabretooth board. Hence why you need good airflow (being it simulated or not :p) on the top of the board. So look for a case with top ventilation to suck out the hot air...

peace kiss and hugg otherwise Movieman will be on it's way to give you both a asswhooping. Don't let Dave do that to you as he enjoys it too much :p

[alpha754293]

Most users jumped to the H80/100 cooling solutions as it as the easiest to install with quad channel rams (as the dimm slots are located left and right of the CPU socket) Some jus tuse them coz they look cool :p If you have taller dimms you can have compatibility issues when installing big air coolers. Hence why you need to make sure IF you opt to go for an air cooler that you best opt straight for low profile dimms ( eg Vengeance LP or G.SKill Ares rams )

Guys have a beer :p It's the way of the internet. Some things get out of proportion or completely misunderstood. However quoting individual phrases comes over totally wrong... so let's focus on the build...

FYI I ran prime 95 4500mhz yesterday on the 3960X ES with the H100 at medium fan setting. Reaching 75-80°C average ( hot ES sample, as the 3950K ES is way cooler ) BUT the PWM area got around 90-94°C on the Sabretooth board. Hence why you need good airflow (being it simulated or not :p) on the top of the board. So look for a case with top ventilation to suck out the hot air...

peace kiss and hugg otherwise Movieman will be on it's way to give you both a asswhooping. Don't let Dave do that to you as he enjoys it too much :p

That's kind of what I was afraid of. Except that because I've never actually ran the system, so my best guestimations were pegging the expected temperatures at somewhere between 70-85 C. I remember I think that the T_junction limit for the 990X was 90 C (I think it was that, I don't remember, and I don't remember if I have any CPU-Z screenshots that I've saved) , and so going on the assumption that it's going to be the same or very similar, I was worried that it would start approaching the recommend/design limits.

And given all of the discussions about how it's going to be pretty much a random lottery in terms of which CPU I get (whether one that I would be able to hit the 4.5 GHz mark with a lower voltage or one that would require a slightly higher voltage to get there and stay there), in my mind (perhaps yet another incorrect assumption) is that the possiblity that it might see those temperatures (even if it had perfect airflow through the case); is still within the reach of possibility. It might be a low possibility, but a possibility nonetheless. And I can't imagine running a CPU at 85 C for a month can be doing great wonders to/for it.

Speaking of airflows, would it be a BAD idea for me to mount the radiator behind a row of 80 mm fans in a 3U rackmount? I hadn't even actually thought about putting it in a rackmount, but since I mentioned it...hmmm...food for thought. Most of the Chenbro or Supermicro cases have at least three 80 mm fans in 3U. Some have 5. And the air will flow from front-to-back. (And some of them will also have two additional 80 mm exhaust fans in the back).

[Bun-Bun]

Noctua has an extensive list of dimm's that work and don't work with the NH-D14.

[BeepBeep2]

That's kind of what I was afraid of. Except that because I've never actually ran the system, so my best guestimations were pegging the expected temperatures at somewhere between 70-85 C. I remember I think that the T_junction limit for the 990X was 90 C (I think it was that, I don't remember, and I don't remember if I have any CPU-Z screenshots that I've saved) , and so going on the assumption that it's going to be the same or very similar, I was worried that it would start approaching the recommend/design limits.

And given all of the discussions about how it's going to be pretty much a random lottery in terms of which CPU I get (whether one that I would be able to hit the 4.5 GHz mark with a lower voltage or one that would require a slightly higher voltage to get there and stay there), in my mind (perhaps yet another incorrect assumption) is that the possiblity that it might see those temperatures (even if it had perfect airflow through the case); is still within the reach of possibility. It might be a low possibility, but a possibility nonetheless. And I can't imagine running a CPU at 85 C for a month can be doing great wonders to/for it.

Speaking of airflows, would it be a BAD idea for me to mount the radiator behind a row of 80 mm fans in a 3U rackmount? I hadn't even actually thought about putting it in a rackmount, but since I mentioned it...hmmm...food for thought. Most of the Chenbro or Supermicro cases have at least three 80 mm fans in 3U. Some have 5. And the air will flow from front-to-back. (And some of them will also have two additional 80 mm exhaust fans in the back).

It is a mixture of voltage and temperatures that kill a CPU. 85 is rather high, but if you combine it with a relatively low voltage I don't think there is much chance to hurt the processor. However, if you were hitting 85c with relatively low voltage (close to stock) you would want a better cooling solution.

[Leeghoofd]

No idea bout the rack fans : 8cm might be noisy or lacking airflow if they run at lower rpm. Secondly no idea if those 8cm fans would do for appropriate airflow for eg the H100 rad.

Each CPU has it's sweetspot and it's up to you to discover it : most 3950K wil run 4500 at around 1.3Vcore. Some might be hotties, some might not. Some might clock higher then 4500, some might have to settle at 4300 or 4400mhz to keep an exceptable balance between heat and/or voltages. And plz don't forget the mobo out of the equation...

Question : Does that app load all cores (included the HT ones) 100% all the time ? Folding at home here scales between 70-100% all the time. Temps when Folding are far lower then my Prime95 test (65-75°C average)

[Bun-Bun]

Leeghoofd raises a good point. Depending on the application HT may or may not even benefit. OC'ing with HT disabled is easier and less stressful then with HT on.

[alpha754293]

Yea, I know that it's an unpredictable function between voltage and clock speed. (sigh...) The fact that it CAN potentially get quite hot (and the fact that the voltages CAN be quite high) is why I had asked about the premature death question. Couldn't put numbers to it before, but the cause for concern still remain the same.

I ended up finding a 4U rackmount that has 120 mm fans inside. I forget who makes the 80 mm fans that's in my 4U. It's a pity that the watercooling isn't the one that's sending the feedback signal to the PWM controller for the fans.

I can make the app use as many processors as I want. However, to answer your question, for this particular type/class of analysis, HTT is currently showing a 3-12% DECREASE in performance (vs. without HTT). I haven't tested it with CFD yet (either using the same program or using a different program). Also note that the HTT tests were performed on current generation Xeons.

I also just realized that if I get a rackmount, I might actually have to flip the air flow direction around so that it actually goes back-to-front. Hmmm....

Bun-Bun - are you volunteering to do the testing for me?

BeepBeep2 - No, a little bit smaller still.
re: faster waterflow - sort of. upto a certain point.
re: noise - be clearer next time (about what it is that you're talking about). To make generic statements means that it is subject to the test of universality.
re: $hit straight - considering that I've been able to highlight some blantant flaws in your points of argument...the rest is self explanatory.
re: rackmount - it was a tongue-in-cheek commentary on the absurdity of some of the comments/replies. Constantly beating on "it depends on voltage/speed/what processor you get" is about as useful as "no freakin' clue." So, while yes, the statement is true - how would presenting the null or void hypothesis be useful at all? In fact, I've already mentioned SEVERAL times that "yes, it depends on voltage and speed" (perhaps not as blantantly spelled out for you, but here it is, blantantly spelled out), but I've also already told you that if you have a bunch of people OCing to 4.5 GHz, you're going to end up with some kind of statistical distribution on both voltage, speed, AND temperature.

Considering that leeghoofd is able to say that 4.5 GHz (+/- 0.1 GHz) @ 1.3 V (+/- some value V) and running at 77.5 C (+/- 2.5 C) average just proves that. Granted, that's just his sample data point. I'm going to guess that he's not the only person that's ever OC'd to 4.5 GHz and so there's going to be more data and you can fully form the 3D statistical surface. (some higher voltage, some lower voltage, some faster, some slower, some hotter, some cooler). But it's going to be within some range on some kind of statistical distribution. (I hate stats. Grrr...)

You do realize that they make ATX compatible rackmounts right? And that the ATX spec only refers to the size of the motherboard, the mounting hole locations and stuff. For all practical intents and purposes of the discussion, I could have easily said "I'm going to mount it on ATX spec plywood". (I won't) but the point is that whether it's rackmount or tower - it doesn't matter. ATX spec does not govern what form the enclosure comes in. Therefore; the part where you said "use ATX specification part" has nothing to do with whether the people here are familiar with rackmounts or not. Here's a list of ATX compatible rackmount enclosures from Newegg (211 results): http://www.newegg.com/Product/Produc...=1&PageSize=20

re: "pretty darn close"
crash-sim-vid-overlay.jpg
(Not my work, NCAC model).

I'm trying to find the SAE paper or the AEI article comparing the physical crushing of a tube to the simulation and how close the simulation has gotten. (Not that you would actually care... - I mean...it only protects your LIFE in your car. Yes, I am assuming that you drive. Or been in one.)

[Bun-Bun]

I can do testing on 2600K's. I have three. However the 3930K I referenced earlier is a mission critical server that is not available for any other usage.

[alpha754293]

Bun-Bun: So? OC it anyways. (j/k. sorta. ) It won't hurt it if you OC it right?