Building for gaming 2500k or 3570k Ivy

[2is]

Not the same as any other intel cpu change.


the die's transistor shrink is the biggest ever percentage wise in one swap. 32nm to 22nm

Still doesn't invalidate my point. To simply assume there may be problem based on nothingness is baseless. There's nothing to suggest the shrink is going to cause failures in a year. You have as much evidence as me claiming the shrink is good for a minimum of 5 years. Actually you have less.

 

[blastingcap]

Still doesn't invalidate my point. To simply assume there may be problem based on nothingness is baseless. There's nothing to suggest the shrink is going to cause failures in a year. You have as much evidence as me claiming the shrink is good for a minimum of 5 years. Actually you have less.

Speaking as a i5-3570K owner I hope there are no problems, but he does have a valid point in that if you don't buy the bleeding edge but rather, bleeding edge minus 1, you are less likely to run into problems. Let others do the beta testing of a product for you. That's actually kind of true when it comes to tech.

The downside to this is that if you do that, you will usually be behind the curve on performance, and possibly efficiency and perf/watt and perf/dollar as well. (In the case of SB vs IB this is unclear because other than the iGPU they aren't THAT far apart in any metric.)

In this case, I was willing to gamble that Intel wouldn't have another foul-up. I would be more nervous if it were an AMD CPU or something, but Intel runs a pretty tight ship when it comes to reliability. Besides, nothing in life is risk free, and sometimes ya gotta live a little.

 

[2is]

If you're talking about probability sure it's mathematically possible. But what about odds? How likely is it that intel developed a CPU that's going to fail in a year? Doesn't really sound like much of a gamble to me.

 

[philipma1957]

Still doesn't invalidate my point. To simply assume there may be problem based on nothingness is baseless. There's nothing to suggest the shrink is going to cause failures in a year. You have as much evidence as me claiming the shrink is good for a minimum of 5 years. Actually you have less.

No evidence is correct as I state in my post. But there are laws that govern electrical flow.

If you're talking about probability sure it's mathematically possible. But what about odds? How likely is it that intel developed a CPU that's going to fail in a year? Doesn't really sound like much of a gamble to me.

Electrons move and make friction thus heat.

Since electrons don't shrink Small trans mean more friction.

Fact ^

Some limit exists as for flow > vs heat is it 22nm is it 16nm is it 10 nm.

every year to 14 months a new cpu comes out.

Buying the ivb for overclocking is fine if you don't mind it aging too quickly.

Lots of us can afford to do it. I may buy one this Summer / Fall.


For now how much better off is someone with an ivb + mobo like the asus maximus v vs a sb + mobo like the asus maximus iv.

I will hold on to the discounted i5 2500k + mobo I picked up this year for the next 18 months or so.

 

[2is]

Except I'm not arguing the facts. Ok smaller transistors means more heat. What are the odds of it failing in a year? Much ado about nothing.

 

[Ferzerp]

Electrons move and make friction thus heat.

Since electrons don't shrink Small trans mean more friction.

I think you're a bit confused here. suggesting a "size" of an electron is pretty much meaningless. They aren't particles with dimensions.

 

[philipma1957]

thus the statement electrons don't shrink. They are a given factor and they cause heat when passing through conductors.

My original statement stands we do not have facts on wear and tear for ivb.

on apr next year we will know how they hold up.

since you can get an upgraded cpu every 10 to 16 months why take a chance on top of the line cutting edge cpu when it is not much faster then the prior model.

If you choose an ivb you can do it if the cash is not important.

I never said they would break

 

[Ferzerp]

thus the statement electrons don't shrink. They are a given factor and they cause heat when passing through conductors.

While technically you were correct in that an electron won't "shrink", it had everything to do with luck and nothing to do with actually understanding what was going on. If you realized that they don't really have dimensions, it's not like you would have brought up size in regards to an electron to begin with.

You're basing your entire view on some sort of model of electricity that you've made up in your mind.

I only skimmed the thread before replying, so I'm not really sure what your point is, but I felt the need to clarify a bit. Electrons are an elementary particle. We can't really discuss them like they're tiny little negative bb's and actually mean much of anything.

 

[2is]

thus the statement electrons don't shrink. They are a given factor and they cause heat when passing through conductors.

My original statement stands we do not have facts on wear and tear for ivb.

on apr next year we will know how they hold up.

since you can get an upgraded cpu every 10 to 16 months why take a chance on top of the line cutting edge cpu when it is not much faster then the prior model.

If you choose an ivb you can do it if the cash is not important.

I never said they would break

I agree you don't have facts on how it will hold up. So let's rely on common sense. Intels not going to develop a processor that's going to fail in a year. Your fears are completely baseless.

 

[2is]

thus the statement electrons don't shrink. They are a given factor and they cause heat when passing through conductors.

My original statement stands we do not have facts on wear and tear for ivb.

on apr next year we will know how they hold up.

since you can get an upgraded cpu every 10 to 16 months why take a chance on top of the line cutting edge cpu when it is not much faster then the prior model.

If you choose an ivb you can do it if the cash is not important.

I never said they would break

Double post

 

[AtenRa]

Everyone can get the $350 SB combo I got.. It was last week.. I'm sure it'll come up again soon. mine was from NCIX Online site.

i5 2500k + asus z68 pro g3 for $350 after $20MIR.

I cant get that combo and i believe everyone outside US as well.
If everyone in US can get that combo then US is the only market that SB may be considered against IB at this time (with this combo deal). In the rest of the world there is no point even thinking about SB over IB.

 

[philipma1957]

While technically you were correct in that an electron won't "shrink", it had everything to do with luck and nothing to do with actually understanding what was going on. If you realized that they don't really have dimensions, it's not like you would have brought up size in regards to an electron to begin with.

You're basing your entire view on some sort of model of electricity that you've made up in your mind.

I only skimmed the thread before replying, so I'm not really sure what your point is, but I felt the need to clarify a bit. Electrons are an elementary particle. We can't really discuss them like they're tiny little negative bb's and actually mean much of anything.

what is your model of electricity?

mine is water model.

http://www.google.com/search?client=safari&rls=en&q=water+model+of+electricity&ie=UTF-8&oe=UTF-8

I did not make it up.

transistor size would be pipe size. small pipes result in more restriction of electric flow. there is more stress on a smaller pipe.

I am not saying that ivb cpus will fall over dead in a year.

I am saying why risk using ivb for small cpu gains when you are over clocking. you only need wait to see if they are any good.


When ssd's came out they were an amazing improvement over hdds in iops. so the speed gain is worth the risk if you do a lot of iop based work.

the ivb is not 10 or 20 x gain over Sandy b. so waiting to see if they last is not a bad idea.

If you go out today you can buy this mobo

http://www.newegg.com/Product/Produc...82E16813131806 for 170 with the discount

http://www.newegg.com/Product/Produc...82E16819115072 buy this cpu for 204 = 374 total.

It is a known factor that oc this to 4.2 and a year later it will still work.


or buy this for 240-10 =230

http://www.newegg.com/Product/Produc...Tpk=i5%203570k


and this for 200


http://www.newegg.com/Product/Produc...mus%20v%20gene

total 430 and no long term data on reliability with a 5% to 10% improvement in performance.

Mind you if you own Sandy now you can wait and see if oc'ing the ivb to 4.2 is a killer of the chip.

My guess is the ivb will proof less reliable then the sb will when you oc it.

BTW what is Intel's warranty policy on a chip that dies from oc.

(I know what it is and if need be Intel could be restrictive about rma'd chips)


This is speculation but sooner or later this will be true if not on a 22 nm chip then on a smaller chip. Or intel will just stop making k chips.

 

[rgallant]

just saying

or the 3rd option for me is to keep my 3 year old 920\E760 , cost =$$00.00

-and is still better[bang for the buck] for multi gpu's than your 2500k [even @$$350.00] and why I did not waste money buying sb last year , so why would I buy it this year when it does not have PCI-E 3.0 a feature some of us want for the next round of cards..

 

[2is]

Philipma1957, my 3770k is at 4.2Ghz. I'll bet you a Haswell CPU it will be fine a year from now.

 

[philipma1957]

Philipma1957, my 3770k is at 4.2Ghz. I'll bet you a Haswell CPU it will be fine a year from now.

I would most likely lose the bet!

 

[borisvodofsky]

just saying

or the 3rd option for me is to keep my 3 year old 920\E760 , cost =$$00.00

-and is still better[bang for the buck] for multi gpu's than your 2500k [even @$$350.00] and why I did not waste money buying sb last year , so why would I buy it this year when it does not have PCI-E 3.0 a feature some of us want for the next round of cards..

LOL, it "does" have pci-e, the combo includes an asus z68 pro g3


LOL, n000bs

 

[borisvodofsky]

I cant get that combo and i believe everyone outside US as well.
If everyone in US can get that combo then US is the only market that SB may be considered against IB at this time (with this combo deal). In the rest of the world there is no point even thinking about SB over IB.

friends in America :thumbsup:

people come and go from the USA all the time..

 

[Renegade_09]

Just put together a i5 3570k with my Hyper 212+ CPU cooler. I used the thermal paste that came with the cooler. Not sure if it made difference but I'm getting pretty high temps. This was my first build but I have taken apart a whole computer and put them back together before. I've also replaced parts here and there when my components die. Hopefully it isn't a install error.

Here are my temps using HWMonitor.

 

[blastingcap]

Just put together a i5 3570k with my Hyper 212+ CPU cooler. I used the thermal paste that came with the cooler. Not sure if it made difference but I'm getting pretty high temps. This was my first build but I have taken apart a whole computer and put them back together before. I've also replaced parts here and there when my components die. Hopefully it isn't a install error.

Here are my temps using HWMonitor.

Congrats on your first build!

What method of paste spreading did you use (e.g., pea in the center, triple-stripe, pea-with-triple stripe, finger-in-plastic bag, etc.)? What is the ambient air temperature outside of your case, in your room? What case are you using and additional fans if you installed any? what are your peak temperatures on maximum Intel Burn Test /LINX?

The paste that comes with the 212+ is pretty decent btw so that is probably not the culprit unless you used too much or little of it.

 

[Renegade_09]

Not sure what the temperature in my room but I'll say it's pretty hot compared to the rest of my house. I think it's because of the computer and other electronics I have in the room.
I'm using the two stock 120mm fans in the Fractal Design R3 no additional fans.

Well I reapplied the thermal paste but forgot to take a picture of it. I used the pea in the center method. Still got the same from what I can tell.

 

[blastingcap]

Not sure what the temperature in my room but I'll say it's pretty hot compared to the rest of my house. I think it's because of the computer and other electronics I have in the room.
I'm using the two stock 120mm fans in the Fractal Design R3 no additional fans.

Well I reapplied the thermal paste but forgot to take a picture of it. I used the pea in the center method. Still got the same from what I can tell.

You may want to use the 3-line-and-a-small-ricegrain method. The idea being to fill in the gaps between the pipes, with a little center stuff to help push out air bubbles.

http://www.overclock.net/t/632591/cooler-master-hyper-212-plus-club

You don't want to use too much, so just a small mini-line for each gap on the cooler, plus a tiny dot in the middle of the heatspreader, and that's enough. Even if you do it perfectly you are unlikely to see a massive difference between that and the traditional "dot" method, though.

An R3 at 2x120mm should be decent, plus the 212+ comes with a 120mm fan as well. I'm assuming you aren't running a GPU load during your CPU tests so a side fan should be irrelevant.

I think the real culprit is your room temperature; you may want to measure it more exactly than "pretty hot." I recently had to move my computer from a reasonably well-ventilated area to one that wasn't as well ventilated but still pretty good, and the difference was a few degrees C just from that alone. If I moved it into a room with a bunch of other hardware that was "pretty hot," I'm sure the difference would have been even more.

 

[Renegade_09]

Thanks I'll try that out tomorrow.
What I meant when I said it is pretty hot is probably because of the new computer. If I have everything turned off my room is a couple a degrees higher than room temperature.

Also those temps are just from browsing the net and installing Diablo 3.

 

[Renegade_09]

Well I just tried reapplying the thermal paste and ran out of the stuff paste mid-way lol. So I tried some other thermal paste I had lying around. It was from Vantec. Not even sure where I got it from but it's a white paste compared to the grey thermal paste that came with the cooler.

I tried filling all the gaps between the pipes and put a two blops on the cooler then mounted it on the CPU. This stuff was harder to work with but after I started it up it's running cooler. Not the 25 degrees C I was hoping for.

 

[blastingcap]

Looks normal to me. Pretty similar to what I get with my 3570K + 212 Evo at idle.

 

[Ferzerp]

Idle temps are pretty useless.

The measurement is less and less accurate the farther from tjmax you get (as evidenced by many idle "temperatures" that people cite which are under ambient temp).

Run IBT, see where the temp peak goes, and that's the only temp number you need to care about. Just make sure to run the longer runs. Mine actually peaks on the short runs, but my cooling is quite a bit different than yours.