Official Intel Ivy Bridge Discussion

Page 14 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
@jose: a $20-35 cm hyper cooler will be a massive upgrade over the stock intel cooler. i noticed that you keep mentioning load temperatures and room tempratures. you can calculate temperature delta from those figures. check the thread you opened, proximon explained it quite clearly. i hope you understood what delta meant.
 
Thanks for pointing that out, De5_roy. I didn't understand "delta" no.

-----

EXT64, you're probably right, . I was hoping that Intel's stock cooler would suffice - I've since heard it was "junk." Wish I could buy the CPU without it and save $20 or so. Now, I just wanted to confirm that an aftermarket cooler would be plenty enough and all I needed in those 85 degree Fahrenheit room temps. I was getting concerned with noise too with a hot CPU. I didn't want to create a noise issue in my attempt to deal with the heat issue. I figure I really don't have much of a choice - I have to get something like an Hyper 212+ and really good paste and a case with good air flow.

I don't have as much money as I did a few months ago as things have 'come up.' I'm getting nickle & dimmed every time I turn around and it's adding up to a couple hundred bucks now.

Will the Antec 302 suffice?

Antec Three Hundred Two
http://www.newegg.com/Product/Product.aspx?Item=N82E16811129180

http://www.youtube.com/watch?v=ubX0pKrtZiI
 
I've never used that exact case, however it seems to be a much improved version of the Antec 300, which I have used for numerous builds. I have an i7 970 and 7970 in my 300 and it does a great job cooling them. And remember, Ivy is a low power chip. I have had my brother's 3770K (at 4 GHz, undervolted) running Folding at Home 24 hours a day and I can hardly feel any heat coming out the back fan. It is amazing, I have never felt a high end CPU put out so little heat. Current temps are right around the 50-55C mark. As my ambient is about 8-10 F below yours, I would guess my setup would be running around 55-60C in your environment.
 
^ Oh good, thanks for that, EXT64. I would be fine with those temps. Was that on Intel's stock CPU cooler or an aftermarket? How much noise was there?

Antec pretty much says that the 302 case is an updated version of the 300.
 
That is on a Hyper 212+ I have two fans on it so that they can run slower. Listening to it now, the computer is not silent, but it is definitely not loud. Additionally, I am not sure how much of that noise is the CPU Fans and how much is the 7 (I think ) case fans). And remember, this is running at full load with Hyperthreading, so it is quite a bit hotter than an i5 would be or an i7 would be under more normal loads. With the overclock and undervolt, I think I am running pretty close to the same power and temp as I was at stock, however I am not certain.
 
I plan to use it for my research simulations (some CFD, structural FEA) and Folding at Home. I've got 4 processors on the way, but my immediate plans are fabricating a case for it out of a Norco 470. I plan to post pictures of the build, and of course Folding results, in the Tom's Folding at Home thread.
 


What work would you be using the IVB CPU to do that would have it under full load long enough to warrant needing an aftermarket cooler anyway?

From my understanding an Intel stock cooler is more than sufficient with a good case with sufficient airflow (one active read 120mm fan or so) to manage any workload you could throw at it ... unless your overclocking the CPU and raising the Vcore beyond what has been previously stated.

I just want to point out that there is no need (for stock clocks) to purchase an aftermarket cooler with any Intel CPU after the introduction of the core2 line ... only the latter model Netburst Intel CPU's suffered from thermal throttling under moderate or higher loads with high ambient temperatures.

I thought it important to point this out ... as spreading FUD about IVB is the last thing we need here in the IVB Sticky, or forums in general for that matter.

People purchase aftermarket coolers for:

Overclocking
Noise reduction
Longevity of the CPU ... which is plain silly given it will be old tech by the time electron migration causes instability anyway
Aesthetic reasons
Because their friends did


:)
 
^ thanks for your thoughts, REYNOD.

We work with HD video since we have to make our own product description videos and documentary DVD's and much more. It really does get up to 85 degrees Fahrenheit ambient INDOOR room temps (30 celsius) here since we don't have central air or air con while it's 105 Fahrenheit (41 celsius) outdoors for much of the summer. I'm trying to prepare for the worst case scenario since this new build will be an expensive one to us. We need a long life-span for this build for this much money.

I'm basing these views off of the experience of others many of them from here at Tom's:

"Using the stock heatsink in 80F ambient temps will affect over CPU temps. It may not overheat and shutdown on you, but if the surrounding air that is being pulled into the case is warm/hot then it will inevitably bring up the temps of the chip. Going with an aftermarket cooler would help and water cooling would be even better, but you still need to remember that they will all be sucking in hot air."

People are seeing temps reach 70c, 80c, 90c and beyond and that is normal ambient room temps of around 70 degrees Fahrenheit (21 celsius), which is why I was concerned about 85f INDOOR temps.

Then, I was getting concerned with noise too due to 85f indoor temps. I didn't want to create a noise issue in my attempt to deal with the heat issue as others here have brought to my attention:
"Intel stock coolers are basically junk."

"If you are concerned about noise, DON'T use the stock cooler."

"Keep in mind anything >60C+ is going to be making a lot of noise even on the H100..."

"To 'me' and most Hot or (too) Warm translates into NOISE to cool the CPU, and IF it takes all of the efforts of say an H100 or Noctua NH-D14 to keep the IB (K) cool then IMO the IB (K) will be DOA to most enthusiasts who have any experience with OC'ing because no one wants Earmuffs playing BF3 or whatever...>55C~60C is when things start getting loud."
"From my understanding an Intel stock cooler is more than sufficient with a good case with sufficient airflow..."
REYNOD, you may be right but, nearly everyone says not to waste anytime with it - they say to just get an aftermarket CPU cooler and be done with it.
"spreading FUD about IVB is the last thing we need here in the IVB Sticky, or forums in general for that matter"
I don't think anyone here is purposely spreading FUD (whatever that is - I assume false info). This issue of heat with the Ivy Bridge CPU's comes from nearly every benchmark test review I've seen. If somebody wants to put it to rest once and for all they could perform the benchmark tests with Intel's stock CPU cooler and in 85 degree Fahrenheit ambient INDOOR room temps (30 celsius) and post the results here.

Or perhaps Intel could use better quality TIM paste? - which could apparently help drop CPU temps by up to 20c.

"Ivy Bridge proven to suffer from poor thermal grease"
http://vr-zone.com/articles/ivy-bridge-proven-to-suffer-from-poor-thermal-grease/15844.html
 
I am running a stock cooler on a 2500k at work (same stock cooler as Ivy uses), and even in cool rooms, it gets quite audible at full CPU loads. I can't remember CPU max temps, however I am sure they were adequate (though not great). However, as that computer is rarely if ever loaded fully on all 4 cores, the cooler is acceptable. I would say (my opinion), if you plan to use this computer where it often will have full load (all 4 cores) and you like your computer quiet, then get an aftermarket cooler.

As a side note, these chips are really resilient to high temps. I recovered an i7 920 from a build at work a while back. I found that it had been overheating (throttling at 100C when fully loaded) due to a faulty Intel Stock HSF and had too high of memory voltage for 2 years. The motherboard even died, yet when I pulled that chip out, put it in a new board, and OC'ed it to 3.2 GHz it ran great, and has continued to run great for 6+ months.
 
They're saying, themselves, "IB sucks, but the only people that are reading this are people that already know that, so you guys go ahead and get SB while they're available, and we'll market the "better performance" crap to the "mainstream" market". :lol:
 
Personally, I find extreme OCes a bit pointless. Going from 3.6 GHz (stock) to 4.5 is 25% and can be done at stock volts, while going from 4.5 to 5.0 is only 10% and requires high voltage (even on SB) and thus very high power consumption. But to each their own, and if you must reach 5GHz, then get SB. I more prefer reasonable OCes (4-4.3) and undervolting to get a perf/watt king.
 
Also, I saw that the temps "weren't a problem past 4.5GHz or so"... but were they still hotter, right? Or were they about the same at that point, and IB got hotter as the OC went on?
 


+1 and I'm running both my machines at 4.7 24/7 without a problem.
The first machine 2600k and the second 3770k.
 
My point regarding stock clocks and the boxed air cooler still stands ... look ... the whole point of the tock this time round was to optimise the cpu for the mobile spacem and to compete with Trinity.

The desktop variant is vanilla for the masses ... slight improvements across the board.

Intel has no intention of pushing beyond 4Ghz ... and performance improvements for the foreseeable future hinge on IPC improvements / cache / cores / improving the GPU onboard.

since the core is now quite small they have plenty of opportunity to increase the real estate.

Gee ... I even sound like a tech writer from the Inq ... sorry bout that.


 


No, not really, just in extreme OC's. Assuming you are comparing IB vs. SB as equivalent to Bulldozer vs. Phenom 2, remember that you have to put it in perspective: (1) this is the first IB stepping - SB only averaged around 4.5-4.7 intially too, and (2) IB does have a small IPC increase over SB, whereas BD was typically lower than P2, and (3) IB is more efficient (less energy used) than SB; BD was a power hog under load compared to P2. So adding up all the pluses for IB, including Quick Sync 2, PCIe 3.0, etc, and contrasting the one minus of not being as excreme an overclocker as the mature SB is, then the choice is simple - stay with SB if you plan to oc at 4.5GHz or higher (unless you can get a good IB chip like Earl45 apparently did). For all else, including price considerations, go with IB.

When comparing BD to P2, most people made the opposite choice. So that's why I say Intel did not pull an AMD with IB 😀.

BTW, I am leaving shortly while on my lunch hour to go pick up a 3770K for $290, an Asus P8Z77 Deluxe mobo for $275 and 16 gigs of Corsair Vengence DDR3-1866 CL9 for $150 total. Hafta see what coolers they have, and also will see if they have a GTX-680 sitting around waiting for some lucky buyer, but not holding my breath for that last one. Methinks NV did a paper launch on those 680's 😛..
 
Its a little more complicated than your facts make out, you can prove anything with facts, that's way too easy :kaola:

With BD you need to take into account that the motherboard was out a ways before the chip was so I would think a lot of those that "decided" to go with BD did so because they already had the Motherboard.

If we go back to the silly facts for a moment we also need to take into account that you can get SB chips just about anywhere. the same cant be said of the P2 chips which AMD removed from the channels pretty fast.
For clarity I am in the UK and can only say it as I see it from here, US and other countries may well have had a different transition.

Personally given a free choice I would rather have a P2 or a SB if picking from either side.

Mactronix :)
 
But fazers, comparing SB to IB would HAVE to be like comparing Phenom I to Phenom II, since it was only an arch tweak. Comparing PhII to Zambezi would have to be like comparing IB to Haswell (which, we sure will).

I agree with Mac that Intel actually focused a little more in the graphics department than the CPU itself, just like AMD did on Llano vs Phenom II (tweaked K10 and included an iGPU) or Trinity and Llano (IB to SB, since they're both APUs, like it or not).

Cheers!
 
Intel's Tri-gate is a trapezoid
Platonic solids

Reverse engineering outfit Chipworks has posted microscope cross-sections of parts of the 22-nm Ivy Bridge processor from Intel.

It shows that the much touted FinFETs, which Intel calls tri-gate transistors, are in fact trapezoidal. Chipworks broke up 64-bit, four-core Xeon E3-1230 CPUs intended for the server market, which Chipworks bought in Hong Kong.
What is odd is that Intel appears to have moved away from a rectangular section which it was showing in 2011.

In a statement Gold Standard Simulations has also waded into the debate. Its CEO Professor Ase Asenov said that there is speculation about the possible advantages and disadvantages of the trapezoidal, or almost triangular, shaped 'bulk' FinFET." GSS has performed a simulation analysis of the FinFET using its statistical 3-D TCAD simulator called Garand. Its simulation looked at the dependence of threshold voltage on gate length for the trapezoidal Intel transistor and an equivalent rectangular-fin transistor.

He said that the rectangular fin has better short channel effects. Still, the million-dollar question is if the almost-triangular shape is on-purpose design, or is this, what bulk FinFET technology can achieve in terms of the fin etching? We will only know when Intel lets people come up and see its etching.

now my ignorance is wondering; could the not as good "short channel effects" lead to the symptom of heat when raising voltages? and which way is the shape, towards the point? (x,y) or Z? (flat or vertical)
 
Sandy Bridge | [strike]Ivy Bridge[/strike] | Haswell...

" We want to emphasize the phrase “typical use case” here because the more you overclock, the further away from typical you get.
While we can confirm the claimed power reduction for typical use cases (just wait until we get to the benchmarks), at this point,
it is not entirely clear how these transistors perform at substantially increased clock rates.
Maybe they're just not optimized for the speeds an overclocked chip pushes yet. It might be necessary to wait for Ivy Bridge's successor, Haswell,
to see if the limits of 22 nm manufacturing can be pushed harder. "

and my favorite part:

" Recommendations

We state without any hesitation: air-cooled Ivy Bridge-based processors cannot be overclocked as much as Sandy Bridge-based processors. Overclockers hunting for the latest and greatest overclockable processor, yearning for high-frequency overclocks, should keep that in mind. Maybe a Sandy Bridge-based chip is still the best choice,
even in a world where Ivy Bridge exists.

If scalability isn't as big of a worry for you, Ivy Bridge is the more natural choice. Its performance per clock is a few percent higher, so long as you're looking at the same frequency from both architectures. After all, a 4.5 GHz Ivy Bridge-based CPU wins benchmarks against a Sandy Bridge processor at slightly higher clock rates.
When you limit your overclock of a chip like the Core i7-3770K to 4.2 or 4.3 GHz, you’re completely on the safe side.
There is no temperature issue, and performance remains impressive.
Then again, such a system won’t be significantly faster than a machine running at its stock clocks. "

Is this from intel's own website?