Haswell reviews are out. As 'bad' overclocker as IvyBridge

[Tabs 2,790]

This is probably my longest run without an upgrade. I have an I7 2600K with a GTX480. I'm feeling good about the CPU. I've been tempted to upgrade my video card but I've been told that I wouldn't see a huge performance boost going to say a 680. I guess I'll keep my wallet in my pocket for a while longer.

 

Yeah, I'm on an even older one processor wise (i7 860 @ 4.0 with a GTX570)  and I still really don't see a reason to upgrade here. I was thinking about getting a new system later this summer but I might just get a new GPU (probably the 770) and some more RAM now after reading the Haswell reviews that came out today. Unfortunately I think we're starting to hit a fundamental wall with the speed of these things - you can only push electrons through these circuits so fast...

[ChaoticBeauty 6,692]

While I understand that Haswell did not need a big performance increase because of smaller demand nowadays, but it's too bad that they didn't fix the bad overclockability of Ivy Bridge. It looks as if they actually want to hinder overclocking. Let's hope Broadwell does better.

 

On a positive note, this might give AMD some breathing room to try and make their CPUs more competitive. Steamroller should come out sometime in 2013.

[martin-w 3,234]

 The 3770K or the 4700K would need to max out (4.4 to 4.5) with real heat issues to keep pace.

 

 

Kind regards,

 

Well no, not at all in respect of the 3700k.

 

With a Noctua NH-D14, mine runs very cool at 4.5GHz. So no, no heat issues at 4.5.

 

It's above 4.6, or above about 1.34 volts that mine hits the thermal brick wall and temps shoot up.

 

 

This still puzzles me. I know, the what, at least with IB, but the why escapes me. If Intel has an engineering reason, I can't fathom it. If it is a marketing reason, I can't figure that out either.

 

It may well be that slapping a chunk of TIM under the IHS, is cheaper than going to the trouble of soldering. Money is perhaps the motivator.

In this day and age anyone who gets a new computer for FSX should be getting a 3rd party water cooler (Corsair is a great brand). If you are serious about overclocking then air cooling isn't the way to go.

 

Sorry to disagree, but that's not true at all.

 

Those of us with an NH-D14, will tell you unequivocally that it cools beautifully. As well, or almost as well as the all-in-one water solutions, especially when you consider fan noise.

 

And in fact, the D14 also matches the latest Corsair H110.

 

The NH-D14 runs my Ivy Bridge in the 50's in FSX. That's at 4.5 GHz.

 

And more importantly, the D14 is very, very quiet. Many of the all-in-one solutions are noisy.

 

Don't write off air cooling just yet.

 

 

2600K/2700K cannot be beat by anything out there unless perhaps you could run TWO of them on a server board...  My opinion.

 

I disagree.

 

You are forgetting that Ivy Bridge is a faster architecture. You are also forgetting that Ivy Bridge can be de-lidded and pushed as high as 2600K/2700K. And when you do that, the superior Ivy Bridge architecture renders it faster.

[MarkW 1,490]

Looks like we will have to wait for someone to delid and see what kind of performance they can get out of it.  If delidding gets you 5Ghz then it may be worth it if you have steady hands and nerves!

 

By the way my 2500K runs at 4.9 on the D14 with very acceptable temps. 

[Jim Young 2,356]

-as opposed to the speculation of super chip status with 7-8Ghz and massive increase in performance per Ghz. What a difference between what was hoped for vs what apparently will be delivered. Still it's better than no improvement whatsoever...or is it?

Hi Stephen!  I never did think much of the 7-8GHz overclocking but I was hoping for an easy 5GHz overclocking.  That may be still possible and I'll wait and see what further reviews have to offer.  I saw that most of the reviews were done quickly and with little or no overclocking and I may change my mind if they can be easily overclocked around 5GHz and 5.4GHz. 

 

I saw too that Intel didn't listen to those who removed the Integrated Heat Spreader (de-lidding procedure) from the cpu (as shown in the pinned topic here in this forum).  It still has it in the Haswell products.... 

 

Best regards,

Jim

 

 

 

 

 

[idahosurge 2]

The below is quoted from Page 6 of the Hardocp review, these are Asus OC tests and their findings.  One in ten chance of getting a CPU that will hit 4.8GHz and be stable.

 

It seems that out of the box a high overclock (close to 5.0GHz) on air with Haswell is a bust.  They also state that the higher your OC the lower your memory speed may have to be.  So whats the use of 2400HMz memory if you can only run it at 1600MHz!

____________________________________________________________________

Luck of the Draw?

From all the feedback that I am getting through motherboard makers about Haswell overclocking, the basic thread at this time in terms of getting "high overclocks" seems to be, "You need to have a good CPU." ASUS has tested a couple hundred Haswell processors at this time and this is ASUS’ specific feedback from that overclock testing.

 

• 70% of CPUs can clock to 4.5GHz
   
• 30% of CPUs can clock to 4.6GHz
   
• 20% of CPUs can clock to 4.7GHz
   
• 10% of CPUs can to 4.8GHz
   
  Overall you will find most CPUs capable of reaching 44x to 45x with varying levels of voltage.

 

These ASUS results were obtained with sealed water cooling systems that are comparable to a Corsair H80/H100 configuration or extremely efficient air cooling with 120mm push/pull fans while applying a maximum core voltage of 1.275v under full thread load conditions.

 

Obviously it is very early in terms of Haswell retail processor production, and we are still not sure what we will see in terms of i7-4770K retail purchased Haswell overclocking headroom. We will be buying a couple of Core i7-4770K processors this weekend and see what comes out of that. Going on the above information, it would seem that getting a 4.8GHz-worthy Haswell sample will be about a one in ten shot with "normal" sealed water cooling systems.

 

Overclocking the Haswell i7-4770K processor that was sampled to HardOCP directly from Intel was extremely simple using two ASUS Z87 series motherboards and we will describe on that on the following pages. From the information above we can assume that we got a "golden" sample directly from Intel, or were extremely lucky when it came to the luck of the draw.

All about the Cooling

As most of you reading this have already guessed, overclocking Haswell is going to be very dependent on your cooling system. Many of you that have put off water cooling in the past are very likely to be about to take the plunge. If you are going to stick with air cooling, you will likely need a very efficient system. ASUS went as far as to make cooler recommendations based on its overclock testing, which at this time I have to consider it the most informed opinion. Given the wide variety of voltages ASUS has seen at various clock speeds, it made some recommendations based on CPU core voltage needs.

 

For voltages up to 1.250-1.265 a Cooling solution meeting a minimum of a Corsair H80i is advised, superior performance can be offered by moving to the H90 or higher performing dual fan closed loop solutions.

 

For voltages up to 1.275-1.300 a cooling solution meeting a minimum of a Corsair H100i is advised.

 

 

For voltages up to or greater than 1.300v a high performance water cooling system is recommended. Minimum recommendation would be a unit such as Koolance EX2-755.

 

For voltages up to or greater than 1.350 a high performance water cooling system is recommended. Minimum recommendation would be a unit such as a Koolance EX2-1055.

 

Keep in mind these recommendations are based on controlling peak temperatures and loads under synthetic stress applications. For nominal real world usage load temperatures will be considerably under that of stress tests.

The last paragraph above is very interesting about synthetic stress applications and I will be addressing this more on the next page.

 

To put it in a nutshell, ASUS’ experience was that any core voltage over 1.2v accompanied by heavy workloads puts the enthusiast into water cooling territory and ASUS has experienced certain 4770K sample CPUs that have required 1.2v vCore just to hit 4.3GHz. From what we have seen, "high end" overclocking will require a good water cooling system. Surely there will be some exceptions to the rule, but I would not expect to be part of that crowd.

_______________________________________________________________________________________________

 

From the same article, Page 6 regarding memory speed.

 

Your specific processor may be capable of doing 4.8GHz, but only with a memory clock speed of 1600MHz, or luck of the draw might allow you to do 2400MHz with a 4.8GHz CPU clock.

___________________________________________________________________________________________________________________

 

Link to the Hardocp article

http://www.hardocp.com/article/2013/06/01/intel_haswell_i74770k_ipc_overclocking_review/6#.UatFUZyboxI

 

It looks like I will be skipping Haswell like I skipped IB.  I guess I will wait and see what IB-E does!

[Kaboki 437]

 

 

He was getting temps of 80-90 degrees Celsius on full load, which is admittedly high for any CPU, but not bad at 5.6.

 

Hmm, how long would a chip last with temps like that? Reason I'm asking is because I'm so terrible at overclocking and I have the same temps on load with a 4.5Ghz overclock..how do I get lower temps..?

[idahosurge 2]

Hmm, how long would a chip last with temps like that? Reason I'm asking is because I'm so terrible at overclocking and I have the same temps on load with a 4.5Ghz overclock..how do I get lower temps..?

 

Better cooling, lower CPU voltage.  If you can not lower your CPU voltage and be stable then better cooling.  Also make sure that the case you have has great airflow.  Poor airflow in a case can greatly affect your temps.  Dust build up in your case, on your PCB's and on your HSF (or radiator if you are on water) can also greatly affect temps.

[Kaboki 437]

Better cooling, lower CPU voltage. If you can not lower your CPU voltage and be stable then better cooling

 

Thanks for your suggestions, I couldn't lower the volts without getting unstable, so I guess I have to look into better cooling solutions. I haven't de'lidded the chip yet as I'm afraid to screw things up but maybe I have to look into that too...

 

Sorry for derailing the thread....back to the haswell ^_^

 

 

Edit: Currently have the Corsair H60 Hydro series, would upgrading too Corsair H100i help?

[Kattz 132]

I disagree.

 

You are forgetting that Ivy Bridge is a faster architecture. You are also forgetting that Ivy Bridge can be de-lidded and pushed as high as 2600K/2700K. And when you do that, the superior Ivy Bridge architecture renders it faster.

 

A buddy of mine likes to try to out-do everything I do with my computers.  He de-lidded his IB 3770K CPU and cranked up the speed to match mine.  At 5.15GHz he melted part of the ZIF socket on his mobo.  Still has too much heat.  2600K/2700K still rocks.  No argument about architecture efficiency, the IB is improved over SB.  But I don't have to dink with my CPU to make it run very VERY fast.

[martin-w 3,234]

A buddy of mine likes to try to out-do everything I do with my computers.  He de-lidded his IB 3770K CPU and cranked up the speed to match mine.  At 5.15GHz he melted part of the ZIF socket on his mobo.  Still has too much heat.  2600K/2700K still rocks.  No argument about architecture efficiency, the IB is improved over SB.  But I don't have to dink with my CPU to make it run very VERY fast.

Well, "one" persons experience isn't really relevant, simply because all CPU's vary, some are good, some are bad overclockers. Average results are relevant, not one persons results.

 

Just because your friend failed, doesn't mean everyone else fails too. There are many that have successfully de-lidded their Ivy Bridge CPU's and achieving 4.9 - 5Ghz. And those individuals will be out performing your 2600K/2700K at the same frequency.

 

In addition, to "match" your 5.1 GHz, your friend didn't need the same frequency. As Ivy is a faster architecture, 4.8 would have been enough to match your 5.1 GHz in terms of performance. And most de-lidded Ivy Bridge CPU's achieve that frequency. So the statement that the 2600K/2700K remains superior is inaccurate.

 

It's also important to remember that there are memory advantages with Ivy Bridge.

 

But I don't have to dink with my CPU to make it run very VERY fast.

True, you don't have to delid your CPU... However, fast isn't just about frequency. As we said, Ivy is a faster architecture, and with memory advantages as well.

 

Therefore a 3700k NOT de-lidded running at 4.6 GHz, at just 1.3 volts, is only a tad less powerful than your 2600K/2700K pushed to the limit with very high voltage running at at 5.1Ghz.

 

And when you consider that just a few hundred megahertz, equates to only one or two frames per second in FSX, it's hardly anything to be too impressed about.

[SAAB340 28]

The reports of poor memory overclocking is not sounding promising. The memory overclocking helped IvyBridge to get an edge over SandyBridge whilst the IPC improvements made up for the lower clockspeed.

 

We need to remember that we are still in early days of both the manufacturing and people's experience overclocking the chip. It might still improve.

 

However at the moment it seems like we won't get any significant improvements when it comes to FSX performance for us regular overclockers for quite a while. Broadwell is only to come soldered to the motherboard so it won't be for us. Skylake, the next socketed CPU architecture would be at least 2 years away from now. That would mean that performance has been stagnant for 4 and a half years from when SandyBridge launched. Go back 4 and a half years before Sandy and we end up in Pentium 4/Athlon64 days. There was a lot of steps with significantly increased performance between P4 and Sandy...

[idahosurge 2]

Quote from Page 10 of the HotHardware Review.

"We should also note that 4th Gen Core processors are using the same TIM between its integrated heat spreader and the surface of the chip as Ivy Bridge, which isn't terribly high performing."

 

A link to the review.

http://hothardware.com/Reviews/Intel-Core-i74770K-Review-Haswell-Has-Landed/?page=1

 

In the next couple of weeks we should see what these can do with the tim replaced.

[Kattz 132]

Well, "one" persons experience... ...anything to be too impressed about.

 

We will just have to agree to disagree, won't we?  All I know is that I don't struggle to get 30fps consistently and sharp performance.  It wasn't that way originally, but it is now.  And in case you didn't notice, I did concede your point that the architecture used in IB is better and more efficient.

 

I've got a pure hoss of a system sitting here on the desk.  It's going to take more than what Haswell and IB brings to the table to make me spend my $. 

 

Peace out, to each their own.

[SpiritFlyer 987]

All I know is that I don't struggle to get 30fps consistently and sharp performance.

Yes sir! Well said.