Intel Core i7-8700 Review: Stock Cooler Falls Flat
-
Page 1:The Core i7-8700 Review
-
Page 2:The Stock Cooler Dilemma & Test Setup
-
Page 3:VRMark, 3DMark & AotS: Escalation
-
Page 4:Civilization VI Graphics & AI, Dawn of War III
-
Page 5:Far Cry Primal, GTA: V & Hitman
-
Page 6:Shadow Of War & Project CARS 2
-
Page 7:Office & Productivity
-
Page 8:Rendering, Encoding & Compression
-
Page 9:Final Analysis
The Stock Cooler Dilemma & Test Setup
The Stock Cooler Dilemma
Intel ships its multiplier-locked CPUs with basic thermal solutions. Although they are notoriously flimsy and usually pretty noisy, they've proven sufficient for cooling previous-generation models. That changes with Core i7-8700.
In the not-too-distant past, Intel's stock heat sink employed a copper core. More recently, however, the company switched to all-aluminum designs. That means Core i7-8700 comes with the same cooler as quad-core Kaby Lake-based Core i5s.
Frankly, we're surprised that Intel carried over the same heat sinks from those seventh-generation Core CPUs. Despite the -8700's 65W TDP, it's still based on a notably more complex die.
On paper, the -8700's 65W TDP fits nicely within the low-profile cooler's 73W rating. But remember that Intel specs the CPU's thermal design power according to its base frequency. Its chips actually exceed the TDP when they dynamically increase voltage and frequency through their Turbo Boost algorithms. As noted on Intel's Turbo Boost 2.0 informational page:
Note: Intel Turbo Boost Technology 2.0 allows the processor to operate at a power level that is higher than its TDP configuration and data sheet specified power for short durations to maximize performance.
According to our measurements, Core i7-8700 peaks at up to 126W during taxing all-core workloads. With that data in-hand, the stock cooler does appear insufficient.
While Intel guarantees base frequencies during normal operation, the company doesn't make promises about Turbo Boost clock rates because its processors only shift to higher P-states (pre-defined frequencies and voltages) when they're running below certain temperature, voltage, power, and current limits. Above them, the opportunistic algorithms are reigned in to keep the CPU in-spec.
As a general rule, Turbo Boost targets lower frequency bins as more cores become active. Intel does still advertise its maximum single-core clock rates, but it no longer divulges the multi-core clock rates (even though you can expose them through the company's XTU software).
| Frequencies | Base | 1 | 2 | 4 | 6 |
| Intel Core i7-8700K | 3.7 GHz | 4.7 GHz | 4.6 GHz | 4.4 GHz | 4.3 GHz |
| Intel Core i7-8700 | 3.2 GHz | 4.6 GHz | 4.5 GHz | 4.4 GHz | 4.3 GHz |
We've seen lots of speculation that Intel stopped disclosing multi-core Turbo Boost frequencies with Coffee Lake-based processors because its stock heat sink and fan couldn't fully facilitate those clock rates. Sure enough, a quick online search reveals several reports from owners claiming that their Core i7-8700s hit the maximum safe temperature of 100°C (TJ Max) during extended workloads. Once the processor reaches TJ Max, it throttles back voltage and frequency (along with power and heat) as a protection mechanism. Of course, throttling also results in lower performance.
Measuring The Impact
To investigate the claims, we observed a Core i7-8700 and its stock cooler during our x265 HandBrake benchmark. This real-world application is optimized to utilize all available cores. Moreover, it employs AVX instructions, which tend to increase power consumption considerably. We opened AIDA's system stability test window to monitor our -8700 during the workload.
- THROTTLE1
- handbrake throtte2
As you can see in the second slide, Intel's cooler was quickly overwhelmed, causing the processor to repeatedly bounce off of its 100°C temperature limit and throttle performance to protect itself (charted in red in the lower window).
With the bundled fan manually set to 100% duty cycle, we logged frequency throughout our test run (first album image). Even with the thermal solution working as hard as possible, the -8700 regularly throttled back from its 4.3 GHz all-core bin into lower ranges.
We also monitored VRM temperature during the test to ensure our motherboard's power delivery subsystem wasn't responsible for the throttling. Those measurements landed within the range we expected.
- THROTTLE2
- h100vi
Next, we ran the same test using a beefy Corsair Hydro Series H115i "Extreme Performance" all-in-one liquid-cooler. Manually cranking the two 140mm fans and pump up to 100% helped ensure that thermal output had no impact on our test results.
The difference is night and day. Intel's Core i7-8700 never exceeded 67°C, and the processor's frequency remained at a pleasingly-constant 4.3 GHz (though we did notice a few spikes higher during brief periods of lighter utilization). AIDA's system monitor confirmed that the CPU didn't throttle. Again, we see that our motherboard's power delivery subsystem satisfied the Core i7-8700's power and current requirements.
Comparing the difference between Intel's stock cooler (HS&F) and the all-in-one makes it clear that thermals clearly affect the -8700's performance. Turbo Boost is clearly designed to minimize the impact of thermal throttling: we only observe a 72-second delta over the course of our ~35-minute test.
A mere 3.4% separating those results may seem insignificant, but remember that we tested these configurations on an open test bench. A closed case would change the outcome almost assuredly. Also, we benchmarked at maximum fan settings, generating quite a bit of noise. It's far more common to use the motherboard's default fan curve, or to dial in an optimized fan profile that ramps up gradually. Unfortunately, those algorithms don't respond to control temperature fast enough to mitigate wild spikes and dips. As a result of the normal delay in fan speed adjustments, plus the less-than-ideal airflow in most PC cases, you could see larger slow-downs than what we recorded from a best-case test environment. This doesn't bode well for builders working with small form factors.
We did experiment with various thermal compounds between the stock cooler and heat spreader, but they didn't help much. Thus, we pin the negative outcome of our experimentation on Intel's paltry heat sink and fan combination.
Bear in mind that our benchmarks are run with the stock cooler and Corsair's all-in-one to highlight the difference in thermal performance. Some tests are short, while others take longer to complete. Some are single-threaded, while other are fully parallelized. Thus, the effects of heat influence each result in a unique way. As noted, we kept the fan speed at maximum and tested on an open-air bench, so our results represent a best-case scenario for Intel's stock cooler.
Comparison Products
Test Systems
| Test System & Configuration | |
| Hardware | AMD Socket AM4 (400-Series) AMD Ryzen 7 2700, Ryzen 7 2700X, Ryzen 5 2600X, Ryzen 5 2600 MSI X470 Gaming M7 AC 2x 8GB G.Skill FlareX DDR4-3200 @ DDR4-2933, DDR4-3466 Intel LGA 1151 (Z370): Intel Core i7-8700K, Core i5-8600K, Core i5-8400, Core i7-8700 MSI Z370 Gaming Pro Carbon AC 2x 8GB G.Skill FlareX DDR4-3200 @ DDR4-2400, DDR4-2667, DDR4-3466 All EVGA GeForce GTX 1080 FE 1TB Samsung PM863 SilverStone ST1500-TI, 1500W Windows 10 Creators Update Version 1703 - All Spectre and Meltdown mitigations |
| Cooling | Corsair H115i Intel stock thermal solution (Core i7-8700) |
MORE: Best CPUs
MORE: Intel & AMD Processor Hierarchy
MORE: All CPUs Content
In what context does it make sense to buy this CPU?
stock cooling is always less than prime.
I wouldn't sweat it, RyanTodd1. Your graphics card will be the gaming bottleneck before the CPU is.
When I got my first computer in 1997, it came with a Pentium II @ 233MHz. There were 266Mhz and 300Mhz models available at the time that I wished I had instead. Looking back 21 years later, I realized that it never made a difference which one I had. I think you'll feel the same way about your i7 8700.
*cough*
https://www.amd.com/system/files/AM4-Wraith-Cooler-Lineup-1920x631.jpg
http://www.relaxedtech.com/reviews/amd/wraith-max-and-wraith-spire-cooler/2
*cough*
you were saying? … yes those are copper plates on those coolers for the 65 and up lineup, yes they do have led and yes thost are actual copper heatpipes on the cooler that comes with the 2700x.
I know a lot of aftermarket coolers that look and perform a hell of a lot worse than what amd puts in the box.
it is only intel that puts half an ounce of third grade aluminium on top of their cpus (because they are too cheap to provide anything worthwhile I suppose …) and expects people to purchase actual cooling after the fact raising the total price of a system significantly.
I don't really agree. The difference between 8700 and 8700k is almost $100 CAD and yet the performance difference at default clocks is very small. So if I am building a machine for someone that will never overclock, save them some money and/or get the 8700 and get a good cooler instead.
The benchmarks paint a pretty nice picture of the 8700. I believe you, Tom's, when you say that the cooler can be overwhelmed, but your benchmarks don't really seem to indicate much of a loss when/if it is happening, especially in gaming.
Honestly though, why don't they differentiate the designation. Intel should have the 8700 at stock 8700K speeds, but just have the K unlocked. It isn't exactly deserving of the 8700 designation if it is clocked 500MHz lower. Just another thing Intel does that irks me.
Hopefully this is the case, although tech has come a lot further since 1997! I wasnt even born then!
Intel should have not included a cooler at all ... its like buying a car with a turbo that is supposed to make 300 hp, but then when it doesn't run at that speed and you take it back they tell you, "oh, that's just the max hp if you buy a better turbo to put on it than the one we sold it with". Would that be acceptable? No.
If my CPU says it runs at certain frequencies, and I buy it for that reason, I expect it to work as advertised on the cooler it comes with.
For 90% of the people, 90% of the time, it simply won't matter. For the few people in the limited circumstances where it matters, a better cooler is always an option.
Article idea: "Typical" case, with two fans, front intake and rear exhaust; stock mobo fan profiles. Same parts, same clocks. Long scripted workload (8-10 hours' worth; a typical work day). The only difference is the coolers. Is there a real difference?
I think the 8700 has it's place. Plus, you can get an even better deal on it when it goes on sale. It's great for those who don't intend to overclock, and for those on a tight budget. Still a quality CPU, if you ask me. And yes, an aftermarket cooler is the way to go.
My i7-7700K@4.8GHz can barely score 152xx!
That is all true except you are missing one crucial point: Intel knows their buyers of K-series chips already plan on overclocking which is why they stopped including that junk cooler with them, and as we all know, AMD's chips don't even overclock well even with high end aftermarket water AIO or air coolers. But yes, Intel is chumping buyers of non-K chips out with such a lame stock cooler product. VERY lame for Intel. The root fail here is that Intel's previous generation non-K GPUs ran just fine under the stock fans. But the last few generations starting with Skylake with terrible thermal dissipation management showed they needed to step it up for the stock cooler. But Intel didn't give a damn.
Anyway, if I were building a new gaming rig and only wanted Intel, I'd probably opt to save the $50 and go for the non-K version here knowing full well that at 1440p or 4K gaming, the CPU means little compared to the GPU in those resolutions. A $25 Cryorig H7 will cool it just fine with a proper air flow case. If you had a K-series and overclocked, you are looking at nearly a $100 (or more) cooler solution to hit 5+Gz. Which again, offers extreme diminished returns at higher resolution over stock speed (in gaming).
i didn't buy a expensive high end cooler either, i guess normally 40 bucsk i got it for like 26 cool master (same cooler i used for my prior pc).
idle temps range 35-40, don't get above 80-85 when video encoding
i don't do real gaming of anykind so what i do do would be a laughting stock (really its one browser based game).
i do have a radeon rx 560 oc 4gb edition.
other than that cpu fan my case has 5 other fans, 3 intake and 2 exhaust plus the power supply fan.
i am happy with the running temps and that of the pc
it is overclosed to 4.7ghz i think
I wonder how many people buy these and actually use the iGPU, except as a backup in case the dGPU dies. For me, the iGPU has never been a consideration when picking a CPU.
Since I haven't seen any previous descriptions of how the physics works before, I will attempt to lay out an explanation that covers your question, and more:
It should easily be understood that without metallic solder under the lid that the thermals will build up on the die. So far so good.
However, that also provides a singular benefit... the average temperature of the die will quickly be pretty close to temperatures measured at any given point on the die. This is helped by how modern OSes move threads from core to core. But, we should not expect temperature sensors all over the die to confirm that what I have said or to micro-manage the throttling behavior.
Therefore, the throttling will always happen in a reasonably narrow thermal envelope across the entire die. This throttle point is based on careful testing of the die to determine the required throttle-point to guarantee the CPU integrity for years. This should even be good for a thermal overload situation (e.g. CPU fan stopped). Again, so far so good, I hope.
Now enter the normal OCer who is not going to de-lid the CPU. They are already covered, as they haven't brought anything special to the table that would interfere with the throttle-point, except making it harder to reach (e.g. higher GHz).
Finally, we have the hard-core OCer who is going to de-lid the CPU and extract heat directly from the die or add metal paste. This is the special case, as now the temperature measured at different points on the die (if we could measure it) at extreme OC is entirely dependent on the vertical thermal extraction at each point across die. This presents no issue as long as the thermal contact is perfectly even. A good OCer takes care in making sure this is the case.
However, the average person (you and I both) may make mistakes in thermal compound or metallic paste/solder application (de-lidded or not) that would create an uneven extraction of heat into the heatsink. When there is a more direct but uneven thermal path to the CPU die, we can get hot-spots that 'might' cause the die temperatures most remote from the thermal sensor(s) to push the die beyond its safe limits in those locations when a 'seemingly' good OC has been achieved. Increase the voltage too far and now we have a problem, as the transistor gates/junctions begin to break down at an accelerated rate.
So, in effect, we might (jokingly) suggest the current Intel thermal design to be: CPU with very-difficult to remove training-wheels / airbags.
It seems that AMD has taken a different bent, and attempts to put more-than-adequate cooling solutions in the hands of its customers to head off issues that might occur, but still has to carefully manage throttling to prevent damage to the die.
Exactly. Reviewers use open air test benches and thus you need a saltshaker handy for every review. This has become industry standard practice, regardless of the applicability of the results to real-world cases (pun intended).