Discussion So. How about we talk about the extra cost to own and operate the Intel 11600K over the AMD 5600X?

Jan 21, 2021
2
1
15
I liked a lot about the Intel 11600K versus AMD 5600X head-to-head review put out by Tom's Hardware, and I especially appreciated its awareness of the current price-gouging on the AMD chip and how market price must and should impact a fair recommendation. It's all about value for money!

The discussion of value, however, seem to me to have one major missing feature:

"If chip B (in this case the Intel 11600K/KF) requires more electricity to run, and runs hotter than chip B (here, the AMD 5600X), then what's that mean for us customers in actual money? How much more should we value chip B? Give me a number, even a soft one, not just some arm-waving!"

You're a power worker or heavy gamer, and you don't like reading long posts? The discussion below argues for adding a hundred dollars (and for some of us a hundred and fifty) to the effective price of that Intel 11600KF you've got your eye on.


Introduction
A failure to price out the difference in power consumption and heat generation isn't just a problem with TH's review, it's a problem with the major video reviewer's discussions as well. They all mention power consumption, but no review I've read yet helps us understand what to DO with that information. This is offering mere noise, not signal, and I cannot abide that from a professional.

This issue combines with another problem the professional reviewers suffer from: They all have super-great computer cases, coolers, power supplies, and motherboards, so they throw them at whatever CPU (or GPU) they're reviewing ... without any regard for what savings on cases, coolers, power supplies, and motherboard more efficient CPUs (and GPUs) make possible ... and therefore what savings the savvy build-your-own people and the major OEMs pocket every day.

Therefore, the purpose of this thread is to help you estimate how much to add to the effective price of the less efficient CPU, understanding that such a number will be approximate and must be specific to your use case.


1. The Power Supply
If you're building a new computer around your new CPU, you'll need a power supply. The more juice the CPU requires, the beefier and more expensive that power supply will have to be, for a given quality and reliability.

Those of us who build our own computers tend to overestimate the power our rig will use, and therefore pay too much for the power supply. I'm running the AMD 5600X, PBO, +200 MHz all-core OC, alongside a 100W GPU. The total power draw of my full system, under a Prime95 all-component torture test, is 270 watts. Under ordinary use, usage ranges between 110 and 170 watts. A 300-watt power supply from a reputable brand will cover my current rig in any use case, and 450 will cover any probable GPU upgrade. I have a 600-Watt power supply. I therefore paid for capacity I'll likely never use.

If I swap in the Intel 11600K/KF, and drop both chips down to stock, I will need to provide power for about another 70 watts. If I overclock, the difference is, very roughly, 120-150 watts. That's the difference between a $55 and at least a $70 power supply.

CAD professionals and AAA gamers, of course, want beefier power supplies, but the same points apply, just with different (bigger) numbers for required power and cost. In either case, if you choose the Intel 11600K over the AMD 5600X, you should bump up your desired power supply to the next higher capacity. 550W rather than 450W, or 750W instead of 650W. Consider even bumping it up another level if you plan to do serious all-core overclocking on the 11600K with a beefy water cooler and high-performance motherboard.


2. The Electricity Bill
One method of putting a number on the additional electricity cost of a power-hungry chip is to:

1. Estimate the average hours a day your computer will be used, assuming moderate load (you're doing something that definitely uses your CPU, but doesn't ask it to go full-tilt). For an ordinary work and gaming computer, this might be 3 hours (most office and communications work doesn't hit the CPU particularly hard, and such a computer may not be used every day).

2. Estimate your electricity cost per kiloWatt-hour (kWh). The cost of electricity in the US varies from about 9 to >20 cents/kWh, with an average of about 14.

3. Plug in the additional wattage of chip A (the Intel 11600K) under sustained but non-stressed load over chip B (the AMD 5600X). This is about 70 Watts (~).(I assume no overclocking, and minimal use of full-current boost here)

Example: 3 hours-equivalent of medium-heavy usage-equivalent per day, for 365 days, using an additional 70 watts and paying 14 cents/kWh.

(3 hours/day * 365 day/year) * (70 Watts / 1000 Watts/kiloWatt * 0.14 $/kiloWatt-hour) = $10.73/year

When pricing in electricity usage, I tend to be willing to pay for three years of cash savings in advance. So, in this case, I would add $32 to the effective price of the Intel 11600K/KF, if comparing it to the AMD 5600X.

3. The CPU Cooler (this part is where the biggest cost difference is)
The hotter the CPU runs, the more you need to spend on a cooler in order to a) make the CPU operate at spec, b) keep running at spec for the time required without thermally throttling, c) boost as desired, again without throttling early, and possibly d) overclock and therefore do a, b, and c more effectively, in both single and multi-threaded loads.

The AMD 5600X is extraordinarily efficient. I have mine overclocked (PBO, +200 MHz all-core OC). I use an Arctic Freezer 7X, a 25-dollar cooler (that comes with its own thermal paste, for a few extra $ of savings). For this setup, under real-world loads, thermal throttling is effectively a non-issue.

The Intel 11600K/KF is a rather different animal. In order to make the following statements, I've had to read between the lines of a number of reviews, which bear no responsibility for any errors I make. So, these points are made under correction - don't take them at face value unless validated!

The word from the testers and reviewers appears to be something like:

a) If you just want baseline stock performance, and you're also using an low-end motherboard that only supplies Intel's recommended base-line current to the CPU, then maybe a mid-range air cooler might avoid thermal throttling if full boost isn't active? Maybe spend 40-50$?

b) If you want full stock performance, including allowing the CPU full boost capability on demand - the situation pretty much all 11600K reviews cover - then you will benefit from something like a high-quality air cooler, or a low-end water cooler. Spend about $60?

c) If you want to overclock this CPU and get the equivalent boost that the AMD 5600X gets from PBO, +200 MHz all-core OC, then you want a very good air cooler or a solid water cooler. Nothing less will "tame this beast", to quote the testers. Spend $80-120?

Corrections to the above statements are welcome!


4. Heat Generation - your Living or Office Space
If chip A uses more energy, it also generates more heat. What's that extra heat in your space actually costing you? I did some back-of-the-envelope calculations and came up with the following rules of thumb. Again, these are offered under correction:

"If you're mostly heating the space (you live in a cold climate), and are using an only moderately cost-efficient form of heat (natural gas, say) then subtract one-fifth from the effective electricity bill."

"If you're mostly cooling the space (you live in a warm climate, or the room gets hot), and your air conditioning is modern and reasonably efficient, add one-third to the effective electricity bill."


Conclusions
So! Given what should be an ordinary level of usage, and understanding the rigs the reviewers actually put both the Intel 11600K and AMD 5600X into in order to figure out how these chips compare in stock and overclocked conditions, we wrap up this post with a rather bold statement.
"If your intended use case is fairly ordinary, and you do not plan to overclock, then: The effective price of the Intel 11600K/KF, primarily taking the cost of electricity and CPU cooler into account, is roughly 80-120 dollars above list price, if compared to the AMD 5600X."

So, if the Intel 11600KF is selling for $235, and those ****

[Moderator edit to remove profanity. Remember, per rules, this is a family friendly forum.]

are price-gouging the AMD 5600X for $350, what you'll actually pay, after budgeting for equally effective coolers for both chips under stock loads, and after three years of electricity usage for stock usage, is not $115 difference, but something closer to $15.
If you want to overclock whichever chip you buy, the hidden "gotcha" cost of the Intel 11600KF is quite a bit higher.
 
Last edited by a moderator:

jtk2515

Distinguished
Good write up! Made me think. I always enjoy some good discussion content.

Best thing the 11600k has going for it is you can run Dota 2 on it and some other games at 1080p or 720p. That is the only reason I can recommend it until we see price drops. If only they had put 96 Eu's in the xe graphics instead of cutting it down to 32.
 
Yeah, you summed up that pretty good and I quite agree.
That is, (in)efficiency should have more impact on "final verdict" in reviews. I can read in forums many times "..eh, is only 15$ more per year.. five beers.." and similar, which is very short sighted view. We must keep in mind, that there are millions of PC's running all the time. And even a company, that maybe only runs 50 PC's all the time, can save quite a money over the year, by making smart decision. Same comes down to a normal household: big saving is usually made by small savings here, small savings there.
 

USAFRet

Titan
Moderator
Couple of quibbles:

Electricity cost.
$10/year is almost just a rounding error in the context of a whole house annual power consumption.
There are many other ways to make up that $10.

If ALL of those have already been considered and done, then the CPU choice can make a difference.

PC power consumption is small in comparison to all the other devices in the house.

Now...if you scale that up to a data center, then it CAN make a big difference.

Heat generation:
This might be considered a wash.
Extra room heat in the summer that you have to get rid of.
Extra room heat in the winter that your house system does not have to do.
 
  • Like
Reactions: SamirD
I don't think most gamers care about the added power consumption. While it is interesting to look at the added electricity cost, $10 a year is around 3 cents a day. You probably could find 3 cents on the ground if you looked around.

The cooler cost is noteworthy, but still not significant. You COULD run a 5600x on the stock cooler, making the delta in cooler cost larger. But, you wouldn't want to.
My opinion is that if you are spending $300+ for a CPU, spending $25 or $30 on a cooler to make the computer considerably quieter is a good idea.

It would require testing, but this $50 cooler below SHOULD be more than plenty for an 11600k in realistic workloads (non avx512). This would make the difference in cooler cost closer to $25. Scythe Mugen 5 Rev. B 51.17 CFM CPU Cooler (SCMG-5100) - PCPartPicker
I remember LTT tested that when they enabled MCE on their 11600k to automatically overclock it, the voltage and temperatures actually lowered, despite the extra performance. I would assume some of the high stock temps are caused by the voltage being not tuned very well, and if you do tune the voltage well, a small overclock might not increase heat output.

I do think the added cost of Intel motherboards is the biggest "hidden cost" of the 11600k.

A solid ATX B550 motherboard is around $125, but even the cheapest ATX Z590 is $180. Sure, you might argue X570 is more feature competitive with Z590, but B550 basically has no real downsides for most people, while any chipset below Z590 does have real downsides.

Honestly both the 11600k and 5600x are good CPUs imo.
 
May 17, 2020
11
3
15
your forgetting the real downside of the 5600x. The fact you cant get the 5600x at MSRP anymore all retailers has it costing 400.00 and the availability is non existant.
 
I have seen it in stock at MSRP on numerous occasions, but stock is still very limited and nowhere near as good as Intel.

That being said, this is a demand issue, not a stock issue. If you look at amazon's most popular cpus, the 5600x is selling in higher numbers than any of Intels CPUs, even at inflated prices, which is silly.

Nobody should spend $350+ on a 5600x.
 
  • Like
Reactions: Nemesia

carocuore

Respectable
Jan 24, 2021
392
95
1,840
About heat... the amount of heat a computer can produce is negligible for your room, even if runs 24/7 the room temperature will most likely remain the same or increase by a fraction of a degree.
 
About heat... the amount of heat a computer can produce is negligible for your room, even if runs 24/7 the room temperature will most likely remain the same or increase by a fraction of a degree.

Have to disagree there depending on how much power your components draw. My RTX 3090 can heat up a room very quickly if left running at full power for extended periods. Now for CPUs this might not be a huge issue as few CPUs are drawing more than 350 Watts (looking at you overclocked 10980XE) but higher end GPUs can literally be space heaters, particularly if you run them for extended periods. There was even a story where someone apparently died from heatstroke because they left their mining rig running overnight in the same room they slept in and the GPUs heated the room up enough to kill this person.

Bottom line is your computer can significantly increase a room's temperature, but that's usually caused more by GPUs than CPUs. The difference between an 11600k and 5600x is probably not that significant in this regard.
 
D

Deleted member 2720853

Guest
AMD is currently complete trash at availability, making my 3900X that I got at $410 in 2020 look like a fantastic deal. Intel has the upper hand when it comes to stock.
 
If you actually have read the article you would have seen a few things.
https://www.tomshardware.com/news/intel-core-i9-11900k-and-i5-11600k-review
We're sticking with our standard policy of allowing the motherboard to exceed Intel's recommended power limits, provided the chip remains within warrantied operating conditions. Our tested settings reflect lifted PL1 and PL2 restrictions, which essentially removes the Tau limitation. Almost all enthusiast-class motherboards come with similar settings, so this reflects the out-of-box experience with a high-end motherboard. Naturally, these lifted power limits equate to more power consumption, and thus more heat, as we'll cover in detail later in the review.
This raises the max usable power from 250 to 295 an almost 20% increase.
We're not big fans of most over-the-top stress-testing utilities: There's a stark difference between the power consumed during stress testing applications that serve as a power virus and the power consumption you'll see during everyday use with the majority of 'normal' applications. In fact, we often don't include Prime95 power measurements in our standard CPU reviews, largely because there is a massive disconnect between this extremely rigorous stress test and the power consumption and thermal load generated by most real-world applications. But, curiosity strikes.
Translation: We know these numbers are complete lies and heavily overblown but saying these words legitimizes us showing these numbers.
Stress test is another 30% higher usage than anything sane.
doQfNIZ.jpg

You can compare tom's results with the results that gamer nexus got and will see barely any difference in performance and GN sticks to 100% intel suggested settings so 125W max after 56 sec IF any software can use that much and only heavily multithreaded stuff can use that much, so stuff that most people won't even touch.

Also ryzen requires 30% more power on the socket than on the CPU itself, it's called PPT, does that power magically disappear, do you somehow not need to provide that power. does that not heat up your room? so comparing the heavily overblown intel power numbers with ryzens power numbers that are 30% reduced.

https://www.gamersnexus.net/guides/3491-explaining-precision-boost-overdrive-benchmarks-auto-oc
Package Power Tracking (“PPT”): The PPT threshold is the allowed socket power consumption permitted across the voltage rails supplying the socket. Applications with high thread counts, and/or “heavy” threads, can encounter PPT limits that can be alleviated with a raised PPT limit.
  1. Default for Socket AM4 is at least 142W on motherboards rated for 105W TDP processors.
  2. Default for Socket AM4 is at least 88W on motherboards rated for 65W TDP processors.
 
The problem with trying to attach "how much you'll spend in electricity" into the value proposition is that the cost of electricity and use cases are highly variable. For instance, you may use the "average cost per KwH in the US", but...
It's simply better to report the measured power consumption against some practical cases (e.g., Handbrake encoding), worst case scenarios (e.g., Prime95), and idle. These values are independent of geographic location. Though I would like some measurements with "light" tasks like web browsing or YouTube video watching, like tests they do on laptops to measure battery life because these are far more common use cases for the average consumer.

Besides, "more energy consumed per unit of work done" doesn't need a monetary value attached to it. We already know energy is money.
 
Last edited: