Conclusion

In today’s performance review, we have taken the Intel Core i7-12700K vs AMD Ryzen 7 5800X CPUs and put them head-to-head, mano e mano, to find out who the champion is in PC performance today. Intel’s 12th Gen Core Procesossors are the new kids on the block, and they certainly have a lot to prove. Based on a hybrid core approach with performance and efficiency cores, Intel had to prove to everyone that this would work on the PC platform, and tell everyone why they needed this approach on the PC platform.

We have been working our way up the CPU product stack, starting with the Intel Core i5-12600K on DDR4. Then thanks to MSI we got the chance to test out directly DDR4 versus DDR5 performance on the 12600K with the MSI MEG Z690 UNIFY motherboard. MSI also sent us a retail Intel Core i7-12700K CPU so that we can properly see how Alder Lake compares on the desktop and test high-end motherboards. Therefore we put it on our test bench today with DDR5 and found out how the i7-12700K compares to the Ryzen 7 5800X, which has proven itself over time.

The Intel Core i7-12700K CPU has an RCP pricing of $409 and generally can be found online between $400-$450. This is a “12 Core” CPU, consisting of 8 Performance Cores and 4 Efficiency Cores (8P+4E) for a total of 20 threads. This is in comparison to the Ryzen 7 5800X which has 8 cores and 16 threads.

The 12700K also has 8 “Cores” but then the addition of the 4 Efficiency Cores, which are very powerful, and adds another 4 threads worth of multi-threading performance to the CPU over the 5800X. The 5800X launched with an MSRP of $449, therefore by MSRP the two CPUs are comparable as they are marketed around a similar price point. Since its launch, the 5800X has come down in price online and is now available between $350-$400.

Synthetic and Rendering Performance

Performance on the Intel Core i7-12700K started out great in the system benchmarks. In PCMark’s standard benchmark, in office-type workloads, the 12700K was on top of the 5800X in performance by 10%. In its Application Benchmark which looks specifically at Microsoft Office performance, the 12700K was up by 12% over the 5800X. This means even in just standard office workloads the 12700K is faster.

Testing continued in 3DMark’s CPU Profile and Geekbench 5’s testing, looking at multi-threaded and single-threaded performance in multiple different types of workloads. 3DMark’s Max Threads test put the 12700K at 30% faster than the 5800X, more threads, better multi-threading performance, simple as that. However, even in single-threaded performance, the 12700K was on top by 10%. This was all backed up in Geekbench which showed a whopping 55% improvement for the 12700K over the 5800X in multi-threading. In single-threading it was 13% faster.

When we loaded up SiSoftware Sandra 2021 we looked specifically at integer and floating-point performance in multi-threading and single-threading. The 12700K smashed integer performance either way. In integer multi-threading, it was 25% faster, and in integer single-threading, it was 15% faster. In floating-point multi-threading, it was 34% faster. In floating-point single-threading, both CPUs were tied.

Then when it came to the rendering portion of our testing the 12700K continued to be a beast. In Cinebench R23 multi-core it was 47% faster than the 5800X, and in single-core, it was 20% faster. In V-Ray 5 it was 33% faster. The 12700K also saved us time in Blender and HandBrake. In Blender, it shaved off 2 minutes of render time over the 5800X, and in HandBrake, it shaved off over 3 minutes of render time transcoding video.

Gaming Performance

Moving on to Game Performance there were a few surprises here as well. We tested at 1080p, 1440p, and 4K on a GeForce RTX 3080 Ti in several games. Starting with Battlefield 2042 we found that at 1080p there was a 9% performance advantage towards the 12700K. Even at 1440p, we saw a smaller 2% difference. Far Cry 6 also showed a very large advantage at 1080p with the 12700K being 22% faster than the 5800X. Even at 1440p, there was a 15% difference. Those kinds of numbers are big enough to affect gameplay and to make or break a fast refresh rate gaming setup.

Watch Dogs Legion was another game where at 1080p there was a 20% difference in favor of the 12700K with smaller differences at 1440p and even 4K. There was a small difference in Cyberpunk 2077 at 1080p, but otherwise, it was GPU limited at 1440p and 4K. MS Flight Sim seemed to be pretty much on par between the CPUs in the new DX12 mode. Crysis Remastered did show some slight differences as well.

We were surprised how big some of the differences were in games at 1080p. The Intel Core i7-12700K is clearly the superior CPU for gaming any way you look at it compared to the 5800X. When you are playing games with a faster GPU, the 12700K will be the CPU that is able to stand out more. If you have a slower GPU or play at 4K, you will more than likely always be bound by the GPU. However, if you have an RTX 3080 Ti, or 3090, or in the future get something even faster when it comes out, the 12700K is going to be the CPU that performance favors. This will be evident even at 1440p in some games. This is something important to think about for the future.

Power and Temp Summary

Alder Lake has been getting a bad rap lately for being called a power-hungry CPU. The thing is, as we have shown, this depends on the scenario. This is why we tested two different scenarios in our power testing today. On one hand, you have the scenario where you are running the CPU at full performance and utilization, think along the lines of rendering video or 3D modeling. In this scenario, our testing reveals that yes, the 12700K is going to consume more power than the 5800X, about 30% more. However, in terms of temperature, we actually found it was slightly cooler than the 5800X, which is helpful.

There is another scenario though, and one that is encountered more frequently, and that is the scenario that you are not always running your CPU at full utilization. This is for something like playing games, your CPU is not going to be at 100% utilization. When we actually ran this real-world scenario the 12700K was pulling the same amount of power as the 5800X. They were both equal in terms of power demand. Yet, we know the 12700K is allowing higher framerates in games. Therefore, it’s doing the same amount of power demand, but also providing fastester performance. That’s actually good efficiency. In terms of temperature, it is also a win at running 10% cooler than the 5800X at the same power load.

Final Points

Intel has come a long way since the past few CPU launches. Real improvement has been made here, real advancements, and real advantages over the competition have been realized. We think the Intel Core i7-12700K is the sweet spot for gaming performance.

This is the gaming CPU you want. It will ensure that when you get faster GPUs you become less CPU bottlenecked than the 5800X for comparison. The clock frequency of the 12700K is a real advantage for gaming performance. Its ability to boost to 5GHz is good for games.

It has just the right balance of cores for gaming performance with plenty of threads to spare for games. The 8 Performance Cores with Hyperthreading are plenty for games, that’s 16 threads for high-performance gaming. The Efficiency Cores can then work in the background to manage background tasks, keeping the CPU from bottlenecking its Performance Cores while gaming. Those Performance Cores can then boost up as high as possible while gaming.

When gaming, the CPU draws no more power than the 5800X, and it actually runs cooler. Therefore power draw while gaming should not be a concern. It is only when at 100% utilization it will demand more power than the 5800X. However, it also delivers by providing a massive boost to multi-threading performance in these scenarios. In multi-threaded applications, it just simply do more with its total of 20 threads.

It can boost render times in anything that renders, well, anything. If you do a lot of video encoding/decoding/transcoding, this CPU will be better than the 5800X for these tasks. We found out that even on a single-core/single-thread function the 12700K is about 10-20% faster than the 5800X. Therefore on a per-core basis, the 12700K is faster, to begin with.

Honestly, AMD should be worried about this. Isn’t competition great? The price of the Intel Core i7-12700K is a bit more than the 5800X right now, but the 12700K offers more and is very forward-looking for gaming performance with faster GPUs in the future. It’s not that much more in price and can be found around $400-$420 right now, which is competitive.

Discussion

Don’t Miss Out on More FPS Review Content!

Our weekly newsletter includes a recap of our reviews and a run down of the most popular tech news that we published.

Brent Justice

Brent Justice has been reviewing computer components for 20+ years, educated in the art and method of the computer hardware review he brings experience, knowledge, and hands-on testing with a gamer oriented...

Join the Conversation

17 Comments

  1. If I were doing a fresh build today, I’d probably get a 12600K or 12700K and a DDR4 board. Since I’m already on AM4, I’ll be looking at the next Zen with the 3D cache.

    The 12900K makes no sense, but the mainstream parts look great.

  2. Good writeup. Not exactly shocked that Intel’s brand new part beats on the 2 year old 5800x though. There are price differences too… the 5800x has been on sale a couple times for $299, and compatible ddr4 mobos are cheap. Am4 Upgradeability is also nice. None of that can be said for Intel platform.

    We’ll see what Zen 4 looks like soon, but on the downside new ddr5 mobos are very likely.

    Even though it is outclassed now, I’m still thinking about picking up a 5800x if I can catch a deal… a last hurrah upgrade for my X470 board. I can probably milk that on my prod system for another couple years until prices and availability chill out on ddr5. The 2700x has been a workhorse for several years now.

    OTOH…. if the i7 could dramatically speed up Civilization, I’m curious. That game becomes so much of a cpu beatdown in the late stages. It’s also poorly threaded.

  3. The Intel chip is clearly NOT a direct competitor with the 5800x. This would only be the case if you wilfully ignore the cost of the parts.

    For the same cost as the Intel chip, you can purchase a 5900x, which is therefore kthe direct competitor to this chip. Core count is not relevant, only cost.

    I’m willing to bet this chip does not perform nearly as well against the 5900x.

    Also to use DDR5 in Comparison with DDR4 is just not good practice, again for the prohibitive cost of DDR5 memory and boards. The AMD setup with DDR4 will be MUCH cheaper in comparison with little overall performance difference and therefore AMD would come out on top in realistic review…

  4. The Intel chip is clearly NOT a direct competitor with the 5800x. This would only be the case if you wilfully ignore the cost of the parts.

    For the same cost as the Intel chip, you can purchase a 5900x, which is therefore kthe direct competitor to this chip. Core count is not relevant, only cost.

    I’m willing to bet this chip does not perform nearly as well against the 5900x.

    Also to use DDR5 in Comparison with DDR4 is just not good practice, again for the prohibitive cost of DDR5 memory and boards. The AMD setup with DDR4 will be MUCH cheaper in comparison with little overall performance difference and therefore AMD would come out on top in realistic review…

  5. [QUOTE=”Gunsmoke, post: 45793, member: 5177″]
    The Intel chip is clearly NOT a direct competitor with the 5800x. This would only be the case if you wilfully ignore the cost of the parts.
    For the same cost as the Intel chip, you can purchase a 5900x, which is therefore kthe direct competitor to this chip. Core count is not relevant, only cost.
    I’m willing to bet this chip does not perform nearly as well against the 5900x.
    Also to use DDR5 in Comparison with DDR4 is just not good practice, again for the prohibitive cost of DDR5 memory and boards. The AMD setup with DDR4 will be MUCH cheaper in comparison with little overall performance difference and therefore AMD would come out on top in realistic review…
    [/QUOTE]

    Based on the cost numbers I’m seeing right now, the 5900X’s MSRP is $549 and is listed at NewEgg and BestBuy for that, $569 at B&H. The 5800X’s MSRP is $449, and is listed at that on BestBuy and B&H, but is ~$400 on NewEgg. The 12700K is listed at $404 on NewEgg, $419 on BestBuy and $380 (after a coupon) on B&H. If you’re looking at only the cost of the chip, I’d say we’re pretty spot on.

    With respect to the platform, we did find it did not make a material difference to performance in [URL=’https://www.thefpsreview.com/2021/12/13/intel-core-i5-12600k-alder-lake-ddr4-vs-ddr5-performance/’]a recent article comparing DDR4 to DDR5[/URL]. It’s an interesting debate on which platform to use – in theory, for a CPU review, one _should_ use the platform that its manufacturer intends for it to be in its best light, and that’s with DDR5. The other factor here is that there’s a bit of a gulf between the quality of the DDR4 and DDR5 boards out there – the DDR5 boards have far better power delivery circuitry which can certainly impact the overall performance of a chip.

    Either way, over the next month or so, we’ll be working on covering all combinations of performance write ups, and that will certainly include a comparison to the 5900x for your reading enjoyment.

  6. I’d agree that testing Intel with DDR5 is appropriate: it supports the standard, and to do the testing with anything less might artificially be limiting the platform.

    In this case, it doesn’t look like it makes any difference, but I support the testing methodology. It would have a significant impact on total system cost, but I don’t think that is a point the article is trying to make right now.

  7. My previous system was a 5800x and I just upgraded to a 12700k a month or so ago.

    For my use cases there’s not much difference performance wise but I was having some stability issues recently with the 5800x build. In hindsight it may have been bios related as all the components seem to be working in other machines after the fact and I believe I’d done a bios update around the time I started getting blue screens and whatnot.

    I like the 12700k and have no regrets other than some getting used to windows 11. I’d recommend it to anybody.

    It’s not obvious which cores in task manager are the efficiency cores or what clock speeds the different cores are using. I’m sure there is another app that can tell you but I wish that was available in task manager.

    I suspect the best value will be a 12600k or 12400 and a b or h series motherboard when they come out but I didn’t want wait.

  8. The review was pretty much what I expected.

    12900k vs 5950x

    12700k vs 5900x

    12600k vs 5800x

    The 12700k vs 5800x has a core count advantage on the intel side.

  9. Interesting.

    I knew it would only be a matter of time before Intel would wrest back the performance crown.

    That said, they are doing so using significantly more power at load. Even without any architecture improvements, it would seem like AMD could narrow the gap by just increasing power use to where Intel is. They also have new improvements coming in the next gen, like that new high performing cache, which should be interesting.

    I’m truly hoping we get back to where things were in 1999 through out the early 2000’s, with the two leapfrogging eachother with every release. That could make things fun again!

    For my purposes though, ehilr Intel is pushing much higher numbers in most cases, they are practically insignificant for me, as being a “4k Ultra” kind of guy, I will spend all of my game time GPU limited, and as for general desktop productivity type of stuff, even a dual core Haswell Celeron is fast enough to not be a problem, so all this extra power is mostly wasted on me.

  10. [QUOTE=”Zarathustra, post: 45828, member: 203″]
    That said, they are doing so using significantly more power at load. Even without any architecture improvements, it would seem like AMD could narrow the gap by just increasing power use to where Intel is. They also have new improvements coming in the next gen, like that new high performing cache, which should be interesting.
    [/QUOTE]
    In most scenarios – basically everything except synthetics designed to push all core — it looks like Intel is about the same as AMD on power use.

  11. [QUOTE=”Zarathustra, post: 45828, member: 203″]
    For my purposes though, ehilr Intel is pushing much higher numbers in most cases, they are practically insignificant for me, as being a “4k Ultra” kind of guy, I will spend all of my game time GPU limited
    [/QUOTE]
    While this is generally a non-issue with Zen 3 [I]today[/I], higher 1080p results means that when GPUs get faster, Alder Lake’s performance advantage will become more significant. Same pattern we’ve seen decades over.

    Obviously this only matters for the games you or other prospective buyers would play 😎

  12. [QUOTE=”Zarathustra, post: 45828, member: 203″]
    For my purposes though, ehilr Intel is pushing much higher numbers in most cases, they are practically insignificant for me, as being a “4k Ultra” kind of guy, I will spend all of my game time GPU limited,
    [/QUOTE]
    Except, this isn’t necessarily always going to be the case. I’ve done extensive testing on this and found that at 4K, CPU’s like the 2920X showed the same average FPS as a 9900K lending credence to the whole GPU thing, but playing games on the former wasn’t smooth. When I dug into it I found that in a lot of games, the minimum frame rates on certain CPU’s were horrendous and thus, not capable of delivering a good gaming experience regardless of the GPU.

    Here is an example from my [URL=’https://www.thefpsreview.com/2019/08/02/msi-meg-x570-godlike-motherboard-review/’]review[/URL] of the X570 GODLIKE motherboard. In fact, I wrote about this issue in detail [URL=’https://www.thefpsreview.com/2019/08/02/msi-meg-x570-godlike-motherboard-review/12/#cmtoc_anchor_id_0′]here.[/URL]

    [ATTACH type=”full” alt=”1642489389714.png”]1420[/ATTACH]

    The average frame rates would tell you that these perform the same, but if you look the minimum frame rates are almost half what they were on the Intel system. Now, the Ryzen 5000 series has stronger gaming performance and even the 3000 series issues with Destiny 2 were fixed, but initially the 3000 series required a work around of sorts to even run on those CPU’s. This is why it wasn’t tested during the initial CPU review.

    [ATTACH type=”full” alt=”1642489463342.png”]1421[/ATTACH]

    As you can see the averages are all pretty close to each other, but as you can see above, averages don’t tell the whole story. What I didn’t graph out were the minimums for all of the systems, simply opting to showcase the X570/3900X and the 9900K. However, the 2920X at stock speeds only achieved a frame rate of 26FPS minimum and 36FPS minimum when PBO was enabled. While the averages look good, I can assure you that it was not the best experience as I saw routine drops into the 45FPS range whenever things would happen on screen.

    Destiny 2 is admittedly a bit of an outlier, but there are plenty of others too. The moral of the story is that while you are primarily GPU limited at 4K, your CPU does indeed play a role and your minimum FPS and thus, your actual experience can be marred by a sub-par CPU.

  13. [QUOTE=”Dan_D, post: 47054, member: 6″]
    Except, this isn’t necessarily always going to be the case. I’ve done extensive testing on this and found that at 4K, CPU’s like the 2920X showed the same average FPS as a 9900K lending credence to the whole GPU thing, but playing games on the former wasn’t smooth. When I dug into it I found that in a lot of games, the minimum frame rates on certain CPU’s were horrendous and thus, not capable of delivering a good gaming experience regardless of the GPU.

    Here is an example from my [URL=’https://www.thefpsreview.com/2019/08/02/msi-meg-x570-godlike-motherboard-review/’]review[/URL] of the X570 GODLIKE motherboard. In fact, I wrote about this issue in detail [URL=’https://www.thefpsreview.com/2019/08/02/msi-meg-x570-godlike-motherboard-review/12/#cmtoc_anchor_id_0′]here.[/URL]

    [ATTACH=full]1420[/ATTACH]

    The average frame rates would tell you that these perform the same, but if you look the minimum frame rates are almost half what they were on the Intel system. Now, the Ryzen 5000 series has stronger gaming performance and even the 3000 series issues with Destiny 2 were fixed, but initially the 3000 series required a work around of sorts to even run on those CPU’s. This is why it wasn’t tested during the initial CPU review.

    [ATTACH=full]1421[/ATTACH]

    As you can see the averages are all pretty close to each other, but as you can see above, averages don’t tell the whole story. What I didn’t graph out were the minimums for all of the systems, simply opting to showcase the X570/3900X and the 9900K. However, the 2920X at stock speeds only achieved a frame rate of 26FPS minimum and 36FPS minimum when PBO was enabled. While the averages look good, I can assure you that it was not the best experience as I saw routine drops into the 45FPS range whenever things would happen on screen.

    Destiny 2 is admittedly a bit of an outlier, but there are plenty of others too. The moral of the story is that while you are primarily GPU limited at 4K, your CPU does indeed play a role and your minimum FPS and thus, your actual experience can be marred by a sub-par CPU.
    [/QUOTE]

    I’m with you, I usually judge my performance not by average framerate, but by minimum framerate, as that shows the worst case.

    I haven’t seen any bad 1% framerates yet, but who knows, you have to keep up with the times, and at some point I’ll need a new CPU again, but until then, I’m fine 🙂

    I’ve always been a “buy what you need NOW” kind of guy. It makes no sense to try to predict to the future, which system will have th elongest longevity, etc. etc., because the early adopter penalty for something that has long staying power is usually super expensive, and you can never really predict the future.

    For right now, my system does the trick. If it doesn’t tomorrow, I’ll cross that bridge when I get to it.

    Future-proofing in tech is a fools errand.

  14. 1% lows at a very minimum and 0.1% lows if you want to be pedantic, and those if you’re not going to go all out and start digging into frametimes.

    I’d much prefer frametime analysis myself…

  15. Great review, kudo’s for MSI providing a good test sample. I would say it would be definitely Intel build if building new at this time, the 12700K smashes the 5800x into oblivion on this review and is very nice seeing Intel getting serious with a great comeback from where they were at.

    As for Zen 4, it also comes with a new AMD platform. Just think of all past AMD new platform changes and how smoothly (cough cough) or not they went. I don’t remember one AMD major platform change that went smooth. Maybe TRX40 was their best but it was a minor update. X370, B350 had over a year of issues, brick boards etc. Intel for the most part new platform changes go rather smooth out of the gate. I would definitely wait and see with Zen 4 with AM5 and most likely go for attempt 2 for AMD if they are competitive or better overall. Right now Intel I think in general is the better choice for the desktop.

    AMD VCache, lol, one CPU incoming -> That to me is not a serious attempt to maintain performance leadership and with a frequency cut to boot. A 5900x/5950x Vcache versions I would have expected right out of the gate as well. Its maybe more due to AMD being limited by how much their products can be manufactured, manufactured limited, stunting their growth other than their price growth. At current AMD CPU pricing, I would not recommend any of their desktop CPUs over Intel except maybe in specific use cases. Not only that, AMD has virtually adbandom their HEDT CPUs with Zen 2 unless you go Pro and now can get a over 1 year old design Zen 3.

    Anyways the 12700k looks to kicks AMD ass at this time, while if pertinent to someone that is a case by case condition. For me the 5800x and 3080Ti is very much proficient for the 4K Bigscreen in the playroom and probably for the next several years. The 3900x good enough for the 6900XT and 4K. The Zen2 3060x and 3090, I could use an upgrade maybe.

Leave a comment