News Intel Core i5-13600K vs AMD Ryzen 7 7700X and Ryzen 5 7600X Face Off

[Admin]

We put the Core i5-13600K and Ryzen 7 7700X and Ryzen 5 7600X through a six-round fight to see who comes out on top.

Intel Core i5-13600K vs AMD Ryzen 7 7700X and Ryzen 5 7600X Face Off : Read more

 

[ManDaddio]

Wish these would be a good pair with a 4090. The performance is nice compared to last gen. The 13600k looks great at this price point.

 

[lmcnabney]

It is interesting why the gaming results were so much different in this article versus the exact same head2head done here. The other site also shows the speed of the RAM used (6000 for AMD and 6400 for Intel).

 

[zecoeco]

It is interesting why the gaming results were so much different in this article versus the exact same head2head done here. The other site also shows the speed of the RAM used (6000 for AMD and 6400 for Intel).

Yep.. and that's the reason why I don't trust Tom's benchmarks.
Its either a high error margin or a bias, favoring intel in most benchmarks as I noticed.
If you're planning to buy a CPU, the best thing to do is to look for more than one review (at least 3) and build up a geomean in your mind, observing the relativity.
So this way you'll get more accurate results on how the CPU actually performs., narrowing the bias and other factors involved which could impact the accuracy of these benchmarks.

 

[jp7189]

I noted something similar before here in another article, where the Intel platform used DDR5 in the performance section, and DDR4 in the price section.

 

[PaulAlcorn]

It is interesting why the gaming results were so much different in this article versus the exact same head2head done here. The other site also shows the speed of the RAM used (6000 for AMD and 6400 for Intel).

Thanks for reading. There are a myriad of reasons that we can see differences between various websites, including test setups, titles, and specific sequences used for comparison. Other factors, like drivers, power settings, and cooling can all impact test results. That's part of the reason why it is best to check several websites before you pull the trigger.

Also, you can check roundups where data from multiple reviews is combined to create a broader view. For instance, this one is comprised of 28 reviews and 6,710 benchmarks:



3D Center Review Roundup
This is the link to the source of the image, and you'll find a link to the reddit thread in the second tweet of that thread, as well. That will provide you with the raw data from each website in a table format.

I think 3DCenter does the community a great service with these comparisons, and I encourage folks to reference them as a general indicator of what to expect. You can also use them to benchmark our benchmarks, as it were.

We disclose our test hardware and procedures in the table at the end of the article. For our standard config, all chips are tested at stock memory settings, using the warrantied data rate to assure a level playing field and the best view of what you would see if you use the chip out of the box. I think the review cited only uses overclocked memory configurations.

We also test in overclocked configurations that use faster memory, and they're in our charts as well. As listed in our test setup section, the Ryzen chips tap out at DDR5-6000, as per AMD's own guidance that's the sweet spot, while we stopped at DDR5-6800 for Intel to remain in the general pricing vicinity for OC. However, Intel can hit over DDR5-8000 with XMP kits that we have full access to. They are just too expensive to be considered rational IMO.

All of our testing uses DDR5, though we do link to Raptor Lake DDR4 testing and mention multiple times that this can cost around a 4% performance loss, on average. We also reference both DDR4 and DDR5 pricing in all the relevant areas.

 

[lmcnabney]

Yep.. and that's the reason why I don't trust Tom's benchmarks.
Its either a high error margin or a bias, favoring intel in most benchmarks as I noticed.
If you're planning to buy a CPU, the best thing to do is to look for more than one review (at least 3) and build up a geomean in your mind, observing the relativity.
So this way you'll get more accurate results on how the CPU actually performs., narrowing the bias and other factors involved which could impact the accuracy of these benchmarks.

There are just so many different ways to inadvertently (or deliberately) cripple a benchmark run. Something as easily missed as not enabling XMP in the board will result in the motherboard running speedy RAM at much lower than rated speeds / latencies. Is the main board introducing some inefficiencies? The benchmarks here don't list the full criteria of the test setup, so you can't see if there is something unexpected going on. Do AMD CPUs perform poorly here because the RAM is DDR 4800 when everyone knows that 6000 is the sweet spot? I don't know. I would like to trust the review sites more than I do now.

 

[PaulAlcorn]

There are just so many different ways to inadvertently (or deliberately) cripple a benchmark run. Something as easily missed as not enabling XMP in the board will result in the motherboard running speedy RAM at much lower than rated speeds / latencies. Is the main board introducing some inefficiencies? The benchmarks here don't list the full criteria of the test setup, so you can't see if there is something unexpected going on. Do AMD CPUs perform poorly here because the RAM is DDR 4800 when everyone knows that 6000 is the sweet spot? I don't know. I would like to trust the review sites more than I do now.

Please read my response in the post above yours. Also, we test with both stock and OC'd RAM, and those results are in our performance charts as well. The settings are listed in the test setup table in the article.

 

[DavidLejdar]

Nice comparison of the cheaper current-gen CPUs. And Intel sure offers more options at the moment, such as being able to use a Z690 MB, one may already have with DDR4, to upgrade the CPU with.

Myself, I recently upgraded from a DDR3 rig, and went for DDR5. Which is not specifically needed for gaming (as even DDR3 still works). But from what I understand, DDR5 should work better with a many-cores CPU than DDR4 does. And my gaming time goes mostly on simulations, open-world games, and strategies, with use-cases such as modding Cities: Skylines to allow for more than 9 tiles to build in, where even 9 tiles can be quite demanding on hardware (possibly leading to lag, which has nothing to do with the GPU).

So I went for DDR5, with "just" a Ryzen 5 7600X for the time being though - and I will eventually upgrade the CPU to possibly one with 3D V-Cache later.

The other option would have been for me to save a bit by going for a DDR4 rig instead, and to use the saving on a better GPU. But as there are some GPUs which can handle 1440p gaming above 100 FPS, at way below the current top prices for some other GPUs, I didn't see much a point in going for a higher tier. And as can be seen in the benchmarks, other hardware does have an impact on the performance of the GPU. And I didn't want to end up getting the GPU bottlenecked, to then throw out a DDR4 MB already after a few months or so.

Sure often boils down to personal preference in any case. I.e. someone wanting top-notch now, they will likely go for Intel, and probably also 4090, to be able to enjoy the reflections on the walls and floors in Portal with RTX. Or someone, who may not care much for the newest game releases, they have a lot of cheaper previous-gen options to build a rig with. And so on.

It is interesting why the gaming results were so much different in this article versus the exact same head2head done here. The other site also shows the speed of the RAM used (6000 for AMD and 6400 for Intel).

E.g. for Red Dead Redemption 2, 7600X came out on top here in the benchmarks too, while it performed worse for Watch Dogs Legion in both cases. The sample of games isn't as broad here though. And one could argue that TH having used DDR5-6800 (mentioned here) for the Intel CPU, in comparison to DDR5-6400 on the linked website, that such was a factor about test results, and that such factor speaks for Intel currently, as it goes beyond what AM5 offers (which would also imply that with DDR4, it would show worse results). And the linked article does mention better productivity performance of the 13600K, which is in accordance with what it says here. So I don't really see a problem. Nothing wrong with checking more than one review of course though.

 

[lmcnabney]

Thank you so much for your reply. This is an example of why I do want to trust your reviews because you are clearly trying to meet the needs of your audience. A lot of work goes into providing benchmark details. Countless hours of fairly tedious work. That effort is appreciated.

The chart above shows a current mystery which isn't really clear. The 7XXX CPUs are not scaling with gaming performance. I understand how increasing the cores wouldn't help and I do understand how there is an issue with only a single chiplet being used in some circumstances, but AMDs chips are raising the boost as they go up in bin just like Intel but they aren't showing increases in performance going above 7700X. Is there some type of cap being introduced on the AM5 platform which some setups are being impacted, but other setups are not?

 

[Thunder64]

This one seems more fair then articles like this in the past. Intel benefits on the lower end because of the E-cores. I am glad that there is a new thread here to post in. Maybe you heard about my complaints in another thread, one of which was that I could not not have any input as to what was wrong. Here is what I wrote, and for the article I cited I think it was pretty damning. As I said, this one appears more fair.

With Ryzen I do. Articles like this pop up fairly often, and I don't like the methodology. I'll go through it.

CPU Pricing and Value: Intel. No problems wit this. Though with the recent price cuts, It's probably a tie.

Gaming Performance: Intel. I'm fine with this for now.

Content Creation/Productivity: Intel. Should be a tie. Intel wins single thread, AMD multi thread. After those first graphs there are six (!) graphs comparing Alder Lake to Zen 3, which is not relevant anymore.

Specifications: Intel. This looks like nothing more than a repeat of CPU Pricing and Value. In other words, just a way to give Intel another win.

Overclocking: Intel. This one is just a mess. First it is ALL Alder Lake vs Zen 3. Alder Lake can eat tons of power, so of course there is more headroom. It should be compared to Zen 4 since they have much closer TDP's. Also, all AMD CPU's are unlocked whereas with Intel you need a K CPU and Z chipset. That alone makes me say this one should go to AMD.

Power Consumption: AMD. It's closer now with Zen 4 but still a win for AMD.

Drivers and Software: Intel. Not sure what this is supposed to even mean. I do know I've never seen AMD win or tie in it though so it seems it's there just to get another win. Both are generally just fine with the occasional problem now and then. I'd call it a tie.

Process Node: AMD. I agree. Not sure why it really matters though. Maybe to throw AMD a bone?

Architecture: Tie. I'm fine with that.

Security: AMD. I don't think it really matters to most people. A lot of the exploits are more theory it seems. But, AMD does have less of them.

So by my count, I'd give Intel 5 because I'd remove Specifications as a duplicate. AMD would get 7 or 8 depending on whether I included pricing. I would consider throwing away security to make it 6 or 7.

Here's the best part though (and by that I mean worst), the comments section starts over two years ago and is closed! Therefore no one can point out the flaws like I just did. Intel wins 7-4. It's the truth. You can't challenge it. Content Creation and Overclocking are the worst by far, IMHO.

 

[-Fran-]

Thanks for putting up the head to head among those three. I do agree with the overall conclusion, but I have 2 nitpicks:

1.- The mention of DDR4 support with Alder/Raptor Lake and no data to show the difference (or loss) of performance by choosing that route. This is the same exact caveat you apply openly with the 5800X3D and its cache and it seems unfair to not point it out explicitly and with data to back it up in there.
2.- The gaming (related to the above) and overclocking are ties in my book. I consider lazy to say "just turn on PBO with AMD, but do tweak with Intel because reasons". This has been demonstrated by many other outlets that all AMD CPUs, going back to Zen1 have plenty of headroom if you can provide the power and cooling. OC scaling not withstanding, if you manually tweak them there's plenty of gas left in the tank. Just because PBO is really convenient and almost a "1 click OC solution" (which is also interesting you don't mention it as a pro for AMD and a con for Intel) is not explored further.

Anyway, as I said, I still consider the i5 the overall better CPU this gen, but as a platform: I'd like to say AMD has a small upper hand.

Regards.

 

[Thunder64]

Thanks for putting up the head to head among those three. I do agree with the overall conclusion, but I have 2 nitpicks:

1.- The mention of DDR4 support with Alder/Raptor Lake and no data to show the difference (or loss) of performance by choosing that route. This is the same exact caveat you apply openly with the 5800X3D and its cache and it seems unfair to not point it out explicitly and with data to back it up in there.
2.- The gaming (related to the above) and overclocking are ties in my book. I consider lazy to say "just turn on PBO with AMD, but do tweak with Intel because reasons". This has been demonstrated by many other outlets that all AMD CPUs, going back to Zen1 have plenty of headroom if you can provide the power and cooling. OC scaling not withstanding, if you manually tweak them there's plenty of gas left in the tank. Just because PBO is really convenient and almost a "1 click OC solution" (which is also interesting you don't mention it as a pro for AMD and a con for Intel) is not explored further.

Anyway, as I said, I still consider the i5 the overall better CPU this gen, but as a platform: I'd like to say AMD has a small upper hand.

Regards.

Read my post just above. It seems clear there is an Intel bias. This one was more fair, but the one I commented on was just horrible.

 

[-Fran-]

Read my post just above. It seems clear there is an Intel bias. This one was more fair, but the one I commented on was just horrible.

I disagree with the premise that Paul or any of the Tom's reviewers have an explicit bias towards nVidia, Intel or AMD, but I do believe in missing things due to the sheer amount of things to take into consideration.

Regards.

 

[Thunder64]

I disagree with the premise that Paul or any of the Tom's reviewers have an explicit bias towards nVidia, Intel or AMD, but I do believe in missing things due to the sheer amount of things to take into consideration.

Regards.

Well there might be bias or they may be half assing it, like I said above lacking correct comparisons. I'd like to think they are on the level, and Jarred certainly seems to be regarding NVIDIA and AMD. As far as Paul goes I'm fine with some of his picks as I stated above. Some others puzzle me though. I was just trying to bring attention to them as I would love this to be a "go to" site like Anandtech was back in the day.

 

[TerryLaze]

Thank you so much for your reply. This is an example of why I do want to trust your reviews because you are clearly trying to meet the needs of your audience. A lot of work goes into providing benchmark details. Countless hours of fairly tedious work. That effort is appreciated.

The chart above shows a current mystery which isn't really clear. The 7XXX CPUs are not scaling with gaming performance. I understand how increasing the cores wouldn't help and I do understand how there is an issue with only a single chiplet being used in some circumstances, but AMDs chips are raising the boost as they go up in bin just like Intel but they aren't showing increases in performance going above 7700X. Is there some type of cap being introduced on the AM5 platform which some setups are being impacted, but other setups are not?

For intel the max boost difference is 300Mhz between the i5 and i7 and 400Mhz between the i7 and i9, because they have access to 240W and only 8 full cores max and games don't use that much power all full cores can run at close to that top speed, so there is plenty of difference to show up as FPS.

For ryzen it's 5680Mhz vs 5608Mhz and 5550Mhz if we take a 6 core game as baseline, so basically 50-80Mhz difference between each tier.
It's barely a difference at all and that shows up in the FPS.
https://www.techpowerup.com/review/amd-ryzen-9-7950x/26.html

AMD Ryzen 9 7900X Review - Creator Might, Priced Right

The Ryzen 9 7900X is a creator's dream. It offers 12-cores and 24-threads that perform fantastic not only in highly-threaded workloads, but single-threaded performance is also there. Priced at $550 it sits right in the middle between Ryzen 9 7950X and Ryzen 7 7700X.

www.techpowerup.com

https://www.techpowerup.com/review/amd-ryzen-7-7700x/26.html

 

[BogdanH]

I generally agree with Thunder64 observations, however I won't use word "biased" -because some might say that I'm biased in favor of AMD

I would just add another thing that "disturbs" me in many performance comparisons. I am talking about Content creation & Productivity performance. Most of application (if not all) that I'm using are fully multi-thread capable: Blender, DaVinci Resolve, ... etc. -heck, even 7zip and WinRAR are multi-threaded.
And now I ask: why would someone run these application in single thread? I mean, for Content creation & Productivity, why even testing CPU performance in single thread? Just so Intel gets points here? For what? I ask, because I'm pretty sure that those who own Intel CPU, run these apps multithreaded anyway.

 

[M42]

I generally agree with Thunder64 observations, however I won't use word "biased" -because some might say that I'm biased in favor of AMD

I would just add another thing that "disturbs" me in many performance comparisons. I am talking about Content creation & Productivity performance. Most of application (if not all) that I'm using are fully multi-thread capable: Blender, DaVinci Resolve, ... etc. -heck, even 7zip and WinRAR are multi-threaded.
And now I ask: why would someone run these application in single thread? I mean, for Content creation & Productivity, why even testing CPU performance in single thread? Just so Intel gets points here? For what? I ask, because I'm pretty sure that those who own Intel CPU, run these apps multithreaded anyway.

You might be disappointed with Zen 4's overall performance vs Raptor Lake for content creation. Puget Systems has a lot of useful performance comparisons. Here's their analysis of 13th Gen Intel:

13th Gen Intel Core Processors Content Creation Review

Following AMDs recent release of their Ryzen 7000 Series desktop processors, Intel is fighting back with their own launch of the Core 13th Gen processors (code named "Raptor Lake"). Unlike AMD's launch which moved to a new socket and added support for DDR5, the 13th Gen CPUs are a drop-in...

www.pugetsystems.com

 

[Thunder64]

You might be disappointed with Zen 4's overall performance vs Raptor Lake for content creation. Puget Systems has a lot of useful performance comparisons. Here's their analysis of 13th Gen Intel:

13th Gen Intel Core Processors Content Creation Review

Following AMDs recent release of their Ryzen 7000 Series desktop processors, Intel is fighting back with their own launch of the Core 13th Gen processors (code named "Raptor Lake"). Unlike AMD's launch which moved to a new socket and added support for DDR5, the 13th Gen CPUs are a drop-in...

www.pugetsystems.com

Puget Systems is not a reliable source.

View: https://www.reddit.com/r/Amd/comments/e4zig7/puget_systems_said_that_intel_has_every_right_to/

 

[bit_user]

he 13600K takes an 8.5% lead over the Ryzen 7 7700X and an 18.7% lead over the Ryzen 5 7600X in threaded applications

Because E-cores, but some people still won't hear it!

 

[bit_user]

Puget Systems is not a reliable source.

Yes, as a system builder, they have vested interests. That's why we like independent sites like Toms, which aren't trying to sell you anything.

 

[bit_user]

I would just add another thing that "disturbs" me in many performance comparisons. I am talking about Content creation & Productivity performance. Most of application (if not all) that I'm using are fully multi-thread capable: Blender, DaVinci Resolve, ... etc. -heck, even 7zip and WinRAR are multi-threaded.
And now I ask: why would someone run these application in single thread? I mean, for Content creation & Productivity, why even testing CPU performance in single thread? Just so Intel gets points here? For what?

I think it's useful for 2 reasons:

1. To explore single-threaded performance on different kinds of workloads. And here, those programs are being used as a substitute for something similar a user might run that's not multithreaded.
2. To show multi-core scaling of the CPU, if the same program is benchmarked in multithreaded mode.

 

[TerryLaze]

I generally agree with Thunder64 observations, however I won't use word "biased" -because some might say that I'm biased in favor of AMD

I would just add another thing that "disturbs" me in many performance comparisons. I am talking about Content creation & Productivity performance. Most of application (if not all) that I'm using are fully multi-thread capable: Blender, DaVinci Resolve, ... etc. -heck, even 7zip and WinRAR are multi-threaded.
And now I ask: why would someone run these application in single thread? I mean, for Content creation & Productivity, why even testing CPU performance in single thread? Just so Intel gets points here? For what? I ask, because I'm pretty sure that those who own Intel CPU, run these apps multithreaded anyway.

If they would use software that can only run in single thread then intel would get an even bigger point because they would be able to clock even higher with lighter threads and get even higher scores, that's why they use the widest possible software threads for single thread testing so that AMD can utilize their very wide cores and look better than they are in comparison.

Or they are just lazy and use the benches they already have to use for multi.

 

[MergleBergle]

One thing I've noticed missing from the 7000 series reviews here on Tom's has been a dive into "eco mode"(specifically in regards to the 7700 and 7600s). Another specifically gaming based website (don't want to mention it in case it would violate some kind of term or condition) speaks very highly of the 7700x, and describes eco mode as "ace" in its summary.

 

[bit_user]

don't want to mention it in case it would violate some kind of term or condition

I'm not aware of any forum policy preventing people from citing other sites or publications. If it's done in the context of the discussion, I think it's entirely fair. Where it would likely cross a line is if someone were making posts that are blatantly promoting another site and have little other value.

There are certain sites I avoid linking or referring to, and I'll call out WCCFTech as one of them. They're rumor mills and a lot of their articles are shamelessly reporting on what other sites publish, with little added value or original content. I consider them parasites, so I don't want to drive any more traffic to them. Unfortunately, because they're indexed in Google News, they do tend to come up in my news search results fairly often.