Intel Core i7-13700K

[W1zzard]

I really wish the MMOs and older games that are still super popular and have a large player base would be tested to see the effect of L3 cache too, in the games section. Like Finsl Fantasy, WoW, etc. They are not niche use cases for gamers. I see plenty of reports from people upgrading to 5800X3D and seeing large fps jumps in scenarios where lots of data from other players in the same zones is being processed. At the moment there’s no info on how the new AMD and Intel processors fare against 5800X3D for those types of games.

With MMOs the problem is that they get patched regularly, which means I have to retest 30 CPUs at any unknown point in time, to have accurate comparison data.

If you had to pick one "older game", single-player offline, what would you choose?

 

[Taraquin]

I must say I keep wondering why Raptor lake outpermforms Zen4 by this large margin in gaming here on TPU, while other reviewsites generally finds pariry or RTL slightly ahead. Is it the game selection? I see that 2 games kinda skews the results: Age 4 and FC6 13900K is 31 and 38% faster than 7950X here). Age 4 very few reviewers use, I give TPU cred for using it though, great game. In other reviews the difference in FC6 is far lower using a 4090 (1080p on HWUB 13900K is only 10% faster), and HM3 favours AMD but TPU is not using that game. Perhaps upping testsystem to 4090 would yield a better picture as even 720p is GPU bound in several games here using 3080? Not meaning to critize here as I love TPU in general

 

[Garrus]

Thanks for those Switch emulation results. I'm very interested in that actually Intel back in the running. So hard to pick one tomorrow.

 

[Forza.Milan]

not cool (literally..)

 

[VeqIR]

With MMOs the problem is that they get patched regularly, which means I have to retest 30 CPUs at any unknown point in time, to have accurate comparison data.

If you had to pick one "older game", single-player offline, what would you choose?

This person does some limited testing for WoW, and in a recent article they pointed out that there’s a free benchmark test for FFXIV 14:

Performance and hardware comparison of Final Fantasy XIV and World of Warcraft

There is a lot of comparison between WoW and FF14 lately. You can compare Y'shtola to Y'shaarj and so on but today I've tried to compare the performance and hardware/software characteristics of both game. Let's do some benchmarks.

rk.edu.pl

I’m not sure how representative that benchmark would be for a WoW capital city or raid scenario, but perhaps it could be roughly tested once on A version of WoW to get some idea.

https://www.reddit.com/r/Amd/comments/uxl0yp

 

[Garrus]

Was AVX512 enabled for Ryzen 7000 series with your RPCS3 results? Could mention it in the description of the chart. I had thought that AMD improved a lot there. Maybe not.

 

[W1zzard]

I must say I keep wondering why Raptor lake outpermforms Zen4 by this large margin in gaming here on TPU, while other reviewsites generally finds pariry or RTL slightly ahead. Is it the game selection? I see that 2 games kinda skews the results: Age 4 and FC6 13900K is 31 and 38% faster than 7950X here). Age 4 very few reviewers use, I give TPU cred for using it though, great game. In other reviews the difference in FC6 is far lower using a 4090 (1080p on HWUB 13900K is only 10% faster), and HM3 favours AMD but TPU is not using that game. Perhaps upping testsystem to 4090 would yield a better picture as even 720p is GPU bound in several games here using 3080? Not meaning to critize here as I love TPU in general

It's most probably game selection, also lots of reviewers use the integrated benchmarks, which often aren't realistic at all, I use actual gameplay, with hand-picked test-scenes. Also I use VBS enabled, which is the Windows 11 default on all systems less than 10+ years old, maybe that makes a difference. Next round of rebenches will use 4090, takes like two weeks of full-time non-stop back-to-back testing.

free benchmark test for FFXIV 14

Unfortunately not representative of actual gameplay, also 2021 article, so today's performance situation will probably be completely different in the game

The site that you linked to does great testing indeed. Probably the best WOW performance testing I've ever seen. I did test WOW for a few years in my GPU benchmarks and it was a huge PITA, and reader interest was surprisingly low. It goes up when new expansions come out, and then drops back down.

 

[ratirt]

13700k is like 12900K with a slight frequency bump and $50 cheaper. The problem is that previous 12700k was $409? Not much of a price bump for the 12900K's performance but the question is:
Is the 13700K rebranded 12900K or does it have some RL improvement? It would seem like, Intel did not sell all 12900K and have rebranded those to 13700K with a clock bump. Still I'd pick this one from Intel's portfolio and a more rounded chip with good performance.

 

[VeqIR]

It's most probably game selection, also lots of reviewers use the integrated benchmarks, which often aren't realistic at all, I use actual gameplay, with hand-picked test-scenes. Also I use VBS enabled, which is the Windows 11 default on all systems less than 10+ years old, maybe that makes a difference. Next round of rebenches will use 4090, takes like two weeks of full-time non-stop back-to-back testing.


Unfortunately not representative of actual gameplay, also 2021 article, so today's performance situation will probably be completely different in the game

The site that you linked to does great testing indeed. Probably the best WOW performance testing I've ever seen. I did test WOW for a few years in my GPU benchmarks and it was a huge PITA, and reader interest was surprisingly low. It goes up when new expansions come out, and then drops back down.

Yeah, just frustrating to not see that use case represented in any reviews, particularly because the effect of larger L3 cache is significant. I’m guessing that 5800X3D might still reign supreme there, based on that reddit post comparing it with an i7 12700k. I doubt the latest generations are going to be that much better. Some of us only or mostly play those older unoptimized games and use their computers for other office or, say, image processing software otherwise.

 

[W1zzard]

Is the 13700K rebranded 12900K or does it have some RL improvement?

This is definitely not a rebranded CPU, it has all the improvements of Raptor Lake

 

[nguyen]

Using only ~65W in PUBG pumping out 300-400FPS, I wouldn't call this CPU a power-hog . My previous 9900K @ 5Ghz is using >100W while pumping out 200FPS (rtx3090)

Using a 30usd air-cooler (ID-Cooling SE-226-XT) with 30usd fan (Phanteks T30), along with 5usd Thermalright ILM, I'm getting 67C highest temp in games with 30C ambient

 

[P4-630]

Really tempting. I wish Intel extend MB support 2 more generations. Doing so would render X3D and buying AMD flagship pointless since every gamers would jump on Intel without a second thought.
With Z690/Z790 being a dead end platform, I'm forced to go with 7800X3D instead of 13600K/13700K.

If you would do 4~5 years with one build and just upgrading the GPU later on this wouldn't be a huge issue.

 

[ratirt]

This is definitely not a rebranded CPU, it has all the improvements of Raptor Lake

OK but there is not a lot of a difference in performance between the two. So where is that improvement if you do not consider frequency bump as a RL improvements over AL. Is that the only one improvement?

 

[riklaunim]

I’m not sure how representative that benchmark would be for a WoW capital city or raid scenario, but perhaps it could be roughly tested once on A version of WoW to get some idea.

Assuming it will be done in game with all game mechanics then it would fit. If it would be just a render then it's kind of pointless as it would only benchmark a narrow side of the game (widefield views of new zones that need more raw GPU power than other aspects of the game).

Like to test mass combat in a repeatable way I use old Karazan and a large pull of mobs to create enough "world state" complexity to behave like in an actual raid or BG. Then some actual raids from time to time to see how the data corelates.

For "complex geometry" benchmarks I use laps around Legion Dalaran and Oribos (which is quite horrid for some reason) while for raw raster - Ardenweald and Bastion widefield static views. That pretty much covers all of WoW minus performance bugs specific to given instance or boss which do happen, usually after a patch

Unfortunately not representative of actual gameplay, also 2021 article, so today's performance situation will probably be completely different in the game

Hey, my benchmarks don't age that quickly Aside of pre-patch regressions/bugs in WoW right now not much changed for now. FF14 benchmark apps do somewhat represent the game. The only thing they don't test is Limsa with 1 000 000 000 cat girls killing FPS.

Hardware Numb3rs did benchmark 5800X3D and 12th gen K parts and if you aren't planning at dumping cash on a high end DDR5 memory kit the 5800X3D will win. If you dump way more cash on top of the line Intel it will win. I'm also trying to get 5800X3D probably/maybe so I'll be able to add some of mine benchmarks. 7000 X3D are also incoming, allegedly to be announced at CES... curious if they will do a laptop one, one of their quotes was kind of alluding to that.

 

[lightning70]

I am using 12600k. In the 10min R23 test, I use a model (Fuma 2) similar to the air cooling used here. It runs at 72 degrees in tests, with a maximum average of 69-70 degrees moreover, my room is a bit hot (26.5 degrees). think these new processors are all designed for 90 degrees and above, amd intel doesn't matter.

 

[P4-630]

This makes me love my i7-12700k even more

You should update your system specs!

 

[zmeul]

With 240 mm AIO, sounds about right to get 70°C. What's your temps on air in R23?

I don't have a air cooler on hand, but two colleagues of mine do - they're hovering around 80deg, one has a Noctua NH-U12S and the other has a Deepcool AK620 but a crappy old case with not that much room in it

we have 3 13700K s in house

 

[Ayhamb99]

With all of the reviews for Intel Raptor Lake and AMD Zen 4 CPUs now out, It makes me very happy to see that how competition has been strong in the CPU market ever since the release of Alder Lake and Intel finally getting their stuff together after the disaster that was Rocket Lake, You can't go wrong with choosing any of them during this time. I just hope the power consumption improves in the later generations but even then not that many people run Heavy Multi-threaded workloads 24/7...


Now about the 13700k, while the power consumption is unfortunately higher than the previous 12700k, its performance is pretty impressive and beats the previous flagship 12900k with same core count and threads, AMD will have to lower the price of the 7900x to stay competitive because with how the AM5 motherboard prices are and the fact that 13th gen can run on last gen Z690/B660 and also still use DDR4 Memory.

 

[VeqIR]

Assuming it will be done in game with all game mechanics then it would fit. If it would be just a render then it's kind of pointless as it would only benchmark a narrow side of the game (widefield views of new zones that need more raw GPU power than other aspects of the game).

Like to test mass combat in a repeatable way I use old Karazan and a large pull of mobs to create enough "world state" complexity to behave like in an actual raid or BG. Then some actual raids from time to time to see how the data corelates.

For "complex geometry" benchmarks I use laps around Legion Dalaran and Oribos (which is quite horrid for some reason) while for raw raster - Ardenweald and Bastion widefield static views. That pretty much covers all of WoW minus performance bugs specific to given instance or boss which do happen, usually after a patch


Hey, my benchmarks don't age that quickly Aside of pre-patch regressions/bugs in WoW right now not much changed for now. FF14 benchmark apps do somewhat represent the game. The only thing they don't test is Limsa with 1 000 000 000 cat girls killing FPS.

Hardware Numb3rs did benchmark 5800X3D and 12th gen K parts and if you aren't planning at dumping cash on a high end DDR5 memory kit the 5800X3D will win. If you dump way more cash on top of the line Intel it will win. I'm also trying to get 5800X3D probably/maybe so I'll be able to add some of mine benchmarks. 7000 X3D are also incoming, allegedly to be announced at CES... curious if they will do a laptop one, one of their quotes was kind of alluding to that.

Excellent to see you here! Thank you very much for your thorough efforts in presenting WoW benchmarks—oddly one if the only places for them, considering the huge player base over the years.
Yes, very curious about the possible 7800X3D.

 

[wolf]

Another CPU that I won't be recommending to average users due to its insane cooling requirements at stock.

For productivity I'd agree, for gaming, not at all. As a gaming CPU the temps are absolutely fine, and as a production CPU.... this isn't the CPU of choice anyway, you'd clearly be recommending others irrespective of thermals.

 

[The red spirit]

Another CPU that I won't be recommending to average users due to its insane cooling requirements at stock.

I wish Intel sold T version directly to customers, too bad that it would make their K version look bad.

 

[Colddecked]

The biggest selling point of Z790 is support for M.2 NVMe SSDs—an ability that AMD natively offers on their Zen 4 platform.

From the conclusion. I think you left out "Gen5" which was clarified later on in the paragraph.

 

[mb194dc]

Another hot power hungry monstrosity. Same as AMD 7x also.

Pretty much every use case is better off on prior gen.

If you're buying latest kit and using 1080p ok go for it I guess.

 

[W1zzard]

From the conclusion. I think you left out "Gen5" which was clarified later on in the paragraph.

Whoops .. lol .. fixed in all 3 reviews .. so many people missed that during reading

 

[P4-630]

If you're buying latest kit and using 1080p ok go for it I guess.

What are you talking about.