13600K vs 12600K – P Cores vs E Cores Benchmarked

One of the things I tested on Intel’s 12th generation chips was their P core and E core performance independently, so I think it’s worth seeing how the new 13th gen chips fare, starting with this 13600K! I’ve also included a full set of benchmarks for the 12600K to see how much of an improvement the new 13th gen chip offers – and spoilers… it’s a lot.

I want to preface these numbers with a disclaimer. These hybrid CPUs still have plenty of bugs. None were complete deal-breakers, but take my 12600K results for example. Here’s the peak power consumption figures I recorded. The power limit for the 12600K is 150W, and yet it’s drawing just 105W (while still getting full performance). The same goes for the 13600K, which peaked at 150W despite its power limit now being 181W. There’s a number of little bugs like that, and performance related issues, that still exist on this relatively new platform. Also to make it clear, I’m using an Asus Z790 STRIX E motherboard with 32GB of DDR5-5600 RAM, an RTX 3060 and a Corsair H150i cooler for my testing here.

Right, I’ll start with Cinebench R23 multi thread, which has plenty of interesting results. The 13600K’s minor core and major clock speed bump means the P cores are only just shy of the entire 12600K’s result. The 13th gen E cores are a touch faster than the 12th gen’s even if you double their results, in fact a comparably specced 12th gen chip would need 17% more performance to match the 13600K, although chip for chip the 13600K is 134% faster in the E core department. Something that’s also interesting is if you add up the P and E core performance of each chip, on the 12th gen chip you get almost identical performance to stock with all cores enabled at once, but on 13th gen you get about 3% less performance. Another bug perhaps?

Moving onto single threaded, it seems like there is only a 3% improvement from 12th to 13th generation P cores here. This is a lot smaller of a gap than I’d expect, especially since in stock vs stock configurations it’s more like a 6% gap. But on the E core front there seems to be a more considerable 11% improvement.

In Blender and the BMW scene the E cores in the 13600K help decrease the render time by a full minute compared to P cores alone, versus a more modest 40 second gap on the 12600K. The same over twice as fast gap applies to the E core results too, with the 12600K running 132% slower. In gooseberry the spread closes a little, with the 13600K’s P cores coming within 5% of the 12600K’s stock performance.

Premiere Pro suffers considerably, running on just the P cores on either chip, with the 13600K dropping from 953 points down to just 802. Surprisingly though, the 13600K’s E cores aren’t far off the 12600K’s P core result with 612 vs 551 points there. After Effects refused to complete on either E core only run for me, but its performance wasn’t down much from stock on the P cores alone. That makes me think After Effects isn’t making much use of those E cores in regular operation. Finally in Photoshop there was a slight drop going from stock to P core only on both, although again a healthy lead for the 13th gen. In fact the 13600K’s P cores outperformed the 12600K in stock 10 core configuration which is mighty impressive.

As for gaming, I’m going to stick with 1080p results here, with CSGO being the best example of lower performance on the slower cores. The 13600K’s P cores performed on par with the stock configuration – actually slightly exceeding stock – whereas the 12600K isn’t quite as good when just running on its P cores. Interestingly, even on the 13th gen E cores you can still get a pretty decent average result, albeit with a pretty stuttery experience shown nicely by the 1% low figures being almost a third of the P core’s result. Compare that to the 12600K’s E cores though and you get almost half again. Ouch.

Cyberpunk delivers a frankly incredible result, which is that the 8 E cores on the 13600K offer more performance than the stock 12600K, the same as the stock 13600K, and the same as the 12600K’s P cores. In fact, looking at the results I wouldn’t be too surprised to find the E cores are doing a lot of the work at stock here as both the 12th and 13th gen chips see a healthy performance improvement with their E cores disabled – the 13600K jumping almost 10FPS up and the 12600K with a more reasonably 3 FPS improvement – although still seeing 10 FPS higher in the 1% lows.

If you were doubting the accuracy of that result, Shadow of the Tomb Raider is here to confirm it, as the 13600K’s E cores again ran at functionally stock performance while the P cores alone ran faster. Testing with a 3060 might also be limiting the gains here, but the fact the P cores can eke out more performance is impressive. That isn’t the same on the 12600K though, with its E cores suffering considerably, although the P cores alone do still net a better experience overall.

Microsoft Flight Simulator finally breaks the trend of the E cores matching stock performance with a significant – but not catastrophic – dip from them. The P cores alone though do still offer more performance than stock in both chips – so it seems like disabling the E cores is still a useful tool to get more gaming performance out of these hybrid chips. Makes you wonder why they bother to include them for these gaming focused chips!

So to wrap up, the 13600K is a considerably faster chip than last gen. Even excluding the doubled E cores, across the board you get more performance. Even in gaming, with a relatively limiting RTX 3060 GPU you still get more performance than last gen. Hell some of the E core only results were just as fast here. That’s impressive for sure. In compute performance the P cores are around 12% faster on their own, and the E cores are not only twice as fast, but more like 130% faster. And that’s possible with the older Z690 and B650 boards, and hell even with DDR4 RAM if you want. Of course the faster the GPU you have, the bigger a difference you’ll see too.

As a final note on this launch, I want to mention that until literally seconds before filming this video I had none of AMD’s new Ryzen 7000 series CPUs, nor a high enough end GPU to do overly meaningful results, hence choosing to go down the tuning and RTX 3060 route. I now have a system from Cyberpower complete with an RTX 4090 so you can look forward to plenty more videos on especially the higher end chips with an unimpeded GPU, so as always make sure you are subscribed so you don’t miss that.