Is AMD in trouble? R5 5600X vs i5-12600K meta analysis...

 

Yeah, I splurged on a stock photo license for this piece...

Dear friends, 

We are gathered here today to lay to rest our dear friend, Zen 3. Zen 3 was received in this world in a mostly positive way, despite their parent, AMD, charging more for them than was expected or deemed seemly. However, Zen 3's life was cut dramatically short on the 04th November 2021 after just one year  on the market, when Intel released their 12th generation of processors, previously codenamed "Alder Lake".

And so, we pay our respects to Zen 3. Ashes to ashes, dust to- OMG, it's alive! It's STILL alive!

Yes, there are quite a few outlets and commentators stating that Alder Lake has destroyed Zen 3 but I think that really only applies at the high end.... and it sort of only applies if you compare Intel's tray pricing to AMD's launch MSRP. Neither of which reflect the current reality for these parts or the reality over the last half a year.

For this comparison, I'm going to focus on the mid-range: R5 5600X versus the i5-12600K.

Now, I'm not really one for anything other than gaming, so I'm going to leave the "content creator" angle and the "professional" workloads off the table because I think mixing the three together, as a lot of review outlets have done, just really doesn't make sense. The use-cases are non-overlapping and for anyone who was a budget content creator in the recent past, it made more sense to purchase either the R7 3700X/3800X, i7-10700K, or the i9-10850K - the first couple of which had pricings below*300 (I've seen them down as low as *250) and the 10850K I've seen below *400, around *350-375. I doubt anyone was choosing the R5 5600X for those use-cases, so that leaves us on a purely gaming-perspective.

*insert currency here: $/€/£

 

Alder Lake comes in three SoC flavours, with the 12600K being a cut-down version of the full-fat die...

As I said above, the high-end is practically lost to Intel's chips if you don't care about power draw. However, the gaming power draw for both the i5 and i7 parts is pretty comparable to the Ryzen SKUs, so that's not really a factor and, it seems that the Intel parts have a lower idle state than their Ryzen competitors, meaning that you will save some power if you're using your PC in an idle state a lot. Similarly, if you're looking at long-lasting multicore performance tasks, sure - the Intel parts consume more power but they complete the tasks in a lower amount of time, meaning that, overall, they consume less power than the Ryzen equivalents.

But it's not a huge win or anything, we're talking around a 10% improvement in power usage for a given task from the numbers I've seen in various reviews (and, again, I'll remind you that's heavily task-dependent!).

So, looking at gaming.

I gathered the data from the following reviews: PC World, GamersNexus, Tom's Hardware, Digital Foundry, and Hardware Unboxed. I then also gathered real world pricing data from the following sites (accurate as of producing this blogpost): Overclockers UK, Mindfactory, Caseking, Amazon DE, Ebuyer.

The reasons for choosing the above are my familiarity with the outlets and also, in the case of the online stores, the availability of pricing information. There were other stores that did not yet list the price of the 12600K, meaning that a comparison was impossible. All data was gathered from Windows 11 testing with the AMD bugfixes in place at whatever 1080p settings each outlet used.

I will make one important note - all the data is a little incomparable since each review outlet used different cooling solutions and different RAM setups. Where DDR5 was used, I didn't specify which was used but they vary from 4400 - 5200 MT/s. Also, specifically, Hardware Unboxed is the only outlet being covered in this data that did not run DDR5. I will note that other outlets saw mostly negative gaming performance on DDR5 compared to DDR4 (e.g. Tom's Hardware) and this slight difference is enough to bring the performance slightly in the favour of the 12600K over the 5600X.

However, these positive differences in performance are smaller than the CPU price difference, as we will now see...

After gathering the review data, I averaged the results - comparing against the listed prices (I also compared the geometric mean but this did not alter the conclusions, only absolute results). I also looked at the median result and compared that with the prices.

Prices in local currency of each store...

The reason I looked at the median is that I believe that this skews the results towards the more commonly experienced range of performance. Looking at the arithmetic mean (average) and the geometric mean, the results lean heavily towards the R5's favour, showing that this is the more cost effective CPU purchase for the performance.

Switching to the median shows that there's more of an equilibrium in play than the average would suggest: when you're looking at purely gaming performance, and only CPU cost, if the price is close enough, the 12600K wins out. And, at the time of writing this piece, the price is quite close between the two competing parts.

However, there's more to this story than the current situation and comparing only CPU pricing.

Table showing ratio of price to average gaming performance. Higher is better - Red text indicates the better option in each store...

Table showing ratio of price to median gaming performance. Higher is better - Red text indicates the better option in each store...

If we take for granted that DDR4 will continue to be better for the 12600K than DDR5 in gaming performance then we can remove that price difference. However we cannot remove the difference in pricing for the available motherboards. Nor can we remove the fact that the 5600X comes with a perfectly serviceable stock cooler for gaming tasks (I'm using it on my 5600X). These two aspects alone mean that, ignoring the CPU cost, the price/performance ratio of the i5 is worse than the R5... much worse.

Yes, of course, we can then turn around and say that in the near-ish future, there will be cheaper motherboards in place for users to plug their i5s into. However, I don't care about the future. We can play that game and there's a very simple counter-example to this logical argument: the R5 5600X has been much cheaper historically than it is at this point in time.

Looking at the prices listed above, we're not far off from the introductory pricing that the 5600X was introduced at before prices were hiked immediately after launch when availability was low. However, it has been down as low as £240 and €250 to 270-ish.

I don't even need to draw up a table showing the ratio of price/performance taking that into account. The R5 part smashes the i5 out of the water in CPU cost alone and when taking into account the platform cost, obliterates it. For gaming, the 12th gen i5 is not a better buy. That's it, that's the take-home message.

As to why the the 5600X is higher than historical pricing right now? Well, it likely has had its price increased on purpose before the launch of the 12th gen processors so that it can be dropped with "sales prices" shortly after. A bigger % drop looks better to consumers and they will choose that over the 12600K, especially with Ryzen's brand recognition and platform mindshare at this point in time. I'd be looking for some great stock-clearing Black Friday and Christmas deals!

The Zen 3 parts will most likely continue to be better value for gamers than Intel's 12th gen processors, especially once price cuts are implemented. This is a good thing for gamers!

Conclusion...

So no, Zen 3 isn't dead. Realistically, it's still the best platform for gaming, especially at the mid-range. At the lower-end, AMD hasn't been competitive since the release of the R5 3600. Both 10th and 11th gen processors from Intel have provided amazing price to performance, way in excess of AMD's offerings (have they actually had any?). If you want a better part than a 5600X in terms of price/performance, don't get a 12600K - get a 10400 or 11400 chip and platform for lower prices.

You won't realistically notice the performance difference and you will essentially be able to plow that extra money back into getting a better graphics card - which is where you're currently better off spending some money right now.