AMD Ryzen 5 2600 / 2600X review: The Intel Core-i5 Coffee Lake killers

AMD Ryzen 5 2600X

The Ryzen 5 2600 and 2600X are AMD’s new mid-range desktop CPUs, and they’re primed and ready to take on Intel’s 8th Gen Core i5 Coffee Lake processors. With six cores and 12 threads apiece, plus respective base clock speeds of 3.4GHz and 3.6GHz, they may not look like huge improvements over their 1600 and 1600X Ryzen predecessors on paper, but this time it’s what’s inside that counts, as both chips now have a faster, more efficient architecture behind them and better tech to help them reach their improved max boost clock speeds of 3.9GHz and 4.2GHz more regularly.

Today, I’ll be looking at both the Ryzen 5 2600 and its X-rated sibling together in one big mid-range face off, pitching them against each other and seeing how they compare to help you decide which one is worth buying.

The good news first. If you’re an existing Ryzen owner, or own an AM4 motherboard, then you won’t need to get a new one to start using the Ryzen 5 2600 or 2600X. You’ll probably need to perform a BIOS update to ensure your board is definitely compatible and has the right support, but fundamentally they’ll work with every AM4 socket board that’s currently available – a much more enticing and wallet-friendly prospect than upgrading to Coffee Lake and having to get a whole new motherboard in the process.

You also get a cooler for each processor in the box, which is another nice thing that helps save you a bit of money when you come to upgrade. In the case of the Ryzen 5 2600, you get an AMD Wraith Stealth cooler, while the Ryzen 5 2600X gets an AMD Wraith Spire.

Ryzen 5 2600X cooler

You can read more about all the new features that come with these new 2nd Gen Ryzen chips in our big AMD Ryzen+ article, but the biggest improvements all revolve around getting the best out of your processor’s clock speed, letting each chip run faster for longer without having to worry about the faff of overclocking. This is largely down to AMD’s improved Precision Boost 2 tech, which will now run each CPU core as fast as it can (where temperatures allow) whenever needed, bringing improved performance to multitasking situations where the CPU isn’t completely overloaded, such as playing and streaming games simultaneously.

AMD’s XFR 2, or Extended Frequency Range 2, tech also comes into play with Ryzen+, helping to improve each processor’s multi-thread performance when CPUs have good cooling. You’ll probably need a better cooler than the stock model that comes in the box to really take advantage of this, all told, but it essentially allows a processor to keep boosting up to 100MHz over its max limit where conditions allow – which is handy for anyone into water-cooling or you have a bit of money to spare on a beefier CPU fan and want to eek out the best possible performance from your respective chip.

Ryzen 5 2600X pins

Let’s see what that all means in practice. To test the Ryzen 5 2600 and 2600X, I stuck both of them into the following system: an Asus ROG Strix X470-F Gaming motherboard, 16GB of G.Skill Sniper X RAM, an Nvidia GeForce GTX 1070Ti graphics card, and a Samsung 850 Evo SSD.

Starting with Cinebench R15, which uses Maxon’s real-world Cinema 4D engine to render a complex, photo-realistic 3D scene of orbs and baubles (whatever floats your boat, I guess), the 2600 is more or less on par with last year’s Ryzen 5 1600X, which is pretty good going considering its lower clock speeds and more energy-efficient 65W thermal design power (TDP) compared to the 1600X’s 95W TDP.

That’s still not a massive increase over the regular 1600, mind, but you can still expect around a 5% increase in single core tasks and a boost of roughly 8% when it comes to more demanding multicore scenarios.

The 2600X, meanwhile, was around 5% faster than the 2600 in both of Cinebench’s single and multicore tests, and by extension the same gap applies to the 1600X. That doesn’t sound like a lot, but compared to the competition over at Intel, it’s surprisingly significant – at least when it comes to multicore performance.

When I put my six-core 3.6GHz Intel Core i5-8600K through Cinebench as well, for instance, its single core score came in 15% faster than the 2600 and 11% faster than the 2600X, but in the multicore test, the 2600 came in 16% faster and the 2600X was 20% faster.

I saw similar results in Geekbench 4 as well. Once again, the Ryzen 2600 either matched or surpassed the 1600X in single core and multicore performance, while the Ryzen 5 2600X was about 5% in front on both counts. Here, the multicore gap between new Ryzen 5s and my Core i5-8600K was less pronounced, but the Intel chip could still only match the Ryzen 2600X rather than surpass it – which, considering the difference in price (around £25 and $15 more over the £194 / $230 Ryzen 5 2600X, and the fact you get a cooler with AMD), is fairly significant.

That’s all well and good for general computing tasks, but the impact on playing games is much, much smaller. Of course, a lot of your gaming performance is going to be dependent on what type of graphics card you have rather than your CPU, but your processor is still responsible for certain tasks such as physics bits and pieces and all that streaming jazz you may or may not occasionally partake in.

At higher resolutions, you can expect to see naff-all difference between having a Ryzen 5 2600 and 2600X, as they both produced an overall average of around 40fps in Rise of the Tomb Raider (40.4fps and 40.7fps if you really want to get specific) when I ran its internal benchmark at 2560×1440 on Very High graphics settings. For comparison’s sake, my Core i5-8600K managed a whole extra frame (41.2fps), but that’s a small enough difference to basically call it a draw.

Things got more interesting at 1080p, but only slightly. While the Ryzen 5s produced respective overall averages of 62.8fps and 63.7fps, the Core i5-8500K positively leapt ahead with… 66.5fps. So still not much of a difference, but it would seem that the age-old wisdom of Intel being better than AMD for gaming still holds true, if only just.

Ryzen 5 2600

Still, one thing is absolutely clear. If you’re contemplating upgrading to a 2nd Gen Ryzen 5 CPU, then you’re probably better off going with the regular £169 / $200 Ryzen 5 2600 unless you’re the photo and video-editing type who regularly uses their PC for lots of demanding creative stuff as well as playing games. Otherwise, you’re just wasting money, as gaming-wise you’re not really getting any real, tangible benefit by opting for the 2600X whatsoever.

Then, of course, there’s the battle between the Ryzen 5 2600 and Intel’s suite of 8th Gen Core i5 chips. The Core i5-8500 is closest to the Ryzen 5 2600 in price (coming in at £170 / $205), but has a slower base clock speed of 3.0GHz and can’t be overclocked, so you’re really looking at comparing it with the £210 / $260 Core i5-8600K if you want a fairer bit of competition.

And really, if it were me building a new PC from scratch today, I think I’d be seriously tempted to go with the Ryzen, as I not only get to save a significant chunk of change (£50 / $60) compared to the 8600K, but I’m also getting better multicore performance, nigh on identical gaming chops and I don’t have to worry about spending even more of my hard-earned cash on a cooler, either, making it even better value for money.

There’s still that pesky Spectre security flaw to put up with, of course, but at least going with Ryzen means you’re also saved the hassle of dealing with the Intel-only Meltdown problem at the same time. AMD may have spent years walking in Intel’s shadow, but with the Ryzen 5 2600 at their disposal, their mid-range game is definitely back in business.

20 Comments

  1. Chorltonwheelie says:

    “it would seem that the age-old wisdom of Intel being better than AMD for gaming still holds true, if only just.” unless you realise you’ve payed over the odds for a K series and overclock the shit out of it then the old i5 wipes the floor with the brand new hope.
    “Killer” sigh.

    • causticnl says:

      “overclocking” on an i5, hahahahahaha

      hahahahaha

      hahahahah

      hahahaha

      hah
      thanks, I needed that.

      • eqzitara says:

        Wtf you talking about?
        i5 OC just as well as i7 and even better since hyperthreading raises temps. AMDs are known to OC less.

        I have my i5 8600k at 5.0ghz easily on a $50 fan.

      • brucethemoose says:

        Yeah, I’m not sure what you’re on about.

        The i5 8600k has more OC headroom than any consumer Intel chip from the past few years (except the G3258, maybe, or one of the older locked chips with a hacked chipset).

        Ryzen, on the other hand, is already clocked near its limit, unless you have a sub-ambient cooling setup.

        • aircool says:

          Yep, 5Ghz is quite pedestrian for the 8600K and easily achievable with a decent fan cooler.

          As for AMD/ATI, I’ve had their stuff in the past, and whilst it’s decent hardware, there always seem to be problems with new games not performing as intended on AMD/ATI hardware. A casual glance through most gaming forums indicates that this is still the case.

          • brucethemoose says:

            Maybe. I think more QA is done on Nvidia cards b/c of bigger marketshare, hence you often get better day-1 performance… But playing day 1 AAA releases these days is a PITA anyway.

            I just switched to an Nvidia card, and I’m having alot more (non day-1) driver issues with it than my 7950.

            Anyway, CPUs don’t need driver-side optimizations, so there should be no issues at all.

        • KenTWOu says:

          Ryzen, on the other hand, is already clocked near its limit…

          Most of the reviews I’ve seen say that XFR 2 in these new Ryzen models is so good, so they recommend buying either 2600X or 2700x and forget about manual overclocking altogether.

    • dr.denton says:

      I love how people assume that everyone wants to overclock and when they do, they win in the silicon lottery.

      Running a product outside its specifications is a non-factor in the assessment of its value for regular customers.
      Most just want a PC that works. Period.

      But lets totally fuck around with voltages and cooling and turbo modes and spend hours of testing, just to have your PC randomly crash a week later, because all your nice little stability tests didn’t QUITE find that last kink, so you increase voltage by 0,025V and do another 48h burn-test …
      All that for what? 20% more performance? Maybe 30% if you’re really lucky, maybe just 10% if you’re most people …

      By all means, go and play around with clocks and voltages and memory timings to your heart’s content. I know it can be fun.

      But understand that you are a tiny minority and therefore you don’t matter to either AMDs or Intels bottom line.

      • sion12 says:

        That is only if you want to push OC to the limit. you can easily just select auto OC in the bios and OC to 3.7ghz to about 4.2ghz on a 8600k.

        • Sakkura says:

          The 8600K turbos to 4.3 GHz by default…

        • dr.denton says:

          /this is in reply to sion12’s post

          True. But seeing as both CFL and Ryzen can easily hit 4Ghz, Intel’s advantage at frequencies around this mark are practically zero.

          My point was: if you really want to take advantage of CFL’s potentially higher single core(!) clock speeds at a given price point, you need to go way beyond 4,3Ghz and that means overclocking. Which most people don’t want to bother with.

      • Chorltonwheelie says:

        We are pc enthusiasts with quite deep knowledge of our systems. Sorry if that upsets you mate.

    • Cooe says:

      You’d think this would be the case, but you’d actually be wrong. Tons of reviewers tested overclocked results as well as stock, (Hardware Unboxed/TechSpot [YouTube channel/website of the same group] as one of many examples) with the Coffee Lake chips even pushed as far as 5.1GHz (which generally requires a killer chip, along with a delid, & pricey monster cooling bc Intel refuses to solder; all of which absolutely kills any “value argument” it had for a gaming rig), and it really doesn’t change the gaming results anywhere NEAR what you seem to think, and definitely nowhere near enough for said reviewers to give the i5 the nod over R5 2nd Gen. And in games/engines that can properly utilize more than 6-concurrent threads (more & more each day in our 8-core/thread console world), the 12-thread Ryzen 5’s put up better all important 1% & .1% lows vs the i5’s, no matter how high the 8600K is OC’d. And if the i5-8600K is already experiencing symptoms of thread saturation & the microstutter, thread-hitching, and frame-pacing/timing issues in specific games today, what does that say about it’s potential longevity? Especially with 8-core/16-thread consoles expectated in 2020 & Z370 being a totally dead platform.

      Also, for the i5’s to have any real gaming advantage AT ALL, one must be notably CPU bound, which means 1080p 144Hz with a GTX 1080 / Vega 64 or faster GPU, which is both an uncommon setup to begin with, and ESPECIALLY for the mid-range market these CPU’s compete in.

      And if your wondering why the i5K doesn’t gain all that much in gaming performance despite throwing a delid, crazy cooling, and a monster core OC at it, there’s a large number of reasons but one of the biggest is that the chip quickly gets bottlenecked elsewhere outside the cores themselves. This is because Intel’s got the core clock multipler totally divorced from the rest of the chip beyond the actual cores themselves. Things like the cache & IMC, and the rest of the uncore all have totally separate clock gen’s, so overclocking the cores w/o overclocking the rest has very limited gains vs the actual MHz increase in many real-world dynamic, cache sensitive workloads like gaming. And if you attempt to notably overclock the cache as well, you’ll quickly run face first into CL’s horrible thermal & power wall; expensive delid + custom loop be darned.

      Ryzen otoh, uses just 2x clocks, one for the entire CCX’s (cores & cache, meaning a higher core-clock actually increases Ryzen’s IPC by increasing cache bandwidth & reducing latency; producing across the board gains), and another for the entire uncore (IMC, Infinity Fabric buses, I/O, etc…) which it simply takes from the DDR4 clock gen. Aka, whatever clock your ram is at (1/2 listed speed bc “Double Data Rate”) that’s what the uncore’s clocked at as well. Both of these mean that AMD gains VASTLY more from increased core AND memory clock-speeds than Intel does, and with gaming as crazy cache latency sensitive as it is, Coffee Lake’s gaming performance quickly plateau’s with higher core clocks. And when you throw in all the extra cost & effort requried to even clock Coffee Lake that high to begin with, the i5’s look even less appealing for a bang/$ gaming rig then they already did.

      The 6c/12t i7-8700/i7-8700K still have a valid place in the market for those 1080p 144/240Hz CPU bound enthusiast gamers discussed above that really want every single frame that they can squeeze out, but the 6c/6t i5’s, and ESPECIALLY the i5-8600K (the cheaper i5-8400 is much easier to argue) look really poor from a big picture perspective vs Ryzen 5 2nd Gen, be it for a primarily gaming focused rig or not.

  2. brucethemoose says:

    The great thing about Ryzen is upgradability. A few years down the road, Zen 2/2.1 (and just maybe 3?) should work on an AM4 motherboard you buy now. Intel’s consumer chips, on the other hand, seem to require a whole new chipset every generation, despite using almost the same architecture. That means you gotta buy a whole new motherboard to upgrade, and go through the hassle of re-installing Windows.

    That being said, while it may not show in Tomb Raider, Intel’s CPUs are absolutely better for faster turn times/performance in late-game Civ or Stellaris, and will give you more FPS in those janky early access games.

    In fact, that’s prime territory for an RPS review. Instead of focusing on AAA games (which are threaded enough to be GPU bound), run late game strategy save, or modded Minecraft, or something else more “obscure”. Something the 100 other hardware review sites don’t do, because they have to stick to deterministic tests for graphs.

    • Don Reba says:

      That means you gotta buy a whole new motherboard to upgrade, and go through the hassle of re-installing Windows.

      You don’t need to reinstall Windows for a motherboard upgrade. Even for something as drastic as switching between Intel and AMD, it’s going to work just fine.

  3. DEspresso says:

    Got my eye on a 2600, it’s quite unfortunate only the 470 boards are available yet. Any word when the first 450s will appear?

    • Sakkura says:

      No idea, but you can use a B350 (or X370 etc) board. The BIOS versions that support the APUs released a few months ago also support the new Ryzen 2nd gen CPUs, so getting an up to date board shouldn’t be too hard.

      And if you should end up with a board that hasn’t been updated, there’s still the AMD boot kit offer. They lend you a compatible CPU for free, just so you can get the BIOS update done.

Comment on this story

HTML: Allowed code: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>