@iratefrenchfry

♬ Sad violin ballad - Inumori

It's not at random, though. It's controlled by their Twitch chat—his seemingly maniacal Twitch chat. Every time IrateFrenchFry is trying to be a good teammate on the objective or in a tense stand-off with enemy players, down comes a visor that blinds him.

If you're wondering how it works, here's what IrateFrenchFry told me:

"I programmed a microcontroller board that monitors my chat and looks for certain words. If it sees one of those words it activates whatever function I want.

"For example, one of my scripts monitors the chat for '!blind', if it reads that word then it activates a servo on my headset that lowers blinders."

The gaming headset is nothing special: just your run-of-the-mill Logitech G733, with the obvious exception of the servo powered gizmo on the right ear-cup.

"Mechanically the glasses work through a linear actuator that I 3D printed. The blinder (just glasses that I bought and painted) is attached to that linear actuator to allow them to raise and lower."

IrateFrenchFry limits how often the activation words can be said by having it so only his own account can say them in the chat. That means it's not just a dreadful free-for-all. Any more so than intended, anyways. You might be wondering if that means he's just having us on and they're pressing the button for the visor to drop, but no, it's still the Twitch chat, but they have to earn channel points the ol' fashioned way to activate it.

What's better than making your favourite Twitch streamer do a dab (or whatever it is people use channel points for these days), or blinding them temporarily during an intense moment in a competitive game?

An IRL jumpscare.

@iratefrenchfry

♬ original sound - IrateFrenchfry

Using their mechanical know-how, IrateFrenchFry also rigged together a system that flashes the lights, unleashes a scary puppet in front of them, and plays a blood-curdling scream.

Combined with a horror game and the visor, it's a pretty overwhelming amount of stuff going on at any one time, both in real-life and the game—for me, anyways. IrateFrench used to be in helicopter search and rescue in the United States Coast Guard, so he might be used to loud noises and tough working conditions.

You can check out more of their clips on their Twitch channel or TikTok.

The reason being that Notebookcheck has discovered EEC filings (via VideoCardz) for a Legion Go S 8ARP1, 8AHP2, and 8ASP2—three separate models. There's also speculation over what these enigmatic model names might mean, and they might hint at a mess of upcoming AMD Z2 processors spanning three different generations.

AMD's Z1 Processors (the Z1 and the Z1 Extreme) adorn a couple of the best handheld gaming PCs today: the Asus ROG Ally and Lenovo Legion Go. The Z1 Extreme processor is, for all intents and purposes, the same as the mobile 7840U (Phoenix Point architecture) or 8840U (Hawk Point) processors, but made specifically for handhelds.

There's therefore been a lot of excitement surrounding possible AMD Ryzen Z2 or Z2 Extreme processors, primarily because these could boast next-gen Strix Point architecture that we already find in some gaming laptops. Along with Intel's new Lunar Lake processors, Strix Point chips provide improved performance and battery life compared to previous-gen processors, and many are hoping they will soon find their way into new handheld devices.

Notebookcheck rightly points out that the listings for these three new Lenovo Legion Go handhelds could very likely refer to different AMD architectural generations being used in their different AMD Z2 processors. The "8AHP2" could refer to a Legion Go with a Hawk Point processor—possibly functionally equivalent to the Z1 Extreme processor that's found in the current-gen Legion Go—and "8ASP2" could refer to a Legion Go with a new Strix Point Z2 processor.

As for "8ARP1"? I'm not sure. Notebookcheck suggests that it could be powered by an AMD Rembrandt APU. However, Rembrandt processors are pretty old now, and I'm not so sure "RP" would be chosen for this codename—where's the "P" in "Rembrandt"? There is reason to think this might make sense, however, as we'll see.

The question is why we'd think all of these AMD processors of different generations that will supposedly feature in Lenovo Legion Go handhelds are Z2 processors. The answer comes from another leak, this time from well-known Billibilli forum leaker Golden Pig (via VideoCardz).

According to Golden Pig, there will be a few Ryzen Z2 series processors, with the top-end Z2 Extreme featuring 16x RDNA 3.5 CUs (four more than the Z1 Extreme's 12 CUs), this being essentially the same as the Strix Point 890M mobile GPU.

In addition, AMD's also expected to launch a Z2 (non-Extreme) APU using Zen 4 architecture with RDNA 3 graphics and a Z2G APU using Zen 3+ architecture and RDNA 2 graphics.

Put two and two (or perhaps what feels more like 50 and 50) together and you have three AMD Z2 processors spanning three different architectural generations, and three upcoming Lenovo Legion Go handhelds with codenames that seem to refer to these very same architectural generations.

If these leaks are correct, this would suggest that there will be three AMD Z2 handheld processors, each one using a different architectural generation of processor, and three different upcoming Legion Go handheld gaming PCs that will each house a different one of these Z2 processors.

Whew, that was a lot. But once you've sifted through the ridiculous naming schemes it's quite simple. And given there's rumour of a Legion Go Lite—and even possible pictures of it—it would make sense for there to be one sporting a current-gen (Hawk Point) or older-gen (Rembrandt) APU.

As for why AMD would call these older-gen processors "Z2", well, it has to call them something, doesn't it? And technically the Z1 Extreme is Phoenix, not Hawk Point, even if they are essentially functionally equivalent.

The joy of naming schemes, eh? Kinda gives me flashbacks to AMD's wheel-o-laptop decoder chip…

You may have already heard a rumour-mill version of this story, something along the lines of, 'intern injects malicious code into AI model, sabotaging 8000 GPUs and causing ByteDance to lose tens of millions of dollars.' Besides TikTok, ByteDance have also created Doubao, an incredibly popular AI chatbot Bloomberg called "China's answer to ChatGPT."

Though in a recent social media post ByteDance said that none of their commercial projects were affected by the rumoured sabotage. That said, their statement reveals that there's more to this story (via Ars Technica).

ByteDance have confirmed the intern was fired back in august for "serious disciplinary violations," including "maliciously [interfering] with the model training tasks" for at least one research project. ByteDance also says that the situation was serious enough to warrant reporting the intern's behaviour to their university, in addition to industry contacts besides that.

According to ByteDance, the intern in question was part of the commercial technology team, not their AI lab. The implication of some available translations of ByteDance's statement suggest that the intern was part of an advertising team rather than a technical team, though commenters under the social post dispute the distinction, claiming that the commercial technology team was part of the AI lab.

Ultimately, ByteDance claims the intern misrepresented some details on their social media profile, and that a number of resulting reports have overstated what happened as a result. Those 8000 GPUs and alleged millions lost? The company says that this was "seriously exaggerated" but then don't elaborate on the actual figures beyond that. Maybe as a result of this ambiguity, commenters took umbrage with this too, accusing ByteDance of downplaying the damage done.

To summarise, a ByteDance intern—who has since been fired—did interfere with the company's AI model training. The company claims the damage is not as far-reaching as rumours suggest, though is unwilling to plainly state how consequential it really was, but still got the former intern's school involved in disciplinary action. And you thought your work experience was a disaster.

However, in games, the Core Ultra 9 285K and its siblings are behind their 14th-generation Core predecessors, as well as equivalent Zen 5 and Zen 4 processors from AMD. Unless you're a diehard Intel fan or favour content creation over gaming, there's little reason to buy one.

However, you might be getting one in a new prebuilt gaming PC and if that's the case, you may also be wondering if it's worth spending more cash on a high-speed DDR5 RAM kit.

Intel's Core Ultra 200S series of chips all support DDR5-6400 by default, which is a pretty big jump up from the DDR5-5600 that Raptor Lake officially topped out at. That said, 14th Gen Core chips can generally take much faster memory than that, though a lot depends on what motherboard you have.

So to see if it's worth splashing out on a super-fast, super-expensive memory kit to go with a new Arrow Lake chip, I tested a Core Ultra 9 285K with six different RAM kits, using our full suite of CPU benchmarks to see what impact RAM has on Arrow Lake's performance.

Gaming performance

Generally speaking, the performance of a game is determined by the graphics card but if it's very CPU-heavy (e.g. a flight or racing sim), then there's a chance that RAM may play a part.

This is why we use DDR5-6000 as the standard RAM kit for testing CPUs, or something that's very close to it. It's a speed that AMD Zen 4 and Zen 5 chips are both generally happy with and most of Intel's recent processors too.

In the case of the Core Ultra 9 285 and Ultra 5 245K, I used a dual-channel kit (two 16 GB sticks) running at 6,000 MT/s with a CAS latency of 32 cycles.

The above results for the gaming benchmarks in our CPU testing suite show a best-case difference of just 6 fps on average, between our standard DDR5-6000 kit and the fastest one I had, a dual-channel 48 GB set of Kingston Renegade Fury DDR5-8200.

This particular set isn't available in stores just yet but a pretty similar 48GB Kingston DDR5-8000 set costs $340 at Newegg—more than twice the price of the DDR5-6000 we normally use.

I'm not being negative about Kingston's memory or prices here, as that's the going rate for ultra-fast 48 GB DDR5 RAM kits.

Our Factorio benchmark uses a custom map, packed to the rafters with belts and machines, and we run a script that generates no graphics—the test runs the underlying engine to see how well the CPU copes with an enormously complicated map. It's by far the best gaming test of a processor's L3 cache and memory sub-system around.

Arrow Lake really struggles in this benchmark, being much slower to process 5,000 updates than Raptor Lake and Zen 5, but one can see that the two fastest RAM kits speed things up considerably. Proportionally, though, it's still only an 8% reduction in the processing time.

Baldur's Gate 3 demonstrated the best gains, with a 10% higher average frame rate, though Cyberpunk 2077 produced the best improvements to the 1% low frame rate (an 8% gain).

Content creation performance

Rather than going through the full gamut of our non-gaming CPU tests, I'm only going to show the results for the 7zip file compression and decompression benchmark. The reason for this is that there was little to no difference between any of the RAM kits in Cinebench, Blender, Handbrake, and Procyon photo editing.

However, 7zip demonstrated some remarkable gains in performance when using DDR5-7800 and DDR5-8200 compared to the standard 6,000 kit. Well, for compression, at least—one can see that decompression isn't affected at all by RAM speed.

That makes sense, though, as compression of a file into a smaller package involves lots of memory writes, with the program generating 24 threads (the maximum supported by the Core Ultra 9 285K), each taking on a compression task.

Decompression is mostly about CPU processing ability and memory bandwidth takes a back seat here, something that's true of all our content creation benchmarks.

To buy or not to buy?

I don't need to provide a protracted analysis of the results here, as the above figures speak enough by themselves. Unless one has a very specific or niche use case for an Arrow Lake chip, there's simply no point in spending a lot of money on getting a high-speed RAM kit.

Of course, I only tested a handful of games and I do not doubt that some out there will respond more with quicker memory, but a lot of that depends on what graphics card one has. If it's an RTX 4090, for example, you may get a bigger increase than I did with an RTX 4070.

But I don't that changes the obvious point here and that's you're better off sticking with a nice DDR5-6000 kit for AMD chips and a DDR5-6400 one for Intel (or even the same 6000). Instead, look at how much RAM you're getting for your money and try to get something with a CAS latency that's around 32 cycles or lower.

For gaming, Arrow Lake needs all the help it can get to improve matters, but it's not going to come from a super-speed memory kit.

Still, when it comes to ultimate sound quality, many point to wired headsets as delivering the goods better than their wireless counterparts. That means any wired gaming headset is going to have one main purpose—delivering great audio while you're sitting in front of your PC. And if you're asking $330/€300 for one, like the Beyerdynamic MMX 330 Pro I'm wearing right now, it'd better be something special. I wouldn't normally frontload a review with an overall opinion, but in this case I'll make an exception: This one is special.

So what do you get for your moolah? Well, this particular version of the MMX 330 Pro is described as an open-backed headset, meaning there are vents in the ear cups that expose the rear of the drivers to the open air. Proper open-backs expose the rear of the drivers almost entirely and are the sort of cans that will cause audio nerds to writhe in ecstasy at their mere mention, thanks to a perceived "airier" soundstage and superior audio positioning.

However, that design will leak a significant amount of audio into the space around you, and have less than ideal passive noise isolation—whereas closed-backs provide both, but usually have a more limited, narrower soundstage.

Semi-open-back headphones are something of a compromise, and I'd say the Beyerdynamics definitely fit better into this category. The small earcup vents widen the soundstage significantly, but they're not so open as to leak out huge amounts of audio or provide zero passive noise isolation. It's a bit of a halfway house solution, and this headset is all the better for it.

The drivers at play here are Beyerdynamic's 45 mm Stellar.45 units, the very same you'll find in the DT900 Pro X and DT700 Pro X studio headphones. Pay attention to that first one—it's currently the set of headphones sitting at the top of our best audiophile headphones for gaming guide.

Yep, the very same drivers are here, in this much more gaming-focussed headset. And they're fabulous. I've used a fair few studio headphone sets over the years, and while they can be stunningly accurate, they can also become quite fatiguing over long listening periods. There are no such worries here. Everything is pinpoint accurate and beautifully well positioned in the soundstage, but the overall effect is still warm, pleasant, and comforting.

Bullets whip past your ears in Grey Zone Warfare with a spine-tingling sizzle, while the grinds and crunches of my nightmare Satisfactory build have a level of depth to them that I hadn't noticed before. It's often said that the most important thing in a gaming headset is a brilliant set of drivers, even if they're not gaming-specific, and that's brought into sharp relief here.

Musically, those drivers really get to work. What's most impressive here is the beautiful, golden quality of the staging and the reproduction. That minimal earcup venting really does widen out the soundstage, and narrows up a little if you cover them with a bit of electrical tape (yes I did put cheap tape on a $330 set of headphones, out of sheer curiosity). This headset is simply a gorgeous canvas to paint your tunes across, with a shiny, self-assured quality from low end bass to high end treble.

Speaking of low end, the MMX 300 Pro is bassier than I was expecting for an open-back set. Not in a horribly overblown, Beats-like way (I could write several articles on my hatred for the Beats, but this is not the venue), but more that, should your choice in music have some significant sub-bass, like Fritz Kalkbrenner's "Facing the Sun", this set can stage it exactly where it needs to be.

Lesser drivers would either ignore that ultra-low end entirely, muddy it, or boost it to high heaven. Nope, here it sits exactly where you want it in the mix, low, dark, jaw-tingling and smooth, while the guitar part shimmers above it. Darned impressive that, and on first listen caused me to exclaim out loud to my partner:

"Man these things are good."

And then there's comfort. Put simply, I think this might be the most plush, cushy feeling headset I've ever worn. The velour earpads cosset your ears in the same way your posterior gets cosseted by your favourite couch, while the memory foam headband is supportive yet comfortable in all the best ways.

There's even a little cutout in the middle of the interior of the band to prevent fontanelle pressure, making this set an absolute pleasure to wear all day. It's a subtle thing to behold too, so you'd have no problem wearing it in a professional environment, unlike some other, zanily-coloured headsets I could name. As an all-day office warrior, I can think of few better headsets to attach to your bonce.

That being said, as a semi-open-back set (Beyerdynamic might not agree, but I'm sticking to it), you may slightly annoy your colleagues. Again, the sound leakage is here, but it's nowhere near as bad as a fully open-backed set. A good compromise, but it's worth bearing in mind that those in the near vicinity will still be able to hear what you're listening to at high volumes.

Unlike the DT900 Pro X, the MMX 300 Pro has a proper, flip down gaming microphone. This is a cardioid condenser mic, and bad news, it's not detachable. The good news is, it sounds excellent, capturing a rich, warm tone with plenty of crispy detail. There's no noise reduction software included here, so you will get a tiny amount of background hiss, and if you put it too close to your mouth it does get a little plosive. That being said, keeping it at a reasonable distance results in truly excellent headset vocal capture.

Beyerdynamic describes it as "broadcast quality", and I have a tendency to agree. I'd still put it through some very light noise reduction before using it on anything professionally recorded, so it's a shame there isn't any available as part of the package here. But the audio quality is excellent regardless, and any background hiss is minimal for a raw mic signal. It's certainly a lot better than almost all other gaming headsets when it comes to recording rich, properly weighted vocals, so it's another win for the Beyerdynamics overall.

You do have a bit of a bulky cable to contend with though. There's an included adapter to merge the mic and headphone jacks into one 3.5 mm output, and while it's excellently made and reassuringly thick, it does feel a bit overdone.

I'd also prefer something braided to prevent tangling. Still, there's a handy inline mute switch and volume dial, and it's decently long, so running it around the back of a floor-dwelling PC shouldn't be an issue.

So then, this is a supremely comfortable gaming headset, with proper studio-grade drivers, a stunning sound profile, an excellent microphone, and build quality that feels like it'll last. The only thing that gives me slight pause is the price, as one of our favourite audiophile gaming headsets, the wireless, planar magnetic, and also truly brilliant Audeze Maxwell can be had for roughly the same money.

Would I go for the Maxwell over the MMX 300 Pro? Yes. Just. That's mainly for the freedom of choice that wireless headset provides, and those staggeringly good planar magnetic drivers. The Beyerdynamic set is a truly worthy competitor though, and if I cared not a jot for taking my cans on the move, I think it'd be a matter of flipping a coin to choose between the two.

Given the audio bliss on offer here, and the sheer quality of the headset overall, I actually think that $330 price tag is just about reasonable—even for a wired set. It's a luxury, highly-engineered, and boutique piece of gear, and it demonstrates that in every single aspect it provides. It's not often I come across a headset that really makes my jaw drop with its sound quality, comfort, and overall design, but the Beyerdynamic MX 330 Pro is just that. It's nice to have nice things, ey?

That's not the only change on show here and we've covered the important changes in detail elsewhere. It's suffice to say that the new Core Ultra 200S series of CPUs is a new world for Intel as they are so different to previous desktop designs.

The Core Ultra 5 245K (and the GPU-less 245KF) is currently the baby of the Arrow Lake line-up and it's basically a Core Ultra 9 285K with two P-cores and eight E-cores disabled. The clocks are a little lower too and there's less L3 cache, all of which reduces the maximum power limit to 159 W.

While it's an obvious successor to the Core i5 14600K, the 245K's $309 MSRP puts it closer to the Ryzen 7 9700X than the Ryzen 5 9600X (currently $249 and $327 on Amazon, respectively). Architecturally, though, it's very different to both as AMD's chips are six/eight-core processors, with support for 16 threads, whereas the Arrow Lake chip has 14 cores, 14 threads.

One further processor that the Core Ultra 5 245K is close to price-wise is the Core i7 14700K. That 20-core chip can be picked for around $50 more than the new Arrow Lake, just 14% more expensive but with twice as many threads on offer.

Gaming performance

Whether any of these processors are worth buying naturally depends almost entirely on how well they perform in a variety of scenarios, so let's begin with an examination of the Core Ultra 5 245's gaming chops, tested at 1080p with a GeForce RTX 4070 graphics card.

If you've already seen the gaming results for the Core Ultra 9 285K, then the above figures shouldn't come as any surprise. While the 245K is on par with the 14600K in Homeworld 3 and Metro Exodus, it is noticeably worse in all the other game tests—15% slower in Cyberpunk 2077 and Warhammer 3, and 12% slower in Baldur's Gate 3.

The Factorio test, which is an excellent judge of L3 cache performance, runs 22% slower on the 245K than the 14700K. And it's even worse when compared to the Ryzen 7 9700X, where the 245K takes 40% longer to process the updates than the Zen 5 chip.

In fact, the CPU with the closest matching performance is the Ryzen 5 7600X, a processor that has fewer cores and threads, lower clocks, and launched over two years ago.

Where the Core Ultra 5 245K does beat the competition is power consumption—using just 83 W in Baldur's Gate 3, it uses 46% less energy than the 14600K and 24% less than the 9700X. Even though the gaming performance is worse, it's not 46% worse, and it's a near-miraculous achievement by Intel.

In previous architectures, like Raptor Lake and Alder Lake, great gaming performance came at a cost of very high power consumption, and this was an aspect that AMD easily had the measure of Intel in. Arrow Lake doesn't completely counter the power advantages that Zen 4 and 5 have, but it's a big step in the right direction.

And there is something else that the little Arrow Lake is good at.

Content creation performance

If you use your PC for nothing but gaming, then it's very obvious what the final conclusion in this review of the Core Ultra 5 245K is going to be. On the other hand, if you do 3D rendering, design work, video editing, and so on then you're in for a pleasant surprise.

The baby Arrow chip is generally better than the 14600K in the majority of our content creation tests and even beats the 24-thread 14700K, despite only having 14 threads available. Compared to AMD's processors, the 245K is substantially better, being 39% and 29% faster in Blender and Handbrake video encoding respectively.

Where it's not so good is in 7zip decompression—better than Zen 5 but worse than Raptor Lake. That's probably because the ring bus that transfers all the data between the L3 cache and memory controllers runs much slower in Arrow Lake than it does in the previous architecture, and the data has to be transferred to another tile before it can be sent to DRAM.

Speaking of memory, Core Ultra 200S processors support up to DDR5-6400 natively but you gain very little performance over DDR5-6000. Only when you step up to the likes of DDR5-7800 do you see any notable gains, though not in any of the content creation tests, other than 7zip compression (which involves a lot of memory writes).

While Arrow Lake has a significant power consumption advantage in gaming, the multi-tile chip demands a lot more energy when processing heavily multithreaded tasks. That said, where the 14600K and 14700K will bounce off their maximum power limits in such scenarios, the Core Ultra 5 245K never does, staying 20 W under the 159 W cap at all times.

This makes the 245K very easy to cool and a potential choice for those looking to build an SFF (small form factor) PC that's entirely air-cooled. However, while one saves a bit of money by not getting an AIO liquid cooler, there aren't many cheap Arrow Lake motherboards at the moment.

Conclusion

Summarising the relative merits of the Core Ultra 5 245K, compared to the 14600K, is a straightforward affair. Gaming is poor, content creation is good, and power consumption is great. But Intel isn't the only company making desktop CPUs and the summary also applies when comparing the 245K to the Ryzen 5 9600X and Ryzen 7 9700X.

The Core Ultra 5 245K isn't as disappointing as the Core Ultra 9 285K, but unless one specifically needs an affordable processor purely for content creation tasks or a low-power chip for an SFF build, there's little to attract one to the Arrow Lake chip.

And while Intel is adamant that Raptor Lake's voltage problem has been fixed, they're somewhat hard to recommend too, even though they're much better in gaming than the 245K.

That's simply because the Ryzen 5 9600X is cheaper and performs really well in games. If content creation is important, then the Ryzen 7 9700X is on par with the 245K and only a few dollars more expensive.

Intel's Core Ultra 200S series aren't bad processors, and the Ultra 5 245K is perhaps the best of the lot right now, but there are much better gaming CPUs on the market for the same money—from AMD and Intel.

In their place, you have the same use of chiplets/tiles that was first introduced with Meteor Lake. TSMC manufactures all of them, too, bar the base tile to which they're all attached. And quite remarkably, we now also have an Intel processor that uses less power than any equivalent AMD chip does in gaming.

We've already covered the architectural changes in detail, along with breaking down the new Core Ultra 200S range. For the moment, that comprises the Ultra 9 285K, the Ultra 7 265K and 265KF (no iGPU version), and the Ultra 5 245K and 245KF.

While halo products rarely sell in numbers anything like their mainstream counterparts, the 285K is going to appear in a lot of gaming and workstation pre-built PCs over the coming months.

It's also the measure of what Arrow Lake's full performance is like and I have no doubt that many people will use it to judge whether it's worth doing a full system upgrade, as Arrow Lake chips do, after all, use a new LGA1851 socket.

Gaming performance

For testing the gaming performance of the Intel Core Ultra 9 285K, I used a ROG Maximus Z890 Hero motherboard, provided by Asus. The rest of the setup comprised 32 GB of Lexar DDR5-6000 CL32 RAM, an Asus Ryujin III 360 Extreme cooler, a Corsair MP700 Gen5 SSD, and a GeForce RTX 4070.

One important thing to note here is that all motherboard vendors have been working hard behind the scenes with BIOS releases, to improve performance, memory compatibility, and stability. I used two such updates from Asus but notably, there was no difference in how well everything ran.

While the above figures aren't terrible by any means, if one compares the average frame rates of the Core Ultra 9 285K to the Core i9 14900K, it's 12% slower in Cyberpunk 2007, 9% slower in Baldur's Gate 3, % slower in Homeworld 3, 3% slower in Metro Exodus, and 15% slower in Warhammer 3. It's also 23% slower in the Factorio test.

It's the same pattern when one compares it to the Core i7 14700K, AMD Ryzen 9 9950X, and 9900X. Other than the odd case where a 1% low value favours the Arrow Lake chip, Intel's new processor is notably slower than Raptor Lake and Zen 5 in gaming. And Zen 4, for that matter.

The Core Ultra 9 285K was tested using Intel's recommended Performance power profile of 250 W for PL1 and PL2, and an Icc max value of 347 A. This is actually the default setting that all Arrow Lake motherboards use, although one can override this, of course. I examined what impact using the Extreme profile (PL1 = 250 W, PL2 = 295 W, Icc max = 400 A) had on the game tests and it practically made no difference.

And that's because Arrow Lake merely sips at power during gaming, as you can see in the last chart in the above collection of results. With an average package power consumption of just 83 W in Baldur's Gate 3, the 24-core Ultra 9 285K uses less energy than a six-core Ryzen 5 9600X.

The world of desktop CPUs is more than just gaming, of course, so all of this could be forgiven if the Core Ultra 9 285K ruled the roost in content creation, rendering, and office productivity.

Content creation performance

Historically, when it comes to content creation workloads, such as image rendering, 3D modelling, and video editing, the processor of choice has nearly always been Intel. That was until AMD released its Zen 5-powered Ryzen 9 9950X—its multithreaded performance makes it an ideal choice for anyone who requires as much of this as possible—so it's natural to expect that Intel would hope to move ahead with Arrow Lake.

Starting with Cinebench 2024, the Core Ultra 9 285K showcases just how much progress Intel has made in improving its cores, especially the E-cores. The fact that the new chip outperforms the 9950X and 14900K when it has fewer threads is quite remarkable. In the nT multithreaded test, the 285K is 4% faster than the 9950X and an incredible 29% better than the 14900K.

During this test, the 285K uses 16% less power than the 14900K, although it does use 7% more than the 9950X. In terms of power efficiency, that does mean AMD's flagship chip is still the best for this kind of workload but the 285K leaves the 14900K far behind.

However, while the good news continues across the rest of our content creation tests when comparing Arrow Lake to Raptor Lake, the Core Ultra 9 285K is notably behind the Ryzen 9 9950X in Blender, 7zip compression and decompression, Handbrake video encoding, and UL Procyon photo editing.

Increasing the power limit does help and it's possible to get the 285K on par with the 9950X, but only by using a PL1 value of 300 W or higher. In other words, by turning it into a Raptor Lake chip.

I also tested the 285K with a range of RAM kits, going all the way up to DDR5-8200, but the faster memory only improved matters by 10% in the very best case in gaming and hardly at all in content creation.

However, there is one exception: 7zip decompression. Using DDR5-8200 CL40, the 285K performed 17% better than with DDR5-6000 CL32 in that particular test, though this is to be expected as the benchmark involves a lot of writes to memory.

Faster RAM also helped iron out the microstutters I experienced benchmarking Cyberpunk 2077 and Baldur's Gate 3, with the DDR5-6000 kit.

Performance analysis

The first thing to ponder is why Arrow Lake handles games worse than its predecessor and competition. There are multiple factors behind it all and it's not a case of one of them being the exclusive reason.

Compared to the Core i9 14900K, the 285K has a much lower maximum P-core clock speed (5.7 vs 6.0 GHz); the ring bus that connects the cores to the shared L3 RAM and rest of the chip is also slower (3.8 GHz vs up to 5.0 GHz).

That's significantly lower and goes some way to explain the relatively low performance in the Factorio test. However, it's also an important factor behind Arrow Lake's reduced energy consumption.

The P-cores no longer have any ability to handle two threads simultaneously, as the much-improved E-cores essentially take over threads that would be otherwise directed to the old P-core HyperThreading.

Not that this impacts gaming very much. I ran through all of the benchmarks with the E-cores all disabled, and apart from in cases where multithreading is everything (Cinebench, Blender, 7zip, and Handbrake), the eight single-threaded P-cores managed the game tests perfectly well by themselves.

Arrow Lake's memory controller is on a separate tile from the compute tile, unlike in Raptor Lake where it's all monolithic, and the connection between the tiles runs at 2.1 GHz.

It's similar to how AMD's CCD and IOD chiplet structure works, where the Infinity Fabric that connects them together runs at 2 GHz. It's possible that a higher clock speed could improve matters but like the ring clock, it may be limited for power reasons.

There are countless other changes underneath the hood of Arrow Lake and figuring out just what exactly is the cause for the performance disparity will no doubt take a lot of investigation, and is likely what's going on at Intel right now as it switches over to work on Nova Lake, its next desktop architecture.

But what really matters right now is whether the Core Ultra 9 285K is worth buying.

Conclusion

For all of Arrow Lake's strengths, if you already have an Intel 13th or 14th Gen PC, or AMD AM5 setup, there's nothing here to make it worth jumping platform. Or at the very least, not yet. It's always possible that more performance can be leveraged out of the architecture via CPU microcode and motherboard BIOS updates. But there is no way to tell when, or even if, that will happen.

However, if you have a much older PC, the performance is good enough to warrant considering a Core Ultra 200S upgrade over a Raptor Lake one. Yes, the latter is a better gaming processor but it achieves this in part through a hefty energy demand, which in turn results in a very toasty processor.

The Ultra 9 285K uses very little power and it's very easy to cool, for gaming at least.

Most of Arrow Lake's sales will be to the OEM and system builder market, of course, and I have no doubt that such vendors have already spent significant sums of money buying chips, motherboards, and RAM ready for the retail launch.

Beyond this market, the largest number of sales will be from the PC enthusiast segment, and this is where Intel has a potential problem.

Is anyone going to spend $586, plus several hundred more on a new Z890 motherboard and DDR5 RAM kit, just for lower power consumption? The simple truth is that if energy demand was a major concern, then an AMD Zen 5 Ryzen setup would be the obvious choice, as these chips are generally better in this respect.

Overall, the Core Ultra 9 285K isn't really a bad CPU. Its performance is relatively near the competition and Intel, with TSMC's help, has crafted sheer magic to reach that low power consumption in games. But paying a top-tier price for second-tier performance doesn't make sense when there are plenty of other options to choose from.

Originally posted on October 20, to the PCMR subreddit, user Vincenz_OB announced they had purchased the popular graphics card for $48.94 from Amazon. 

This was met with disbelief, with one user saying: "On Amazon - a good deal on a Graphics Card is also known as a rock inside a nice box." 

The reason many thought Vincenz would get a 'rock in a box' is because Amazon, being a huge multinational corporation, is believed to often check its returns through weight and seals on the box. Though less common now, there was a problem with the wrong Ryzen chips being sent to customers back in 2017. If you are clever and lucky enough, you could essentially tape up something heavy and send it back—play the odds that a firm dealing with a million packages a day wouldn't notice.

After just a single day (thanks Prime next-day shipping), the 4070 Super had arrived at Vincenz_OB's address, much to the disbelief of many a commenter. 

I reached out to the original poster who told us: "With it being sold and shipped by Amazon, I was hopeful" but that they also tried to not "get too excited until it was delivered and unboxed". 

I asked if they have had the chance to actually plug it in and try it out, to which they said "I tested it and ran a few games last and everything is running great. Couldn't be happier with the performance." I have seen the order details from Amazon and can confirm this very lucky gamer got it for just 56 and 27 cents Canadian. That's $40 USD, which is even less than the noted listing says.

Generally, it's not worth this kind of risk when buying second-hand or from a site you don't trust but, as long as you document the opening process and can verify what you have received, this is relatively low risk to do on Amazon. I'll be using the service quite a bit to check over the best Black Friday gaming PC deals as the big day draws near, though I don't expect to find anything this exciting. 

Unfortunately for me, who just last week purchased the exact same card, the original poster managed to grab the very last one in stock. 

This is the RC1 from Realforce, as noted over at Gigazine. It's a compact, tenkeyless design with a two-tone grey finish across the keycaps. It's what I'd call a very minor departure from the usually quite boring-looking Realforce boards we're used to—though only a little. It's still a pretty plain-looking design compared to some mechanical keyboards these days.

The two-tone grey keycap set and smaller frame might make this keyboard a little more attractive to your average PC gamer, anyways.

If you're wondering what the fuss is about, here's the crux of it. This keyboard uses switches designed by the creator of the Realforce brand, Topre. The Topre Corporation, a Japanese company, specialises in air-conditioning equipment, pressed automobile parts, and, most importantly, fantastic non-capacitive keyboard switches.

The same switch found in the RC1 is reportedly also used inside Japanese Seven Bank ATMs. I'm not sure if that's useful information but I wanted to mention it anyway. The bank's latest ATM buttons do look a lot better than the metal ones we get here in the UK, at least.

It's tough to explain why a Topre switch is good in words. Trust me, I've tried many times to people that don't really care.

The issue is if you look at what a Topre switch most closely resembles… it's a membrane switch. Though it's not really that similar, just in the plunger and rubbery bit. But it's much more than a squishy, flaccid membrane (sorry, membrane switch fans). A Topre is a quiet switch with a satisfying response—that's why I rated the Realforce R2 I have at home as the best mechanical keyboard if you want something quiet. It can have an adjustable actuation too.

Altogether, Topre switches are well-suited for rapidly smashing out heaps of words in a short period. A real typist's board.

There's no greater proof of that than the Realforce Typing Championship. I've embedded the video of the latest event below. It's the company's own event put on each year to find the best typist in Japan. This year's has already taken place in Tokyo, and the winner was 三山羊.

Now, Google Translate suggests 三山羊 means 'Three Goats'. Presumably a mistranslation or a phrase without a direct counterpart in English, there's also the small chance it's three actual goats wearing a trenchcoat. Regardless, they're really good at typing extremely quickly.

In the final battle between 三山羊 (Three Goats) and くわな (Kwana), both sides are typing at over 800 kpm. That's keystrokes per minute. They occasionally hit over 900 kpm. At one point, 三山羊 manages over 1000 kpm.

They're using an application called Weather Typing 4.3 (RTC2024 version) for the event. You can download it from the Realforce website, though it's entirely in Japanese. They're also only allowed to use Qwerty and Dvorak layouts.

You can watch all five hours of furious clacking and commentary on YouTube. It's a really well-put-together event, with the hosts running a tight ship with interviews, ceremonies, and esports-grade shoutcasting.

It's been going on for five years, too, and I have vague memories of hearing of its existence before, but I guess my brain unfairly pushed it out to make way for other information, like my PIN code and home address. There's also the Ultimate Typing Championship, sponsored by Cherry, which may have actually been what I was thinking of.

If you are crowned the best typist in Japan you are awarded a golden Realforce keyboard as a trophy, and nothing has made me want to perfect my touch typing more than the sight of this spectacular creation. Not only is it a gold-encrusted (gold painted) Realforce but the keys appear to actually work.

Anyways, I didn't expect to end up here, I was only interested in the RC1. If you are, too, you might have to wait a while outside of Japan to get your hands on one. They're usually imported elsewhere, and while the supply has become far more robust in recent years in the UK and US, costs can be high and we might not get the shiny thing quite as quickly.

Worth keeping an eye out, anyways. I will be.

In its release notes (PDF), Nvidia reports (via Overclock3D) "higher than normal CPU usage" as an "open issue" for iCUE and Bluestacks software. The former is software that allows for centralised control for all your Corsair hardware—you know, for that sweet, sweet RGB sync—and the latter is Android emulator software.

It's a little strange that a GPU driver update could cause high CPU usage, but all these components are interconnected, I suppose. We don't know exactly how much "higher than normal" the CPU usage will be for users of these programs, either. However, if you use either of them and you won't be playing the new games the update's targeting, it might be best to skip the 556.03 driver or wait for a hotfix.

According to Nvidia, the latest DLSS 3-supporting games that this driver should benefit include:

• Dragon Age: The Veilguard
• Alan Wake 2: The Lake House
• Call of Duty: Black Ops 6
• Horizon Zero Dawn Remastered
• No More Room In Hell 2
• Red Dead Redemption
• The Axis Unseen

Of biggest note might be the first three of these. Black Ops 6 is set to launch tomorrow and will be coming to Game Pass, Dragon Age: The Veilguard will launch on the spookiest of dates (October 31), and the Alan Wake expansion is available right now.

Gamers wanting to play any of these titles who also use Bluestacks or iCUE might have a decision to make, then. After all, the driver could increase performance in these games more than enough to offset any increase in CPU usage alongside iCUE or Bluestacks.

You could, of course, install the driver and then roll it back to uninstall it if it's causing too many problems. Or you could temporarily uninstall iCUE or Bluestacks, depending on how necessary it is. Uninstalling the former might be difficult for some systems, though, given that Corsair can sometimes make its software a requirement for control of its fans and so on.

Whatever the case, if you use either of these pieces of software, at least take the time to decide whether the latest Nvidia driver update is worth it, because it might not be.