Rock Pi X vs Raspberry Pi 4: x86 & ARM Showdown

[@martkt10]

The power consumption of either board would have been interesting too, though im guessing the arm cpu would be less

[@johnkelly6861]

Also interested in power consumtion and cpu temp (expected as Christopher pointed out that both using similar cooling).

[@wurstwas]

Yes, I think that would have been very interesting, since it might have been a decision point for the comparison.

[@tubegor]

@KS Kiran That is right as well. kzbin.info/www/bejne/b6DJoKR6mJuFl6c 2 watts in idle and about 4 watts full load. On my x86 PC only fan consumes 6 watts. 😭

[@sergeydrozhzhin9263]

It would be interesting to compare power consumption when performing the same jobs.

[@jackmclane1826]

@KS Kiran That's pretty much obvious. ARM was already tuned for energy efficiency and that's a whole world of requirements that's not compatible with existing chip designs. The x86 branch has grown a plethora of variants in the last decade. And chips like the Atom don't have much in common with their bigger brethren. But we now see that - when tuned to the same energy consumption - Arm is not faster per watt than x86. They all fight the same battle.

[@DallasMike424]

Chris could read the the Nottingham phone book every week and I would tune in just to listen to his calm voice.

[@KameraShy]

Don't give him such ideas.

[@nathanmead140]

I would do that too

[@ThePsychologyPod3430]

How can i get started with raspberry

[@TheTechieScientist]

Me too

[@ronjohnson7263]

Dallas stop lying you know this man sound like the dry eye guy lol!!! His voice is dry and boring just like you if you like the sound of it.

[@jeyendeoso]

5:10 "why are scroll bars a ridiculous size these days?" Not only ridiculous size, but sometimes very well camouflaged as a slightly darker grey on a grey background.

[@deanharrysmith718]

Really difficult to find sometimes as some I have noticed on windows to be only 1 pixel wide and only when you have the mouse directly over it does it expand out. These GUI designers are really bad at times and why do they feel everything needs a new redesign evey 2 years or so. My Parents (70's) struggle every time a refresh happens which must be the same for a large proportion of the world user base.

[@DallasMike424]

The older my eyes get, the smaller the scroll bars get. This can't end well, can it?

[@DallasMike424]

Someone needs to develop a GeezerOS Linux distro for people like me. Large scroll bars, large fonts, and the left turn signal always flashing as I motor down the information superhighway at 30 Baud.

[@6581punk]

Flat design, it sort of appeared with the whole Web 2.0 approach to design. Make things flatter and simpler.

[@shakeyourbunny]

@@6581punk flat design is the bane of every user, where consistent user interface elements are hidden in plain sight and you never can be sure if something is a button or not.

[@cholesterol6703]

6:45 "Veeeery exciting," states Chris while watching the computer equivalent to paint drying.

[@TheTechieScientist]

Kinda knew the results before by looking at the thumbnail but came to be one of the first and a huge fan of Explaining Computers, love the topic and hoping for more videos on SBC comparisons

[@marcdraco2189]

"Honey, dinner's ready!" "I'm coming right away." (Sees new upload from Explaining Computers) "Damn, something's come up. I'll be about 20 minutes!"

[@ExplainingComputers]

:)

[@Cl4r1ty_]

Isn’t this everyone right now?

[@e1woqf]

First things first! 🤣

[@johndripper]

lol

[@drsmith3411]

@@Cl4r1ty_ In California, it's breakfast ;)

[@RoboNuggie]

Nice, I like the real world tests and not just the synthetic testing, both have their place of course and show how things have moved on. I'm biased in the Raspberry Pi's favour, but good results from both... Thank you Christopher!

[@MartinBgelund]

I would really like to see comparison of the power consumption in this kind of testing videos. Especially when it's a race between two different CPU architectures.

[@russellhltn1396]

Same here. I'm looking at an application where they'd run off of battery power. Power consumption suddenly becomes important.

[@flandrble]

both are basically identical. The Z8350 is a 2W CPU.

[@EdwardRLyons]

"Elsewise". What a great word!

[@europeansovietunion7372]

He has the best words.

[@EdwardRLyons]

@Tony B Yes!

[@ryanmalin]

Happy new year Chris! Always a good Sunday morning waking up to one of your videos. Cheers from Florida!

[@maptastik]

This was great, thanks for comparing these two SBCs! I do a lot of work with spatial data analysis in Python and love the idea of using the RaspberryPi for tinkering with geographic data. That said, ARM support for several key libraries is lacking for ARM. While the RockPi X may not be blowing the RaspberryPi 4 out of the water on performance, for me its x86 processor provides a key advantage in support for the types of libraries I need to be able to work with spatial data in Python.

[@sannidhyabalkote9536]

Boy , do these SBC's have come so far I couldn't have imagined this amount of growth in this small time

[@ali_new_world]

Almost every comment has the word SBC Can someone tell me what it means lol

[@FlyboyHelosim]

@@ali_new_world Single Board Computer.

[@brianwesley28]

I can't wait until they have eight cores with L2 and L3, 32GB of RAM, 256GB of eMMC, M.2 NVMe, and USB 3.1 Gen 2, with a comparable power consumption.

[@williamg7776]

Thanks for making the video, I have been looking for this comparison for awhile.

[@leeoliver2969]

Another great video, it got me off my seat to upgrade my rpi4 firmware. Still collecting outside weather. Our rain was clearly predicted (by more than 24 hours) with a pressure drop from 1000hPa to 975hPa. I'd like to see that beautiful 1000hPa weather again.

[@watsoft70]

A little unfair to call the last two tests a tie as a win is a win (and the KZbin playback was noticeably better on the Rock Pi X), but I agree that the tests show marginal difference and so the Pi definitely "holds its own".

[@yllib2012]

For some reason when I saw the title I was thinking a new Rpi 4 was out with X86. Now that would be NICE. Great video.

[@srtcsb]

Really like the comparison videos. Your testing gives an objective view of each board's capabilities. Thanks for another great video.

[@ExplainingComputers]

Thanks Steve.

[@pedroramirez2215]

Great comparison, there are very few arm vs x86 comparisons this was gr8 to watch

[@Cl4r1ty_]

Just came from looking at x86 SBCs!!!

[@ExplainingComputers]

Ah! :) Then the results here may surprise you. They surprised me . . .

[@drescherjm]

I am waiting for AMD 7nm or 5nm offerings.

[@Byc845]

Me too!

[@davidglenn6468]

Another great video, with informative, but surprising results. And yes, I have to agree that Chris’s voice is particularly soothing to the ears. Nice

[@cleanycloth]

I would love to get a Rock Pi X and pop all of my old 90s/early 00s games on it. Would be a fantastic little machine!

[@perrymcclusky4695]

WOW, close results. I didn’t expect that. I appreciate the video and looking forward to your next video!

[@ExplainingComputers]

Thanks Perry. :) Hope all is OK with you.

[@perrymcclusky4695]

@@ExplainingComputers Thank you. Doing okay. Old age isn’t for sissies. LOL. Hope your chest has finally healed completely and your dedicated filming room is coming along okay. I’d probably follow your videos no matter what your content, but glad you are doing what you are doing.

[@ElmerFuddGun]

Scroll bars... OMG why do they have to hide them??? Often in Linux Mint (and others) I haven't even known there were more options because the bar was hidden! Please developers *_STOP HIDING SCROLL BARS!_* Arrrrgggghhhh!

[@ExplainingComputers]

I eventually found that in Linux Mint you can set the scroll-bar width in Preferences, System Settings, Themes, Settings. So in Mint I have a very easy to use (24 wide) scrollbar.

[@ElmerFuddGun]

@@ExplainingComputers - That's good but hiding scroll bars are the worst especially for new users to an OS. It can be very frustrating. All OSes should *DEFAULT* to NOT hiding scroll bars and to have a normal easy to use width. Advanced users can then change those settings. These OSes are trying to get new users but all these kinds of things are major put offs and keep new users away.

[@johnm2012]

@@ElmerFuddGun Hiding scroll bars is a thing on phones and tablets, presumably because they take up space. Since, for some reason, there's an obsession with making desktop operating systems look and behave like mobile operating systems the latter's compromises are adopted on the desktop too.

[@robsku1]

@@ExplainingComputers Do you know if the same configurations are readily available in XFCE desktop? I mostly use Linux Mint with XFCE as my default OS... This computer currently runs Ubuntu though, which I don't really even like.

[@ElmerFuddGun]

​@@johnm2012 - Tell me about it! That's why Windows 10 looks so ugly compared to Windows 7. It looks like it was designed for low power phones and not "modern" PCs from the last 10+ years that have GPUs. They take away from usability by hiding "small" stuff like scroll bars because they think it looks fashionable but it actually wastes everyone's time.

[@lollerich]

Great video as always! You also look a lot healthier, I'm happy to see that you're on your way to a full recovery or maybe even completely recovered (I didn't want to presume).

[@renobodyrenobody]

Just a message to say: thank you. Everything is clearly explained as usual.

[@ExplainingComputers]

Thanks.

[@Simpaulme]

Love the passion at the close of this vid 👌

[@stevefreeman730]

A very fair comparison. I like that you used the same os so it was an apples to apples comparison.

[@benywiratmaka8636]

Nope, Kden live part is a bit misleading to consider a tie. It did shows 1s difference, yes you might think it is a miniscule difference and can be dismissed as a tie. But consider this: the clip that was being rendered is only 10 seconds long. It is simply too short to be able to showcase the difference of performance between rock pi x and raspi 4.

[@mikes333]

I think one of the greatest benefits of ARM processors over x86 is its significantly lesser power consumption requirements. Please include power consumption as a standard test for any future x86 vs ARM processor comparisons.

[@ExplainingComputers]

I compared a number of SBCs at idle and load in this video: kzbin.info/www/bejne/b6DJoKR6mJuFl6c The Pi 4 4GB was 2W/4W, the Rock Pi X 2W/5W

[@delphicdescant]

Just thought I'd, uh, put in a word of encouragement for RISCV while we're at it.

[@RollerCoasterLineProductions]

Because of your videos I'm gonna order a Pi4 2 gig and the Pi ZeroW and fiddle around with them, gonna run a PiHole on the 4 and a camera on the ZeroW. Your videos are VERY informative and helpful.

[@opiniondiscarded6650]

I would love to see a GPU comparison between the two

[@deanlawson6880]

Wow! What a wonderful and fascinating video - Thanks so much for this Prof. Chris! It sure seems that we now have an environment, with computing platforms, between the ARM and X86 where the hardware is roughly comparable in terms of speed and performance. So, your choice of hardware is now really based on what you want to do, and the availability of your choice of software and tools, as it should be!! This is very very interesting! I look forward to the time that software and apps, and tools (..etc..) really catch up on the ARM processing end, as I think this (the ARM platform) is a better more efficient platform ultimately. Thanks again for this really interesting and fascinating video.. (Yes, I watched this and am writing this on my RasPi 4B 8Gb "desktop" in the DeskPi Pro case!)

[@davidknight247]

The Sysbench performance was suspiciously 4x faster on the pi. Did you check that all four cores were number crunching on the RockPiX?

[@ExplainingComputers]

I didn't but should have done! :) Now you mention it, it does suggest only one core was in use on the Rock Pi X, even with threads set to 4.

[@igorpecovnik]

Results are completely wrong since sysbench return more or less random numbers forum.armbian.com/topic/7763-benchmarking-cpus/ One could not use more broken tool.

[@hankhulator5007]

@@ExplainingComputers To be the right sysbench or not to be… There is something rotten in the kingdom of sysbench : I installed it from debian buster (armhf) on my Raspberry pi 4 overclocked to 2,000 MHz and noticed two big differences : * switches syntaxes are different, * time taken for this test is wayyy beyond yours (almost 15× more !) The buster package is 0.4.12, when the git latest stable release tarball is 1.0.20, so I downloaded and compiled this source. This time the syntax was the same as yours, but the numbers are the same I've got from the official packaged version of sysbench. The same exact switches as yours are giving : * Rpi4 O.C.@2,000 MHz total time: 58.5359s So I also compiled and ran it on my laptop and got : * Laptop Haswell intel core i5-4210M total time: 8.6295s In order to get the same total time as you, I have to go down to : --cpu-max-prime=2800 & --events=12000. Unless I missed something huge, your version of sysbench has a big problem (I suppose you're on a arm64 OS, but this doesn't make such a big difference).

[@jaimieandash9484]

Thank you so much for your kind and patient sharing of information, Christopher. :-)

[@IsaFuie]

I've been longing for a PC upgrade, but in these trying times, finding these components has been a bit of a challenge. At this rate, I might well be a Pi convert!! Who knows, one of these days I might give these things a go.

[@horseradishpower9947]

If you look at what is on offer now, there is a huge amount of choice now. Point of interest, there are now DACs being sold by the Raspberry Pi Foundation, so there us the chance for really nice sound, too. Have used a Raspberry Pi 4B since they first came out, and it has come leaps and bounds. Well worth an investigation, at the very least.

[@woldemunster9244]

It would be more fun to get Pi as "second" computer, while you wait for hardware availability.... If you find yourself using the Pi primarily, just sell the old hardware and you can buy deluxe case for the Pi? :D

[@IsaFuie]

@@horseradishpower9947 I did actually check a few websites that sell Raspberry Pi computers right after watching this video. I'm certainly intrigued. As much as I love to treat myself to more costly computer parts from time to time, I'll always appreciate computers for what they are; computers. Not every computer has to be water cooled with some state-of-the-art graphics card that needs to be powered by several hydroelectric power plants, just so some lout can play Fortnite at over 9,000 fps. Besides, in a world of big corporations, I feel we, as consumers, should be more supportive of smaller businesses. The Raspberry Pi, I feel, provides that "grassroots joy" which is lost on many people nowadays. Sometimes less is more, and the fact that the components of a desktop can effectively be held in the palm of a hand, at a very affordable price for a many, is testament to how far technology has come.

[@IsaFuie]

@@woldemunster9244 lol I won't be using Pi as my 'main', I'll certainly use it as a second computer. I was thinking, if I do get a Pi, I'd like to use it to familiarize myself with some of the Linux distros. I'm not thinking of any particular distros right now though. I'll cross that bridge when I get there.

[@woldemunster9244]

@@IsaFuie I recommend "not Ubuntu", i used it myself for 7years on laptops/desktop as my main OS. The hardware support and ease of use made me think that it's "best distro" but now i've used Mint and couple "special" distros. Ubuntu begs you to make an Canonical account for some restore reason and it can mess your passwords if not careful... :D But Mint opened my eyes to the fact that Ubuntu is almost as bad as Microsoft-product, they want you to use M$-store wannabe with their own repositories and you get a lot of bloat software... I got Steam /Proton working on Mint really easy but heavy titles kinda need a KVM to get full Gpu power and i'll be doing that when i get a bigger boot drive. :D Take the plunge, Rufus and any distro, liveboot from USB. Or Virtualbox to run really wild distros. :DDD

[@hawaiianryan1890]

Some surprising results. Thanks for the comparison 👍

[@ExplainingComputers]

No problem 👍

[@NexXxus86]

not to mention that the Intel HD 405 should be slightly faster than the Videocore VI

[@tayntedmemories]

Intel HD 405 is just a underclocked HD 4000

[@NexXxus86]

@@tayntedmemories Gen8 based on broadwell (HD5000) www.notebookcheck.net/Intel-Atom-x7-Z8700-Benchmarks.140906.0.html

[@michaelclement1337]

Great presentation. I get the feeling that the winner between the 2 is more so to what software you want to run or what education you want to derive

[@VictorNoelCoryPaz]

The new Beagle V looks interesting. Maybe, a review later on? 😅 Great video, btw!🐾

[@ExplainingComputers]

Yes, I will certainly cover the BeagleV here when a final board is available . . .

[@VictorNoelCoryPaz]

@@ExplainingComputers Great! Thank you so much! 🤗🐾

[@AndrewGulak]

Great video Chris! You're so obviously all in for Raspberry Pi! I've been watching for years, so it's pretty apparent. I am as well a fan of the Pi and the companies original mission.

[@catcatcatcatcatcatcatcatcatca]

It seems this more so highlights the good work raspberry pi has seen in optimization and drivers, as for purely comparing the architectures Apples latest stunt with their own processors provides much wider and more impressing proof. Today, in my opinion, the only thing holding ARM back is code. How much raspberry pi has seen improvement on that front is pleasantly surprising, but there is not much of a debate if ARM as architecture will continue to outperform x86 on low powered applications - and maybe even in more power heavy ones too.

[@phonewithoutquestion80]

Very entertaining test. I am glad I picked up the 400, by the way, this had been the best low budget computing decision ever, a lot more flexible than picking up a burner laptop for even more and feeling burned later down the pipeline.

[@horseradishpower9947]

This will be interesting, as this could give an indicator as to the direction of processors in computing for the future. As in, whether ARM will start being much more a standard in regular build PC devices or not.

[@ExplainingComputers]

Very true. Making this video made me rethink a few things, as you will see.

[@horseradishpower9947]

@@ExplainingComputers I am in agreement with your assessment. In addition, there is the reduced power consumption to consider as well, along with products like the Raspberry Pi 400 being available. Having the computer inside the keyboard saves space, and makes for a more tidy desk space. General office use doesn't need the fastest, most powerful systems, so ARM can easily be an attractive prospect, in terms of initial cost, and running costs. Really good video, thanks for making it for us all.

[@rageagainstthebath]

The fact that RISC CPUs will eventually outperform legacy, bloated x86 per unit of energy or time has been known ever since RISC CPUs emerged in the first place. It just took too many years for consumers to understand that, mostly due to lack of compatibility with "The OS". There are already supercomputers built using ARM chips, and I wouldn't dare to question the authority of high performance computing scientists. Finally, ARM RISC CPUs show their dominance in consumer toys market. It took too long, but it's already here.

[@John_Locke_108]

Highlight of Sunday is getting a new video from you.

[@ExplainingComputers]

:)

[@DylanMatthewTurner]

You should do a video on the new BeagleV

[@ExplainingComputers]

When released, certainly. Although I understand that the first batch will lack the GPU. I've been sent a lot of materials on the BeagleV, and am still uncertain what is actually going on! :)

[@MrHitman1966]

Splendid stuff - really enjoy watching your videos - no flowery nonsense and straight to the point - refreshing 👍

[@richardeadon6396]

13:20 'elsewise' not sure I've heard that word before!

[@tinkmarshino]

Very cool results Chris.. I, like you, thought the arm would be blown away.. amazing things are happening all the time in this world of little things..

[@RadicDotkey]

I have a weird feeling that Sysbench was using 1 core only on the Rock. Are you able to redo the test with htop opened?

[@ExplainingComputers]

Good idea -- I'll try and do this!

[@retroretiree2086]

I too am suspicious of sysbench's core usage on x86. I have 8 core/16 thread ryzen 7 CPU and no matter how many threads I created the result was very similar.

[@jachtk893]

I have the exact same feeling, other tests are pretty similar while sysbench are about 4 times as fast as z8350

[@allesklarklaus147]

@@retroretiree2086 I just did a test on my rather old intel i7 something dualcore in my laptop. The cpu has hyperthreading enabled and if I run sysbench with 4 threads it does max out all the 4 (symbolic) cpus in htop. If I run 2 threads it maxes out two cpus, but sometimes two on the same hardware core. However, I get about half the time (8.2s vs 13.6s) with 4 instead of 2 threads. 8 threads does only marginally decrease the time to 8.1s.

[@allesklarklaus147]

Oh and 3 threads gives me 10.2s. So yeah this test does benefit quite nicely from hyperthreading I would say

[@brent8783]

Hi Christopher. Like many others, the actions taken against Parler were the tipping point for me. The challenge for me has been availability of content on Rumble, Vimeo, etc, but I'm willing to accept a content drought in order to break my dependence on Amazon and Google (Alphabet). I love what you do and will stay subscribed, but won't see your content anymore on KZbin. Please keep up the great work, passion for technology and sharing what you do :) . Sorry to post anything political-ish, but wanted to let you know.

[@MichaelGrantham18]

While not a performance metric, I think it's also interesting to note the size of the user bases of either board. I think it's not too much of a stretch to say that the Raspberry Pi has a much larger user base than the Rock Pi and therefore will have more support online when you run into problems. But the Rock Pi has full size HDMI... Rock Pi wins

[@ExplainingComputers]

True.

[@Rick-vm8bl]

Have to agree. Really not a fan of the pi's micro hdmi ports, especially when most of the adaptors are great bit chunky things that tend to break.

[@rageagainstthebath]

The Raspberry Pi community is huge, but I'd argue it's not really a knowledgeable bunch. Google any problem with your Raspberry Pi and you'll only find answers that you're using a wrong power adapter, bad SD card, unofficial OS, and if all that is exhausted, they deem your board faulty and recommend getting a new one. Personally I buy broken RasPis for cheap and fix them for my cluster, but nobody on the forum seems to be interested in giving any other advice than four mentioned above. Compare it with Odroid community or Armbian forum.

[@beauslim]

A thorough and objective comparison, as usual. One thing that might matter to some is that the Atom has hardware crypto acceleration while no Pi does. So, I would expect VPNs, file transfers using encrypted protocols (scp, rsync), and disk encryption to be significantly faster on the Rock Pi X. I would love to see a test, though.

[@igorpecovnik]

Sysbench doesn't give correct numbers. forum.armbian.com/topic/7763-benchmarking-cpus/ Proper comparisons? github.com/ThomasKaiser/sbc-bench

[@ElmerFuddGun]

4:19 - Is your Rock Pi X running on batteries? Top right menu shows battery time remaining but I didn't see any smart battery connected.

[@ExplainingComputers]

It was running on mains power, but for some reason displays the battery indicator.

[@ElmerFuddGun]

@@ExplainingComputers - Count on me to notice those things! ;-) I wonder where it is getting the percentage and time then?

[@ExplainingComputers]

@@dank9561 Wired mouse and keyboard. Linux just gets confused! :)

[@MLGPRO-dx8fg]

Was pretty surprised by the results! I too have a Intel atom board laying around and mistakenly thought that it was more powerful than my rpi 4. Good to see low power chips reaching x86 levels of performance

[@MLGPRO-dx8fg]

@J Fz Yeah I am aware that the Intel atom here is ancient by today's standards but even with that in mind I had just assumed that the cpu in the pi would be awful. Excited to see what the pi 5 will have whenever it releases because the pi 4 makes a great server for discord bots and other things

[@drewswoods]

Using an rpi4 as a primary desktop, I mostly find myself wanting an x86 due to gaming or needing specific applications.

[@horseradishpower9947]

That is why I keep my laptop around still. Gaming. The rest? It is all Raspberry Pi, just like you.

[@notthesameman]

Same here waiting fo pi4 to get on its real shoes, sonocs super shoes

[@amihartz]

I used a Pi 4 as a desktop for about a year and a half in uni, with the 4 GB RAM model and booting from an SSD it's surprisingly decent. Not much games but I am a console pleb anyways so I didn't care much.

[@BilisNegra]

For a primary desktop you can get way better options than these SBCs if size/ extremely low power consumption is not vital. I got several SFF HP desktops with 3rd gen i3 on eBay for 40-45 quid each (and that was two years ago). Such CPUs (i3-3220) while not particularly powerful, score more than twice as much in Passmark than the Atom in the PI X

[@rogerkoh1979]

It is really very different between arm and x86 back a few years ago. Remember the handheld pc can only do so little. Now it is work neck to neck in some applications.

[@bertblankenstein3738]

Thinking about the future. Windows likely is at the end of the road, even though it does run on ARM. I guess this will be slightly reminiscent of the 80s where there were a lot of micros, but this would be two contenders where there will be choices for OS for both. Intel currently isn't doing very well, ARM on the rise, and what will happen to AMD?

[@Rick-vm8bl]

Interestingly Microsoft is actively working to improve its ARM support. They are reportedly desining their own ARM chips at the moment, and have been working with Apple to bring Windows to the new M1 macs. It's interesting that we're coming full circle back to arguably better suited chips for the job. Now if only we could get a new Acorn Archimedes released!

[@АлексейДмитриев-ш1ф]

The CPUs and OSes market is a little bit more complicated than you think. Intel is not limited to x86 as well as x86 is not limited to Intel. Intel also does network solutions, FPGAs, GPUs, and their chip manufacturing plants don't vaporize just because some competitors got success. AMD is pretty much the same, although they don't have their own plants, which is good and bad simultaneously, but I don't want to get deeply into this now. Both of these manufacturers can roll out new solutions soon, and I won't be surprised if these solutions would be heterogeneous or completely different from what we have now. Big Little is already a thing in x86, why won't there be x86+ARM, or x86+RISC-V, or CPGPU, or CPNPU, or whatever. The main advantage of x86 is available software, of course. The main disadvantage is that it had a long way and got excess and outdated instructions (and other internal stuff), which are probably not needed in most cases, but they can not be omitted because backward compatibility would be lost. The main advantage of ARM is that it can omit anything the manufacturer wants just because it would not break much. And the disadvantage is that there is not much software available. Because x86 is bulky it will always be more expensive, generate more heat and require more power than ARM competitors. Thus ARM can achieve more frequency in the same thermal and power packages. BUT! Think of CPU instructions not as some mystery thing, but as a flow of electrons. It is a flow of electrons that limits performance of a chip. It does not matter if the instruction is performed on x86 or on ARM, the theoretical performance maximum will be the same. You can accelerate flow by applying more voltage, or shorten path by applying more thin manufacturing process. But manufacturing process is already almost reached its limit. Someone from Intel said a long time ago that going below 10 nm will be very complicated, because it is close to a charge size. The electrons just don't go in such a small traces, meaning a conductor becomes more like an isolator. And indeed it is now, but surprisingly only for Intel. TMSC and others claim that they achieved 7 nm or less, which I don't kinda believe. So ARM and x86 will be pretty much the same in the near future, introducing more parallelism and heterogeneous computing in each new generation. The real breakthrough would be changing silicone to something else (GaN?). And Windows is faaaar away from its death, there is just no competition for it. Today it is not an OS, it is a whole ecosystem with its own development tools, apps, etc. There is no any other OS that you can install and use your PC as effectively as with Windows.

[@6581punk]

Microsoft needs to design its own chips, it seems ARM's own core designs are too general purpose for desktop. Apple for instance added all sorts of tricks to the M1 to make running apps designed for X86 run fast.

[@robsku1]

@@АлексейДмитриев-ш1ф _"There is no any other OS that you can install and use your PC as effectively as with Windows."_ This is just plainly false, I work much more effectively with Linux than with Windows on literally everything I use my computer for. Maybe for some use cases t can be, but most of the time and for most use cases out there: it just isn't true. And even in those cases where Windows is better choice, it's always only because some specific 3rd party application is not available for Linux (or OS X).

[@robsku1]

@sbcontt YT I guess we know very different kind of regular folks... Or maybe you're just too much in the past with how easy Linux distributions have made it for an average user to just jump in and start using it - I've had a number of friends who have requested me to install Linux on their computers, and I've done it and shown where they can look for and install additional applications from and they haven't went back. When they get a new computer they still call me, that's the only step they need help with unless the system gets messed up after a power loss or something. And to quote one of my friends whose system got really messed up (this was over a decade ago and I managed to get it sorted) "I wouldn't know how to fix that, but if this was Windows my only choice would have been to take it to the store where they just reinstall the whole damn OS"). I could have done the reinstall faster though, but I really wanted to show that while their system had broken beyond their level of knowledge that doesn't mean you can't fix it. Most people who I've had the luck to turn into Linux have one thing they are thankful of: It's easy. They were frustrated with Windows and now they are not frustrated anymore. They just needed someone to help them install it. A basic Chrome machine? Hardly - you get more than that from just the basic installation of any desktop oriented distribution.

[@markharrisllb]

I find this channel is excellent at explaining computers…who would have thought it.

[@ExplainingComputers]

:)

[@DumahBrazorf]

The pi is 4 times faster on 4threads on a 4core cpu... Maybe there's some problem and the multithread are all executed by the same core on the atom. Can you execute a 1thread bench?

[@ExplainingComputers]

A very interesting thought. Could well be the issue.

[@alifilhan9745]

@@ExplainingComputers Look, here the firmware of the rock pi a.k.a the good old BIOS or UEFI is a locked down cage for this CPU, tablets or laptops with exact same hardware runs nearly twice faster than this. maybe Rock Pi makers did this to reduce power consumption or to promote sales for other coming boards. I'm not saying Atoms are good, but at least they aren't this weak. And for the multi-tasking, yes it is possibly the issue too, alongside locked firmware

[@antaeusdk]

Nice video, one comment though: Power draw could have been a relevant comparison point.

[@thebeststooge]

So many choices in SBC these days.

[@joefish6091]

anything X86 has a massive price premium. people start eyeing up a product, the middle men resellers instantly +$100 to the shelf price. junk CPU tech that sits on shelves gathering dust waiting to be boxed up for the landfill or shipment to China India, suddenly becomes as valuable as 2020 CPUs.

[@thebeststooge]

@@joefish6091 You notice that as well? Sad on so many levels.

[@harshbarj]

Would be nice to see an X86 board with a more recent CPU. That z8350 is 6 years old, compared to the BCM2711 at a year and 6 months. Though sadly nearly all lower cost x86 boards seem to be using these vastly out of date processors. Would love to see something with the Atom 6000 line which was just released.

[@arkeynserhayn8370]

Pray for ryzen embedded. There is rumored V2000 but that not SBC-friendly.

[@shaytal100]

This is surprising. At the same clock speed and the same number of cores, x64 should be faster than arm64! While arm64 should beat x64 in lower power consumption. But Atoms are designed for lower power consumption. They were made for Laptops. Would have been nice to see how these 2 boards compare in power consumption. There must be something wrong with the command line benchmark program you used. I could have believed if both boards would tie there, but the arm 4 times faster makes no sense! In general the x64 should beat the arm. The only explanation for these results that come to my mind is, the arm may have faster and/or more cache? This could well be a big advantage depending on the benchmark.

[@andrewgrant788]

I was expect the ARM board to be much faster than the Atom. Perhaps because I have seen how fast the ARM based Macs are or perhaps I just remember the Atom based Netbook I used to own.

[@ThatFellowOnline]

If BT Sport did matches like this, they'd have a chance of convincing me to buy into a sports channel for the first time in my life.

[@deadadam666]

Video idea - id love to see a more in depth investigation into the difference between risc and cisc - ideally with as close to a direct comparison as possible , same age chip , same clock speed , ram speed and storage speed ..... etc etc. I mean with Apple moving over and phones and mobile devices getting more and more powerful are we coming to the end of cisc's life ? Keep up the great work !

[@stuartg40]

Surprisingly comparable but it would be good to see an explanation of the sysbench discrepancy.

[@juanignacioaschura9437]

It comes down to the Instruction Set.

[@Brainstorm4300]

Excellent video as always Chris! I've predicted ARM outperforming/competing x86 in consumer laptop market back in 2015 and I think it is finally happening with Apple's M1 powered laptops. It's an exciting time and we may see a major change in the coming years.

[@boyblue66]

Would have been nice to see the results with an overclocked ARM cpu. There was a big heatsink on the Pi so it wouldn't have been too much of a stretch. That said i don't run my Pi4's overclocked.

[@theo_jeff]

Can the Rock Pi be overclocked? Prob not.

[@boyblue66]

@@theo_jeff that’s why I said ARM cpu. The pi can be over locked to 2.0 GHz

[@theo_jeff]

@@boyblue66 yes but it kind of shows that these two boards aren't as equally matched as it seems.

[@boyblue66]

@@theo_jeff hi Jeff, I don’t think these boards are evenly matched. That was my point. I don’t think the author does either. The only real advantage would be software availability, but that’s only if you need a specific package.

[@hankhillsnrrwurethra]

@@boyblue66 the issue of cooling in a small case comes up as well. Provided heat sink may or may not cut it.

[@Matt_moran78]

I think you summed the end of the video well Christopher. I think people are going to be torn between x86 and ARM processors before long. The question will be what one wants to do with the computer.

[@cokeforever]

Hello, Chris. In addition to sysbench one can use hardinfo (benchmarks section) as it offers several more cpu tests (blowfish algorithm, fibonaci numbers, crypt/decrypt async etc)

[@SidebandSamurai]

@4:30 you mentioned that you could not get WiFi working. I believe you need to recompile the kernel to include the necessary driver support, or you can execute a modeprobe command that will load a specified driver.

[@derekp2674]

...and that is exactly why Linux is a real PITA on some hardware. I did once have a Linx tablet PC with the same Atom chipset and never found any Linux distros that would work its wifi. Downloading additional drivers is tricky if you cannot connect to the 'net. At least these SBC's let you use a wired connection.

[@pnnytx]

Time to run Crysis...

[@wesss9353]

You need to try cyberpunk 2077.

[@System64-Fumo]

Crysis runs on the pi both in windows and linux but idk about rock pi x

[@wesss9353]

@J Fz use stadia to launch cyberpunk on the pi!

[@wesss9353]

@J Fz it's not cheating! It's fair and balanced!

[@NattaTRAINS]

@J Fz cyberpunk 2077 seems to have native linux support iirc. Sorry if I am wrong.

[@erikdekoster4137]

Thanks Chris for your time and effort as always much appreciated, hope you are well. I'm surprised about there not being a Linux driver available for the Rock Pi with it being an on board bit of hardware. It wound have been nice to see some more items on the showdown list e.g. power consumption and temperatures. Not complaining though I'm always looking forward to your posts.

[@kataseiko]

The WiFi doesn't work on Raspberry Pi 4 with Windows 10 ARM either.

[@boink800]

wiki.radxa.com/RockpiX/downloads

[@bazurk_dot_com]

I love your content brother. Thank you for all your hard work.

[@ExplainingComputers]

I appreciate that.

[@PS_Tube]

Just got the notification. Were still 3 comments and 30 views in. Yay !! It's PI day though. ♥️

[@wesss9353]

Tau is 2x better!

[@musthakheemkhanpatan7739]

You Are The Best KZbinr For SBC in Yt

[@ExplainingComputers]

:)

[@NexXxus86]

you could of showed some 3D benchmarks.

[@ExplainingComputers]

True. :)

[@matneu27]

Yup, let's launch blender and render a complex illuminated scene on it 😁

[@chavescrave]

Another great video cheers Chris. Happy New Year

[@whatamalike]

More nourishing than a Sunday roast!

[@theeiledon]

Hi Chris, your videos have become a part of my Monday morning ritual now. Thanks for all that you have done. Have you ever covered, or are you planning to cover, a video on making tablets out of SBCs? I've been working on a portable Pi 4 tablet with 10" screen with the problem currently being an elegant rechargeable power source.

[@lorderectus1849]

I’m simply board!

[@rajdeepbhattacharrya1752]

That Atom CPU surely performs better than I thought. Would have loved to see a power usage and temperature comparison too. Great video as always :-)

[@AstroTechGuy]

A SBC with Apples M1 would be amazing.. But yeah, i know, that it's not possible.

[@amrrahmy123]

that's a mac mini? it's much more than other SBCs, but it is an M1 in a box

[@shakeyourbunny]

The very main problem of ARM-based computers is that they don't have a standardized boot scheme like UEFI oder BIOS and their hardware is usually described in a static tree, which makes dealing with the boot process at least very difficult.

[@winsomehax]

It's hard work trying to explain to ARM superfans that their baby isn't perfect - but here goes. A modern processor is a relatively small instruction set that floats on a massive cache of super-fast RAM and multiple cores. The instruction set means little these days. ARM came from the low power world and let people buy its design for SOCs. x86 came from the performance world and didn't care about power (until the mid 2000s). x86 chips ALWAYS hammer ARM on performance at the top end. This isn't a criticism of ARM... its priority of power usage rather than performance is reflected in this. Every attempt by ARM designs to challenge the top of x86 on performance has always resulted in gigantic increases in power usage. Why? because as I stated up front, performance in these processors is not determined by the instruction set for the most part. It's all about massive caches, and cores... all of which drain huge amounts of power. What you see more these days is x86 chips scaling down performance to challenge ARM in these small environments (often by killing caches/cores) - while retaining compatibility with previous x86 software - which is becoming a bit futile now. What you don't get with X86 yet is the the ability to buy the design to incorporate into SOCs. I can't see this situation changing much for a while. X86 will retain the top end performance crown, but ARM represents a much better choice for these small computers in which high performance isn't the critical thing.

[@Seatux]

Baby steps. The M1 Macs are just the beginning towards similar performance or better on less power. I am eager to see a 12 - 16 processor core M chip next on the iMac next. Here's to hoping there would be a Windows ARM CPU that can compete or else we all be stuck using Apple hardware.

[@susugar3338]

I should choose x86 CPU (Intel Atom N2940) for a mini server? I am surprised by CPU benchmark in video.

[@Rick-vm8bl]

We're likely heading for a future where 90% of applications will be pefrect for ARM chips. Desktop computers, mobile phones, tablets, household devices etc. x86 will then be pretty much exclusively for high cpu intensive systems such as servers, transcoders, etc. Given the performance of the M1 is currently making Intel look extremely silly I expect we'll see some staggering performance improvements in ARM chips over the next few years, and thankfully it looks like a few players are stepping up their game to compete with Apple and Qualcomm.

[@winsomehax]

@@susugar3338 Does it perform at the level you need. How much does it cost. What's the power usage. Do you need to run x86 instructions? Add all that lot of up and you have your answer.

[@susugar3338]

@@winsomehax It is a mini PC (Atom N2940 -4GB DDR3 - 120GB SSD NVMe) and very cheap, only about $800. Power consumption is about 18W. It's running Debian 10 server, Hass IO and Zonemineder in docker. CPU usage is about 50 percent. Is it enduring when running 24/7? Sorry about lacking of information and my poor English.😟

[@bluephreakr]

Ties at the end? If we're being fairly strict about the results the Rock Pi X is a clear winner on ARM.

[@marksmithcollins]

To consider Rock Pi X as a representative of X86 world, Z8350 is too much horribly outdated

[@ExplainingComputers]

I never suggested that the Rock Pi X is representative of the x86 world in general. But it the pretty much the only x86 SBC to sell for the same price (equipped) as an ARM SBC like the Pi 4. So this video is representative of performance from x86 and ARM SBCs with the same price (and form factor).

[@marksmithcollins]

@@ExplainingComputersI agree, but is there enough candidate pool of x86 AIO SBC for ~$80 range ? I also didn't think you chose Rock Pi X because it is the king of x86 realm, but it didn't deserve your surprise of 'not overwhelm ARM'. Even in their release season Z8350 devices heard a lot of complains because of its performance.

[@SidebandSamurai]

Chris, this is a great video. Brilliant work, can you also try using an application like VLC media player, playing from a local file on both computers. I think it will more closely show media playback capabilities between both SBCs. I know these tests are a royal pain (yes pun intended) but you do such a through job, you try to keep everything as sterile, everything fair. Brilliant job again. Also is there a way you might be able to do the history of ARM processors? In the UK you had the Acorn PC which was ARM based. I heard this was a fantastic PC.

[@auser1484]

The music for this video should have been Survivor - Eye of The Tiger

[@noamw3841]

Learnt something new today, didn't know that arm and x86 can be equivalent, thanks for the video!

[@Chris.Brisson]

The King is dead (x86)! Long live the King (ARM)! As RISC-V approaches the throne.

[@PilotAtInception]

BeagleV?

[@chriholt]

Great showdown Chris! I agree with you that I have always considered ARM processors sort of minor-league novelties, but quite obviously that is not the case any more!

[@mikeymaiku]

i made it to get wooshed, im one of the first few

[@ExplainingComputers]

You made it indeed! And second. Silver medal for you today! :)

[@mikeymaiku]

@@ExplainingComputers i accept it with great pride and will hang it on my imaginary wall

[@Kenny-bw2cz]

Chris you didn't compare the power usage of the 2 boards. That is another important factor. You can get a kill-o-watt power meter or something similar and compare them in idle and under load. That would be interesting

[@ExplainingComputers]

I did a video on SBC power consumption recently, which included these two boards. The idle/load was 2W/4W for the Pi, and 2W/5W for the Rock Pi X: kzbin.info/www/bejne/b6DJoKR6mJuFl6c

[@Kenny-bw2cz]

@@ExplainingComputers thank you Chris.. thanks for pointing me towards that.

[@avejst]

Interesting SBC Thanks for sharing your knowledge to all of us 👍😊

[@reptilez13]

Fantastic video! Wouldn't the render of the 10 second duck clip still be a win (if there were multiple test passes especially) for the Rock Pi since its a 10 second clip and would be much more exxagerated in a longer video render? 1 second faster per 10 seconds of video would add up, or am I missing something?

[@ExplainingComputers]

I did not think of this -- excellent point on longer renders.

[@tralphstreet]

Chris I'd love to see how they compare in emulation. You should definitely integrate that into the testing, specially N64 and the likes which seem to give these SBCs a run for their money.

[@devjock]

Whenever I see "SBC" and "x86" in the title, I get a bit excited, but what I'm looking for in an x86 mobile system is never what the manufacturers are aiming for; A truly sick MS-DOS / Win98 rig in my pocket. With all the bells and whistles. Great soundblaster sound, Decent "videocard", fast boot time, etc.. Just a nice little deck to make some tracker modules in dos, pimp out Norton Commander with all the external programs to view any extention under the dos-sun.. I guess it's still a pipedream for now..