There is definitely a selection bias at play here: your run-of-the-mill personal computer owner does not even know what a benchmark is. Interesting (but not surprising) trend among enthusiasts though.
Is it though? There is historical data for that, and even when AMD was crushing Intel for years on all benchmarks including price/performance and early adopter usage, Intel ultimately ended up with a larger market share. Prescott was a (litteral) hot mess and the Athlon 64 3xxx series was amazing, but it did not take long until things came back to normal with the introduction of the Core series. I expect the same to happen here. AMD is great at slipping through the cracks between Intel's CPU generations, but it cannot seem to be able to become a market leader.
This is precisely my point: there were numerous moments in the past when AMD was trailing closely Intel among enthusiasts because of a superior product, but ultimately it never led to AMD overtaking Intel in the overall consumer market: Intel has always been able to catch up within a generation. I see no reason why this time would be different. And thus, the enthusiast market has little to no bearing on the overall market share.
The overall consumer market is decided by what OEMs ship in their laptops and desktops. Normal people cares about Intel and AMD as much as they care about the brand of the gasoline in their car, possibly less. Convince Dell & Co to buy from AMD and AMD is going to have the upper hand in consumer market.
Dell & Co do buy from AMD, so I think it's not quite that. Admittedly right now their laptop search tool indicates that they sell 130 models with Intel processors, while AMD is relegated to the 15" Inspirons and Alienware laptops. But still. It seems like maybe there's got to be some more subtle reason why you're not seeing them in the Chromebooks or XPS 13s than just that they're devoted to one manufacturer and shutting the other one out.
I think availability and price are the most important factors for the consumer market. If AMD can deliver a lot of chips with lower prices and same performance they might take a larger part of the market.
There is a huge difference this time: Intel no longer has a process lead and is actually now far behind the best fab (TSMC). Intel was able to come back last time because of Core + best fab in the industry. Without that advantage this time, it will be much more difficult for them.
They also used shady practices like paying OEMs to not use AMD CPUs. They were heavily fined for this. If they try to do it again, the government will likely give a greater punishment this time since they would be repeating the same crime. Especially now with the anti-trust friendly sentiments within the government.
I’m surprised AMD’s laptop share increased that much (even in benchmarks) — when I went looking for an AMD laptop a couple months ago, I could only find a handful that seemed good.
Intels drop is almost exactly equal to AMDs gain (20% market share between the most recent two laptop datapoints for both). If it was just intel laptops dropping out and converting to apple silicon you wouldn't expect AMD to grow.
I suspect that macs just aren't represented in this graph, probably very few people benchmarks macs when they're identical to all the other macs.
I know plenty of people who run Windows on intel Macs (particularly the laptops) in various ways and they actually do play with benchmarks to see how well it works.
Yeah, that's what I'm saying. Apple Silicon doesn't show up anywhere in the benchmark. If we were seeing a chart of totals rather than percentages, I'd expect to see a void where Apple Silicon entered.
> As the PerformanceTest software only runs on Windows OS and counts on user submitting their benchmarks. This chart may be non reflective of non Windows user base.
Then the intel percentage is (1+2)/(1+2+3) and the AMD percentage is 2/(1+2+3). But if Apple hardware is a significant portion of the benchmarks, then when intel Macs disappear, the fraction of systems that can submit the benchmark decreases. So the intel percentage decreases and the AMD percentage increases.
The denominator changes identically for intel and AMD.
I doubt it. Consider that Apple is the #4 personal computer vendor after Lenovo, Dell, and HP. They have less than 10% market share and many of their products are still shipping in Intel versions.
Tim Cook have said Apple sold more M1 Macs than Intel-based Macs. In a recent Gartner report, Apple held 8% of the market share worldwide. This implies Intel has lost over 4% of the market in just one quarter.
Relatively few people are probably looking for the perfect laptop. The average user might just want something cheap that does the basic tasks for them, and this is exactly where you will find more and more AMD CPUs.
Funny that you mention that AMD CPUs are the 'cheap and basic' option when it seems like over the past 5 years Intel has become the inferior chip manufacturer.
I didn't mean to imply that the CPUs are inferior, quite the contrary - they give you more bang for the buck, which is why you often find them in cheap offers. When the AMD CPUs are cheaper than the Intel CPUs, of course more computer makers will put them into their cheap builds. The fact that they are better CPUs is a nice bonus.
I am not buying it either. AMD has been the performance leader for a while. OEMs putting the fastest CPU into a junk chassis because "they are cheaper" makes zero sense unless we're talking shenanigans
Until the 4000 series of Ryzen mobile Intel had a pretty clear lead in 15-25w part market. With wider availability of Ryzen 5000 series mobile CPUs I expect to see far more options by Q12022.
I purchased a first gen Zen CPU around when they first came out and noticed a lot of random issues on Linux that often resulted in full system crashes; usually due to C-State stuff and other times due to various motherboard firmware issues that only Linux users seemed to experience.
They didn't take too long to get fixed but you needed the latest versions of the kernel to keep up with them. However other than those first few months my experiences match yours and I have experienced zero issues.
I could see it being possible someone possibly still using an older version of the kernel would still experience these issues, especially if they haven't flashed their bios.
I had to stop the installer from booting, add this hard to find line:
ivrs_ioapic[4]=00:14.0 ivrs_ioapic[5]=00:00.2
to the kernel load line, just to be able to install it.
And then, after finally managing to install it, the stability issues. Even with the latest Liquorix kernel, after some hours, the system just kept randomly freezing. It was a total infuriating experience.
Latest would be Zen 3 or Ryzen 3000 series. I've got Zen 2 chips, Ryzen 2000 series, same as you.
yesco's comment above yours is probably useful here - I can definitely believe there were some issues with Ryzen when it first came out. How long ago was this and what kernel version were you using? The links you posted seem to reference Linux 4.x and 5.1.
If you tried again with some recent 5.8/5.10 kernel I bet you'd see a marked improvement. Linux moves quickly and support for new hardware is generally established early, though not without initial bugs ;)
I tried running Linux for one full year, since I got the laptop, until I got fed up. Even if Linux in Ryzen works flawlessly now, I got an Intel laptop from my employer and that one has all the Linux stuff.
Another honest review of issues with Ryzen, hopefully he is someone you can believe:
https://youtu.be/Fmj2kUk-xO4
I am somewhat surprised even though anecdotally over the past 18 months, everyone I know that got a new laptop went AMD. But I'm not an "average" person, so I've got tech-knowledgeable friends and family that takes my advice.
Asus Zephyrus G14, HP Omen 15, Lenovo Legion 5, Lenovo Thinkpad T14.
These are the laptops people around me bought. Ryzen 4000 proved AMD can make good laptop CPUs (APUs) and Ryzen 5000 has been welcomed by OEMs in a way that hasn't happened before for AMD.
But in business it's still much more likely for someone to go Intel...
You should keep in mind that these stats are not per machine where steam is installed. It's per account. It causes distortions when multiple accounts are used on the same computers. The prominent case would be internet cafes, which are considerably popular in China and Korea
ARM is missing from the list for now. With Apple moving to ARM, game developers will need to fix their remaining memory ordering bugs in their multithreaded codebases.
Intel stopped using Pentium 4's NetBurst architecture and switched to Core derived from Pentium III
High power consumption and heat intensity, the resulting inability to effectively increase clock rate, and other shortcomings such as an inefficient pipeline were the primary reasons why Intel abandoned the NetBurst microarchitecture and switched to a different architectural design, delivering high efficiency through a small pipeline rather than high clock rates.
I wonder how they determine the server market share. It seems way too low given that over the last year, I see offers for AMD-based servers and vServers pop up everywhere at least in Germany. Many hosters support AMD as an option or as the only choice.
The longevity is a good point, but if those numbers are derived from users running PassMark on those servers, I would suspect that they don't keep running the software on their server they purchased some years ago: From my understanding, people would be more likely to run this software on a recently-purchased computer to evaluate its performance.
Yeah, EPYC CPUs kicks ass while Xeons cost an arm and a leg, so I really don't understand how Intel could be selling more server CPUs than AMD right now, unless it is due to some OEM shenanigans (like on laptops)
A lot of server customers care about more than just how good the CPU itself is. Having a reliable support ecosystem and platform continuity is an important consideration in the server market. Intel has proven for decades that they can provide these things, while AMD burned a lot of people when the Opteron line was left to rot. It's going to take time for AMD to prove itself a reliable supplier in the server market.
maybe the "reviews" should never be read? I can't say I've ever read any. I just go there to compare hardware benchmarks as the website name indicates. if you go there looking for reviews then more fool you. they're are plenty of good hardware review websites. this website is an excellent collection of hardware benchmarks, nothing more
Its also the results produced by their "benchmark" that are skewed in Intel's favor and changes every time amd release something better. Just never use this site for anything, read reviews from trusted hardware review sites.
I'm sure anything Intel can do AMD can match or better, the numbers can only skew so far. plus that's just some random guy on youtube's spin. the website also holds masses of aggregated data for loads of other hardware, which is extremely useful and valuable for me
Fascinating that AMD is not dominating in the arena of servers: That's where I would've expected the biggest adoption. They're generally more affordable than Intel especially in the arena of performance per watt, and the differences in feature parity for things like media playback have less of an impact there.
Can anyone give me insight into why Intel is so far ahead in this space?
It’s pretty common actually. A server buy can easily be an 8 figure deal, getting servers that turn out to be stupidly configured for your apps that your stuck with for 5 years is exceptionally dumb.
I should've emphasized lab before of homelab. Agree, doing a full stack hardware validation (managed to even get a full multi-vendor stack deployed into a Dell validation colo to do our internal performance test). Even then, we only ran out own performance tests, as we cared about our workload, not a generic workload suite.
Edit: Additionally, we only did that work on a single set of systems prior to purchase, not the systems post purchase, so 1 of N, instead of rerunning on N. It was presumed that performance would remain the same on same hardware configuration etc.
Larger companies have no problems finding people to pay to support Linux, or if they are really large (say, Amazon), supporting it themselves and being who people (including big companies) pay to support and host Linux systems.
it all depends on size. some larger (typically older) companies with smaller engineering teams will make the tradeoff you are describing. But the big players today, who manage hundreds of thousands to millions of servers are not going to pay per machine / per core license counts for support. They build and maintain most of their own software for their own needs.
it says it right in the description, which you don't need a subscription to read
> In 2019, the Windows operating system was used on 72.1 percent of servers worldwide, whilst the Linux operating system accounted for 13.6 percent of servers. Compared to 2018, both companies experienced an increase to their overall market share.
There's often dozens of Windows servers installed on site in millions of offices around the world, doing various tasks (email, active directory, file storage, etc).
I'm sure if you looked at "servers installed in a data centre" then it would be closer to what you expect. But there are so many servers doing "boring" jobs in cupboards youll never see it totally skews the figures.
This is so true! Most of the in-house data centers I had seen had Windows and they didn't really bother to even think about changing. It's totally different from cloud service providers, that some of them don't even provide a Windows server option.
Many of the VPS providers don't even have windows options because the demand is so low it's not even worth offering as a product.
Server farm sysadmins I know habitually lament they are rusty in windows whenever it's in front of them because it's completely vanished from their daily experience
If this was a survey of companies where the weight of say, a liquor store with a 20 year old beater in the back closet that they'll call a server is 1 vote, and say, all of Amazon is also 1 vote, then maybe you can get that number, but that's basically the only way
There's a difference again. Is "server" a chassis in a rack with redundant power and some RAID array or is it some standard computer that was christened as such? It's not a useful category because it allows for arbitrary segmenting to pump the numbers. Are we counting by service, machine, company? Depends what result you want.
If you're going to count windows machines (as in cases) by hand at a Noc, farm or center you won't be needing a calculator or likely a second hand.
Windows is certainly more prominent than Solaris or any commercial traditional unix but outside of desktops, which have a cleaner definition, it's been going in same direction for 20 years.
Microsoft has also been pivoting away from windows as a revenue source for a while. Video consoles, media companies, Azure, LinkedIn, GitHub, they don't necessarily see windows in their future and we shouldn't either
Not all servers are web facing or live in data centres. There are tons of small businesses that run servers in a backroom. Those are predominantly Windows.
right but that analysis isn't useful. If I used the same logic to determine a solution for hauling a large load by vehicle and looked at how most cubic meters of things are hauled, I'd probably be forced to conclude that a sedan trunk is the best option. However that's only due to the preponderance of sedans.
It's a very careful way to state the question to make it look like the numbers back what's clearly a wrong answer.
It puts the framing of "well I guess I'm using a sedan" and resituates the problem to solve all its shortcomings. It's a classic implicit framing propaganda technique.
Not only does Microsoft offer comprehensive support packages for their products, but in fact the TCO of a Windows Server system is almost always lower than a Linix system while also having the better user productivity. Considering this, any business that is serious about digital transformation would do well to indeed reach for Microsoft’s portfolio.
Of course it‘s easy to make up statistics to show that the OS made by some Sovjet hacker in his mother’s basement (probably using lots of stolen MS IP) is supposedly more popular than the products made in the US by a reputable firm such as Microsoft, but who is going to believe that?
I used to have a small startup with over 50 servers per engineer. Tell me all about TCO and digital transformation, I really need to learn about that stuff.
I work for a hosting provider. Windows exists and it's used for corporate infrastructure, but the vast majority of actual applications and support services are running on Red Hat, CentOS or Ubuntu.
Windows trends to be reserved for AD, jump hosts (much to my chagrin, I prefer SSH), virtual desktops and stuff that end-users access, because user management is where Windows is still ahead.
I suppose in small businesses they will have all the user management bits but no need for more than one or two "real" servers, so it doesn't matter if it's Windows too.
In bulk, Linux servers are much easier to manage though and since most application platforms will work just fine if not better on Linux, it's just the sensible choice unless you have something specific that requires Windows. Generally easier licensing certainly does not hurt.
The fact that on any Linux distribution you get trivial, easily-managed access to thousands of software packages free of licensing worries out-of-the-box makes for a large advantage over Windows where the concept of using a package manager and central repository seems to be a fairly new thing and you have to set everything up yourself or just deal with developers and users installing random crap from who knows where because by default there's no better option. There's also the fact that Linux server software tends not to have GUIs because servers are headless by default, which also makes them easier to manage once you run out of fingers to count them.
Are these real numbers or are they counting SMB instances running on commodity hardware and doing other weird junk to play with the stats?
Windows isn't even the most used OS on Azure. Certainly large deployments at Google, Facebook, Amazon, etc. accounting for a huge proportion of the population aren't running Windows either.
Intel CPUs are also a bit more flexible on the high end: if you want 12T of RAM in one machine, you still need an Intel 8-socket machine. I wonder if this is costing AMD a bit in the server space.
I don't think that's too important. Scaling that much on a single server is economically rather bad and we're talking about mid 6 digits per Server. If it's in any way posse, and it usually is, you'll probably want to scale horizontally. These servers sure exist, but I highly doubt they make up a large share of the market.
There's a lot of caveats on this data. It's only showing the ratio of people submitting reports through their benchmark software, which apparently only runs on Windows. So, that makes it sound like the entire space of Linux servers is excluded.
I'm just speculating here, but it feels to me like they're just not accurately measuring the market share for servers.
Purely anecdotally, it's been a very long time that I encountered a non-Intel server. I actually wouldn't be surprised if the numbers were roughly in the right ballpark, ±5% or so.
In AWS I use the AMD whenever I can (faster and cheaper) but they aren’t as available so my terraform either has to get complex logic about A-Z or default to Intel or just use two AZs which is less available.
Where I last worked they said it would be hard to switch to AMD quickly because Intel CPUs had a feature set they were dependent on and would take some time/effort to switch.
The two main reasons AMD is not dominating is servers right now are:
1) 5 year support contracts
2) OEM unavailability.
Even now getting an AMD server is a "go out of your way" situation. Add to the chip shortage and you have people holding on to hardware a bit longer or going intel for availability (those are minor reasons).
Must be the inertia. CPUs live long in server space: where I'm looking, server providers offer hardware for rent that is up to 10 years old, and AMD only started to rule CPUs in 2017.
E.g. the CPU on the VPS I'm renting is from 2013, and it makes perfect sense that it is form Intel.
It will take a long time, but you can still see the uptick. Zen has the potential to take over Intel at least, considering density and pricing, they are exceptional value.
Intel adds a bunch of software value to their processors that allow them to compete on a different plane than AMD. Kinda like how Apple turns ordinary laptops into objects of desire with proprietary software: Intel does the same thing to datacenters.
As a user of many servers, I use very little Intel code other than what they've contributed to Linux. Am I missing out? I think I'm a pretty typical hyperscaler.
Intel’s MKL + their compiler + their CPUs with the latest SIMD extensions is still one of the best stacks for Scientific programming. Get a scientist to write some single-core kernel in C or C++ language, and that compiler will work serious miracles on code that needs to scale up to hundreds if not thousands of machines (think supercomputers). The game only changed here recently when Nvidia went all in on the server GPU offerings in the past few years, but writing fast CUDA can get much messier than your average biologist would care for.
Intel has more software engineers that AMD has employees, so their support for their math libraries, compilers, their GPU drivers, their benchmarking tools, and their contributions to the LLVM ecosystem, the Linux kernel, and other projects have helped them get an iron grip on academic/scientific side of high-performance computing.
There is a lot of talk of non-performance related drivers of market share here, but I think there is a lot to be said for ISA extensions and libraries. A lot of HPC code relies on AVX512 and MKL.
- According to the text below the chart, the metric is "CPUs in use rather than CPUs purchased". Assuming servers are replaced after 5 years, and assuming unit sales have not changed massively over that period (big assumptions I know, but just hand waving here...), that would mean each quarter of sales is only 5% of the total being reported here. Now in the past quarter AMD went from 2.5% total market share to 5% total market share. That would mean that AMD had 50% market share of sales in the past quarter (assuming they were selling barely any servers 5 years ago).
- Second metric, consider core count rather than CPUs. AMD's biggest CPUs have over the past year or so had 2x the core count of Intel's (64 vs 28) and hyperscalers at least tend to buy the highest core count machines, so if comparing in terms of cores, the weighting shifts even more in AMDs favor.
On one hand AMD stuff might still not be too easy to get.
In my case I had to wait ~1 month to get a server having a Ryzen 9 5950X (the AX101 here https://www.hetzner.com/dedicated-rootserver/matrix-ax ). I was willing to pre-order & wait indefinitely, other people might want a server "now".
On the other hand during the pre-Zen3 era I personally didn't feel an urge to swap from Intel to AMD (I didn't feel like there was a lot of difference of performance/price). This has changed for me with Zen3 so maybe things will change during the next months... .
Apple has a 15% computer market share, of which only a fraction is the new ARM processors. There is a 0% chance of them having a 30% market share in the next 5 years.
Every percentage is a fraction so that statement doesn't mean anything. But Tim Cook has said Apple sold more M1 Macs than Intel-based Macs. This is an indication that fraction is over 50% already.
Kinda surprised on the server chart, consider most of my Linode and Vultr instance are running on AMD epyc CPUs for a years for now (at least since I last checked). I believe the market shares for AMD is much higher than the chart shows consider how low the cost per core is for a CPU slot.
The server market share shows just how slow that sales channel is & why Intel isn't too panicked whenever AMD overtakes them for a bit (IIRC server sales are a significantly larger share of revenue followed by laptop sales). AMD would need to keep up their advantage for ~5-10 years to make a meaningful dent in replacing those old Intel servers that are still chugging along for now.
Server replacement rate is actually fairly high in the industry (5 years at most and usually less than 4 for tech oriented companies last time I read an industry report). The main problem is you can't buy AMD based servers. They are completely unavailable, even to hyperscalers judging from the graphs. The only reason Intel is keeping the lead is because they can actually fulfill the demand.
Right. Server replacement is ~5 years but it's staggered. So you'll gradually roll out new servers & gradually decommission old ones (with a rapid decommissioning phase towards the end of the server lifetime). However, not every company will adopt AMD processors when they first become available OR they may hedge their bets & do a mixed blend (e.g. Amazon offers Intel & AMD for their customers). Both of these are why it'll probably take 5 years to start seeing the initial impact & longer to see the full impact (e.g. AMD start exceeding Intel marketshare) for the early majority/later majority & laggards to adopt.
As far as being unable to buy AMD based servers I don't know. There's certainly a lot of announcements about it:
Obviously it's hard to know how many they've rolled out but I think it's more than a trivial amount. I'm suspicious of the graph because I don't know how they're capturing that data.
I've never used an AMD CPU and am surprised at these numbers! I do a lot of scientific programming (Matlab, etc) and wanted to be sure I was on the happy path when it comes to the vector extensions, etc.
Looks like the sample size has some bias built into it given the data is from those who use their performance software. I would trust the numbers from Firefox much more (higher scale, much more ubiquitous product).
52% of computers here are with 2 cores. 40% are with 4gb of RAM or less. 20% are 32bit, which haven't been sold for god knows how many years. These stats do not represent new PC that are being sold.
I'd be more open to AMD if Linux support wasn't so spotty. I haven't installed Linux on an AMD desktop in years so I'd like to hear if that's changed. While I don't mind Intel, I'd certainly prefer paying for performance at a lower price.
Archlinux. Didn't really need to consider drivers; things just worked out of the box mostly with exception of the GPU where I've installed the AMD driver via pacman (a one liner).
I think x86 is an outdated design that is about to be phased out.
2021-06-14 Prediction: Intel will use "RISC-V plus x86 compatibility layer" or "RISC-V plus x86 heterogeneous computing architecture" to develop a new generation of "warehouse/workshop model" CPU
Regarding AMD laptop CPU market share, I'll offer an anecdote why haven't I joined AMD.
I've been trying to find a mini-workstation grade laptop and it seems there's almost no choice. Main limitation is actually the amount of maximum RAM and whether it's dual-channel.
If one doesn't care about that then there's actually plenty of excellent AMD CPU laptops.
- Lenovo seems to only offer the X13 that allows for 32GB of RAM in dual channel. Are there others? The Yoga Slim 7 Pro looks amazing but it seems to only have 1 slot for RAM, capped at 16GB maximum?
- HP has the 255 G7.
Can somebody offer some more of these, if there are any? Also, is the Lenovo Yoga Slim 7 Pro actually capped at 16GB RAM single channel? The information on the internet has been contradictory -- some folks on Reddit claim they use 32GB in two slots, the sellers in AliExpress told me in a chat that the laptop only has 1 RAM slot... Confusing.
How can a programmer get a semi-workstation grade AMD laptop these days?
I've got an HP Omen 2020 with 8c/16t zen2 4800, 64GB ram, 3TB nvme ssd, Nvidia 1660Ti and AMD Renior graphics. I could have opted for 128GB but I usually don't run enough on the machine to tax even the 64G.
Display isn't UHD, but a very crisp 1920x1280. I run it, lid closed, to drive an displayport 4k monitor connected via 1 mini DP plug.
I bought it off of Amazon late last year w 64GB ram and upgraded the SSDs myself.
Positively blows the doors off the similar age Dell unit with an i9 10k something sitting next to it. A huge part of that may be that I run Linux on mine, and the day job runs an ultra locked down w10. Dell only has 32GB ram and 500GB Data ssd. Has UHD 4k display, but I can't see that well. Retina displays are easier on the eyes IMO.
Well, I'm also thinking of abandoning the idea of a Ryzen laptop, buy a Threadripper desktop workstation (and put Linux on it) and then also buy a MacBook Air M1 and remote into the workstation through it -- exactly because the retina screens are sadly still unbeaten.
Such a plan can get very expensive though... So jury is still in recess.
Thanks a lot for the product name! I'll definitely look it up.
Do you have any remarks on fan noise and thermals? Additionally, is the display good enough to work on a sunny day while being in the shade? And can you choose integrated AMD graphics only?
Fans do wind up on heavy CPU/GPU load. The HP laptop cooler is quite a bit better than many I've had before, and the fan winds down about as quickly.
I've brought this laptop to outdoor (!) karate class when we had some zoom students on. I manually adjusted the brightness, but it was fine.
By default, mint installs the AMD driver, and nouveau. I wrote a simple switcher a few months ago, so I can toggle between AMD and NVidia. Next gen mint (due in a few weeks) apparently has the AMD/Nvidia optimus like switcher built in.
For this laptop, one downside is that I could not get the AMD to drive the external displays as well as the laptop display. Nvidia can drive both. It looks like this is a switcher config, that xrandr can't handle (at least the versions I have on there). If this is fixed in the next version, I'll happily use AMD for most display and NVidia for my GPU work (computing).
171 comments
[ 3.0 ms ] story [ 226 ms ] threadThere is definitely a selection bias at play here: your run-of-the-mill personal computer owner does not even know what a benchmark is. Interesting (but not surprising) trend among enthusiasts though.
They also used shady practices like paying OEMs to not use AMD CPUs. They were heavily fined for this. If they try to do it again, the government will likely give a greater punishment this time since they would be repeating the same crime. Especially now with the anti-trust friendly sentiments within the government.
I suspect that macs just aren't represented in this graph, probably very few people benchmarks macs when they're identical to all the other macs.
It looks like their software doesn't support macOS so it's pretty safe to say there are no Macs in the dataset.
Their software does support Andriod & iOS. It would be interested to see how these changed if they were included.
> As the PerformanceTest software only runs on Windows OS and counts on user submitting their benchmarks. This chart may be non reflective of non Windows user base.
1. intel running Windows OS
2. AMD running Windows OS
3. intel Mac running Windows (various ways)
Then the intel percentage is (1+2)/(1+2+3) and the AMD percentage is 2/(1+2+3). But if Apple hardware is a significant portion of the benchmarks, then when intel Macs disappear, the fraction of systems that can submit the benchmark decreases. So the intel percentage decreases and the AMD percentage increases.
The denominator changes identically for intel and AMD.
This graph more-or-less only measures the enthusiast Windows user market.
Reason: Linux drivers. Ryzen Linux support is just that bad.
I have not ever heard of any issues with Linux on Zen architectures, nor have I experienced any on either of my Zen 2 machines.
They didn't take too long to get fixed but you needed the latest versions of the kernel to keep up with them. However other than those first few months my experiences match yours and I have experienced zero issues.
I could see it being possible someone possibly still using an older version of the kernel would still experience these issues, especially if they haven't flashed their bios.
Plenty of issues starting from the installation.
I had to stop the installer from booting, add this hard to find line: ivrs_ioapic[4]=00:14.0 ivrs_ioapic[5]=00:00.2 to the kernel load line, just to be able to install it.
And then, after finally managing to install it, the stability issues. Even with the latest Liquorix kernel, after some hours, the system just kept randomly freezing. It was a total infuriating experience.
And I was not the only one with these issues:
https://www.linuxquestions.org/questions/linux-newbie-8/can%...
https://bbs.archlinux.org/viewtopic.php?id=250297
I nuked Linux, and now the machine runs only Windows. And it works fine. No more randomly freezing.
That's what I get for being an AMD fanboy and an early adopter. I'm glad you have the latest Zen processors, I am not so lucky.
yesco's comment above yours is probably useful here - I can definitely believe there were some issues with Ryzen when it first came out. How long ago was this and what kernel version were you using? The links you posted seem to reference Linux 4.x and 5.1.
If you tried again with some recent 5.8/5.10 kernel I bet you'd see a marked improvement. Linux moves quickly and support for new hardware is generally established early, though not without initial bugs ;)
Another honest review of issues with Ryzen, hopefully he is someone you can believe: https://youtu.be/Fmj2kUk-xO4
Just trying to express that as valid as your experiences are, equally valid is the fact that the issues you experienced are no longer present.
Asus Zephyrus G14, HP Omen 15, Lenovo Legion 5, Lenovo Thinkpad T14.
These are the laptops people around me bought. Ryzen 4000 proved AMD can make good laptop CPUs (APUs) and Ryzen 5000 has been welcomed by OEMs in a way that hasn't happened before for AMD.
But in business it's still much more likely for someone to go Intel...
AMD has gone from 20% to 30% in the past 18 months among PC gamers.
https://store.steampowered.com/hwsurvey
From the chart, it looks like Intel lost 10% of the PC gaming market share in just 1 year.
High power consumption and heat intensity, the resulting inability to effectively increase clock rate, and other shortcomings such as an inefficient pipeline were the primary reasons why Intel abandoned the NetBurst microarchitecture and switched to a different architectural design, delivering high efficiency through a small pipeline rather than high clock rates.
https://en.wikipedia.org/wiki/NetBurst_(microarchitecture)
https://en.wikipedia.org/wiki/Intel_Core_(microarchitecture)
tl;dr we're not seeing much change over the past 18 months.
Also, VPS/hosting are far from the entirety of the server market: don't forget all the on-prem, enterprise, and SOHO boxes.
https://cpu.userbenchmark.com/
Can anyone give me insight into why Intel is so far ahead in this space?
I think this excludes huge population of servers as we know Windows is not dominant there.
This may be more useful information for the server market, even if old.
https://www.tomshardware.com/news/amd-vs-intel-q3-2020-cpu-m...
Edit: Additionally, we only did that work on a single set of systems prior to purchase, not the systems post purchase, so 1 of N, instead of rerunning on N. It was presumed that performance would remain the same on same hardware configuration etc.
Edit: Because people seem to think that the numbers are made up, here's some more sources with not quite as high, but similarly dominant numbers:
Red Hat: https://www.redhat.com/en/blog/red-hat-continues-lead-linux-... - microsoft @ 49% in 2017
t4: https://www.t4.ai/industry/server-operating-system-market-sh... - microsoft @ 47% in 2018
If it's going to run Linux, you tick that option when ordering. Or you tick "no OS".
> In 2019, the Windows operating system was used on 72.1 percent of servers worldwide, whilst the Linux operating system accounted for 13.6 percent of servers. Compared to 2018, both companies experienced an increase to their overall market share.
(And since it describes both “Windows” and “Linux” as “companies” that gained marketshare, I’m somewhat suspicious of their information gathering.)
But not being registered with them, I don’t have access to what passes for their detailed source information for that claim.
I'm sure if you looked at "servers installed in a data centre" then it would be closer to what you expect. But there are so many servers doing "boring" jobs in cupboards youll never see it totally skews the figures.
It'd be the easiest $10,000 I'd ever make
I don't have access to the full numbers, but I'd bet on the linux side being majority red hat for the same reason.
Look at top500, it's literally 0% windows, https://www.top500.org/statistics/list/
Many of the VPS providers don't even have windows options because the demand is so low it's not even worth offering as a product.
Server farm sysadmins I know habitually lament they are rusty in windows whenever it's in front of them because it's completely vanished from their daily experience
If this was a survey of companies where the weight of say, a liquor store with a 20 year old beater in the back closet that they'll call a server is 1 vote, and say, all of Amazon is also 1 vote, then maybe you can get that number, but that's basically the only way
you could even look at Red Hat, one of microsoft's competitors, who showed microsoft's market share at 49.6% in 2017: https://www.redhat.com/en/blog/red-hat-continues-lead-linux-...
So the position moved 71 -> 49 -> 24
There's a difference again. Is "server" a chassis in a rack with redundant power and some RAID array or is it some standard computer that was christened as such? It's not a useful category because it allows for arbitrary segmenting to pump the numbers. Are we counting by service, machine, company? Depends what result you want.
If you're going to count windows machines (as in cases) by hand at a Noc, farm or center you won't be needing a calculator or likely a second hand.
Windows is certainly more prominent than Solaris or any commercial traditional unix but outside of desktops, which have a cleaner definition, it's been going in same direction for 20 years.
Microsoft has also been pivoting away from windows as a revenue source for a while. Video consoles, media companies, Azure, LinkedIn, GitHub, they don't necessarily see windows in their future and we shouldn't either
It's a very careful way to state the question to make it look like the numbers back what's clearly a wrong answer.
It puts the framing of "well I guess I'm using a sedan" and resituates the problem to solve all its shortcomings. It's a classic implicit framing propaganda technique.
Of course it‘s easy to make up statistics to show that the OS made by some Sovjet hacker in his mother’s basement (probably using lots of stolen MS IP) is supposedly more popular than the products made in the US by a reputable firm such as Microsoft, but who is going to believe that?
Plenty of open source people have offered paid support so corporate clients could expense things.
Windows trends to be reserved for AD, jump hosts (much to my chagrin, I prefer SSH), virtual desktops and stuff that end-users access, because user management is where Windows is still ahead.
I suppose in small businesses they will have all the user management bits but no need for more than one or two "real" servers, so it doesn't matter if it's Windows too.
In bulk, Linux servers are much easier to manage though and since most application platforms will work just fine if not better on Linux, it's just the sensible choice unless you have something specific that requires Windows. Generally easier licensing certainly does not hurt.
The fact that on any Linux distribution you get trivial, easily-managed access to thousands of software packages free of licensing worries out-of-the-box makes for a large advantage over Windows where the concept of using a package manager and central repository seems to be a fairly new thing and you have to set everything up yourself or just deal with developers and users installing random crap from who knows where because by default there's no better option. There's also the fact that Linux server software tends not to have GUIs because servers are headless by default, which also makes them easier to manage once you run out of fingers to count them.
Windows isn't even the most used OS on Azure. Certainly large deployments at Google, Facebook, Amazon, etc. accounting for a huge proportion of the population aren't running Windows either.
So a reasonable guess for today seems > 5 million.
[1] https://en.wikipedia.org/wiki/Google_data_centers
[2] https://www.datacenterknowledge.com/archives/2011/08/01/repo...
- Stable\known CPUs, most vendors are used to Intel.
- Infrastructure\Support, Intel has a lot more infrastructure to support customers\contracts
- Long cycles, servers are more complex to replace and contracts can span multiple years.
- Compatibility, many rely on Intel specific features or are afraid to try their solution on a different CPU vendor.
- Relationships\Market Share, Intel still is a far bigger brand with more prestige
I'm just speculating here, but it feels to me like they're just not accurately measuring the market share for servers.
(in any case I cannot run it - needs the ncurses v5 shared library, I have already v6 installed on my AMD server)
1) 5 year support contracts
2) OEM unavailability.
Even now getting an AMD server is a "go out of your way" situation. Add to the chip shortage and you have people holding on to hardware a bit longer or going intel for availability (those are minor reasons).
Anecdotally I just ordered 50 or so AMD servers.
E.g. the CPU on the VPS I'm renting is from 2013, and it makes perfect sense that it is form Intel.
Intel has more software engineers that AMD has employees, so their support for their math libraries, compilers, their GPU drivers, their benchmarking tools, and their contributions to the LLVM ecosystem, the Linux kernel, and other projects have helped them get an iron grip on academic/scientific side of high-performance computing.
BTW, lots of HPC are (again) going for AMD. Hopefully I don't end up getting subpoenaed twice this round! It was a little painful last time.
- According to the text below the chart, the metric is "CPUs in use rather than CPUs purchased". Assuming servers are replaced after 5 years, and assuming unit sales have not changed massively over that period (big assumptions I know, but just hand waving here...), that would mean each quarter of sales is only 5% of the total being reported here. Now in the past quarter AMD went from 2.5% total market share to 5% total market share. That would mean that AMD had 50% market share of sales in the past quarter (assuming they were selling barely any servers 5 years ago).
- Second metric, consider core count rather than CPUs. AMD's biggest CPUs have over the past year or so had 2x the core count of Intel's (64 vs 28) and hyperscalers at least tend to buy the highest core count machines, so if comparing in terms of cores, the weighting shifts even more in AMDs favor.
Of course there are plenty of caveats here.
In my case I had to wait ~1 month to get a server having a Ryzen 9 5950X (the AX101 here https://www.hetzner.com/dedicated-rootserver/matrix-ax ). I was willing to pre-order & wait indefinitely, other people might want a server "now".
On the other hand during the pre-Zen3 era I personally didn't feel an urge to swap from Intel to AMD (I didn't feel like there was a lot of difference of performance/price). This has changed for me with Zen3 so maybe things will change during the next months... .
They will get there in the next couple of years, but I don't there there yet.
"People who like software want to build their own hardware"
I am surprised, I thought they'd be moving towards the higher margins of software and marketplace even more
As far as being unable to buy AMD based servers I don't know. There's certainly a lot of announcements about it:
https://blog.cloudflare.com/technical-details-of-why-cloudfl...
https://www.crn.com/news/cloud/new-google-cloud-vms-with-amd...
https://aws.amazon.com/ec2/amd/
https://azure.microsoft.com/en-us/blog/announcing-new-amd-ep...
Obviously it's hard to know how many they've rolled out but I think it's more than a trivial amount. I'm suspicious of the graph because I don't know how they're capturing that data.
Looks like the sample size has some bias built into it given the data is from those who use their performance software. I would trust the numbers from Firefox much more (higher scale, much more ubiquitous product).
Firefox shows total CPUs being used. By new and potentially very old computers. My own desktop CPU was made in 2012.
Cpubenchmark.net shows new machines being built or bought.
So, think of the latter (this month sales) as the rate of change for the first (total CPUs being used).
I see no reason at all to benchmark an old computer, just new ones.
Therefore, if current trends keep, Firefox reported AMD market share will grow. If.
2021-06-14 Prediction: Intel will use "RISC-V plus x86 compatibility layer" or "RISC-V plus x86 heterogeneous computing architecture" to develop a new generation of "warehouse/workshop model" CPU
https://github.com/linpengcheng/PurefunctionPipelineDataflow...
Also, are there no higher quality numbers about what these companies actually sell?
I've been trying to find a mini-workstation grade laptop and it seems there's almost no choice. Main limitation is actually the amount of maximum RAM and whether it's dual-channel.
If one doesn't care about that then there's actually plenty of excellent AMD CPU laptops.
- Lenovo seems to only offer the X13 that allows for 32GB of RAM in dual channel. Are there others? The Yoga Slim 7 Pro looks amazing but it seems to only have 1 slot for RAM, capped at 16GB maximum?
- HP has the 255 G7.
Can somebody offer some more of these, if there are any? Also, is the Lenovo Yoga Slim 7 Pro actually capped at 16GB RAM single channel? The information on the internet has been contradictory -- some folks on Reddit claim they use 32GB in two slots, the sellers in AliExpress told me in a chat that the laptop only has 1 RAM slot... Confusing.
How can a programmer get a semi-workstation grade AMD laptop these days?
Display isn't UHD, but a very crisp 1920x1280. I run it, lid closed, to drive an displayport 4k monitor connected via 1 mini DP plug.
I bought it off of Amazon late last year w 64GB ram and upgraded the SSDs myself.
Positively blows the doors off the similar age Dell unit with an i9 10k something sitting next to it. A huge part of that may be that I run Linux on mine, and the day job runs an ultra locked down w10. Dell only has 32GB ram and 500GB Data ssd. Has UHD 4k display, but I can't see that well. Retina displays are easier on the eyes IMO.
Such a plan can get very expensive though... So jury is still in recess.
Thanks a lot for the product name! I'll definitely look it up.
Do you have any remarks on fan noise and thermals? Additionally, is the display good enough to work on a sunny day while being in the shade? And can you choose integrated AMD graphics only?
I've brought this laptop to outdoor (!) karate class when we had some zoom students on. I manually adjusted the brightness, but it was fine.
By default, mint installs the AMD driver, and nouveau. I wrote a simple switcher a few months ago, so I can toggle between AMD and NVidia. Next gen mint (due in a few weeks) apparently has the AMD/Nvidia optimus like switcher built in.
For this laptop, one downside is that I could not get the AMD to drive the external displays as well as the laptop display. Nvidia can drive both. It looks like this is a switcher config, that xrandr can't handle (at least the versions I have on there). If this is fixed in the next version, I'll happily use AMD for most display and NVidia for my GPU work (computing).