As another commenter noted, this is data from the German online shop MindFactory, which could be understood as the German equivalent to NewEgg. MindFactory would probably overrepresent enthusiast and small business customers, and underrepresent your average consumer who would rather buy their laptop from a big box store (MediaMarkt/Saturn) or enterprise customers, who probably buy directly from Dell/Lenovo/HP/Fujitsu.
Also Germans in general are more into alternatives like Firefox, Linux, etcetera as they have a healthy distrust to Big Tech. Even though if AMD is quite big itself it was more liked by a lot of enthousiasts there even when their performance was seriously lagging.
Thanks! Even if this is only the enthusiast crowd, they're typically ahead of the curve, by a year or two, so Intel can't affort to ignore this. At some point the there will be less willingness from Dell/HP/Lenovo/Fujitsu to buy a subpar product.
The hatred towards intels 28x cpus from reviewers seems overblown to me. They are still pretty good chips that beat AMDs single chiplet offerings in cinebench, with decent efficiency. They're kind of just 2nd best at everything, whether thats multicore, single core, efficiency, or gaming... which to me doesn't seem bad, taken as a whole.
I think they meant that compared to specific chips.
AMDs x3D chips are exceptionally good for gaming but are relatively very poor for MT “productivity” stuff (this gen seems to be a lot better at that, though).
13/14th gen also seemingly also had somewhat better price/performance overall than AM5 chips.
I haven't seen much hatred. As you said they are just subpar in every metric, except maybe idle power consumption. Power consumption under load though is far superior on AMDs x3d.
I guess if you look long enough, you'll always find some hate, for example Userbenchmark hates on all AMD CPUs for years and is very biased.
Their latest review says the AMD CPUs are bad, cause nobody needs that much performance.
I've seen a lot of what I would call hatred. "Intel has failed!" "285k is junk!" and so on. Just a bit more harsh and sensational than I think they should be, as opposed to giving a balanced perspective. Like I said they are not the best at any specific thing, but have better efficiency than before, still beat AMD at certain tasks, good memory controllers, and so on. With the right pricing they would be easy to recommend.
With only two relevant brands of PC microprocessors, "second best" means "worst". Intel might be close to AMD, but rational reasons for choosing them appear to be reduced to socket compatibility with the CPU in someone's relatively recent old PC, which should allow an upgrade with the significant cost reduction of keeping the old motherboard and cooling system.
No, what I mean is, 285k beats the 9800x3d at multiprocessing stuff & productivity tasks, but loses to 9950x. It beats the 9950x at gaming but loses to the 9800x3d. It performs slightly worse than 14900k at gaming and some other tasks, and overall price/perf, but does its job much more efficiently. There's no single alternative thats better in every metric.
Yes, the CPUs listed are in the range where customers cannot perform expensive experiments (buying both Intel and AMD to test) and rely on review sites.
Personally, when I first got access to an Epyc, I was underwhelmed by the performance. For numerical performance it was slightly worse than M2 or cheap old Intel processors. I'm now a bit skeptical of reviews.
I listened to a review that said an i7 Thinkpad was cool and quiet with an 8-10 hour battery life. Fans scream at the slightest load, it's 45 degrees constantly and the battery life is 3 hours if you don't touch it and 1 hour if you do. And digging deeper, that's just normal for them. Serves me right for trusting a "real" reviewer.
Should have insisted on a Framework, by the sounds of it, it would actually have at least not worse battery life.
No, creating a throwaway account and making a superfluous claim creates the thread sentiment.
There might be valid comparisons for specific Epyc generations and their Xeon counterparts. But processors are a bit like car engines. Power and torque numbers doesn't tell the whole story.
Datacenter processors are optimized for different scenarios. Use the wrong processor for the wrong job, and you get abysmal performance. We have an AMD system which won't win any speed records, but that thing has enormous number of memory channels and PCIe lanes, so it's basically a semi with an extra long trailer.
Fittingly, that processor lives in a storage cluster and delivers tremendous amount of I/O both in IOPS and throughput. Same processor would look silly in a compute cluster, though.
I'd rather talk in SPEC scores if we're talking about Epyc and Xeon processors, and in TOP500 lists and national supercomputing centers, but you do you.
Remember the Pentium 4/Netburst era? They bounced back pretty well from that with the Core CPUs.
Intel is still tons more successful than 99.9% of companies out there. Specifically, they're still doing a $54 billion/year revenue, making it the ~260th largest company in the US (by revenue). This entire "if you're not leading, then you're losing" type of winner-takes-all mindset is fairly silly.
Well Intel used its massive monopoly and market domination to prevent all the biggest providers from using AMD. To a point where AMD literally couldn't give away chips for free.
In some years Dell profitability basically depended on them not selling any AMD, because if they did, they would get the massive subsidy from Intel. These company were losing market share because they couldn't sell AMD systems, but that was preferable to not getting rebates.
This lead to massive anti-trust style lawsuits in Europe, the US, Japan and other places, and basically it lead to a 10 year legal battle that only ended because AMD was simply out of money to continue to fight the legal fight. So the made a settlement to get some cash.
If intel didn't have such insane market dominance, incredibly balance sheet and high margin for all the legal battle and all the fines, AMD would have really, really hurt them in the 2000s. But they could sustain the fight long enough and when AMD made a misstep, and Intel didn't, AMD crushed.
AMD kind of saved itself with consoles sales and then managed to turn it around with Zen.
The Core were good but also AMD fumbled with bulldozer which left them floundering until the Zen era.
AMD has been doing pretty well from a strategic point of view, specifically in high volume data centre SKUs, and so it's hard to see Intel being given as wide a window this time. And there are now also other players looking to take some market share, most notably Nvidia, but also all the big players are looking at their own in house ARM silicon.
The road back for Intel is not going to be nearly as wide and smooth as it was for Core.
For processors it’s a bit different I believe. We care about cost, efficiency and raw power. If you are not leading in at least one dimension you are completely dead.
You are literally leaving money on the table if you pick a processor not leading one of these (or being very good at 2-3 of these).
Ability to source replacements. Behavioural predictability (or, more specifically, correctness). Backwards-compatibility. Side-channels, or preferably lack thereof. So-called "trusted execution environments".
Sure, Intel isn't leading in any of these either – except maybe backwards-compatibility –, but there are more characteristics we care about than just the three you listed.
The problem is not today but tomorrow where it's competitors are going to smaller and smaller nodes.
Also Tsmc keeps 50+% margin so does Amd so actual cost of processor is less than 20% of sale price. Intel foundry seems uncompetitive let alone it's current processor architecture.
Depends on how you define that, I’d bet that 99.9% of all companies aren’t losing billions of $ every quarter while their revenue is declining even more. Also while simultaneously failing in pretty much every market segment they operate in..
> Anyway, now they are a taxpayer’s problem, for the same reason that Boeing is.
And the same way as Boeing, Intel spent a fucking lot of money doing stock buybacks.
Looking back at the charts it's to the tune of US$ 14bi just in 2020, another US$ 13bi in 2019, US$ 10bi in 2018 and the trend continues into 2011, a brief respite during the 2008-2010 crisis period.
US$ 152bi has been spent by Intel since 1990 in stock buybacks but the largest share of it is from the past 10-15 years.
While that is true, Intel has probably spent more on R&D than all of their competition put together. That they still fell behind is indicative of their failure to execute, but it's not for lack of spending.
People think of Intel as just doing CPUs, but they had a massively broad impact, from scientific computing libraries to compilers, to plenty of other hardware offerings. It's not actually obvious they needed to jump onto mobile, GPU, etc, they just needed to continue to have the best processors around. They don't any more, at least not obviously.
They can 100% survive as the second-best US-based processor producer, esp if they focus and do well, but they cannot be as large as they are and be number 2. Intel has 5x the employees of AMD.
A few years ago while I was working for a large enterprise I was brought in by our CTO to understand what IoT problems Intel could help us solve.
I remember saying ‘unless it has something to do with their connectivity solutions, none, because they no longer have a viable edge compute architecture for our use cases’, but I was assured they were willing to ‘co-develop’ and I may be surprised. Indeed the end of this statement alone was surprising, because why would any company of Intels scale want to co-develop with a mid size enterprise for an internal facing use case?
I won’t go into great detail but it turns out:
- they had no new product in the pipeline, I never did figure out what there was to ‘co-develop’
- their solution was to install $1k NUCs at the edge instead of $100 RPI derivatives
- their busy bodies were willing to discuss just about anything because they evidently did not have a very firm remit, and we had our own busy bodies that also loved to pursue passion projects that had no business value
- they evidently had a lot of these ‘solutions’ teams that were kind of roaming around to… do something… with large companies, that was not really related to selling hardware or services
It just struck me as incredibly odd that a notoriously stubborn company that was so content to focus on ultra high margin products was wasting so many resources talking to customers about point solutions that would never scale into a meaningful business.
I ended up giving the CTO my grim assessment on the initiative and going to focus on other things. A part of me thought maybe I was just missing something and the teams would come back with something useful, but I never heard about any of it again. Good riddance.
Yeah. Process leadership was always Intel’s secret sauce. They could put out mediocre designs tied to the x86 albatross as long as they were consistently ahead of everyone else on process.
When that stopped, they were left with mediocre designs and the albatross, but several process nodes behind others on TSMC. The hidden handicaps became apparent and made the problem worse.
I'm not sure this perception is correct: Intel had diverse products from the earliest days with own series of ROM, SRAM and so on. They never were a cpu boutique.
The last couple of generations of desktop Intel chips are burning out at stock settings. The current leading edge is both faster and cheaper on AMD, despite being on an older TSMC process. 90% market share to AMD in the DIY/component market in that context seems low.
I’ve had trouble verifying that these chips still have reliability problems after the microcode updates. There seem to be a lot of anecdotes, but pc manufacturers are reporting normal rates of warranty return. It’s possible long term reliability is worse but it’s pretty easy to stress test a CPU.
Anyway, would love to hear if anyone has some decent quality data about these chips reliability.
Parent stated "I’ve had trouble verifying that these chips still have reliability problems after the microcode updates"
Your video links to Intel CPUs doing remarkably poorly before the latest microcode update. Parent was presumably trying to determine if Intel has actually fixed (or at least mitigated) the problem, which your video (much as I love Gamers Nexus) doesn't really have an answer for.
I don't follow PC hardware at all but recently I'm looking into replacing my 10 year old gaming PC and would normally just throw in an upper-mid range Intel CPU and be done with it
But with the recent failures of intel CPUs it's had me looking into AMDs to find the equivalent. I don't want to risk my investment not lasting because it just burns out over time.
(Also I like the fact that each generation doesn't have a new socket unlike intel)
AMD had a similar problem which was caused by some motherboards pushing the SoC voltage too high (because it provides better overclocking). If you want to be careful (and it looks like you are), stay away from the latest hardware and avoid overclocking (or learn the theory and configure everything yourself -- if you do any automatic overclocking, voltages set by most motherboards are still higher than what I'm personally comfortable with; usually you can push them way down without affecting stability, but it takes hours of trial and error).
Intels 12000-14000 chips were ironically only having problems when cooled well, because before the latest microcode they would ramp up the voltage too much with sufficient cooling. When run ragged at high temps they were reliable even before the updates.
I got a bit confused by the X3D CPUs as many places listed X3D is the equivalent to the largest i5. But on the flip side the latest X3D is a few years old so I didn't know why it's better than the more recent version of their non-X3D CPUs
The latest X3D cpu was just released, as part of the 9000 series. Though it's probably out of stock because it's currently just about the best processor for cache heavy workloads, especially gaming.
The X3D chips have way larger caches so they punch well above their weight on many workloads that are traditionally cache bound. And with the 9000 series they figured out a better assembly method so they work even better than the 7000 or 5000 series ones.
Thanks for the info. So if I wait a few months for stock to replenish it sounds like the 9000 series X3D may be a good option for a future-proof gaming PC
I recently had a problem that I thought might be the CPU, mobo, or RAM, and was looking at upgrading. For the same reasons as you, I also decided I'd get an AMD this time. The only real choice seemed to be the 7800X3D for a gamer. (Though now the 9800X3D is out, so that's probably the better choice.)
Luckily, after my computer locked up and rebooted enough times, I realized it always happened while saving the game. I replaced the C: drive's NVME and that fixed the problem, so that upgrade can be put off again for a while.
Well, stats are dubious as it's from a specific shop, but in general, Intel has to be punished for their marketing practices of the last several decades based on market dominance.
Every generation of processor requiring unique slot, remote backdoor integrated in motherboards and multiple vulnerabilities found in recent years is what forced me to switch to AMD several years ago, and my AM4 socket is supported up to this day, even new versions of processors are still provided.
You are assuming a lot of things that aren’t relevant at all in this context.
Mindfactory is big in the PC hardware space in Germany, but they are not the only player. The sales from other large players like Alternate can also not be ignored and even Amazon sales are quite significant for PC hardware.
Then there is the elephant in the room: those numbers only account for DIY systems. They don’t account for OEM desktop systems or workstations.
This comment is very misplaced. I'm Dutch and I wrote my comment not because the retailer is German, but because it are desktop CPU sales figures from a single retailer whose customers seem to be primarily gamers.
Just like in 2016-2019? Stefanik, the opposite of an isolationist (to put it mildly) is scheduled for the UN ambassador job. Rubio, a neocon, is floated as Secretary of State.
People need to understand that talk is just talk. Watch the actions.
Long term I wonder how healthy x86 is as a whole. It seems like only a matter of time before someone puts out risc-v aws/gcp/azure instances that compete, or puts out a laptop part on a chromebook that kills it.
The only thing holding back the wave is windows... and microsoft themselves seem somewhat open to the idea of not using x86 anymore.
Oracle is already giving tons of ARM processors (and RAM for that matter) in their free tier, and the performance is not bad for CRUD applications, docker, and a generalist dark server (i.e. a utility box for a sysadmin).
However, for the top tier of computation, vectorization, memory bandwidth and PCIe lanes are still important.
My enterprise Intel based workstation from my job: loud, hot, slow, heavy, inefficient, loaded with security software, malware analyzers, phone home reporting, etc. Horrible experience.
My personal AMD laptop: thin, light, quick, long battery life. Pleasant experience.
Must be in datacenter use, since they've still got a 70% majority [0] in home use and it's barely dropping at all. Maybe gamers are some specific outlier that prefers Intel specifically, but all I hear them talking about are the X3D processors so it doesn't make any sense.
I think lots of people still buy Intel because AMD now charges a notable premium for their understandably vastly better hardware.
If a PC lasts 40 months (might be more, might be less) and new CPUs have been out 2 months, assuming no growth of the market only 5% of PCs have been replaced in that period and therefore the impact on something like the Steam survey cannot exceed an increase/decrease of 5% compared to the previous aggregate results. Arithmetic is your friend!
Well Ryzen (which is when AMD became a better choice) launched about 7 years ago, so AMD should've replaced most Intel CPUs since then given that it's like 84 months if that figure of yours holds, but the market share curve has basically been flat.
Why not? It's a monthly survey, it should be impeccably up to date with what people are using.
It's also not accounting for things like the absolute flood of N100 mini pcs (since they're typically not gaming rigs) that sell for barely anything compared to the cheapest AMD offer. Granted they also suck, but Intel seems to have really invested themselves into the Chinese style quantity over quality, selling lots and lots of cheap crap which will influence any sort of percentage survey massively.
Annual sales are a fraction of the hardware that's already out there. If you dig into the Steam HW surveys you will find positively ancient hardware still being used. Most people don't replace their systems more than one or twice a decade.
Iirc, the Nvidia 1000 series gpus were still the most common ones according to the stream survey even when the 4000 series was being released. And when Ryzen was first released, there was still a sizeable fraction of steam users on core2duo processors.
Well that's true, but even so the delta should indicate what people are buying right this moment. If we take the 132M Steam users and have each one buy something every 5 years, that's 2.2M each month or 1.67%. The 5 year figure is probably on the high end so let's maybe say 1.5%.
The delta this month was +1.2% for AMD, so would that imply that 80% of people bought AMD, and 20% Intel? In August AMD seems to have actually had negative delta, maybe from people selling the effed up 13/14th gen for cheap? Idk.
I'm in programming, not hardware, but people who's business is selling PC's told me that the main reason of continuing Intel dominance in long term manufacturers contracts, so laptops and prebuilt PC's are still dominantly based on Intel hardware, and that's what most people buy.
I’m very curious as to what the American likely pending increased scheme of tariffs will do for intel. It seems like they could plausibly be the vastly cheaper option by this time next year for US residents.
Intel also outsources to TSMC, I believe for most consumer chips, so not really very different from AMD. And I guess the trump tariffs would be mainly aimed at China not to CPUs made in Taiwan.
You may be referring to Intel future plans to use their own factories again with 18A, but I had the impression from the tech press that 18A volume will not be so great until 2026 (assuming things go well).
Besides that, Intel is not known for selling chips cheaply, so I would expect them to also increase prices if tariffs force their competitors to do so.
i've read extensively on trump and watched a lot of interviews with him. he's hard to pin down, his speaking style is... interesting from what i gather he uses 'tariffs' like threats. but does not want to tariff actually.
two guys are bargaining for who gets to use a public fairgrounds. it's trump and some European guy. the fairgrounds are public, there should be no charge. trump will say "i will kick you in the foot every time you use these fairgrounds when i want them". the European will object, "that's not fair! that's bad business! it's a public fairground we all agree it's fair use". trump will reiterate that he does not care, he will kick you. the European can either agree or trump will stomp the foot.
if trump agrees with all the economists that tariffs are bad business and never implements them, he'll never get a deal. he has to credibly threaten tariffs to have bargaining power. he's very convincing doing this
personally, i believe the the amount of tariffs will be smaller than most people expect. but i wouldn't be surprised if i'm wrong.
Well, yes, given Intel's 13th and 14th generation processors self-destruct if you look at them funny, it's not surprising. I am waiting for the Zen5 Threadripper to replace my Xeon.
