PassMark Sees the First Yearly Drop In Average CPU Performance In Its 20 Years (tomshardware.com) 54
For the first time since 2004, PassMark's global CPU benchmark data shows a decline in average processor performance, with laptop CPUs dropping 3.4% and desktop CPUs falling 0.5% year-over-year. Tom's Hardware reports: We see the biggest drop in laptop CPU performance results. PassMark recorded an average result of 14,632 across 101,316 samples last year. But, in 2025, the average score sat at an average of 14,130 points between 25,541 samples, decreasing the average score by 3.4%. The average desktop PC result in 2024 netted 26,436 points for 186,053 samples. But for 2025, the average score currently sits at 26,311 points for over 47,810 samples -- a 0.5% drop from last year. While that drop is small, we should only see a continued progression of faster performance.
[...] Passmark itself mused on X (formerly Twitter) that it could be that people are switching to more affordable machines that deliver lower power and performance. Or maybe Windows 11 is depressing performance scores versus Windows 10, especially as people transition to it with the upcoming demise of the latter. We've certainly seen plenty of examples of reduced performance in gaming with some of the newer versions of Windows 11, particularly as Intel and AMD struggled to upstream needed updates into the OS. [...] PassMark also muses that bloatware could contribute to the sudden decline in performance, but that seems like a longshot.
[...] Passmark itself mused on X (formerly Twitter) that it could be that people are switching to more affordable machines that deliver lower power and performance. Or maybe Windows 11 is depressing performance scores versus Windows 10, especially as people transition to it with the upcoming demise of the latter. We've certainly seen plenty of examples of reduced performance in gaming with some of the newer versions of Windows 11, particularly as Intel and AMD struggled to upstream needed updates into the OS. [...] PassMark also muses that bloatware could contribute to the sudden decline in performance, but that seems like a longshot.
Reasons (Score:5, Insightful)
Generally, computers have been "fast enough" for quite a while. Especially true for me, since everything runs Linux. So focus has been more on saving energy and being efficient. This is true especially on portable devices- to extend battery life and make machines smaller and lighter. But also true on desktops to cut heat and wasted energy. Even servers are affected somewhat (although there is usually always going to be demand for more and more performance on those).
Re:Reasons (Score:4, Informative)
Re:Reasons (Score:4, Interesting)
Four CPU threads and 8GB RAM
Still running Devuan Linux fine on my 2008 Dell with Core 2 Duo and 4GB RAM. I only had slowdowns when using Firefox with several standard tabs open because it has become even more of a resource hog.
But, the aging hardware is coming with small issues so now I'm thinking of buying a used Thinkpad T15, and an upgrade from 8 to 16 or maybe even splurge on 32 GB RAM while I'm at it. I could eliminate just about all swapping with zswap then.
Re:Reasons (Score:4, Informative)
Re: (Score:3)
Quite a change from back on my earlier projects involving a Microvax II - it used 8 hours of CPU time every 24 hours and still served multiple users satisfactory. All it came down to was how to fine tune priorities and run background jobs with the heavy loads.
Amazing what you could do with that and only 64MB of RAM.
Re: (Score:2)
While it no doubt will see you out the judgement of average performance over time is not a good judgement of what capabilities you need in general. My CPU spends most of its life doing nothing as well. Until it is doing something, and when I do that something it takes a LONG TIME which is why a fast CPU was needed in the first place.
Analogy (and I just looked this up) my car's average speed over the past 500km has been 32km/h. That doesn't mean I am okay owning a car that can't do 130km/h even if I only dri
Re:Reasons (Score:4, Insightful)
I think computers have been "way faster than anyone needs" for quite a while, now.
New CPUs are all simply fucking amazing.
I recently counseled a friend into getting a bargain bin CPU after demonstrating that no game known to man would be affected unless he turned off VSYNC and tried to test how many hundreds of frames per second he could get. His first impulse was to get an i9, to which I said, "why?"
There are of course plenty of things to do where a person might want a faster CPU, but the reality has been for a long time- that people (the average) don't really need them.
The only nit I'll pick, is that efficiency and performance aren't necessarily linked. They can be, when the market achieves that performance by pumping clocks and other thing that reduce the overall efficiency of the part, but my M4 Max has the highest performance CPU core you can buy, and is also highly efficient.
Re:Reasons (Score:4, Interesting)
There are times when more CPU is needed. For example, for enterprise applications required for essential security for business like Crowdstrike and other MDRs, you need at least 1-2 cores dedicated to those, otherwise one winds up with a bottleneck. This doesn't mean you have to use i9s, but i7 is probably the best court, so users can have enough single core performance and multiple cores to do items, not to mention keeping enough headroom for future AI apps (MS Recall, etc.) RAM, similar. Those Chrome tabs are not getting any smaller.
The bottleneck these days after RAM seems to be GPU, and after that disk I/O (ironically, Optane would have made a perfect solution for this... RIP.)
Re:Reasons (Score:5, Informative)
As you mention, RAM is the bigger bottleneck.
We found long ago that 8GB machines could have all the CPU in the world, their bottleneck is going to be the speed of their disk.
GPU- definitely not a bottleneck.
Re:Reasons (Score:5, Interesting)
Re: Reasons (Score:3)
I don't want my computer to only be fast at gaming. I want it to be fast at other things as well. I doubled my cores recently to that end. Gaming indeed did not benefit, but other tasks (like compilation) did - 1.8 times the performance on the phoronix kernel benchmark for example. The system is palpably faster in general and of course I can also now run more virtual machines.
For someone who only games, six to eight cores are plenty.
I could have had more cores even than the 12 went to from 6, but not only a
Re: Reasons (Score:4, Insightful)
Compilation is generally good at doing that, but really the things I need to compile on a regular basis are small enough that if you blink, you'll miss the difference between 4 and 8 cores.
Intel was my choice for a long time for that reason. Then I moved to Macs for the M1, and am now on an M4.
AMD, for its part, long the underdog in per-core performance, really kicked up its game with Zen4. Whenever I'm in the market for an x86 again (PC laptop is getting a bit long in the tooth), I'll be giving them their due.
The point of all this is- even with an i3 or a low end AMD laptop Zen4 APU, the single core performance is really just phenomenal. So as long as you've got a few cores for base workload and one left over for an interactive task, the machine tends to feel snappy as hell. CPUs really have outpaced the base software load these days- even at the budget end. That didn't used to be true.
Re: (Score:3)
Re: (Score:2)
The article I read yesterday ended with the authors wondering if Win 11 was the cause as all the bench marking was on Windows and the percentage on Win 11 had increased. Makes sense to me as Windows always has got slower as hardware got faster.
Re:Reasons (Score:4, Informative)
Generally, computers have been "fast enough" for quite a while. Especially true for me, since everything runs Linux.
This has limits though, especially on Linux desktop/laptop devices where a lot takes place in the browser. Modern browsers use OODLES of memory and CPU time to do things that would have taken a fraction o the resources with a native application.
I'm seeing things like 3-10% idle CPU usage and 500+MB of RAM used by individual browser tabs for "professional" web applications such as Confluence, Office365 (Outlook, Excel and Powerpoint mostly) or Teams. I'm running out of RAM just using work apps in a browser - on a Linux device that has 16 gigs of RAM total. If I need to fire up a VM locally at any point I'm screwed.
Luckily I also have access to devices with 64+GB of RAM for work, but who wants to carry around a 16" workstation with a 200+W power brick? Running all the heavy stuff remotely is pretty annoying when it comes to working collaboratively - not being able to screen-share individual windows from the remote machine is problematic, for instance.
Re: (Score:2)
The times I have been most wanting more performance lately is when I install the cumulative OS patches. Especially after an OS reinstall.
Takes less than an hour to install the OS, takes the rest of the day to install the cumulative OS patch.
Re: (Score:2)
Is this Windows you're talking about? Even my many-cored workstation with oodles of RAM and SSDs takes ages to install Windows updates. You can throw as much hardware at that as you want and it'll still take forever.
Re: (Score:3)
Definitely Windows. Only OS I have experienced that was worse was AIX in the beginning of the 90's.
It's only to run bloat (Score:5, Insightful)
A lot of games and applications run fine. The only reason CPU hardware needed to increase as much as it has is simply that Windows and other applications started collecting more and more telemetry.
Excel even is a beast these days, but it's still an effing spreadsheet. You still cells and values in it. If you're not running mega spreadsheets, back in the day what you were using took 20 MB of ram, now just launching the program takes 100 MB. 5X as much, and that's still a relatively light weight application.
Browsers, what a nightmare of data collection. Early 2000s, watch videos, comment on forums. 2025. Watch videos, comment on forums. few hundred megabytes to gigabytes of ram, way more processing power.
And why? Exceptions are high def video and more, but for regular surfing? You're kidding me.
Re:It's only to run bloat (Score:4, Informative)
Re: (Score:2)
Re: (Score:2)
As for Office, I still use 2007 SR3. I don't need the newer bloated features that use Internet. I barely use Office too like I used to do at work and school. I also have updated LibreOffice as a back up.
Re: (Score:3)
The only reason CPU hardware needed to increase as much as it has is simply that Windows and other applications started collecting more and more telemetry.
Nice conspiracy theory, but the reality is that no, telemetry doesn't require any meaningful CPU hardware in the slightest. That is well and truly reflected in people's CPU use and nothing to do with your examples. Let's have a look shall we?
Excel even is a beast these days, but it's still an effing spreadsheet.
What? Excel is a spreadsheet? I thought it was a python Interpreter. I thought it was a fully Turing complete system. I thought it was a database connection tool, a front end for complex scripted macros, a group ware platform which interfaces with multiple other apps an
Re: (Score:2)
Nice conspiracy theory, but the reality is that no, telemetry doesn't require any meaningful CPU hardware in the slightest. That is well and truly reflected in people's CPU use and nothing to do with your examples. Let's have a look shall we?
Okay data collection paid shrill. Explain to me what benefits I have as an end user performing the same browsing tasks for something like Windows 7 and Windows 10 or 11. Explain to me what features are there that justify the increase in resource.
Explain to me why I can render websites and use an OS like Linux, say Mint or Ubuntu to keep it simple and it takes a fraction of these resources? Same web pages.
Also, they DO use a lot of cpu, we just have a lot more cores these days. If you knew how to check perfo
Re: (Score:2)
At no point did I say data collection doesn't happen. At no point did I say browsers don't hoover up everything about you. The only thing I refuted is your absurd notion that this is what causes high resource requirements in modern applications. And no your Linux browser that comparably follows the same web and security standards do not use any less resources than the hoovering Chrome.
If I'm shilling for data collection, you're shilling for illiteracy. Try and follow the conversation.
Re: It's only to run bloat (Score:2)
"telemetry doesn't require any meaningful CPU hardware in the slightest"
It may not require it, but if Microsoft have coded it, it probably uses all the resources it can grab. They probably dropped Minesweeper because they couldn't get it to run in less than 2Pb ram.
Re: (Score:2)
CPU is not a resource, it's unit which does work. You can't grab it, you can only give it work to do.
RAM is a resource. It makes sense to grab all resources you can - it speeds up execution of tasks. You can't grab CPU.
Re: (Score:2)
The way Windows reports memory usage changed between Windows XP and Windows 7. I don't know the specifics, but I believe they now count memory shared between processes in each individual process, whereas before it only reported unique memory. If you ran the same versions of software on Windows 7 and Windows XP, you'd see much higher memory usage reported on Windows 7. My explanation probably isn't quite right, but there was a very clear change in the reporting. That's some of the usage, but not all of it.
An
The newer CPUs (Score:3, Interesting)
Re:The newer CPUs (Score:5, Informative)
It's crazy that I now actively avoid pressing the start icon on Windows because it will take 5 seconds to settle down from it's "dynamic" loading of shit and shit ads.
99.9% of the time I open it I want to simply launch an application that is already installed. I don't want to search the internet, that's what my open browser with 100+ tabs is for. I don't want suggestions for software to install, I already have it installed. I don't need sports news I never asked for, the weather, which is on my phone, or anything else. I simply want to be able to easily find my applications and start them.
Re: (Score:2)
Finally... (Score:1)
Maybe, finally, we are that the point where crummy bloatware developers are kicked to the curb and their pleas for mercy ignored, because users demand results now on the hardware they have.
New Laptop (Score:4, Insightful)
Sir, do you want the ten-hour or the sixteen-hour model?
Sixteen!
It could be up to 3% slower.
Don't care.
Re: (Score:2)
Re: (Score:2)
Apple people) also value making the laptop yet another 0.5mm thinner forever. Smaller battery
You mean like the latest Macbook Pro and its 24-hour battery life?
Thin and light is a clear value proposition for a laptop. RAM and storage aren't going to add much weight or bulk nowadays. I'd rather go with a slightly bigger screen and better cooling, but passive cooling helps with reliability and battery life, not only noise. Apple's experience optimizing phone CPUs led to really power-efficient designs, so they don't need the kind of TDP that a typical x86 CPU has.
Re: (Score:2)
I suppose my post did have just a tinge of derision since I personally think laptops have become needlessly slim and I'd rather have built-in ports and the extra speed at times. But I completely accept other peoples' different priorities.
Re: (Score:2)
The M1 MacBook Air is thin enough as it is. It's also fast enough, and the battery life is good enough. There isn't any point to faster unless you are doing 4K video. As soon as you press a key or click a button it's done. What more do you need? You used to click on something and have time to reach for the coffee cup while the PC was busy. Not anymore.
Memory and storage are the limiting factors now, and on a desktop USB ports.
Re: (Score:2)
Apple is going all out on performance and battery life, and sacrificing upgradability and the ability to fix things yourself to get it.
They integrated the RAM into the CPU because it allows them to run the RAM at significantly faster speeds than they could if it was on the motherboard. That was a huge performance win.
They have gone all out on maximizing the battery and that's a key part of their marketing. The approach they chose was to go with a LOT of small batteries instead of one big one. They put small
All those speculative execution patches, probably (Score:5, Interesting)
Seems like every month or so there is a new patch for a supposedly dire speculative execution vulnerability (that in reality, is not a concerning issue for 99.9% of users). These patches really do nothing but hurt performance in the name of protection, and it's probably a big part of why we're seeing this.
Re: (Score:3)
Erm no. In fact in most cases the patches are optional and not rolled out. Speculative execution has had virtually no impact on anyone and in most cases people's systems aren't actually patched against it.
On the flip side I just bought a laptop where I prioritised battery life over completely unneeded performance. I suspect that is an incredibly common use case and far more likely to be the reason. People stopped chasing faster CPUs a long time ago, more recently the industry has started chasing lower power
Re: (Score:1)
Re: (Score:2)
Don't believe me? Check your processor's microcode revision.
Is this like "do your own research" for COVID vaccines? Microcode revisions tell you nothing about what the microcode does or includes. You will absolutely have a recent microcode. In fact if you own an Intel 13 series the last 6 or so microcode revisions you recevied have zero to do with security or Spectre at all and everything to do with their stability.
By the way actual performance tests on recent microcodes vs launch state for the CPU shows that all results are completely within margin of error. E.g. t
It's the number of cores (Score:3)
From the graph, the single-core performance is increasing year on year on both laptops and desktops. Yet the total performance (multithreaded) is lower. Meaning it's that people have bought laptops and desktops with fewer cores, not that people have selected the less powerful cores. But it's all speculation unless PassMark discloses more detailed data (e.g. sales of i9/R9 vs i5/R9, average number of cores, or average clock frequency)
I just don’t need it. (Score:2)
I don’t play games on a computer. I use my computer for work, and I’m a designer working entirely with 2D static graphics, so pretty much any laptop CPU on the market has more horsepower than I need. I’m doing fine with my M1 Mac. I do need 32 gigs of RAM to get through the day (because I’m lazy and don’t ever close anything) but that will last me for years until Adobe finally needs even more RAM just to keep Indesign, Illustrator, and Photoshop all running at once.
Re: I just don’t need it. (Score:2)
Even just running Lightroom classic, I have been able to utilize 64 GB of RAM on my PC. My X570 motherboard allows 128GB officially with my AMD 5950X. Doesn't seem like regular desktop chips can go much beyond that.
Win11 virtualization based security impact (Score:2)
Is more than 3%. VBS includes memory integrity checking.
I can't enable the feature to measure the impact, because of some old unsigned Firewire and MIDI device drivers, that prevent VBS from being turned on.
hdd based performance drop? (Score:3)
Re: hdd based performance drop? (Score:3)
To reduce price, most (all?) manufacturers of spinning HDDs enshittified them. Google/Bing/DDG "Shingled magnetic recording".
Re: (Score:2)
Wait, when did we stop caring about CPU speed? (Score:2)
All the comments I see are, "Nah, I don't need any faster CPU"
I guess the marketing now is that we can't double speed in X months anymore, so no one should care.
When did that stop happening, BTW? When did we fall off that curve?
Lol. Oops, another exponential that's bit the dust, like all the rest.
At least this has a "soft landing" in that most compute intensive tasks appear to be done with GPUs now.
Will it go up after the forced W11 upgrade cycle? (Score:2)
I wonder whether the pending upgrade cycle for W11 will result in an average performance increase as a lot of people are forced to finally upgrade their ancient hardware, or whether a lot of people will just trade in for the lowest spec systems possible, depressing average scores further.
It seems to me that the data is highly suspect anyway, as the "average" user isn't running benchmarks and uploading the results to Passmark.
Not really a CPU score? (Score:2)
If the performance metrics of a CPU can be influenced by the OS (Win11 or otherwise,) then is it really a measure of the performance of the CPU, or is it a measure of the performance of the ecosystem? Those are two very different things, and I think relevant to the discussion.