Hacker News Reader

Comment by CrazedGeek | Hacker News Reader

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.

View on HN

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...

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.

The average user might not a really know what a CPU is, keep on mind.

It's also exactly where you will find absolutely nobody who makes a habit of running PassMark on their computer.

This graph more-or-less only measures the enthusiast Windows user market.

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 have an AMD laptop, and my next one will definitely be Intel.

Reason: Linux drivers. Ryzen Linux support is just that bad.

Any source on this claim? Care to elaborate with any specifics?

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.

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.

My laptop is an AMD Ryzen 5 2500U.

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.

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 don't disbelieve you, nor am I attempting to invalidate your experiences.

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.

Fine for me. (K)Ubuntu on a Thinkpad T14.

I would suspect it's a denominator effect of Apple dropping intel chips.

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 suspect that macs just aren't represented in this graph

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.

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.

They wouldn’t be running windows on the Apple Silicon macs though, just the intel ones.

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.

Apple says only between 2 and 4 % of its users use Bootcamp to run Windows.

From the notes at the bottom:

> 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.

Yes, but consider if you have three laptops:

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.

Apple says only between 2 and 4 % of its users use Bootcamp to run Windows.

and also AMD's new integrated graphics

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.

Not to AMD, though.

Reply on news.ycombinator.com