A few years ago I was working at a place that needed to do builds targeting the Jetson platform, and was somewhat allergic to tossing it into the cloud due to cost. We ran the numbers and the Altra paid for itself pretty quickly. Great thing, it ripped through our C++ builds + Docker image creation - I think we ended up with a 64 core version (don't remember through who, but we needed a server form factor). Ended up still moving our release builds to the cloud, due to some dicey internet situations but for local builds this thing was A+. I hope they're still using it.
> And the latest one, an Apple MacBook Pro, is nice and fast but has some limits — does not support 64k page size. Which I need for my work.
I wonder where this requirement comes from ...
I would guess to develop and test software that will ultimately run on a system with 64k page size.
Is there a fundamental advantage over other page sizes, other than the convenience of 64k == 2^16?
64k is the largest page size that the ARM architecture supports. The large page size provides advantages for applications which allocate large amounts of memory.
Yes there are
(as a starting point 4k is a "page size for ants" in 2025 - 4MB might be too much however)
But the bigger the page the less TLB entries you need, and less entries in your OS data structures managing memory, etc
I realize the Ampere Altra Q features the Armv8.2-a ISA. Does anybody know if there are chips with Armv8.6-a (or above) or even SVE that one can buy? I did some research but couldn't find any.
Radxa orion o6 claims to be the first arm v9 system available.
I've always wondered why there isn't a bigger market / offering for dedicated servers with Ampere at their heart (apart from Hetzner)
If anyone knows of any, let me know!
A lot of software is built and optimized for x86 and EPYC processors are really good so it is hard to get into arm, don’t think that many companies use it.
> A lot of software is built and optimized for x86 and EPYC processors are really good so it is hard to get into arm, don’t think that many companies use it.
That is just not true.
Nowadays, most OSS software and most server side software will run without any hinch on armv8.
A tremendous amount of work has been done to speed up common software on armv8, partially due to popularity of mobile as a platform but also and to the emergence of ARM servers (Graviton / Neoverses) in the major Cloud providers infrastructure.
However, it's hard to get into ARM other than using cloud offerings.
Because those cloud offerings have handled for you the problematic case of ARM generally operating as "closed platform" even when everything is open source.
On a PC server, usually you only hit any issues if you want to play with something more exotic on either software or hardware. Bog-standard linux setup is trivial to integrate.
On ARM, even though finally there's UEFI available, I recall that even few years ago there were issues with things like interrupt controller support - and that kind of reputation persists and definitely makes it harder to percolate on-prem ARM.
It also does not help that you need to go for pretty pricy systems to avoid vendor lock-in at firmware compatibility level - or had to, until quite recently.
Why is it hard to get a mac or a pi?
Pi is relatively underpowered (quite underpowered, even), has proprietary boot system, and similarly isn't exactly good with things you might want in professional server (there are some boards using compute modules that provide it as add on, but it's generally not a given). Also, I/O starved.
Mac is similarly an issue of proprietary system with no BMC support. Running one in a rack is always going to be at least partially half-baked solution. Additionally, you're heavily limited in OS support (for all that I love what Asahi has done, it does not mean you can install let's say RHEL on it, even in virtual machine - because M-series chips do not support 64kB page size which became the standard on ARM64 installs in the cloud, for example RHEL defaults to it and it was quite a pain to deal with in a company using Macbooks).
So you end up "shopping" for something that actually matches server hardware and it gets expensive and sometimes non-trivial, because ARM server market was (probably still is) not quite friendly to casually buying a rackmount server with ARM CPUs for affordable prices. Hyperscalers have completely different setups where they can easily tank the complexity costs because they can bother with customized hardware all the way to custom ASICs that provide management, I/O, and hypervisor boot and control path (like AWS Nitro).
One option is to find a VAR that actually sells ARM servers and not just appliances that happen to use ARM inside, but that's honestly a level of complexity (and pricing) above what many smaller companies want.
So if you're on a budget it's either cloud(-ish) solutions or maybe one your engineers can be spared to spend considerable amount of time to build a server from parts that will resemble something production quality.
I am running an ARM64 build of Ubuntu on my MacBook Air using Multipass. I've never had a problem due to missing support/optimisation for ARM - at least I didn't notice any. I even noticed that build times were faster on this virtualised machine than they were natively on my previous Tuxedo laptop which had an Intel i7 that was a couple of years old. Although, I blame this speed mostly on the sheer horsepower of the newest Apple chips
If you use AWS, lots of software can easily be run on Graviton, and lots of companies do that.
https://www.theregister.com/2023/08/08/amazon_arm_servers/:
“Bernstein's report estimates that Graviton represented about 20 percent of AWS CPU instances by mid-2022“
And that’s three years ago. Graviton instances are cheaper than (more or less) equivalent x86 ones on AWS, so I think it’s a safe bet that number has gone up since.
yeah if you're running a node backend, the changes are cosmetic at best (unless you're running chrome to generate pdfs or whatever). easiest 20% saved ever. if I were Intel or AMD I would be very afraid of this... years ago.
I don’t think a lot of companies realize they are using it. At three companies now, I’ve witnessed core microservices migrate to ARM seamlessly, due to engineering being direct pressure to “reduce cloud spend”. The terrifying (and amazing) bit is that moving to ARM was enough to get finance off engineering’s back in all cases.
They're slow and the arch is less compatible? Arm cores in web hosting are typically known as the shit-tier.
I think the main use case for these is some sort of Android build farm, as a CI/CD pipeline with testing of different OS versions and general app building, since they don't have to emulate arm.
Well, I've run some Ampere Altra ARM machines in my studio so I can speak to this;
A) No you can't use ARM as android build farms, as androids build tools only work on x86 (go figure).
B) Ampere Altra runs faster for throughput than x86 on the same lithography and clock frequency; I can't imagine how they'd be slower for web, it's not my experience with these machines under test. Maybe virtualisation has issues (I ran bare-metal containers - as you should).
My original intent was to use these machines as build/test clusters for our go microservices (and I'd run ARM on GCP) but GCP was a bit too slow to roll out and now we're far into feature locking any migrations of that.
So I added the machines to the general pool of compute and they run bots, internal webservices etc; with Kubernetes.
The performance is extremely good, only limited by the fact we can't use them as build machines for the game due to the architecture difference - however for storage or heavy compute they really outperform the EPYC Milan machines which are also on a 7nm lithography.
1341 PLN / 371 USD isn't "cheap" for 25% more cores. That's almost double the price.
Q64-22 on eBay (US) for $150-200 USD / 542-723 PLN.
25% more cores and 36% more clock. It amounts to paying 85% more for 70% more perf. Not too bad.
He clearly refers to that and states they did not respond.
Also, CPU was hardly the biggest cost here.
I never bought used computer parts. Are these parts generally reliable for ~2 years when bought used?
I haven’t bought new hardware since I was a teenager. Second hand is cheap and good for the environment. I never received a broken part and everything has worked reliably for me.
2-3 years is not a lot. My daily driver laptop is from 2011 and still going strong.
Sure, there are “lemons” out there, but there are also a lot of people who just replace their hardware often.
I concur. Doing this for almost all my technical equipment and mobile phones and never had a problem. For important/expensive things you can buy on refurbished stores that offer a 1-year warranty in EU.
Are you still using the same battery?
I work at an e-waste recycling company. People throwing out old but still working servers, desktops, and laptops is pretty common. Companies regularly decommission and throw out their IT assets after some number of years, even if the stuff still works (which most of it still does).
My current computer is from more than 10 years ago, and I found it in a dumpster. Works fine.
I regularly buy used hardware. It fails when it fails, same as the new stuff. Is there a higher probability? Possibly, but at the small sample sizes I'm at I can't feel the difference. Feels random either way.
Built a gaming desktop for a friend almost 2 years ago; used GPU and CPU (maybe a few others things too), everything's going great. It helps that our local Cragslist offers efficient buyer protection.
Server-side, I also bought used Xeons for an old box and recertified 10TB Exos. No issues there neither.
The HDDs are a bit of a gamble, but for anything else I can only encourage you to buy used!
> It helps that our local Cragslist offers efficient buyer protection.
Used professional hardware (servers, workstations) are made with higher quality standards so they last fairly long.
Plus if they're from a data center, they will have been in a cooled, filtered, and stable space for their lifetime, vs a desktop that may have been in a dusty room getting moved or kicked from time to time.
Also they're good for heating your home.
those are server-grade stuff, it's normal for them to work 10 years continuously.
[flagged]
Offtopic, I'm so confused why this is top1 on my HN? Just a pretty normal build?
It's not "your" HN, HN doesn't do algorithmic/per-user ranking. (Ed.: Actually a refreshing breath of wide social cohesion on a platform, IMHO. We have enough platforms that create bubbles for you.)
It's top1 on everyone's HN because a sufficient number of people (including myself) thought it a nice writeup about fat ARM systems.
I haven’t been following hardware for a while, granted, but this is the first time I see a desktop build with an arm64 cpu. Didn’t know you can just… buy one.
I guess this post proves you still can't :)
Not that much changed since this:
https://marcin.juszkiewicz.com.pl/2019/10/23/what-is-wrong-w...
For what it's worth, I've been using a Lenovo X13s for some 3 months now. It's not a desktop, and it took years for core components to be supported in mainline Linux, but I do use it as a daily driver now. The only thing that's still not working is the webcam.
Normal ARM64 80 core system with $1000 EATX motherboard? How is this typical?
EATX is a pretty standard server motherboard form factor.
It's not even a multiple CPU board...
This is indeed a pretty standard (and weak) ARM server build.
You can get the same CPU M128-30 with 128 3ghz cores for under $800 USD.
You can throw two into a Gigabyte MP72-HB0 and fit it into a full tower case easily.
That'd only cost like $3,200 USD for 256 cores.
RAM is cheap, and that board could take 16 DIMMs.
If you used 16 GB DIMM like OP that's only 256 GB of RAM, in a server, it is not that much... only one gig per core... for like $500 USD.
Maybe for a personal build this seems extravagant but it's nothing special for a server.