geerlingguy

I've used a number of mini PCs and Arm SBCs with 8 GB of RAM. Most casual use (browsing, writing, media playback) is no problem, but using "creative" software (image editors, NLEs, DAWs, etc.) quickly gets annoying.

I had three YouTube titles, this one got about 2x the watch time as the other two, so I let YT pick it.

For technical readers, I write my blog posts. I don't have to (I lose money hosting that thing and all it's mostly just free training data for Anthropic, OpenAI, et all), but I enjoy writing, so I do it anyway. If you don't like the YT game, please subscribe to the blog's RSS feed. I much prefer text over video.

I run a large number of benchmarks; the Hackaday article of course focuses on the headline findings, but you can see the rest here: https://github.com/geerlingguy/sbc-reviews/issues/111

Yeah, not sure why Dell has such a spread on US vs rest-of-world pricing here.

I know some have mentioned the $699/$599 price could also change a bit in the US come November. The internal assembly and parts seem like they'd cost Dell a little more than Apple's cost building the Neo, so maybe in the US they're selling them closer to at cost, and subsidizing it with the Windows exclusivity and price elsewhere? Not sure, but it stinks the price only applies here. It's a good deal at $699. Harder to justify at $1,000+.

No, this laptop was loaned to me for a couple weeks (it's boxed up and ready to ship back now), and I fully disclose any and all sponsorship (this was not sponsored in any way); see my policies here: https://github.com/geerlingguy/youtube#sponsorships

Further, I'm basing my opinion on a number of benchmarks and power measurements besides just HPL. I just favor that because it's one of the few that push a CPU hard, fills system memory, and can't easily be gamed/made irrelevant with new architecture.

The main point of my blog post is to say this is the first Intel chip that's even punched in the same weight class as any Apple Silicon, wrt efficiency. It quite surprised me.

Can you please post an issue on the repo? Long on my todo list has been trying to get the Mac tests more repeatable/consistent and optimized (always easier on Linux).

They still eat battery, but at least the midrange ones, not nearly as much as they used to.

With only two USB-C ports that means you get just headphones and power if you're not using a hub or USB dock.

These tests were done on the CPU cores, not on the GPU. In every test I ran the A18 chip blew past Intel's Core 5 on GPU performance.

There were a series of other benchmarks too, but the Hackaday article only highlights HPL as that was the benchmark that changed my perspective on Intel a bit. This was the first X86 chip to trade blows in efficiency (and idle power draw) with Apple's M3/M4 (I don't have any M5 to test with right now).

Yeah, the first step is they realize pushing clocks is not the most important thing; second step is to apply that to other chips they make too! I want to see more desktop and server chips getting closer to Arm efficiency.

I run a number of other benchmarks for a better comparison (linked from my blog post: https://github.com/geerlingguy/sbc-reviews/issues/111).

The Intel chip holds its own for multicore in less bursty workloads, where the fans help.

Not only that, the efficiency (FP64, HPL) is also beating all the M4/M3/M2/M1 systems I've tested besides the M4 Mac mini: https://github.com/geerlingguy/top500-benchmark#results

I think the bigger news is Intel's made a chip that isn't just pushing clocks to hit some performance number, they've actually tuned clocks for efficiency... which also leads to much better battery life.

And the fan on the system in the original article[1] also allows the Intel system to run substantially faster for longer tasks (where the Neo thermally saturates and has to stay throttled).

(The iGPU is another matter... sadly still a bit lacking.)

I'm hoping Intel makes a follow-up to the N100-series chips with just four of their newer low power E-cores.

[1] https://www.jeffgeerling.com/blog/2026/excited-for-intel-eff...

There are still big leaps, like their optical lattice clock that's sensitive to the gravitational difference caused by tidal effects on the earth's crust.

It can detect gravitational differences of a few cm!

Sadly, no major miniaturization breakthroughs since. And those CSACs still eat up a lot of power (as do good OCXOs), making them rough for battery powered wearables.

Technically it will run on Pis — someone in my YouTube comments already has it going on top of basic Pi OS 13 (disabled NetworkManager and cloud-init to get it going). Someone else on Mastodon already has it running via a custom Pi 4 UEFI image.

I'm adding more details in my blog post here: https://www.jeffgeerling.com/blog/2026/proxmox-ve-arm-offici...

Yeah I clicked through and just see hundreds of "this account limits who can view their posts" messages.

Also a RAIDZ array (or just mdadm or BTRFS) with a cache in front can easily hit a couple GB/sec even on more consumer level hardware.

Jellyfin: https://www.jeffgeerling.com/tags/jellyfin/

LLMs: https://www.jeffgeerling.com/tags/llm/

Etc... (I don't do any crypto)

Apple used to even sell Xserve RAID and SAN software specifically to cater to the higher end users who needed fast networked storage (research, broadcast, media).

They kinda hit a standstill on the networking side as they moved to the trash can Mac, and I think that had killed momentum towards any high speed IO besides the Thunderbolt ports moving forward.

I was hopeful they'd keep the Mac Pro going into the M-series era but after one version... it's gone.

Yeah the best option for cost is to buy one of the many TB4 to PCIe adapters on the market now (I've tried two), that are less than $100. They require external PSU power though, and they don't look nearly as nice on a desk.

The Mac Studio in most of my videos is my own workstation, purchased a while back.

The 4 Mac Studio cluster I tested with Exo and Thunderbolt RDMA was a loaner cluster from Apple. There's no universe in which I could justify buying that.

I always mention any equipment that's provided / loaned / etc. and the nature of my relationships with vendors—I try to also keep this document updated for transparency's sake: https://github.com/geerlingguy/youtube/blob/master/Sponsors....

Normal 4K editing workflows are fine with even 2.5 Gbps wired networking (wired to make latency like when scrubbing not-horrid)... 10 Gbps more necessary if editing ProRes RAW and doing multicam (4-8 streams simultaneously).

I'm in a never-ending pursuit of cutting time when transferring large files though — AI models is one thing, but I have 100-200 GB projects I transfer across for backup or import (using an external TB5 drive capable of 3-5 GB/sec), and if I can do that in 1 or 2 minutes per project instead of 3 or 4, that's a noticeable speedup.

Definitely overkill for 98% of my workflow though. 10 Gbps is fine.

Yeah, raw RDMA is one thing, making sure other protocols can use it is another... hopefully Apple gets there, but right now SMB with RDMA is probably a tiny, tiny niche.

I do wonder what Apple's custom-built servers are running for networking. I wish they still made Mac Pro-level chassis specifically for better IO. Thunderbolt 5 is nice, but far inferior to a PCIe x16 slot (which can hit 400 Gbps networking) nowadays.

Projects like TenFourFox and PowerFox have ported modern FireFox to late PowerPC Macs running Mac OS X. And some of those Macs are at the bleeding edge (back then) of modern device support (PCIe, WiFi, etc.) so they can also run surprisingly modern hardware and software via Linux. Getting harder with 32 bit support vanishing though.

I've had the same struggle. There are some decent LED clocks with up to 1" displays (high), but any further and you get weird colors or glaringly-bright displays.

The best options are expensive clock displays from like Time Machines, MasterClock, etc... but they are not cheap and not meant for consumers.

I've run it on a couple Macs, and it looks like one of the issues they're trying to fix is poor multi-core scaling that affected Geekbench 6 on bigger multicore systems (with 32, 64, or 128+ CPU cores). I'll hopefully get time to test on my Ampere Altra Max systems later, to see how scaling works.

Geekbench 5 is famously much better than Geekbench 6 for benchmarking these workstation-class CPUs (though many other benchmarks are better still, especially if you can grab the source code and compile them yourself).

I still like having a set of benchmarks that run across Android, iOS, Windows, macOS, Linux, and on Arm, X86, RISC-V, etc... even if imperfect, it's a point of reference to get a general feel. And the single core scores are a great representation of a 'feel' against baseline in day-to-day use.

The one feature I would love to see is WebSerial support, now that Firefox has it too.

I think it was a Maker's Muse video where I learned a ton about different 3D printable materials' shear strength when he was testing them in tiny battlebot-style robots. Some things feel unintuitive but I'm glad people take the time to test them and share their results!

I'd love to be able to add 1U or 2U of blinkenlights to my rack, even if the pattern was completely random.

Yeah, and since it seems there's some competition there, the prices are decent. You can get milled metal, 3D prints, populated PCBs... crazy what you can do now and how quickly it can be turned around, if you are willing to pay.

I wish there were more regional places like PCBWay and JLCPCB in US, EU, etc (with similar pricing) so shipping didn't require circumnavigating the globe.