sbc-reviews

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

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 quite in the same depth, but there are some more general benchmarks across all cores and latencies here: https://github.com/geerlingguy/sbc-reviews/issues/92

But it's not really a good board, sadly: https://github.com/geerlingguy/sbc-reviews/issues/62#issueco... https://github.com/System64fumo/linux/blob/main/hardware/dev...

it has a few issues, I think jeffgeerling sums it up fairly well. https://github.com/geerlingguy/sbc-reviews/issues/82 "Like the Pi 4, I think this system is the first RISC-V desktop environment that isn't painful to use, just inconvenient. Actions still have delays, but the delays are more reasonable, and don't make me constantly question if the computer's frozen." also some really odd choices by Eswin for the eic7702x, which is essentially 2 p550 chips glued together.

Nit: It's the Pi 500+ (the + was eaten up by HN's automated title sensationalism-removal, I guess) And I've posted benchmark data to my sbc-reviews repo here: https://github.com/geerlingguy/sbc-reviews/issues/81 Performance-wise it's pretty much the same as the Pi 5 16GB (and can be slightly faster than the regular Pi 500 depending on the task, if it benefits from faster storage or more RAM...) Since this is the first Pi with built-in NVMe (I'm not counting the Compute Module Developer Kit), I plugged in an eGPU and tested a new 15-line patch for AMD GPU drivers, which seems to support practically all modern AMD graphics cards[1]. [1] https://www.jeffgeerling.com/blog/2025/full-egpu-acceleratio...

Those are in my SBC-reviews repo: https://github.com/geerlingguy/sbc-reviews/issues/80

Of course if you're powering this off solar in some remote location then we're into the niches where an RPi 5 might make sense, but otherwise it's just down to electricity cost. How that works out in practice depends on workload though, while the N100 consumes more at idle, it can also finish workloads much faster so can potentially spend more time at idle. While the RPi 5 idles at around 4W (including a NVMe) and the N100 at around 15W, the RPi 5 uses 12W at full tilt[1], ie close to N100 idle power draw. Alternatively the N100 with 32GB of RAM can replace two or more RPi 5's in terms of performance, so in that regard might even come out ahead. [1]: https://github.com/geerlingguy/sbc-reviews/issues/21#issueco...

Also see: https://github.com/geerlingguy/sbc-reviews/issues/21#issueco...

Ah thank you! That's actually quite a bit lower than I would've expected. Especially how it lags behind the rest in write perf. Also found this geerling review [0] today (which mentions yours as well) which does list slightly higher numbers but he notes that he used actual thermal paste to get better cooling with that gaping heatsink box, which seems to have helped on at least a few fronts. If your cooling theory about the G2 holds, then the memory might be overheating on the X4. [0] https://github.com/geerlingguy/sbc-reviews/issues/48

https://github.com/geerlingguy/sbc-reviews/issues/21 Power Idle power draw (at wall): 1.8 W (See note) Maximum simulated power draw (stress-ng --matrix 0): 9.7 W During Geekbench multicore benchmark: 7.9 W During top500 HPL benchmark: 11 W (2.75 Gflops/W)

I've posted separately some test data: https://github.com/geerlingguy/sbc-reviews/issues/21

See: https://github.com/geerlingguy/sbc-reviews/issues/21 748 single / 1507 multi

I'm also posting some additional benchmarks to my GitHub sbc-reviews repo: https://github.com/geerlingguy/sbc-reviews/issues/21

And a bunch more test data: https://github.com/geerlingguy/sbc-reviews/issues/21 (includes some notes on power consumption, etc.)