Blog

Cores are the new megahertz, at least for enterprise servers. We've gone quickly from 32, to 64, to 80, to 128, and now to 192-cores on a single CPU socket!

Amazon built Graviton 4, Google built Axiom, but if you want your own massive Arm server, Ampere's the only game in town. And fastest Arm CPU in the world is inside the box pictured above.

It has 192 custom Arm cores running at 3.2 Gigahertz, and in some benchmarks, it stays in the ring with AMD's fastest EPYC chip, the 9965 "Turin Dense", which also has 192 cores.

High-core-count servers are the cutting edge in datacenters, and they're so insane, most software doesn't even know how to handle it. btop has to go full screen on the CPU graph just to fit all the cores:

tl;dr Raspberry Pi engineers tweaked SDRAM timings and other memory settings on the Pi, resulting in a 10-20% speed boost at the default 2.4 GHz clock. I of course had to test overclocking, which got me a 32% speedup at 3.2 GHz! Changes may roll out in a firmware update for all Pi 5 and Pi 4 users soon.

My quest for the world record Geekbench 6 score on a Pi 5 continues, as a couple months ago Martin Rowan used cooling and NUMA emulation tricks to beat my then-record score.

In a win for modular, private, local IoT, I just upgraded my Home Assistant Yellow from a Raspberry Pi Compute Module 4 to a Compute Module 5 this morning, and got an instant 2x speed boost.

I first posted about the Yellow in 2022, and walked through my smart-but-private HA Yellow setup in my Studio in a video last year.

Because I was running an eMMC CM4 in the Yellow before, I ran a full backup (and downloaded it), yanked the CM4, flashed HAOS to a new NVMe SSD, and plugged that and the CM5 into my Yellow. After running a Restore (it's a handy option right on the first page that appears when you access homeassistant.local), I was up and running like there was no difference at all—just everything was a little more snappy.

The Raspberry Pi Compute Module 5 is smaller than a credit card, and I already have it gaming in 4K with an eGPU, running a Kubernetes cluster, and I even upgraded my NEC Commercial display from a CM4 to CM5, just swapping the Compute Modules!

The Compute Module 4 was hard to get for years. It launched right after the COVID supply chain crisis, leading to insane scalper pricing.

It was so useful, though, that Raspberry Pi sold every unit they made, and they're inside everything: from commercial 3D printers, to TVs, to IP KVM cards.

After a long journey getting AMD graphics cards working on the Raspberry Pi 5, we finally have a stable patch for the amdgpu Linux kernel driver, and it works on AMD RX 400, 500, 6000, and (current-generation) 7000-series GPUs.

With that, we also have stable Vulkan graphics and compute API support.

When I wrote about getting a Radeon Pro W7700 running on the Pi, I also mentioned AMD is not planning on supporting Arm with their ROCm GPU acceleration framework. At least not anytime soon.

Luckily, the Vulkan SDK can be used in its place, and in some cases even outperforms ROCm—especially on consumer cards where ROCm isn't even supported on x86!

After years of work among a bunch of people in the Pi community (special callout to Coreforge!), we finally have multiple generations of AMD graphics cards working on the Raspberry Pi 5.

We recently got Polaris-era GPUs working (like the RX460), but in the past month we've gotten 6000 and 7000-series GPUs up and running. And many parts of the driver work at full performance—well, as much as can be had on the Raspberry Pi's single PCIe Gen 3 lane (8 GT/sec)!

I've been testing tons of modern AAA games, like Doom Eternal and Crysis Remastered, and can get 10-15 fps at 4K with Ray Tracing on, or 15-20 fps at 4K. Dropping down to 1080p is not enough to overcome the Pi's CPU bottleneck—only at resolutions under 720p does the Pi's CPU and the single PCIe lane not seem to get in the way quite as much.