4 ms·
I’d honestly would like to read bcantrill’s opinion on this
by znpy 1y ago
I’d honestly would like to read bcantrill’s opinion on this
- wmf 1y agoI'm not Bryan but this sounds like the same over-abstracted architecture astronaut approach as UEFI. It's the compete opposite of Oxide's firmware.
- bcantrill 1y agoBingo: forcing UEFI and D-Bus to breed in captivity is antithetical to our approach of holistic boot.[0][1] [0] https://rfd.shared.oxide.computer/rfd/0241 https://rfd.shared.oxide.computer/rfd/0241 [1] https://www.osfc.io/2022/talks/i-have-come-to-bury-the-bios-not-to-open-it-the-need-for-holistic-systems/ https://www.osfc.io/2022/talks/i-have-come-to-bury-the-bios-...
- ryao 1y agoGetting rid of the BIOS sounds awesome. I assume that means that it gets rid of code running the SMM, which can prevent invisible code from sapping performance from the machine for no apparent reason. For example, on recent AMD machines, such as my Zen 3 machine, memory bandwidth measured by ZFS' checksum algorithms will randomly drop by a significant percentage and there is no obvious reason why, although I suspect patrol scrubs are involved. That begs the question. What does the patrol scrubs of ECC memory on your hardware?