7 ms·
My real world usage for a Rpi 4 is pretty close to 5W nominal, 8W peak (but I do run a zigbee/zwave and NVMe ssd). My Zero W is pretty close to 2.5W but it’s ov
by syntaxing 3y ago
My real world usage for a Rpi 4 is pretty close to 5W nominal, 8W peak (but I do run a zigbee/zwave and NVMe ssd). My Zero W is pretty close to 2.5W but it’s overclocked. I’m pretty sure you can underclock the CPU and RAM, and under voltage the RAM, and you’ll get very close numbers to a Rpi. Has to be a N100 though, didn’t look into but a N95 is significantly more inefficient
- megous 3y agoThen Rpi4 is very inefficient even compare to other ARM boards. RK3399 based boards idle at 1.7W or ~1.2W with a binary blob TF-A firmware and suspend to 200mW. My Allwinner H3 based boards idle at similar power levels and I haven't measured suspend to RAM. On A64, suspend is ~60mW and is fast so it can be used regularly and you can wakeup the board in response to work being needed.