I found a nice chart about memory bandwidth on the Crucial Micron website.A great majority of GCC commits over the past year have been for RISCV.We will. I happen to know of multiple companies working on server-grade RV chips, and at least in theory we should be seeing sample silicon this year. There's a long way to go before we start seeing anything like Apple's M-class aarch64 cores, though.There is also the Mango Pi. All are in the bottom half of the chart. It'd be fun to see a fast Risc V processor. I wonder if that will happen
As far as I can tell there's nothing terribly special about Apple M-class processors other than they use the latest (read expensive) TSMC fabrication process. This enables better power efficiency and higher clock speeds, along with lower yields.
- Image may be NSFW.
Clik here to view.
Since common x86 desktop processors have two memory channels, then total bandwidth is double the last line of that table.
Thus, a Ryzen 5800 enjoys about 50 GB/sec total memory bandwidth and a 7800 about 100. On the other hand the M2, M2 Pro, M2 Max and M2 Ultra respectively enjoy 100, 200, 400 and 800 GB/sec total memory bandwidth. The observation is that 800 GB/sec is good for a desktop processor.
What's the memory bandwidth for a Pi 5?
Statistics: Posted by ejolson — Mon Apr 01, 2024 1:28 am
