What about second m.2 slot on top of a board which looks like itās used for wifi/bt card. Will it be usefull for anything else but this card, or itās a proprietary format. And what interfaces are available if user wantās to ditch the wifi/bt card ?
While the above will be maintained through EOY 2026, weāll hopefully also see support for the 5.15 LTS kernel given the problems with KVM support on current ARM64 SBC w/ kernel 5.10.x (x>7) which forced me to switch to v5.13/v5.14 as LXD wasnāt able to get VMs to start any more. (I still wonder what went wrong there.)
@hipboi
I just reserved my rk3588, can you allow access to your kernel source tree for the rk3588?
By the way, i created a PR to SDL in order to have 3D hw accel and would be interesting to know what mali dev is present and if it will be useful for this platform.
Very excited for this! The specs look really good! Will you be providing Android images for this? Also will you be sending out early samples to developers and reviewers again like you did with the Rock Pi X?
Cool, and Iād placed my $5 for 16GB already
Hopefully I can play with it much more than my Pi 4 (4GB), which is lacking here and thereā¦
And it also gives me reason(s) to upgrade my desktop SSD
When it is out, pls. make sure it would come with a reasonable heatsink, hopefully just like the āarmorā fanless one with Pi.
Great news, I canāt wait to get it as soon as they will be available!
Itās great successor for nano-pc t4 board which is best rk3399 IMHO, but with only 4GB RAM This one is far better than anything
I see that m.2 E is only 1x pcie lane, still google dual edge tpu wonāt work, but there is 6T NPU.
This is really nice news! Also your comment on cnx-software about manufacturing in the EU would make future products a lot more accessible to me
I preordered for the 16Gb. A lot of things tick the box for me, more RAM, NPU, 2.5GbE, PoE
I have a Helios64 NAS running on the RK3399 that runs a few docker containers including photoprism. Since it uses tensorflow for family photos, 4Gb on that device is a bit limited. Having an NPU on the Rock 5 would be interesting to offload the tensorflow tasks instead of thrashing the CPU.
Since it also has a 2.5GbE NIC, the Rock 5 model B and the Helios64 would be a great combination.
Would there be a way to add a mini UPS on the Rock 5? That would make a nice ecosystem.
I think there is still no G610 support under bifrost so maybe some samples to Alyssa Rosenzweig & the Mesa crew might be wise as its got a mighty GPU x10 RK3399 perf but no driver