You are viewing a single comment's thread:
RE: PowerUp 💪 November 2026 - 👻 Halo WIN? 😇
Well, now that you've laid down the specs, I think we will have to go up in rankings to afford something like that... With the recent memory pumps, I suppose my machine from 2 years ago is still a "good deal" 😂
Just checked the specs:
CPU: Intel Xeon E3-1270v6
Memory: 32 GB DDR4
Storage: 2x450 GB SSD NVMe
Double tripling the space, and the rest maybe could serve... 😃
0.00003549 BEE
That is very capable for an API node right now (previously was not because you would need more memory).
But now the shared file for the HiveD is so small, that the rest of the memory can easily be used for the PostgreSQL performance and ZFS cache.
Might struggle a bit on the replay, because of the only 8 threads, but it will be able to finish. The heavier part of the replay (populating the DB and creating indexes) is IO bound, so if you SSDs can take a lot of IO/s, that amount of CPU will be fine for serving afterwards. It will flat out during replay, but depending on how much IO/s you can squeeze from ZFS it might handle the job with some time.
You won't be able to serve A ton of users, but will be able to do easy 1 Million calls per day I reckon. And if you are looking like me, for a personal super fast API node, then I can tell you, its MEGA amazing!
That DDR4 is gold! :D LOL I have to convert my 16 core AMD fast 64 DDR4 into a slower 128GB DDR4 mobo :D with some DIMMS I have left. But before that, I have to convince wife to upgrade my personal machine PC into a DDR5 recent CPU thing... and I am not sure if the plafond will be approved ;P
Note, when you are replaying, don't expose things yet, test internally that you see the swagger interface and such... then let me know when you are there, as there are a couple changes you might need to do, depending on how you setup the access to external.