So schnell wie zwei RTX 2070:
> Hello! Has anyone tested rtx 3080 on a 30x384 net?
<@!356759717944754176> I just received mine today in a Dell alienware (with a 10900kf). 10s benchmark, default settings:
PS C:\Users\redoi\Documents\Lc0\263-cuda-j92-190> ./lc0 benchmark --nncache=1000000
``` _
| _ | |
|_ |_ |_| v0.26.3 built Oct 10 2020
Detected 10 core(s) and 20 thread(s) in 1 group(s).
Group 0 has 10 core(s) and 20 thread(s).
Found pb network file: C:\Users\redoi\Documents\Lc0\263-cuda-j92-190/J92-190
Creating backend [cuda-auto]...
Switching to [cuda-fp16]...
CUDA Runtime version: 11.1.0
Latest version of CUDA supported by the driver: 11.1.0
GPU: GeForce RTX 3080
GPU memory: 10 Gb
GPU clock frequency: 1710 MHz
GPU compute capability: 8.6
===========================
Total time (ms) : 340599
Nodes searched : 8601001
Nodes/second : 25252```
```Same thing, with LS 15.0:
===========================
Total time (ms) : 340246
Nodes searched : 22540115
Nodes/second : 66246```
Ich freue mich drauf, die Karte zwei meiner RTX 2070 (Super) hinzuzufügen.
Dann dürfte ich mit einem aktuellen J92-Netz über 50.000 Nodes/second erreichen.
Schöne Aussichten!
Lothar