search

swanchain / go-computing-provider

A golang implementation of computing provider
MIT License
11 stars 15 forks source link

ThomasBlock ThomasBlock-ECP no more ECP jobs, wrong free RAM #95

Closed ThomasBlock closed 2 weeks ago

ThomasBlock commented 2 weeks ago
  1. since 48 hours there are no more ECP jobs. one reason could be that ECP only reports 6 GB RAM of 32 GB. I have no other software running. what can i fo?
  2. The Jobs i got last week are only 512 M GPU. How can i get 32G Jobs - are there special hardware requirements? What about Aleo?
/usr/local/bin/computing-provider -v
computing-provider version 0.5.1+git.a47e4a3

CP Account Info:                                                                                                                                               
   CP Account Address(2.0): 0xC39230f372Db8ABc2f2bA13d23B8206afB2b1ECC                                                                                          
   Name:                    ThomasBlock-ECP                                                                                                                     
   Owner:                   0xaC43f4ed670daaaB06391ffa0FDe48407972B4BD                                                                                          
   Node ID:                 04d03fc936bdb1f0f879b3d2b77bbe783d258090ccd4ce7ed2bc36bdeb7677c0602e1e0de5aa7d671e972871d4957518fe1656c0f1466c8f41ffa8505892c881d5

Capabilities:               
   Task Types:              Fil-C2-512M,Aleo,Fil-C2-32G

image

image

Bildschirmfoto vom 2024-06-17 10-43-17

ThomasBlock commented 2 weeks ago

update: it fixed itself.. ?

time="2024-06-17 10:51:04.277" level=info msg="collect hardware resource, freeCpu:24, freeMemory: 6 GiB, freeStorage: 226 GiB, freeGpu: map[NVIDIA 4070 SUPER:1]" func=reportClusterResourceForDocker file="ubi.go:952"
time="2024-06-17 10:57:04.278" level=info msg="collect hardware resource, freeCpu:24, freeMemory: 31.63 GiB, freeStorage: 226.51 GiB, freeGpu: map[NVIDIA 4070 SUPER:1]" func=reportClusterResourceForDocker file="ubi.go:952"
sonic-chain commented 2 weeks ago

update: it fixed itself.. ?

time="2024-06-17 10:51:04.277" level=info msg="collect hardware resource, freeCpu:24, freeMemory: 6 GiB, freeStorage: 226 GiB, freeGpu: map[NVIDIA 4070 SUPER:1]" func=reportClusterResourceForDocker file="ubi.go:952"
time="2024-06-17 10:57:04.278" level=info msg="collect hardware resource, freeCpu:24, freeMemory: 31.63 GiB, freeStorage: 226.51 GiB, freeGpu: map[NVIDIA 4070 SUPER:1]" func=reportClusterResourceForDocker file="ubi.go:952"

update: it fixed itself.. ?

time="2024-06-17 10:51:04.277" level=info msg="collect hardware resource, freeCpu:24, freeMemory: 6 GiB, freeStorage: 226 GiB, freeGpu: map[NVIDIA 4070 SUPER:1]" func=reportClusterResourceForDocker file="ubi.go:952"
time="2024-06-17 10:57:04.278" level=info msg="collect hardware resource, freeCpu:24, freeMemory: 31.63 GiB, freeStorage: 226.51 GiB, freeGpu: map[NVIDIA 4070 SUPER:1]" func=reportClusterResourceForDocker file="ubi.go:952"

The bug of resource acquisition has been fixed

sonic-chain commented 2 weeks ago

The ECP engine is being repaired because of the instability of rpc, which leads to the failure of task sending. Currently, the 32G task and aleo task are not enabled, but will be enabled in the future.

ThomasBlock commented 2 weeks ago

ok thank you. i can confirm that jobs are working now

image