mcdull

Members
  • Content Count

    112
  • Joined

  • Last visited

Community Reputation

0 Neutral

About mcdull

  • Rank
    Advanced Member

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. also try 3615 if you are in 3617.
  2. My setting. E3-1275 ESXi, CPU=4 Cores per Socket = 4: Sockets:1
  3. It happens again. It runs fine for a week and getting slower and slower than no response.
  4. My box loaded with 3615 image and 1.03b loader under ESXi 6.7. It starts and runs fine for days, but it will suddenly disconnect from network. And it will consume quite some GHz of the host CPU. Force power cycle the VM will make it run fine for another few days. And I have no idea on what caused it.
  5. Yup. try different firmware for your card. And also try different virtual controller. Most should work with SATA.
  6. mcdull

    DSM 6.2 Loader

    My DSM is getting into trouble. It is setup on top of ESXi 6.7. And it recently stopped working after a few days. Ping cannot reach the host but the VM keep using 3.0Ghz of CPU. I have no idea if it is some malware that infact the device but I am not quite sure on where should I check. Serial log has nothing special. The DSM is 6.2.1 with latest update. Any help will be appreciated.
  7. mcdull

    DSM 6.2 Loader

    That's bad. So I am stopped at 3615xs. Hope it can still be supported.
  8. mcdull

    DSM 6.2 Loader

    Is E3 1275v2 a 4th Gen CPU?
  9. I have checked that 3615xs is also a dual core cpu with HT. So enabling HT may already solve your problem. please see cat /proc/cpuinfo if you got 4 cpu cores picked up.
  10. One possible solution is to put everything into VM, your CPU is 2 cores 2 HT, so it could be set as 4 vcpu under ESXi. If you have HT disabled, please enable it first on your baremetal.
  11. Woo... I got it solved.!!!!! thanks for @Designator to tell me that he got the same issue while he got a new CPU. Then I looked into others set up again and checked the log. /run/synophoto/concept.socket] construct failed: connect: Connection refused This line drew my attention and I believe it is all related to the setting of hypervisors. It seems that the software tried to connected to something that failed.. which is in terms of sockets. It turns out to be the "NUMBER OF CORES" has to match with the identified CPU. I always assign 2 cores in previous setting. Now, with 4 cores, subjects now popping out real fast.!!!!!!!!
  12. If that is the case, it is not a CPU generation problem. And how you set up your box? using Esxi or bare metal?
  13. My Bad. The latest version still able to recognize faces. Just that it will not show up until finish the entire database. Subject recognition is still the same as before, only Child being recognized. Photo assistant never returns a result, but it will suggest similar pictures.
  14. I cannot edit the previous reply. I found some resources that documented the similar technology being used. http://hyd23.rssing.com/chan-12487997/all_p164.html
  15. I have a little bit more detail. In moments version 1.1.2-0511, everything was fine. Face, subject, place, etc. In moments version from 1.2.0-0638, facial recognization and place works fine. In moments version 1.3.0-0691, only place works. I looked in the package file, and suspect the following change may account for the difference. in 1.3.0 Under package\models\concept\detector, there are 5 files. cpu_model.bin cpu_model.xml gpu_model.bin gpu_model.xml model.pb cpu_model.??? were introduced in 1.2.0-0638 (so may explain the break of subject recognition) gpu_model.??? were introduced in 1.3.0-0691 (or 1-2 version earlier that may explain the break of facial detection) My xpenology box is builted under ESXi with E3-1275v2 (sandy bridge), thus, there is NO intel GPU visible to the DSM. However, I have no clue on the xml. Hope there are experts that can help to force software calculation. cpu_model.xml gpu_model.xml