You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I've seen others report this as well (always in a scrypt context, if that matters). Just tested it with someone in chat as well:
[20:15] Lowering + gives you hardware errors? That seems odd.
[20:16] Removing it caused that. Running +.01 now with no problem.
[20:16] Maybe your node was trying to give you diff 0 work, I've heard people complain about errors related to that before.
[20:17] Idk. Diff now is 655, no errors or rejects.
[20:19] Remember scrypt numbers are normally 2^16 times smaller than sha256 numbers. So if you want diff 1 work to report you'd use +.000016
[20:23] qu, if you have a sec, would you test if +0 gives you HWs? If it does, see what happens with +.00001
[20:23] Sure.
[20:24] Thousands of HW errors immediately, with +0.
[20:26] Tens of HW errors every second with .00001, on all 4 GPUs.
[20:26] But .000016 should work fine. Final test?
[20:28] Well yes, a few Rejects but that may be startup angst. No HW errors.
[20:28] Ok cool thank you. Nice to confirm that.
[20:28] .0000154xxx is the scrypt equivilent for 1 diff
[20:29] I think your HWs are from trying to target "below 1".
I'd like to suggest pseudoshare target having a floor difficulty of 1/DUMB_SCRYPT_DIFF. (Perhaps ceil to the nearest .0000000x to be safe that it never rounds down below "1 diff".)
The text was updated successfully, but these errors were encountered:
Freshly booted run_p2pool and point my miners to it. They will get work of xx/0 diffs and will get tons and tons of HW Error until diff increase to xx/1 xx/2. The HW error rate slow down but sometime they will not stop, unless I restart the miner by doing S, C , Y,
For luck some time I'll mine at my fail over node for a few hour and come back to my local node. By that time my local node been idle and diff been reset to xx/0 diffs again and give me HW errors, some time that cause my WU rate to 0.00 all day and nothing get done.
Im having this problem too with since in updated my node to the most current version. Althought when using an asic with 28mhs it works fine but the gpu miners still only get hw errors
I've seen others report this as well (always in a scrypt context, if that matters). Just tested it with someone in chat as well:
[20:15] Lowering + gives you hardware errors? That seems odd.
[20:16] Removing it caused that. Running +.01 now with no problem.
[20:16] Maybe your node was trying to give you diff 0 work, I've heard people complain about errors related to that before.
[20:17] Idk. Diff now is 655, no errors or rejects.
[20:19] Remember scrypt numbers are normally 2^16 times smaller than sha256 numbers. So if you want diff 1 work to report you'd use +.000016
[20:23] qu, if you have a sec, would you test if +0 gives you HWs? If it does, see what happens with +.00001
[20:23] Sure.
[20:24] Thousands of HW errors immediately, with +0.
[20:26] Tens of HW errors every second with .00001, on all 4 GPUs.
[20:26] But .000016 should work fine. Final test?
[20:28] Well yes, a few Rejects but that may be startup angst. No HW errors.
[20:28] Ok cool thank you. Nice to confirm that.
[20:28] .0000154xxx is the scrypt equivilent for 1 diff
[20:29] I think your HWs are from trying to target "below 1".
I'd like to suggest pseudoshare target having a floor difficulty of 1/DUMB_SCRYPT_DIFF. (Perhaps ceil to the nearest .0000000x to be safe that it never rounds down below "1 diff".)
The text was updated successfully, but these errors were encountered: