Quote from: jonano on February 27, 2018, 08:40:46 AM

Quote from: rgsnedds on February 27, 2018, 08:11:05 AM

great work... and I know Protovist has been working to try and fix the bugs in the linux miner so although we are still early days, I see a bright future for XDAG.
Still.. more days not mining is a pain... there should be a note in the OP to let people know that we have identified issues with the linux miner...

Code:

raskul - Today at 07:39
good morning! any joy fixing the bugs?
Protovist - Today at 07:43
Not yet. The more I look at it the more it looks like an architectural problem with the miner. Only the "best" (lowest) hash is submitted for each iteration, so adding additional GPUs doesn't help the number of shares submitted.

Yes, my thanks to Protovist too for the porting GPU miner to Linux.

Regarding his answer about GPU miner: architecture of GPU miner is similar to architecture of CPU-miner. After recieving a new task miner has some time to find descent share and then this share is sent to pool. All further shares for this task must be more difficult than previous. Pool anyway ignores all shares with low difficulty (because good shares were recieved previously), but your nopaid share value will decrease, because nopaid shares value is calculated as average value based on your shares. So it is your interest to send each next share for each task with more difficulty than previous.

Using several GPUs has sense. The mining is actually a process of searching of some 64-bit value called nonce, which gives the lowest possible hash of block (nonce is written to block header). So GPU just start from some random nonce and increments this value for each hash. If use several GPUs - each GPU searches nonce in own range. So your chances to find a good nonce increases. One hand is good, but two hands is better.

Hopefully the problems will be fixed soon. I can't mine XDAG until the linux miner is working properly and others will not be getting the correct shares if they are mining using the linux miner.