On the 20th of October, Penguin informed the community of a bug found in the mining code.
Penguin exploited the bug at his own advantage for a few days, for test purposes, then fairly decided to made it public.
This was definitely a good move for both communities, Golos and Steemit, as the bug was also present in steemd code. He deserve all kudos for this exemplary altruist behavior.
For those who are now aware of what happened on Steemit mining, you have to know that steemit mining queue is monopolized for months by 2 miners : rabbit and gxt-1080.
gxt_1080 (aka nobody) has confirmed he is using a GPU miner code, which provides him with a huge advantage in terms of computing power.
rabbit has not discovered himself but is suspected to use the same technology as gxt_1080, with less resources.
If you are interested in the full story, you can read the following posts (in English):
The last with my (prescient) thought on what could happen to Golos mining
Several CPU miners considered this bug fix as an opportunity to (re)start their mining node on the Golos network. But is it worth ? What is the real impact of golosd bug fix?
Let's have a look at the following chart.
This chart shows the distribution ratio of pows found by miners
On Golos launch day, we see some diversity in miners because gtx-1080 was not ready at the very first time. This allowed a group of CPU miners (others in the chart) to grab up to 27,63% of the total daily pows.
Then gtx-1080 released his horses and the game was done. Starting for day 2, he captured 84,72% of the pows. His only competitor was penguin, an unknown challenger. We all thought it was rabbit's Golos avatar. They both dominated the mining queue for 2 days.
On the 20th of October, at the end of the day, Penguin made the mining bug publicly known.
Immediately, some miners patched their node and started to catch pows again.
One interesting thing to notice is the emergence of s_g_m.
Either he also has GPU miner code or enough CPU power to outperform other CPU miners.
I guess the my second assumption is the correct one as he popped out on chart just after the bug fix has been made public.
If we now look closer at numbers, you will see that other CPU miners fight to get a mere average 2.45% of all pows.
In value, this is an average of 33 pows per
Here the detailed pow distribution:
The group of lucky miners who succeeded to found a pow the last 5 days is composed of 18 miners, with an minimum of 6 and a maximum of 9 lucky miners per days. That's not a lot.
I made a test setting up a miner node on a dedicated server with 2 CPU Intel Xeon 5650 - 6 cores @2.67Ghz, providing average 33k hash mining power. I let it run one day and didn't grab a pow.
Personally, given the cost of maintaining a running node in terms of money and time, I feel running a CPU miner is worthless.
К сожалению, я не пишу это на русском. Это немного слишком технической. Перевод разрешено и приветствуется.
Спасибо за чтение!
Вы также можете проголосовать прямо со Голос здесь. Для этого внизу страницы нужно проделать следующее
Если Вам понравился этот пост,
не забудьте проголосовать или подписаться на меня или поделиться