The power of the mining pool is too big? Solved in three seconds

Easy Miner, creating value for users

This article is only for A cards. There is no doubt that N cards use the nbminer kernel. The mining pool computing power error often appears positive. If the N card still has a large gap when using the nbminer kernel, you should check the network and overclocking parameters yourself.

Effect of using nbminer kernel with N card

Recently, we heard some users report that the gap between mining pool computing power and local computing power is quite large. We contacted two users and presented the entire process to everyone in the form of a record flow.

First, let’s talk about the problem that should never occur: ineffectiveness.

The user's computing power is 50MH/s, 365 are accepted, and 143 are invalid. It is not difficult to understand why the mining pool's computing power does not meet the standard.

Especially 30 series N cards are prone to invalidation when using PhoenixMiner

The three most basic data that affect the computing power of a mining pool are acceptance, rejection, and invalidity. As early as a year ago, we wrote " Pay attention to "1 low and 2 high", these 3 situations affect computing power ". Users who love learning can take some time to read it.

A user complained in a QQ group that the average daily computing power error has been exceeding -5% for a long time, and he was so frustrated that he wanted to change the software.

We contacted him as soon as possible, and after obtaining his consent, we replaced the kernel with teamredminer 0.81.

> User gave feedback at 3:37 a.m. on the 18th, got in touch and took screenshots at 7 a.m., error -5%, red

>At 18:00 on the morning of the 19th, the error dropped to -3.97%, showing a significant effect

>At 6:42 AM on the 19th, about 24 hours after the setting was changed, the error was -2.48%

>At 10:44 am on the 20th, about 48 hours after the setting was changed, the error was only -3.07%

According to Spark Mining Pool, anything below -5% is considered normal, and -3% is excellent.

(Screenshot from Spark Mining Pool)

However, there are still 1-2 machines that are unstable (disconnected). I believe the data will be more beautiful after they become stable.

Let’s look at another user’s data.

The 24-hour computing power errors of the four machines were -5.68%, -4.36%, -7.13% and -10.06% respectively.

>February 18, 10:06, 24 hours of using teamredminer-0.8.1

>February 19, 10:30, 24 hours of using lolminer 1.21

>At 11:01 on February 21, PhoenixMiner5.5c has completed 24 hours

>February 22, 12:40, 24 hours of using teamredminer 0.8.1

The raw data above is rather messy and is organized as follows.

In summary, the 24-hour effective computing power of the mining pool of PhoenixMiner (abbreviated as pm) is significantly lower, while the 24-hour computing power of the mining pool of teamredminer (abbreviated as trm) is very good, and may even be positive, which users jokingly call "respiratory" .

So, do you understand? If your machine runs for 24 hours and the mining pool has a large difference in computing power, users with group control can select the machine in the group control and switch with one click, and select teamredminer 0.8.1 as the kernel; users without group control can stop mining, select teamredminer 0.8.1 as the kernel, and then start mining. The problem can be solved in 3 seconds.

After 24 hours, don’t forget to come back and like it. If you are an N card user, the mining pool computing power using the nbminer kernel is also very ideal, please like it in the lower right corner.

Since PM has always been the recommended kernel for A cards, many users are used to using PM to find faults. After replacing TRM, there may be an adaptation period for both the user and the hardware. Because each kernel has different effects on the hardware, there may be incompatibility, which may cause a crash or blue screen. We will summarize these issues later.

trm does not support N cards, which means that users who mix A+N still need to choose pm or NBMiner kernel.

The local hashrate of some cards using trm seems lower than that of pm, but the final mining pool hashrate may still be higher. You can decide for yourself how to make a choice. No matter how good trm is, it cannot offset the losses caused by overclocking, network, hardware and other failures. This should be beyond doubt.

According to TRM's official statement, the card fee for Polaris architecture, such as 4xx and 5xx series, is 0.65%, and the fee for other architectures (such as VII, VEGA, 5XXX and 6XXX series) is 1%. There is no additional fee for using Easy Card.

From now on, the recommended kernel for A cards is changed to TRM. If there is a better kernel in the future, we will make adjustments accordingly.