You can specify configs for the cards like with the cn_config, but it's eth_config instead. To make all cards run A mode, you can use --eth_config=A
You can set a specific mode and intensity with something like so: --eth_config=A300
You can also provide a list with one entry per card, like so: --eth_config=A300,B400,A,B
And it's need to optimize speed... On Phoenix with double kernels my RX 580 gives about 600KH/s (about 2%) more speed. Even if 0,65% dev fee not true, speed on TRM eth too low in compare with PM.
I'm guessing you haven't read any of the last ~20 posts in this thread. You might want to go back and look through them a bit.
TL;DR: Phoenix miner reports hashrate ~2.9% higher than it actually gets. So by your measurement, we're getting about 0.9% better actual hashrate (I.E. your pool hashrate will go up 0.9%).
Number after A or B is the intensity? A200 is less intensive than A400?
I read last several pages and understand that Claymore and Phoenix give fake speeds, but I look into accepted speed on the pool-side and Phoenix gives me the same speed as showed accepted speed in miner. And for now it sometimes a bit more than TRM and sometims a bit lower.
But thanks for answers.
Hi UnclWish!
I truly believe you only mean well, and I _really_ believe you want to run the miner most optimal for yourself, and you (like all miners) would like to be able to test and verify miners to compare for yourselves. Like todxx mentioned, Phoenix miner has been underperforming its displayed hash rates with an avg of -2.9% right now across all our runs. That has now been verified in tests done outside of our control as well, but still mostly with people that we know, and I'd love if people like you, that we only really interact with here in this forum, would run some tests as well.
Therefore, I would like to offer you compensation in XMR for mining air for a few days so you can test for yourself using our released ethash miner testing tool. I've tried to present the math and statistics in the documentation for the tool, and show that 24h poolside tests using a single gpu or even full rig are just way too random to trust. We need a much much larger sample size, and that's why we built and released the tool in the first place.
It is indeed difficult on a single 580 though, it will take maybe 5 days to do a single test, and naturally, it must run 100% stable during that time, no crashes or internet outages etc. Maybe you have more than one 580 to test with? If you're interested, try to get started yourself, then we'll answer any questions you might have here in public to you get going. Then you report back when the test is done.
For people thinking that running a 5+ days long test with a much lower diff than normal pools sounds ridiculous, unfortunately it's what you need if you only do a single gpu. The test is just a very controlled way of keeping track of your mining over a longer period of time, where pools normally only present a sliding window of the last N hours.
Again, the tool and documentation is available here:
https://github.com/Kerney666/trm-ethash-miner-tester/Cheers, Kerney
