327 h/s with XFX RX 550 4GB using xmr-stak-amd. $109 per video card.

MoneroMonero
$114.93
‑2.42%
cryptocurrencyfreak.com5y ago

I recently found the XFX RX 550 4GB model for sale for just $109 shipped via Amazon and couldn’t resist the price tag. I was reading people getting around 300 h/s with the cards and that works out to around 2.75 hashrate per dollar spent. Not bad at all, so I ordered two (and just ordered two more that will be here tomorrow) and started playing with xmr-stak-amd to see what I could get out of them. Needless to say, I am pleased. Disclaimer: Below I have linked to Amazon using my affiliate links. It helps fund my hobby and keeps my writing these kind of articles.

A quick rundown of the mining rig:

Here are the settings I am using for xmr-stak-amd for 2 XFX RX 550 4GB:

/* * Number of GPUs that you have in your system. Each GPU will get its own CPU thread. */ "gpu_thread_num" : 4, /* * GPU configuration. You should play around with intensity and worksize as the fastest settings will vary. * index - GPU index number usually starts from 0 * intensity - Number of parallel GPU threads (nothing to do with CPU threads) * worksize - Number of local GPU threads (nothing to do with CPU threads) * affine_to_cpu - This will affine the thread to a CPU. This can make a GPU miner play along nicer with a CPU miner. */ "gpu_threads_conf" : [ { "index" : 0, "intensity" : 512, "worksize" : 8, "affine_to_cpu" : 0 }, { "index" : 0, "intensity" : 512, "worksize" : 8, "affine_to_cpu" : 0 }, { "index" : 1, "intensity" : 512, "worksize" : 8, "affine_to_cpu" : 1 }, { "index" : 1, "intensity" : 512, "worksize" : 8, "affine_to_cpu" : 1 }, ], /* * Platform index. This will be 0 unless you have different OpenCL platform - eg. AMD and Intel. */ "platform_index" : 2,

To break it down a bit, I am starting 4 GPU threads and assigning 2 threads per card, each with an intensity of 512 and worksize of 8. One video card is set to processor core 1 and the other to processor core 2. The Pentium G4400 is a dual core processor with no hyperthreading.

Platform_index is on a case by case basis depending on your hardware, but 2 is what worked for the above configuration.

The results can be seen below:

Averaging 655 h/s over 15 minutes. That works out to exactly 3 h/s per dollar spent on the video cards at a very low power output.

I would be interested in hearing your thoughts on this. Keep in mind that this is all at stock speeds! I plan to switch to Linux once I reach 8 video cards and have 12 XFX RX 550 running for just under 4 kh/s at a cost of just $1,308 dollars. The cards should retain a decent resell value due to the ram.

Update:

Here is another setup that you may wish to try. It provides just a bit less hashrate per second at stock speeds but may be more stable if you try to overclock.

/* * Number of GPUs that you have in your system. Each GPU will get its own CPU thread. */ "gpu_thread_num" : 4, /* * GPU configuration. You should play around with intensity and worksize as the fastest settings will vary. * index - GPU index number usually starts from 0 * intensity - Number of parallel GPU threads (nothing to do with CPU threads) * worksize - Number of local GPU threads (nothing to do with CPU threads) * affine_to_cpu - This will affine the thread to a CPU. This can make a GPU miner play along nicer with a CPU miner. */ "gpu_threads_conf" : [ { "index" : 0, "intensity" : 256, "worksize" : 8, "affine_to_cpu" : 0 }, { "index" : 0, "intensity" : 256, "worksize" : 8, "affine_to_cpu" : 0 }, { "index" : 1, "intensity" : 256, "worksize" : 8, "affine_to_cpu" : 1 }, { "index" : 1, "intensity" : 256, "worksize" : 8, "affine_to_cpu" : 1 }, ], /* * Platform index. This will be 0 unless you have different OpenCL platform - eg. AMD and Intel. */ "platform_index" : 2,

Related