|
Hello,
any chance for a macos version?
thx
|
|
|
|
|
Hello,
what is the best dual mining algo pair that can be used with this miner?
thx
|
|
|
|
|
Hello,
I just restarted some miners after long time.
I am using 5.5c on 580 8Gb since later versions gave me errors. They are all around 29/31 Mh, can I squeeze some more hashes ?
Also one 580 is a 4Gb, so I am mining ETC, but the Mhs is around 15 Mhs, I remember it suppose to be around 29.
There are the settings I am using:
-proto 2 -coin etc -cdm 0 -stales 0 -worker AMD5804G -gpus 1 -amd -acm -astats 1 -astats 1 -leaveoc -mi 14 -gt 40 -gser 2 -wdog 1 -tt 80 -tmax 80 -fanmin 70 -fanmax 85 -powlim -15 -cclock 1200 -mclock 2050 -cvddc 930 -mvddc 960 -log 0
Also -tmax 80 not working, it sets it to 84 default.
Am I missing something from when I was mining with version 3.0 ?
thx
|
|
|
|
There are 2 config files in the github https://github.com/ganzocrypt/xmrig-macos/releases/tag/v6.13.0-dev-macosI used KawPow on unminable and I played a bit with Intensity: "kawpow": [ { "index": 0, "intensity": 36864, "worksize": 256, "threads": [-1] } ], So you can check that in the OpenCL section of the config. Start with a low number and increase it. The above was for a 580 8gb |
|
|
|
|
Hi,
is there a way to disable API server?
Cannot find if exists in the -help
Answer: --api-bind-telnet 0 --api-bind-http 0 will disable it.
|
|
|
|
|
@miner29
Thanks a lot will update the release note!
|
|
|
|
|
@miner29 did you had a chance to test it?
can you share coin / card / cpu / macOS version so I can add to the list
any issues?
thx
|
|
|
|
I saw that video, my cards are older  1080 TI, 1070 TI and 1060 6Gb. BTW I am on Ubuntu thx |
|
|
|
I recompiled xmrig with latest libs. No feature changes so just get the one on the github. Please report any issues, or success with algo and card so that I can update the list. Download binary from here: https://github.com/ganzocrypt/xmrig-macos/releases |
|
|
|
|
@miner29
Oh cool, great to know that is working, let me know if there are any issue.
So you are mining on rx/0 on GPU, which coin?
Also will update the list of Cards and Algo that is supporting as people using the miner can give some feedback ( it's like trial and error).
Will probably update it soon since I recompiled bunch of libraries that needed to be update it in brew.
Ok RVN needs more memory, great to know!
You can play with intensity and worksize to increase hashrate, just keep in mind that at certain point you might have issues with the miner.
At that point just lower the values.
I tested on RX580 8Gb, getting around 7Mhs and I have the following:
"kawpow": [
{
"index": 0,
"intensity": 36864,
"worksize": 256,
"threads": [-1]
}
],
|
|
|
|
@BenhurMarcel You should mine those 2 coins with the CPU. Only KawPow is supported for GPU mining at this stage You should have only one thread per card. I posted 2 config files samples.Can you post the OpenCL section of your config file? There are still issues in supporting some algos for AMD cards as you might know OpenCL is not supported anymore by Apple. NOTE: Updated the tested Algo, if you test a different algo and it IS/NOT working please let me know. https://github.com/ganzocrypt/xmrig-macos/releases/tag/v6.13.0-dev-macosA suggestions, you can mine with your CPU Keva or Monero and with your 2 GPUs (if GPU memory is enough, since 2Gb is low) you can mine RVN coin. If you want to try this, you need to have 2 instances of xmrig with 2 different configs. So you launch on 2 different terminal: ./xmrig -c configCPU.json ./xmrig -c configGPU.json |
|
|
|
|
Let me know if there are any issues and if you need help with the config file.
|
|
|
|
|
Yeah I know.
The issue is that macos OpenCL is basically not supported and OLD, like version 1.2.
So some dev made the miner but it does not work on all combination since libraries mismatch.
I helped to test and recompile the xmr code.
It was a real pain !
I had to use homebrew, recompiled bunch of libraries.
Modify some make scripts and finally worked.
I tested on KawPow, so you can use unminable or other auto-switching pools service to mine bunch of other algos too.
I know there are a lot of people with hackintosh and AMD cards who can use this miner if they want to.
Looking to see if there are any issues (which probably there are) with other AMD cards, I tested on RX 580 and someone tested on Radeon Pro 5300
|
|
|
|
First release of XMRIG for mac os working for CPU and AMD GPUs. Some algos are still not working using AMD GPUs due to OpenCL not supported by by Apple anymore. For example UPX, which is a CPU mineable coin, gives errors if mined with GPU. All other Algos should work, if not, please provide information in the issue section. All work for making XMRIG working with AMD GPUs is from @Spudz76, I just went through the trouble of compiling the code for macOS which was a pain with all dependencies and Homebrew installs. - You can run one instance of XMRIG for the same algo/pool with CPU and GPU
- If you want to run on different pools, you can run 2 instances of XMRIG with 2 different config.json, one for CPU and the other for GPU and define 2 different algos and pools.
Download binary from here: https://github.com/ganzocrypt/xmrig-macos/releasesIf there are any questions of issue please post here or on the github issue page https://github.com/ganzocrypt/xmrig-macos/issues |
|
|
|
|
Hi,
wanted to know if there is a ETC miner with the new fork for macOS (amd cards), any suggestions?
I am not looking for those automatic ones, like minergate or similar, but a client.
thx
|
|
|
|
As such, we have decided to implement a halving, reducing the block reward by 50%. Thus, rather than 4,100 UPX per block, the block reward will be around 2,050 UPX per block (and will still slowly decrease over time). maybe you should read again and understand what is written pretty clearly ! |
|
|
|
|
Hi,
can you pause the miner while running ?
can't find info.
thx
|
|
|
|
I removed the --preparedata and is working.
But I do not see any significant gain other than getting the same hashrate as with old_mode: true on 1.8.8, without it 1.8.8 was slower for me.
Also CPU utilization went up and since I am mining with CPU lost 600 h/s
- 1.8.8: hash 175K with old_mode true
- 1.8.9: hash 175k without old_mode and tweaks 1-8 and no preparedata at command line
So the small gain 100/200 h/s is canceled by higher CPU usage which cost 600 h/s
my card RX580 8Gb bios modded are at CC 1200/940 MC 1950/900
I fixed this issue and re-uploaded 189, so you can download it again and use old_mode + preparedata together. Work fine 1.8.9. Checked with profiles settings, got a speed increase of ~5.5% on standard VBIOS. A little, but nice.  Algos with small scratchpad size like upx2, turtle have the least benefit from tweaking profiles on RX, if you look at it in %. Others should be there about 3-5%, depending how good is the strap in the modded vbios. I don't understand why are you still on stock vbios Looks like with old mode and preparedata I am back at same GPU and CPU hashrate as before, so no loss CPU % and hashrate Will play a bit with tweak as you mentioned thx |
|
|
|
Just update to 1.8.9, Win 10, 7 RX 580 cards
getting an error:
Error CL_INVALID_WORK_GROUP_SIZE when calling clEnqueueNDRangeKernel for kernel 00 for DeviceID 4 (Thread 9)
any idea?
this is my config:
{
"cryptonight_type" : "upx2",
"double_threads" : true,
"gpu_conf" :
[
{"id" : 0, "intensity" : 107, "worksize" : 8, "threads" : 2, "fragment": 0, "tweak_profile": 1, "old_mode" : true},
{"id" : 1, "intensity" : 107, "worksize" : 8, "threads" : 2, "fragment": 0, "tweak_profile": 1, "old_mode" : true},
{"id" : 2, "intensity" : 107, "worksize" : 8, "threads" : 2, "fragment": 0, "tweak_profile": 1, "old_mode" : true},
{"id" : 3, "intensity" : 107, "worksize" : 8, "threads" : 2, "fragment": 0, "tweak_profile": 1, "old_mode" : true},
{"id" : 4, "intensity" : 107, "worksize" : 8, "threads" : 2, "fragment": 0, "tweak_profile": 1, "old_mode" : true},
{"id" : 5, "intensity" : 107, "worksize" : 8, "threads" : 2, "fragment": 0, "tweak_profile": 1, "old_mode" : true},
{"id" : 6, "intensity" : 105, "worksize" : 8, "threads" : 2, "fragment": 0, "tweak_profile": 1, "old_mode" : true}
]
}
Try adding --nobinaries to the cmd Or just remove the --preparedataedit: yes i see what i have forgot, going to fix and re upload later old_mode + preparedata dont cooperate I removed the --preparedata and is working. But I do not see any significant gain other than getting the same hashrate as with old_mode: true on 1.8.8, without it 1.8.8 was slower for me. Also CPU utilization went up and since I am mining with CPU lost 600 h/s - 1.8.8: hash 175K with old_mode true - 1.8.9: hash 175k without old_mode and tweaks 1-8 and no preparedata at command line So the small gain 100/200 h/s is canceled by higher CPU usage which cost 600 h/s my card RX580 8Gb bios modded are at CC 1200/940 MC 1950/900 |
|
|
|
|
Show Posts
