Thank you!
Topic: [XMR] JCE Miner Cryptonight/forks, now with GPU! - page 105. (Read 90841 times)
As I see, at pool is all ok!
Thank you!
Thank you!
if you force --variation 5 it should work.
The bug is due to Leviarcoin and Loki wallet having a very close format. It confuses my detection. But bypassing the detection with --any --variation 5 should workaround. It will still display Leviarcoin, but will mine Loki. And of course, look at your pool to ensure you get the shares.
The bug is due to Leviarcoin and Loki wallet having a very close format. It confuses my detection. But bypassing the detection with --any --variation 5 should workaround. It will still display Leviarcoin, but will mine Loki. And of course, look at your pool to ensure you get the shares.
21:08:40 | Pool loki.miner.rocks:5555
21:08:40 | Currency Leviarcoin (XLC)
as I understand its not right coin, at thise pool coin is LOKI
It's not a problem?
21:08:40 | Currency Leviarcoin (XLC)
as I understand its not right coin, at thise pool coin is LOKI
It's not a problem?
Sure, i add it to next 0.26
I've finished all combinations of algos (hundreds!) up to hexa-hash, and now in test phase. It's still time to add some coins, I add Loki and Gadcoin.
Note that you can now mine any coin with JCE, even if not listed, example for Loki:
--any unlocks the coin list
--variation 5 tells jce that's CN-Heavy (otherwise jce cannot know, since it's an unlisted coin)
I've finished all combinations of algos (hundreds!) up to hexa-hash, and now in test phase. It's still time to add some coins, I add Loki and Gadcoin.
Note that you can now mine any coin with JCE, even if not listed, example for Loki:
Code:
jce_cn_cpu_miner64 --any --variation 5 ..........
followed by usual params (wallet, pool...).--any unlocks the coin list
--variation 5 tells jce that's CN-Heavy (otherwise jce cannot know, since it's an unlisted coin)
Hi!
Can you add LOKI Coin to JCE Miner it's Heavy algo
THNX!
Can you add LOKI Coin to JCE Miner it's Heavy algo
THNX!
each has 20M of cache and hardware AES, so that's a very good mining base.
i cannot give a Hashrate because i've only Ryzen cpu and old Core2, but it should be high, and it may worth you try using the manual config file to optimize max.
i know best config for ryzen but not on your super xeons sorry
i cannot give a Hashrate because i've only Ryzen cpu and old Core2, but it should be high, and it may worth you try using the manual config file to optimize max.
i know best config for ryzen but not on your super xeons sorry
config file is only for CPU fine tuning. Not for coin configuration. JCE is simpler than XMRStak or xmrig.
let me look at your coin and i tell you how to mine it. i'll add it to the list of known coins in next version, but you can already mine it with JCE 0.25
edit :
the interresting part is the --any to unlock the coin list (means: to mine any unknown coin) and --variation 1 to tell JCE that's a Generation 1 cryptonight algo (that's original cryptonight, like Electroneum for example). Since the coin is unknown, JCE cannot autodetect the Cryptonight algo. Otherwise the --variation parameter is optional.
It works fine this way. Just that's crazy to start a new coin with the old Cryptonight mode, where ASICs kill everything, but it works.
let me look at your coin and i tell you how to mine it. i'll add it to the list of known coins in next version, but you can already mine it with JCE 0.25
edit :
Code:
jce_cn_cpu_miner64.exe --auto --any -p x -u gadxxxxxxxxxxxxxxxxxxxxxxxxxxx -o pool.gadcoin.com.br:5555 --variation 1
the interresting part is the --any to unlock the coin list (means: to mine any unknown coin) and --variation 1 to tell JCE that's a Generation 1 cryptonight algo (that's original cryptonight, like Electroneum for example). Since the coin is unknown, JCE cannot autodetect the Cryptonight algo. Otherwise the --variation parameter is optional.
It works fine this way. Just that's crazy to start a new coin with the old Cryptonight mode, where ASICs kill everything, but it works.
Thank you Admin for CPU mining tool with CryptoNight algo. But I dont really understand to set up this config file. Would you mind providing me some example files? I intend to mine this coin Gadcoin with your software. This coin link is here (https://bitcointalksearch.org/topic/ann-gadcoin-3339476)
Thank in advance.
Thank in advance.
you mean http server API ?
I'm not a big fan of that but i admit it's needed to get integrated to 3rd party tools. i'll do it but i think i'll do the Linux port first. once the multihash is done.
i've several hundreds of assembly to maintain now, just the 6 multihash times the 6 forks are 36, times the cpu archi, times dualthread mode, times 32/64 bits...
edit: i'll go for the http api, better than hacking the log system.
I'm not a big fan of that but i admit it's needed to get integrated to 3rd party tools. i'll do it but i think i'll do the Linux port first. once the multihash is done.
i've several hundreds of assembly to maintain now, just the 6 multihash times the 6 forks are 36, times the cpu archi, times dualthread mode, times 32/64 bits...
edit: i'll go for the http api, better than hacking the log system.
Added miner to Forager - automatic profit switching miner with MSI Afterburner integration
[/end_shameless_ad]
thanks bro, not a shameless ad, mining tools and miners work together [/end_shameless_ad]
i'll add a ref to your tool next time i update the doc
Cheers
If you also add API, would be great. It would add the ability to properly benchmark and monitor the miner.
Or as a simple solution, since I'm basically capturing output to detect hash rate - to show it more often (or add a switch that does that)
nothing gpu-special, but each round of gpu mine xmr needs 88 bytes of input cpu side, for each gpu thread, and there are often ~1000 of them per gpu, time 7 gpu you get ~500K of cache flooded in your cpu, enough to annoy a minig thread.
Oh.... I didn't know that... Thanks for info! Added miner to Forager - automatic profit switching miner with MSI Afterburner integration
[/end_shameless_ad]
thanks bro, not a shameless ad, mining tools and miners work together [/end_shameless_ad]
i'll add a ref to your tool next time i update the doc
nothing gpu-special, but each round of gpu mine xmr needs 88 bytes of input cpu side, for each gpu thread, and there are often ~1000 of them per gpu, time 7 gpu you get ~500K of cache flooded in your cpu, enough to annoy a minig thread.
Added miner to Forager - automatic profit switching miner with MSI Afterburner integration
[/end_shameless_ad]
[/end_shameless_ad]
with so many gpus, you probably cannot use all cache for cpu mining, you may have better perfs with 2x Triple plus nine simple for IPBC
for cn-v7 i'd say seven or six simple on a ryzen 1600/1700. you'll get ~430 I predict on a r5 1600
However the cheap way would be to stay with 8x simple threads and with --low parameter and let the GPU miners take the cache they need. that's what i do, but my ryzen has only 3 GPU. This way JCE drops from 506 (alone) to 497, quite limited drop, and no impact on gpu mining.
What GPU's you're talking about? For wich GPU's it's need CPU cache? More specifically, please. for cn-v7 i'd say seven or six simple on a ryzen 1600/1700. you'll get ~430 I predict on a r5 1600
However the cheap way would be to stay with 8x simple threads and with --low parameter and let the GPU miners take the cache they need. that's what i do, but my ryzen has only 3 GPU. This way JCE drops from 506 (alone) to 497, quite limited drop, and no impact on gpu mining.
with so many gpus, you probably cannot use all cache for cpu mining, you may have better perfs with 2x Triple plus nine simple for IPBC
for cn-v7 i'd say seven or six simple on a ryzen 1600/1700. you'll get ~430 I predict on a r5 1600
However the cheap way would be to stay with 8x simple threads and with --low parameter and let the GPU miners take the cache they need. that's what i do, but my ryzen has only 3 GPU. This way JCE drops from 506 (alone) to 497, quite limited drop, and no impact on gpu mining.
for cn-v7 i'd say seven or six simple on a ryzen 1600/1700. you'll get ~430 I predict on a r5 1600
However the cheap way would be to stay with 8x simple threads and with --low parameter and let the GPU miners take the cache they need. that's what i do, but my ryzen has only 3 GPU. This way JCE drops from 506 (alone) to 497, quite limited drop, and no impact on gpu mining.
@JCE-Miner
Any recommendation for a 6-7 GPU Ryzen motherboard? Thanks.
Any recommendation for a 6-7 GPU Ryzen motherboard? Thanks.
same for me, triple hash is better than double, and even two single is better than double.
but double alone is almost twice faster than simple. ipbc gives strange results...
planned for next release:
triple to hexa hash
graft defaults to cn-v7
more stable Hashrate on double AES 64
but double alone is almost twice faster than simple. ipbc gives strange results...
planned for next release:
triple to hexa hash
graft defaults to cn-v7
more stable Hashrate on double AES 64
@JCE-Miner
I did some tests using double_hash with Ryzen 7 1700 at IPBC, and I think it's worse than single hash. Ya, the cache is the problem.
Ryzen 7 1700 runs with 16 threads around 2200 at 3.6GHz 2933 RAM/L3. With double hash, more than 16MB cache is used. With Ryzen 5 1600, with just 12 threads, there is 4MB free to play
.
Is nice to see R5 1600 shinning. I'm thinking about a second RIG with a R5, but I was reluctant about mining performance.
I did some tests using double_hash with Ryzen 7 1700 at IPBC, and I think it's worse than single hash. Ya, the cache is the problem.
Ryzen 7 1700 runs with 16 threads around 2200 at 3.6GHz 2933 RAM/L3. With double hash, more than 16MB cache is used. With Ryzen 5 1600, with just 12 threads, there is 4MB free to play
.Is nice to see R5 1600 shinning. I'm thinking about a second RIG with a R5, but I was reluctant about mining performance.
problem with graft ? ok i'll focus to test it more. maybe it selects the wrong fork.
what mode do you use (32 or 64 bits, simple or double hash ?)
the best IPBC mode for ryzen 6 Cores is two triple and ten double hash, but triple not avaiaible yet, need more tests. will be in 0.26 to be released soon.
with my assembly i run at 1698 with peaks at 1704, while 0.25 maxed at 1662, on my ryzen 1600
Edit: i just switched my rig to Graft and got only bad shares. when forcing CN V7 with --variation 3 it works fine.
so it seems Graft switched to v7 recently
to mine Graft with JCE add parameter --variation 3
it will be automatic in 0.26
what mode do you use (32 or 64 bits, simple or double hash ?)
the best IPBC mode for ryzen 6 Cores is two triple and ten double hash, but triple not avaiaible yet, need more tests. will be in 0.26 to be released soon.
with my assembly i run at 1698 with peaks at 1704, while 0.25 maxed at 1662, on my ryzen 1600
Edit: i just switched my rig to Graft and got only bad shares. when forcing CN V7 with --variation 3 it works fine.
so it seems Graft switched to v7 recently
to mine Graft with JCE add parameter --variation 3
it will be automatic in 0.26
Jump to: