Larvitar
Jr. Member
Offline
Activity: 196
Merit: 1
|
|
June 15, 2018, 10:10:35 PM |
|
You are doing much progress with GPU. wow man
Sorry if it's boring, but I have to discuss Awesome Miner again.
Well, I tried today to run JCE "inside" Awesome miner. As you said some pages ago, JCE Miner has some similiarities with CPUminer-OPT API, so I tried to "camuflate" JCE as CPUminer-opt. Not success here
But reading log, I found the solution (or just a piece): The problem to run is the option "--api-remote -b 0.0.0.0:4034". I can set -mport easily, but Awesome don't alow to put off the --api-remote option.
To solve this, just ignore the --api-remote option or accept it as --mport (with IP:port form). Could you do this in the next release?
Thanks a lot man! Your miner is I M P R E S S I V E!
|
|
|
|
JCE-Miner (OP)
Member
Offline
Activity: 350
Merit: 22
|
|
June 15, 2018, 10:24:41 PM |
|
no problem, i'll ignore the parameter --api-remote. do you need i ignore -b too, with a dotted ip after ? i think i'll release a last cpu version using the gpu version, but with gpu disabled. then a alpha with gpu+cpu
|
|
|
|
whotheff
|
|
June 16, 2018, 12:56:40 PM |
|
Xmr Stack = 130h/s JCE = 74h/s
huge pages on, avx-aes
on AMD A10-5800K
same pool, same time of day, waited atleast 15 min to confirm hashrate.
|
|
|
|
JCE-Miner (OP)
Member
Offline
Activity: 350
Merit: 22
|
|
June 16, 2018, 05:47:01 PM |
|
Nope, JCE is not 50% slower than stak. Maybe not that faster, but not 50% slower. And that score is very low for a A10, i can get 115 with a core2 quad. I guess you use one thread only, and stak enabled double-hash. It can be enabled on JCE with manual config too.
At least give the exact Stak configuration you used, and i can tell you how to do the same on JCE.
|
|
|
|
whotheff
|
|
June 16, 2018, 08:50:37 PM |
|
Nope, JCE is not 50% slower than stak. Maybe not that faster, but not 50% slower. And that score is very low for a A10, i can get 115 with a core2 quad. I guess you use one thread only, and stak enabled double-hash. It can be enabled on JCE with manual config too.
At least give the exact Stak configuration you used, and i can tell you how to do the same on JCE.
Noticed that XMR stack uses only two cores (so it can supply them with 2Mb cache each=4MB total for that CPU). Then disabled two of them in JCE and it jumped to 130h/s. How to enable double hash? What does it do?
|
|
|
|
Larvitar
Jr. Member
Offline
Activity: 196
Merit: 1
|
|
June 16, 2018, 09:13:24 PM |
|
no problem, i'll ignore the parameter --api-remote. do you need i ignore -b too, with a dotted ip after ? i think i'll release a last cpu version using the gpu version, but with gpu disabled. then a alpha with gpu+cpu
Yup!
|
|
|
|
JCE-Miner (OP)
Member
Offline
Activity: 350
Merit: 22
|
|
June 17, 2018, 08:29:52 AM |
|
thanks for the test, it seems my autoconfig is bad on the A10, it allocates too many threads. I'll fix it, thanks.
The multihash (double hash is the 2-case, you can set from 1- to 6- ) is also called low-power in stak IIRC. It's about using, on one CPU core, twice the register and twice the cache to get sometimes twice the speed. The trick is that it let the other cores free, so it consume less power and allows the Turbo to enable, for CPU with turbo.
Technically, it's good, sure when you want to save power, but also when you run out of cores and not of cache. If you have a CPU with 2 cores but 8M cache, normal config would give only 2x2M = 4M cache used.
you may enable double-hash to use 2x2x2M cache = 8M of cache, and get some extra perf. It works more or less depending on the CPU. It's very efficient on Ryzen, and not at all on Core2.
I looked closer at the A10, and yeah that's a little APU with little cache. I give you an experimental config that could let you get some extra perf, but not sure, i cannot test, i've no A10.
"cpu_threads_conf" : [ { "cpu_architecture" : "auto", "affine_to_cpu" : 0, "use_cache" : true }, { "cpu_architecture" : "auto", "affine_to_cpu" : 1, "use_cache" : false }, { "cpu_architecture" : "auto", "affine_to_cpu" : 2, "use_cache" : true }, { "cpu_architecture" : "auto", "affine_to_cpu" : 3, "use_cache" : false }, ]
now finishing 0.29c, the last CPU-only version, with some updates.
|
|
|
|
Iamtutut
|
|
June 17, 2018, 09:09:34 AM |
|
What would the 0.29C update be ?
For the CPU / GPU miner, will it be possible to split CPU and GPU mining ? I don't mine CN heavy algos at all with the CPU (4MB cache), but I sometimes mine CN heavy with the GPUs. So if possible I'd like to mine CN V7 / lite with the CPU separated from any other CN algo with the GPUs.
|
|
|
|
JCE-Miner (OP)
Member
Offline
Activity: 350
Merit: 22
|
|
June 17, 2018, 09:21:09 AM |
|
good idea. It won't be possible in the earlier versions, but i love the idea. 0.29c updates : Use of the GPU version frontend, but GPU disabled Support of Masari fork --variation 11 Support of Haven fork --variation 12 Stellite defaults to XTL variation ETN defaults to v7 other coins that have forked will default to their respective fork Coin list in alphabetical order new coin : BXB new coin : WOWnero -b parameter ignored for compatibility with AwsomeMiner ditto with parameter --api-remote
|
|
|
|
whotheff
|
|
June 17, 2018, 09:53:12 AM |
|
thanks for the test, it seems my autoconfig is bad on the A10, it allocates too many threads. I'll fix it, thanks.
The multihash (double hash is the 2-case, you can set from 1- to 6- ) is also called low-power in stak IIRC. It's about using, on one CPU core, twice the register and twice the cache to get sometimes twice the speed. The trick is that it let the other cores free, so it consume less power and allows the Turbo to enable, for CPU with turbo.
Technically, it's good, sure when you want to save power, but also when you run out of cores and not of cache. If you have a CPU with 2 cores but 8M cache, normal config would give only 2x2M = 4M cache used.
you may enable double-hash to use 2x2x2M cache = 8M of cache, and get some extra perf. It works more or less depending on the CPU. It's very efficient on Ryzen, and not at all on Core2.
I looked closer at the A10, and yeah that's a little APU with little cache. I give you an experimental config that could let you get some extra perf, but not sure, i cannot test, i've no A10.
"cpu_threads_conf" : [ { "cpu_architecture" : "auto", "affine_to_cpu" : 0, "use_cache" : true }, { "cpu_architecture" : "auto", "affine_to_cpu" : 1, "use_cache" : false }, { "cpu_architecture" : "auto", "affine_to_cpu" : 2, "use_cache" : true }, { "cpu_architecture" : "auto", "affine_to_cpu" : 3, "use_cache" : false }, ]
now finishing 0.29c, the last CPU-only version, with some updates.
12:52:22 | Hashrate Thread 0: 59.71 h/s 12:52:22 | Hashrate Thread 1: 2.57 h/s 12:52:22 | Hashrate Thread 2: 54.01 h/s 12:52:22 | Hashrate Thread 3: 2.44 h/s 12:52:22 | Total: 118.71 h/s 12:52:23 | Pool changes Difficulty to 2634. limited to two threads with -t 2: 12:47:53 | Pool changes Difficulty to 2376. 12:47:59 | Hashrate Thread 0: 62.49 h/s 12:47:59 | Hashrate Thread 1: 60.26 h/s 12:47:59 | Total: 122.74 h/s
|
|
|
|
JCE-Miner (OP)
Member
Offline
Activity: 350
Merit: 22
|
|
June 17, 2018, 10:34:25 AM Last edit: June 17, 2018, 03:39:03 PM by JCE-Miner |
|
so the no-cache is bad on the A10. It's very good on modern Intel, decent on Ryzen and bad on Core2. And so, bad on A10 too. So keep the config you had found before, with two normal threads. I won't have time to fix the A10 autoconfig in next version, will be for 0.30 edit: i'm polishing the GPU version. I see very strange results on my Pitcairn rig (five different Pitcairns : 7870 and 7850, 1G and 2G...) On my 2G, when I use all the memory available, I can reach 526h/s on my 7870 2G, and 495 on CN-v7 on my 7850 2G, which is far above what Claymore 9.7 gave on CN-Classic. It's even the best score ever I reached, period. Maybe i didn't configure SRB well, but so far i didn't reach such an hashrate. But on my little 7850 1G, that's terrible, i peak at 343 where Claymore gave a whooping 450 ! It looks like my code is faster when using the double-memory, but a lot slower otherwise. And double-memory is barely possible on 1G cards. It works on Bonaire 1G because it has a little GPU that can be saturated with two threads using ~512M each, but my Pitcairn 1G are underused by an order of magniture... The total hashrate of the rig, counting the 2G that mine faster and the 1G that mine slower, is 10% lower that claymore 11.3, but with 0 bad share, where Claymore, as I said previously, gives ~10% bad shares on old GPU when configured at max, or is stable at 10% lower hashrate. 0.29c onlineUse of the GPU version frontend, but GPU disabled Support of Masari fork --variation 11 Support of Haven fork --variation 12 Stellite defaults to XTL variation ETN defaults to v7 other coins that have forked will default to their respective fork Coin list in alphabetical order new coin : BXB new coin : WOWnero -b parameter ignored for compatibility with AwsomeMiner ditto with parameter --api-remote
|
|
|
|
4ward
Member
Offline
Activity: 473
Merit: 18
|
|
June 17, 2018, 09:22:03 PM |
|
--variation 12 returns bad shares with multi_hash higher than 2 while --variation 5 works fine tested on Core i5 7600k (6MB cache)
|
|
|
|
Iamtutut
|
|
June 17, 2018, 09:52:39 PM |
|
It means Haven has not forked yet.
|
|
|
|
JCE-Miner (OP)
Member
Offline
Activity: 350
Merit: 22
|
|
June 18, 2018, 05:25:33 AM |
|
it should have forked already, and Masari should fork in a few hours. probably my code is broken in multihash, i've so many assemblies now that it's almost impossible to test all in decent time.
i'll release an emergency fix asap, thanks for the report !
|
|
|
|
gvb
Jr. Member
Offline
Activity: 140
Merit: 9
|
|
June 18, 2018, 08:15:05 AM |
|
thanks for the work on this.
I just notice one problem on the machine I'm currently using.
It only uses 3 of the 4 cores.
Adding the -t 4 right behind the --auto parameter in the start batch file seems to solve the 'problem'.
CPU is an Intel i5-4590 3.3GHz
|
|
|
|
raske
Member
Offline
Activity: 154
Merit: 14
|
|
June 18, 2018, 08:18:57 AM |
|
It means Haven has not forked yet.
of course it is forked, the new algo is CN Haven, asic&NH proof
|
|
|
|
JCE-Miner (OP)
Member
Offline
Activity: 350
Merit: 22
|
|
June 18, 2018, 07:36:07 PM |
|
In most case such an error comes from a fork I enabled in advance, like XTL but here that's a real bug, Heavy algorithm is used in place of Haven when multihash is used. gracefully multihash is rarely used with Heavy/Haven. a hotfix will come soon.
|
|
|
|
Larvitar
Jr. Member
Offline
Activity: 196
Merit: 1
|
|
June 19, 2018, 03:03:23 AM |
|
so the no-cache is bad on the A10. It's very good on modern Intel, decent on Ryzen and bad on Core2. And so, bad on A10 too. So keep the config you had found before, with two normal threads. I won't have time to fix the A10 autoconfig in next version, will be for 0.30 edit: i'm polishing the GPU version. I see very strange results on my Pitcairn rig (five different Pitcairns : 7870 and 7850, 1G and 2G...) On my 2G, when I use all the memory available, I can reach 526h/s on my 7870 2G, and 495 on CN-v7 on my 7850 2G, which is far above what Claymore 9.7 gave on CN-Classic. It's even the best score ever I reached, period. Maybe i didn't configure SRB well, but so far i didn't reach such an hashrate. But on my little 7850 1G, that's terrible, i peak at 343 where Claymore gave a whooping 450 ! It looks like my code is faster when using the double-memory, but a lot slower otherwise. And double-memory is barely possible on 1G cards. It works on Bonaire 1G because it has a little GPU that can be saturated with two threads using ~512M each, but my Pitcairn 1G are underused by an order of magniture... The total hashrate of the rig, counting the 2G that mine faster and the 1G that mine slower, is 10% lower that claymore 11.3, but with 0 bad share, where Claymore, as I said previously, gives ~10% bad shares on old GPU when configured at max, or is stable at 10% lower hashrate. 0.29c onlineUse of the GPU version frontend, but GPU disabled Support of Masari fork --variation 11 Support of Haven fork --variation 12 Stellite defaults to XTL variation ETN defaults to v7 other coins that have forked will default to their respective fork Coin list in alphabetical order new coin : BXB new coin : WOWnero -b parameter ignored for compatibility with AwsomeMiner ditto with parameter --api-remote I have to news, a good and a bad one: The good: JCE is lauching normally by Awesome as CPUMiner-Opt. The bad: API isn't working with Awesome: 18/06/2018 17:26:35.007 [022] [E]Failed to process API request (time: 1001 ms): summary Nenhuma conexão pôde ser feita porque a máquina de destino as recusou ativamente 127.0.0.1:4034 18/06/2018 17:26:41.239 [023] [E]Failed to process API request (time: 1001 ms): summary Nenhuma conexão pôde ser feita porque a máquina de destino as recusou ativamente 127.0.0.1:4034 18/06/2018 17:26:47.440 [027] [E]Failed to process API request (time: 1000 ms): summary Nenhuma conexão pôde ser feita porque a máquina de destino as recusou ativamente 127.0.0.1:4034 18/06/2018 17:26:53.655 [041] [E]Failed to process API request (time: 1000 ms): summary Nenhuma conexão pôde ser feita porque a máquina de destino as recusou ativamente 127.0.0.1:4034
|
|
|
|
JCE-Miner (OP)
Member
Offline
Activity: 350
Merit: 22
|
|
June 19, 2018, 05:23:52 AM |
|
sounds logical, jce api is either its own, or xmr-stak compatible, but not cpumineropt-compatible. Maybe ask Awsomeminer to support jce, it should be simple since it has command line similar to cpuminerot (and to claymore) and api just like stak. i got bad report about 0.29c algo selection, haven is broken, bixbite seems too, need to hot release a fix.
|
|
|
|
gvb
Jr. Member
Offline
Activity: 140
Merit: 9
|
|
June 19, 2018, 06:47:21 AM |
|
JCE,
is it possible to add a max hashspeed storing for the current session and display this in the 'r' report and the light blue/cyan line that reports hash speeds every now and then?
it's easier to know the max speed this way.
|
|
|
|
|