[XMR] JCE Miner Cryptonight/forks, now with GPU! - page 108.

JCE-Miner

member

Activity: 350

Merit: 22

thanks, that's very surprising to me !

first i see the normal speed of one core is about 29-30 on both miners, that's what i pretend : jce on par on intel with aes.
but... the more threads you add the more speed you get. jce max at four threads single hash, by construction, and you get 69, a lot more than the autoconfig default of 2 threads.

stak allows penta-share and with a whooping 15 shares (three penta) you rise to 100 Shocked

i wish i have such a cpu to test. that's so different from the ryzen behavior !

btw dualshare to be released soon, and so you'll be able to do four doublehash but i expect stak still being better with its penta feature. i'll add it to jce but assembly takes a lot more time to write.

tc_

newbie

Activity: 16

Merit: 0

jce here ya go: anymore testing let me know.

xmr-stak HASHRATE REPORT - CPU
| ID | 10s | 60s | 15m | ID | 10s | 60s | 15m |
| 0 | 37.6 | 39.1 | 36.8 | 1 | 28.9 | 30.0 | 30.3 |
| 2 | 29.5 | 30.4 | 30.3 |
Totals (CPU): 96.1 99.5 97.4 H/s
-----------------------------------------------------------------
Totals (ALL): 96.1 99.5 97.4 H/s
Highest: 101.9 H/s
-----------------------------------------------------------------

jce -t 2
22:19:11 | Hashrate Thread 0: 28.62 h/s
22:19:11 | Hashrate Thread 1: 28.99 h/s
22:19:11 | Total: 57.60 h/s

jce -t 3
22:24:11 | Hashrate Thread 0: 11.98 h/s
22:24:11 | Hashrate Thread 1: 23.96 h/s
22:24:11 | Hashrate Thread 2: 25.70 h/s
22:24:11 | Total: 61.62 h/s

jce -t 4
22:20:30 | Hashrate Thread 0: 11.02 h/s
22:20:30 | Hashrate Thread 1: 18.85 h/s
22:20:30 | Hashrate Thread 2: 16.13 h/s
22:20:30 | Hashrate Thread 3: 23.03 h/s
22:20:30 | Total: 69.02 h/s

Iamtutut

full member

Activity: 1120

Merit: 131

Don't bother, you'll make me try different settings and we'll see how it will work. I hope to get 250-300H/s.

JCE-Miner

member

Activity: 350

Merit: 22

you'll get +1h/s about, on all algo including heavy, with simpleshare, and dualshare will be enabled for heavy too, so it won't be an improvement but a whole new feature.
32 bit aes will be improved too, and maybe 32 bits non-aes if my tests are good

it's rarely possible to use all cores with heavy because each share consume 4M of cache.

i don't have a ryzen g yet, no idea how jce works on it. i plan to buy one to test jce gpu.

Iamtutut

full member

Activity: 1120

Merit: 131

Will test in +/- 2 weeks with Ryzen 5 2400G, I will use 3 cores / 6 threads in "heavy duty" mode.

djkice

member

Activity: 99

Merit: 10

will the update help heavy?

yeah i know i cant compare to old CN but is there any optimisations to use all 16 threads for higher hashrate or nah?

JCE-Miner

member

Activity: 350

Merit: 22

thanks, everybody should give the equivalent stak config, it's a lot more understandable.

in short, you cannot yet. max low-power (i call it multi-hash in jce because that's a multihash, not related to power...) is 1 in 0.24d and will be 2 in the upcoming 0.24e
i will raise this limit to 5 or 6 in subsequent releases. if you use any low power >1 current jce cannot win just because it does not support that feature yet. it's not a fair compare btw, as i said in previous post, Hashrate comparison must be done on exact same config.

i plan to release multihash enabled version in a few days, once i did all tests. with max multihash at 2 for now, and there will be a light speed increase on aes version in all cases, even with single-hash, on both 32 and 64 bits
i now reach 504 on my ryzen on CN-v7, while 0.24d peaks at 502

edit:
https://ark.intel.com/products/76616/Intel-Core-i7-4600U-Processor-4M-Cache-up-to-3_30-GHz

a 4MB cache cpu... strange you want to use three x5 thread... i would expect the best config be two threads on core 0 and 2 (or 1 and 3) in x1 mode.
try run jce with no config file, with --auto -t N where N from 2 to 4. i'd be interrested into what N is the best, and what Hashrate you get in each case.

tc_

newbie

Activity: 16

Merit: 0

Quote from: JCE-Miner on April 26, 2018, 01:14:56 AM

yes it supports monero v7, i myself mine with it right now

params for monero v7 :

-o POOL:PORT -u WALLET -p x --variation 3 --any --auto --low

example:
-o xmrpool.eu:3333 -p x -u WALLET -p x --variation 3 --any --auto --low

Where WALLET is your wallet of course. the trick is the --any param to use.
what coin did you want to mine ? on what pool ?

Found out that my wallet (copy and paste) had a unknown character at the end, corrected it. it does work.
with XMR-STAK in windows I get around 90-100, but with JCE i get much lower. I will post my CPU config and maybe i am doing something wrong.
but I do not let it mine using xmr-stak 243 my cpu CPU 0 is not affiliated with it, just 1,2,3 and set to 5x thread.
Not sure how i can do the same with jce.

update: this is what my xmr-stak cou set to: 70-90. highest 110.hashrate

"cpu_threads_conf" :
[
{ "low_power_mode" : 5, "no_prefetch" : true, "affine_to_cpu" : 1 },
{ "low_power_mode" : 5, "no_prefetch" : true, "affine_to_cpu" : 2 },
{ "low_power_mode" : 5, "no_prefetch" : true, "affine_to_cpu" : 3 },

],

how do i set to same with jce config, i7-4600u.

JCE-Miner

member

Activity: 350

Merit: 22

you mean old cryptonight ?
you cannot compare cryptonight classic and cryptonight heavy, the second is something like twice slower, on purpose, to annoy the ASIC. comparison must be made on same algo, same computer, same number of threads. on my ryzen 1600 stak and xmrig are at 250, jce is 252, so slightly faster but very little difference i admit. we're at hardware max both.

edit : i redid the test just for you:

Quote from: djkice on April 25, 2018, 03:28:05 AM

getting 243.5h/s on cryptonight-heavy using Ryzen 1700X 3.84Gh/s

On my Ryzen 1600, stock, no overclock i get :

Code:

Cryptonight Variation: Cryptonight-Heavy

Starting Mining thread 0, affinity: CPU 0
Thread 0 successfully bound to CPU 0
Allocated shared Large Page at: 00000148e7200000
Allocated 4MB Cached Large Page Scratchpad Buffer for CPU 0 of NUMA node 0 at: 00000148e7400000
Starting Mining thread 1, affinity: CPU 2
Thread 1 successfully bound to CPU 2
Allocated 4MB Cached Large Page Scratchpad Buffer for CPU 2 of NUMA node 0 at: 00000148e7800000
Starting Mining thread 2, affinity: CPU 6
Thread 2 successfully bound to CPU 6
Allocated 4MB Cached Large Page Scratchpad Buffer for CPU 6 of NUMA node 0 at: 00000148e7c00000
Starting Mining thread 3, affinity: CPU 8
Thread 3 successfully bound to CPU 8
Allocated 4MB Cached Large Page Scratchpad Buffer for CPU 8 of NUMA node 0 at: 00000148e8000000
Devfee is 1.5%

20:12:36 | Connecting to mining pool pool.sumokoin.hashvault.pro:3333 ...
20:12:36 | Sumokoin (SUMO) Mining session starts!

During mining time, press:
 h    display hashrate for each mining thread.
 r    display full report.
 q    quit.

20:12:36 | Connected to pool. Now logging in...
20:12:36 | Successfuly logged as Sumoo4dRGo3cAZPajai8WbgfUZLatPKrqdsFHmfQ1vvrZrQEhSVPKMK3Jj3Qv16BgGQupDw6vEHh1jJAzYXkNScETFrNFUf354v
20:12:36 | Pool changes Difficulty to 10000.
20:12:43 | Hashrate Thread 0: 62.59 h/s
20:12:43 | Hashrate Thread 1: 62.59 h/s
20:12:43 | Hashrate Thread 2: 63.33 h/s
20:12:43 | Hashrate Thread 3: 63.31 h/s
20:12:43 | Total: 250.81 h/s

(note: the wallet is the CN-Heavy dev donation wallet, clear on purpose)
You cannot have worse perf on a OC Ryzen 1700X than my cheap stock 1600. There's a problem, maybe you run the 32-bits version?

Test on stack, exact same machine:

Code:

[2018-04-26 20:19:00] : Mining coin: sumokoin
[2018-04-26 20:19:00] : Starting 1x thread, affinity: 0.
[2018-04-26 20:19:00] : hwloc: memory pinned
[2018-04-26 20:19:00] : Starting 1x thread, affinity: 4.
[2018-04-26 20:19:00] : hwloc: memory pinned
[2018-04-26 20:19:00] : Starting 1x thread, affinity: 6.
[2018-04-26 20:19:00] : hwloc: memory pinned
[2018-04-26 20:19:00] : Starting 1x thread, affinity: 10.
[2018-04-26 20:19:00] : hwloc: memory pinned
[2018-04-26 20:19:00] : Fast-connecting to pool.sumokoin.hashvault.pro:3333 pool ...
[2018-04-26 20:19:00] : Pool pool.sumokoin.hashvault.pro:3333 connected. Logging in...
[2018-04-26 20:19:01] : Difficulty changed. Now: 10000.
[2018-04-26 20:19:01] : Pool logged in.
HASHRATE REPORT - CPU
| ID |    10s |    60s |    15m | ID |    10s |    60s |    15m |
|  0 |   62.3 |   (na) |   (na) |  1 |   62.3 |   (na) |   (na) |
|  2 |   62.1 |   (na) |   (na) |  3 |   62.1 |   (na) |   (na) |
Totals (CPU):   248.7    0.0    0.0 H/s

jce is +2h/s as expected. Really JCE is slightly (but barely) faster. Not 20% slower.

djkice

member

Activity: 99

Merit: 10

hopefully new version is bit better on heavy xD

I was getting around 640 on stack old cryptonight algo

JCE-Miner

member

Activity: 350

Merit: 22

yes it supports monero v7, i myself mine with it right now

params for monero v7 :

-o POOL:PORT -u WALLET -p x --variation 3 --any --auto --low

example:
-o xmrpool.eu:3333 -p x -u WALLET -p x --variation 3 --any --auto --low

Where WALLET is your wallet of course. the trick is the --any param to use.
what coin did you want to mine ? on what pool ?

Iamtutut

full member

Activity: 1120

Merit: 131

I does support V7. I mine stellite (XTL) with.

tc_

newbie

Activity: 16

Merit: 0

hey i am not sure what i am doing wrong, i took all my info from xmr-stak.

its mining monero-v7

does this product support monero-v7?

I get this error: Wallet does not match any of the supported currencies.
then it lists in red all the coins. so i take it that this program doesn't support monero-v7.
and my wallet it the same copy/pasted from xmr-stak pools.txt file i have (it works its mining right now).

aGeoM

newbie

Activity: 43

Merit: 0

Quote from: JCE-Miner on April 25, 2018, 12:30:33 PM

https://www.amd.com/en/products/cpu/amd-ryzen-7-1700x

close to my Rysen 1600, and yeah max perf is --auto -t 4 --archi ryzen
Heavy is the algo where, honnestly, i beat xmrig by the lowest factor, i'm less than 1% better. on 64 bits aes. on 32 bits and non aes i obliterate it. Cool

edit: all code done for dualshare, now entering test phase.
perfs on one thread, with compared to without doublehash

core2 64 bits : +5%
core2 32 bits +11%
ryzen 64 bits : +92%
ryzen 32 bits +99%

The ryzen pipelining is impressive... Shocked

I can not wait to test the new version Cheesy

JCE-Miner

member

Activity: 350

Merit: 22

https://www.amd.com/en/products/cpu/amd-ryzen-7-1700x

close to my Rysen 1600, and yeah max perf is --auto -t 4 --archi ryzen
Heavy is the algo where, honnestly, i beat xmrig by the lowest factor, i'm less than 1% better. on 64 bits aes. on 32 bits and non aes i obliterate it. Cool

edit: all code done for dualshare, now entering test phase.
perfs on one thread, with compared to without doublehash

core2 64 bits : +5%
core2 32 bits +11%
ryzen 64 bits : +92%
ryzen 32 bits +99%

The ryzen pipelining is impressive... Shocked

djkice

member

Activity: 99

Merit: 10

getting 243.5h/s on cryptonight-heavy using Ryzen 1700X 3.84Gh/s

Any tweaks I can do to make it better?

Running on 4 threads, can I not run 8 or is it not benetifical ?

edit:

tried -t 8
16:32:32 | Hashrate Thread 0: 13.58 h/s
16:32:32 | Hashrate Thread 1: 14.47 h/s
16:32:32 | Hashrate Thread 2: 14.47 h/s
16:32:32 | Hashrate Thread 3: 14.37 h/s
16:32:32 | Hashrate Thread 4: 19.07 h/s
16:32:32 | Hashrate Thread 5: 19.12 h/s
16:32:32 | Hashrate Thread 6: 60.23 h/s
16:32:32 | Hashrate Thread 7: 60.13 h/s
16:32:32 | Total: 215.41 h/s

JCE-Miner

member

Activity: 350

Merit: 22

impressive chip, do jce detect L4 at startup ? i coded it but could never try Cry

Lonnegan64

jr. member

Activity: 37

Merit: 5

Quote from: JCE-Miner on April 24, 2018, 12:21:10 PM

woow, what's that beast ? a knight landing ?

For sure an Intel Broadwell with 128 MB eDRAM, used as L4 cache:
http://www.cpu-world.com/CPUs/Core_i7/Intel-Core%20i7-5775C.html

JCE-Miner

member

Activity: 350

Merit: 22

cpu temp : possible but cannot be high priority
Nicehash : nope dev fees are not to nicehash. i've nicehash specific netcode that should not be enabled if you don't use nicehash, strange. i'll proof read the netcode there may be a side effect somewhere.

what version do you use and what coin do you mine?

finished almost the 32 bits doublehash and it's better than expected, my core2 jumps from 23.2 to 26 on one core.

edit : checked and that's a legit message from pool, not devfees, maybe your pool redirect to nicehash ?
give me your pool and coin please so i can try. on the opposite, run jce with --nicehash param to see if it fixes.

Claymore has nicehash mode always enabled, maybe he's right to do so...

griffoncoin

newbie

Activity: 37

Merit: 0

Hi

I am getting "Invalid nonce; is miner not compatible with NiceHash?" and I am not mining at nicehash, is dev fees on nicehash and its not submitting and mining stays there for some time then goes back to mining on my pool

---
21:21:36 | Hashrate Thread 3: 55.27 h/s
21:21:36 | Total: 227.06 h/s
21:21:57 | Pool sends a new Job.
21:21:59 | Thread 2 finds a Share, value 120001
21:21:59 | Rejected by the pool.
21:21:59 | Message from the pool: Invalid nonce; is miner not compatible with NiceHash?
21:22:20 | Pool sends a new Job.
21:22:26 | Pool sends a new Job.
---

Topic: [XMR] JCE Miner Cryptonight/forks, now with GPU! - page 108. (Read 90841 times)