[JCE]Fast & stable CN/v8/Heavy/Tube/XHV miner, CPU+GPU, Vega56 1800+ RX580 1200+

[JCE-Miner]

forget the b7 and try the new experimental (read: very untested) B-sync GPU
That's the lone .exe in the github page.

1. Stable and fast hashrate
2. new parameter --no-monitor to disable fan/temp monitoring (the purple report)
3. various optims for some cards and some algos, expect about +0.5%
4. Fees still 0.9%

so far it burns like fire on my rigs, i hope it really fixes all the hashrate and warmup problems.

[lebuawu2]

forget the b7 and try the new experimental (read: very untested) B-sync GPU
That's the lone .exe in the github page.

1. Stable and fast hashrate
2. new parameter --no-monitor to disable fan/temp monitoring (the purple report)
3. various optims for some cards and some algos, expect about +0.5%
4. Fees still 0.9%

so far it burns like fire on my rigs, i hope it really fixes all the hashrate and warmup problems.

testing about 5 min, from start until 5 min 8 GPU RX 580 8GB always reach max hashrate, really great.

Edit: testing with --no-warmup param also give max hashrate, so no warm up needed anymore.

[UnclWish]

forget the b7 and try the new experimental (read: very untested) B-sync GPU
That's the lone .exe in the github page.

1. Stable and fast hashrate
2. new parameter --no-monitor to disable fan/temp monitoring (the purple report)
3. various optims for some cards and some algos, expect about +0.5%
4. Fees still 0.9%

so far it burns like fire on my rigs, i hope it really fixes all the hashrate and warmup problems.

Thanks! Now testing new version.

EDIT: It's strange, but one of my RX 580 8Gb just stops sometime to mine. Speed is 0, voltage and clocks on minimum state. It's not stuck, just stopped...

Warmup phase is still needed. On my 580 cards max speed didn't archived after start...

[JCE-Miner]

good, this time i'm on the good way, and that's probably the same as srb found. i released first so no question whether i hacked it.
i think i found the locked thread bug, i'll release a fix soon, i wait for more feedback from other users (unless i really have three).

[Iamtutut]

forget the b7 and try the new experimental (read: very untested) B-sync GPU
That's the lone .exe in the github page.

1. Stable and fast hashrate
2. new parameter --no-monitor to disable fan/temp monitoring (the purple report)
3. various optims for some cards and some algos, expect about +0.5%
4. Fees still 0.9%

so far it burns like fire on my rigs, i hope it really fixes all the hashrate and warmup problems.

Oh F*CK THAT'S FAST !!!!!!

Declared HR on start is almost 3700H/s, with B7 and very same parameters I have 3450H/s after full warm-up (max was 3550 just for once)

I'll post my real HR when I reach the 1st hour of mining.

Could your next step be trying to lower the power drawn on CN V8 ? It's 100W more than with other algos, which makes this algo the more expensive by far (can't talk about profitability right now...).

[UnclWish]

good, this time i'm on the good way, and that's probably the same as srb found. i released first so no question whether i hacked it.
i think i found the locked thread bug, i'll release a fix soon, i wait for more feedback from other users (unless i really have three).

Yes, one of GPU stops to mine... 3rd time it happens...

[JCE-Miner]

I'll plug my last HD6950 into a real rig on Win 8.1 to be able to test, i got too many bad feedbacks about the HD6000 mining, and i want to support them, even if i'll probably get negligible fees from them

the no-monitor option is still good to have, and was easy to add.

I'll post my real HR when I reach the 1st hour of mining.

it should be lower (stale shares and fees) and 1 hour is too short to have the HR converge to its mathematical average, unless you use very small diff like 1000. But the b-sync provided extra speed in all my tests, even on older cards. I reached 480+ (v8) and 430+ (heavy) on my HD7950 for the first time. On CN-classic just for fun, i got 572, just a little under Claymore v9.7 which gave 578. It took me one year but I finally almost reached back its speed. Too late.

[Iamtutut]

I'll plug my last HD6950 into a real rig on Win 8.1 to be able to test, i got too many bad feedbacks about the HD6000 mining, and i want to support them, even if i'll probably get negligible fees from them

the no-monitor option is still good to have, and was easy to add.

I'll post my real HR when I reach the 1st hour of mining.

it should be lower (stale shares and fees) and 1 hour is too short to have the HR converge to its mathematical average, unless you use very small diff like 1000. But the b-sync provided extra speed in all my tests, even on older cards. I reached 480+ (v8) and 430+ (heavy) on my HD7950 for the first time. On CN-classic just for fun, i got 572, just a little under Claymore v9.7 which gave 578. It took me one year but I finally almost reached back its speed. Too late.

Are we supposed to keep the same instensity ?

[UnclWish]

On my 580 8Gb (Hynix) b-sync didn't give stable max speed even after 20-30 minutes after start. Speed became more stable but not max at least at one GPU. On 1st 5-10 minutes during warmup speed sometimes max sometime not... But I can't test more than 30-40 minutes because 1 of GPU's stops mine at all...
Returning to b7 or maybe b4 - it gives lower maxspeed but higher effective hashrate - it's better than all other versions I think...
Will try to do more tests on b7 before returning to b4...

P.S. Still can't find optimal multihash for my 270X 4Gb (3,5Gb OpenCL) on heavy algo. Max I can use is 2 threads with 1 - 448 and 2nd - 432. But speed is always fluctuate from 410 to 500 h/s. If I use 2x432 m.h. speed is almost 415-420 h/s. Is there some way to stabilize speed on 2 threads with different multi_hash?

[fluxy12]

Hello JCE,

I think you finally did it, i'm testing the experimental and it is stable on my first rx580 8g at max speed (20mn at the moment).

I'll test it on other rigs now.

Good job !

[fluxy12]

On my 580 8Gb (Hynix) b-sync didn't give stable max speed even after 20-30 minutes after start. Speed became more stable but not max at least at one GPU. On 1st 5-10 minutes during warmup speed sometimes max sometime not... But I can't test more than 30-40 minutes because 1 of GPU's stops mine at all...
Returning to b7 or maybe b4 - it gives lower maxspeed but higher effective hashrate - it's better than all other versions I think...
Will try to do more tests on b7 before returning to b4...

P.S. Still can't find optimal multihash for my 270X 4Gb (3,5Gb OpenCL) on heavy algo. Max I can use is 2 threads with 1 - 448 and 2nd - 432. But speed is always fluctuate from 410 to 500 h/s. If I use 2x432 m.h. speed is almost 415-420 h/s. Is there some way to stabilize speed on 2 threads with different multi_hash?

On 580Gb did you try to upgrade the Wattage ? on one on my rig i had to up from 910mv/cv to 930mv/cv

[JCE-Miner]

thanks, i think i finally, after some failed attempts with the warmup and the hybrid, found a way to stabilize the hashrate. Sad i haven't found it earlier when the b5 came out, i would have taken the lead by an order of magnitude

Online is the 0.33-sync GPU (note i removed the b) with a fix for stuck GPU threads, and a little optim for RX560, mostly for Unclwish. Otherwise this is the same.

next task: plug back my HD6950 and do new tests on it.

[Iamtutut]

4th trial with the new experimental version. >3h30mn mining.

Still 4X RX574 with bios mod, intensity set to 464 for all threads.

Power draw at the wall: 470W

Max HR: 3709H/s.
Usual is between 3700 and 3707H/S.
Actual reported HR by the miner: 3645,45H/s
HR seen by the pool:
- last 15mn: 3,6KH/s
- last 2Hs: 3,85KH/s


That's quite amazing, just think that 3 months ago to reach such HR would have required a Vega 64 FE + a good Vega 56.

[BS0D]

I've noticed significant hash drop since 33b6 at cnFast. Didn't test others. Maybe normal/lite algos on V64 too.

[JCE-Miner]

possible, i do very little tests on older algos like Fast or Turtle. When i plug my HD6950 i'll redo some retro-tests on those algos. Otherwise i just mine a few Stellite just to check it still works.

[_ap_]

forget the b7 and try the new experimental (read: very untested) B-sync GPU
That's the lone .exe in the github page.

1. Stable and fast hashrate
2. new parameter --no-monitor to disable fan/temp monitoring (the purple report)
3. various optims for some cards and some algos, expect about +0.5%
4. Fees still 0.9%

so far it burns like fire on my rigs, i hope it really fixes all the hashrate and warmup problems.

with:
- cpu: working perfect
- integral gpu: working perfect
- 8gb rx 580: working perfect
- 4gb rx570: after pool login, the miner stopped and i got: "Press any key to continue..." message

my config: "mode" : "GPU", "worksize" : 8, "alpha" : 64, "beta" : 16, "gamma" : 4, "delta" : 4, "epsilon" : 4, "zeta":4, "index" : 3, "multi_hash":944 },

any idea?

*update: algo = bittube

thanks!

[lebuawu2]

forget the b7 and try the new experimental (read: very untested) B-sync GPU
That's the lone .exe in the github page.

1. Stable and fast hashrate
2. new parameter --no-monitor to disable fan/temp monitoring (the purple report)
3. various optims for some cards and some algos, expect about +0.5%
4. Fees still 0.9%

so far it burns like fire on my rigs, i hope it really fixes all the hashrate and warmup problems.

with:
- cpu: working perfect
- integral gpu: working perfect
- 8gb rx 580: working perfect
- 4gb rx570: after pool login, the miner stopped and i got: "Press any key to continue..." message

my config: "mode" : "GPU", "worksize" : 8, "alpha" : 64, "beta" : 16, "gamma" : 4, "delta" : 4, "epsilon" : 4, "zeta":4, "index" : 3, "multi_hash":944 },

any idea?

*update: algo = bittube

thanks!

4gb GPU multi_hash should be lower for 2 thread about 464 or 480.

[UnclWish]

forget the b7 and try the new experimental (read: very untested) B-sync GPU
That's the lone .exe in the github page.

1. Stable and fast hashrate
2. new parameter --no-monitor to disable fan/temp monitoring (the purple report)
3. various optims for some cards and some algos, expect about +0.5%
4. Fees still 0.9%

so far it burns like fire on my rigs, i hope it really fixes all the hashrate and warmup problems.

with:
- cpu: working perfect
- integral gpu: working perfect
- 8gb rx 580: working perfect
- 4gb rx570: after pool login, the miner stopped and i got: "Press any key to continue..." message

my config: "mode" : "GPU", "worksize" : 8, "alpha" : 64, "beta" : 16, "gamma" : 4, "delta" : 4, "epsilon" : 4, "zeta":4, "index" : 3, "multi_hash":944 },

any idea?

*update: algo = bittube

thanks!

4gb GPU multi_hash should be lower for 2 thread about 464 or 480.

Maybe he use 1 thread? He didn't copy 2 strings...

[_ap_]

forget the b7 and try the new experimental (read: very untested) B-sync GPU
That's the lone .exe in the github page.

1. Stable and fast hashrate
2. new parameter --no-monitor to disable fan/temp monitoring (the purple report)
3. various optims for some cards and some algos, expect about +0.5%
4. Fees still 0.9%

so far it burns like fire on my rigs, i hope it really fixes all the hashrate and warmup problems.

with:
- cpu: working perfect
- integral gpu: working perfect
- 8gb rx 580: working perfect
- 4gb rx570: after pool login, the miner stopped and i got: "Press any key to continue..." message

my config: "mode" : "GPU", "worksize" : 8, "alpha" : 64, "beta" : 16, "gamma" : 4, "delta" : 4, "epsilon" : 4, "zeta":4, "index" : 3, "multi_hash":944 },

any idea?

*update: algo = bittube

thanks!

4gb GPU multi_hash should be lower for 2 thread about 464 or 480.

Maybe he use 1 thread? He didn't copy 2 strings...

yes, only 1 thread, and this config worked with all previous version.

[UnclWish]

forget the b7 and try the new experimental (read: very untested) B-sync GPU
That's the lone .exe in the github page.

1. Stable and fast hashrate
2. new parameter --no-monitor to disable fan/temp monitoring (the purple report)
3. various optims for some cards and some algos, expect about +0.5%
4. Fees still 0.9%

so far it burns like fire on my rigs, i hope it really fixes all the hashrate and warmup problems.

with:
- cpu: working perfect
- integral gpu: working perfect
- 8gb rx 580: working perfect
- 4gb rx570: after pool login, the miner stopped and i got: "Press any key to continue..." message

my config: "mode" : "GPU", "worksize" : 8, "alpha" : 64, "beta" : 16, "gamma" : 4, "delta" : 4, "epsilon" : 4, "zeta":4, "index" : 3, "multi_hash":944 },

any idea?

*update: algo = bittube

thanks!

4gb GPU multi_hash should be lower for 2 thread about 464 or 480.

Maybe he use 1 thread? He didn't copy 2 strings...

yes, only 1 thread, and this config worked with all previous version.

Try with 928.