Topic: SRBMiner Cryptonight AMD GPU Miner V1.9.3 - native algo switching - page 74. (Read 237247 times)

read two previous page. webchain work very poor

2 gb and 4 gb different setings

Recommended drivers:
Adrenalin 18.6.1

Hello doctor!

First of all, I would like to express a big thank you for your effort and work on SRB Miner!
It's going really well, I've been working for a long time with SRB Miner and my Vega64 graphics cards, currently 12 pieces in number.

My question currently is, still working or is recommended at most the ADRENALIN driver 18.6.1 to use.

What about the newer drivers? Do these work? Or do you continue to work only on 18.6.1? Should we try it for ourselves?

The 2nd thing:

What bothers me a bit or is upset that with SRB 1.7.0 the power consumption of Vega64 cards was about ~ 185 watts per card!

With SRB 1.7.5. The consumption per card increased approximately to 203 watts / card. Could you perhaps also consider this or perhaps mention it in the information?

Because it brings nothing if graphics cards make 10-30 H / s more, but the power consumption increases every time!
This is not an optimization!

Personally, I am very surprised that nobody here talks about electricity consumption or reports or writes about it, because it´s a big thing for everyones MONEY!.

Here are the current values of my RIG with Vega64 to version 1.7.5 on V8(GRAFTCOIN) .:

[2019-01-04 19:09:05] GPU0 :    2073 H/s [T: 32c, RPM: 2410, CC: 1408 MHz, MC: 1100 MHz][BUS:28]
[2019-01-04 19:09:05] GPU1 :    2076 H/s [T: 30c, RPM: 2289, CC: 1408 MHz, MC: 1100 MHz][BUS:19]
[2019-01-04 19:09:05] GPU2 :    2076 H/s [T: 27c, RPM: 2380, CC: 1408 MHz, MC: 1100 MHz][BUS:37]
[2019-01-04 19:09:05] GPU3 :    2071 H/s [T: 35c, RPM: 3573, CC: 1408 MHz, MC: 1100 MHz][BUS:34]
[2019-01-04 19:09:05] GPU4 :    2069 H/s [T: 39c, RPM: 3677, CC: 1408 MHz, MC: 1100 MHz][BUS:9]
[2019-01-04 19:09:05] GPU5 :    2077 H/s [T: 32c, RPM: 2377, CC: 1408 MHz, MC: 1100 MHz][BUS:13]
[2019-01-04 19:09:05] GPU6 :    2072 H/s [T: 32c, RPM: 3536, CC: 1408 MHz, MC: 1100 MHz][BUS:31]
[2019-01-04 19:09:05] GPU7 :    2079 H/s [T: 34c, RPM: 2349, CC: 1408 MHz, MC: 1100 MHz][BUS:16]
[2019-01-04 19:09:05] GPU8 :    2069 H/s [T: 39c, RPM: 3603, CC: 1408 MHz, MC: 1100 MHz][BUS:25]
[2019-01-04 19:09:05] GPU9 :    2076 H/s [T: 34c, RPM: 2336, CC: 1408 MHz, MC: 1100 MHz][BUS:40]
[2019-01-04 19:09:05] GPU10:    2072 H/s [T: 32c, RPM: 2375, CC: 1408 MHz, MC: 1100 MHz][BUS:22]
[2019-01-04 19:09:05] GPU11:    2074 H/s [T: 25c, RPM: 2337, CC: 1408 MHz, MC: 1100 MHz][BUS:5]

[2019-01-04 19:09:05] Total:    24884 H/s

it runs stable!



best regards & big thanks!  
Patrick

who test webchain algo ?
What are the best intensity and worksize for Radeon RX550 series gpu ?

Awesome, thanks for reporting.
Also stales shown in miner are not a problem because those are accepted shares.
Its just more a 'cosmetical thing'.

Just tested UPX on :
470 4g
570 4g
580 4g

Driver 18.6.1

Ran it for 30 minutes, no compute errors.


Also 3.3khs is low for 580 4gb, it should do 3.7khs without any crazy overclocking.

Feedback on
V. 1.7.5

Now it finally works when I have mix of 270X and 280X video cards.
Hashrates are very good for normal V8 algo:
280X 450 H/s
270X 345 H/s
290 530H/s

I do not OC.
I also sometimes get lots of STALE shares.

Good job!

DOC How to change UPX Intensity?

I have something like you with V8.
I was using 1.7.0 working fine, after switching to 1.7.4-5 my RX 580 4GB is giving about 12 H/s more than before but full of compute errors.
The 8GB 580 are same hash rate & I think they work fine but not tested carefully.

Try to disable the 4GB card pressing the number where it is placed, if compute errors disappear it is a miner problem.
I don't know if 1.7.0 have the uplex algo, if it have try with this version.

Also I have better H/s with 4GB than with 8GB but it is a very old miner problem, I think is better optimized for 4GB cards.

I tried to reduce cc to 1150 mc 1950, got same errors
I am mining uplex algo.

also with another miner, clocks are ok at 1200 and 1950 with Vcc 940/900 with 800 Watts, getting around 21kh/s

so I wanted to test this miner to see if I can get more.

Strange thing is that one of the RX580 is a  4 Gb card, other 6 are 8Gb, and it's getting 3.3kh/s while the others are around 10% less.

What algorithm do you use? When using the webchain, there are always a lot of mistakes and rejected shares.

Lower clocks

Hello,

I am looking for config for 7 RX580, currently I am getting "warning: Too many compute errors"
thx

I agree, its awesome miner for all algorithms. Except webchain only!

You could use this :


basically make a script that does what you want (nothing)

Thanks for reply - but I think that will just allow the miner to continue mining if 1 GPU crashed and others are OK.

Is there any way to have the miner quit / terminate itself when 1 of the GPUs crashes?

How have you tried it out when those are Nvidias and SRBMiner is AMD only?

Can someone tell me if srbminer works on old GPUs like gtx6*** and gtx5*** series ,I've tried it out but the hashrates seems so lower than xmrstack and xmrig

Hi, guys) Could someone connected to f2pool for HYC? Unfortunatly I can't( The pool rejects connection with SRB miner. Only xmrig((( 

U ungrateful SOB, the Miner is awesome, if u hate it dont use it.
But dont critise the dev and his works on your whims and fancies,

Yes you already said this a few times, thank you
It is too complicated for me , use LOUD miner it is working much better that's what i heard.