SRBMiner Cryptonight AMD GPU Miner V1.9.3 - native algo switching

[bruslie]

is it important worksize in config ?

i tried 8 and 16 and 12
i think no any effect. any idea ?

yes me, blockchain driver is RIP about 1200 hash / s on vega 64

[1714896836]

1714896836

[1714896836]

1714896836

[1714896836]

1714896836

[doktor83]

Guys, you won't get back your V7 hashrate, this algo gives less hashrate and consumes more power.
The most affected cards are Vega's.

[doktor83]

heyho, Im new to SRBMiner since today. All my Vega Rigs are running fine so far...except one

Im trying to give each card manual settings, however it doesnt seem to work -> always get parseError. These are the settings im trying to use:

"gpu_conf" :
[
   { "id" : 0, "intensity" : 120, "worksize" : 16, "threads" : 2},
   { "id" : 1, "intensity" : 120, "worksize" : 16, "threads" : 2},
   { "id" : 2, "intensity" : 120, "worksize" : 16, "threads" : 2},
   { "id" : 3, "intensity" : 120, "worksize" : 16, "threads" : 2},
   { "id" : 4, "intensity" : 120, "worksize" : 16, "threads" : 2},
   { "id" : 5, "intensity" : 120, "worksize" : 16, "threads" : 2},
   { "id" : 6, "intensity" : 120, "worksize" : 16, "threads" : 2},
   { "id" : 7, "intensity" : 120, "worksize" : 16, "threads" : 2},
   { "id" : 8, "intensity" : 120, "worksize" : 16, "threads" : 2}
]

what am I doing wrong?

Thanks in advance

Post the whole config, this part looks OK

[RuMiner]

I have 6x vega 56 on v7 11500M/H with worksize8
now on v8 with worksize16, inetnsity default 0
1000M/H less.

I try to play with intensity but ....didnot manage to get back this 1000M/H.

I got about 1000 h/s on old 18.3.4 drivers. updated to 18.6.1 and everything went ok. worksize 16, intens 112 for vega56, 124 for vega64

one rig has a problem after drivers updating from 18.3.4 ( initializing opencl devices and nothing happens on 18.5.1 and 18.6.1 (

[djim123]

I have 6x vega 56 on v7 11500M/H with worksize8
now on v8 with worksize16, inetnsity default 0
1000M/H less.

I try to play with intensity but ....didnot manage to get back this 1000M/H.

I got about 1000 h/s on old 18.3.4 drivers. updated to 18.6.1 and everything went ok. worksize 16, intens 112 for vega56, 124 for vega64

one rig has a problem after drivers updating ( initializing opencl devices and nothing happens (

I will try to re install driver,using 18.6.1 now.

For now using defaults I:120 W:16 double treads.

P7:1407 P61212,Mem 900 mhz

mv 910 ,905, 900

and rig is using more power 1180w befor 1020

[wgestickley]

Hey Guys,

i am just setting up a new RIG and got an error i had never before...

I run 18.5.1 Driver with Windows 10 (last updated) and last SRB Version

When i start the Miner, the Wndows opens and the text says, Initialising open CL driver/cards.... but than nothing happens.... verry strange...

Any ideas?

THX

Oliver

[oxzhor]

I have 6x vega 56 on v7 11500M/H with worksize8
now on v8 with worksize16, inetnsity default 0
1000M/H less.

I try to play with intensity but ....didnot manage to get back this 1000M/H.

I got about 1000 h/s on old 18.3.4 drivers. updated to 18.6.1 and everything went ok. worksize 16, intens 112 for vega56, 124 for vega64

one rig has a problem after drivers updating from 18.3.4 ( initializing opencl devices and nothing happens on 18.5.1 and 18.6.1 (

Can you post your string?

[camchimonline]

V8 more power than v7, hash drop down than v7

[oxzhor]

V8 more power than v7, hash drop down than v7

Hmm see the same here it use 140watt more ...

[vitalyoff]

RX550 2G
"intensity" : 27, "worksize" : 32, "threads" : 2 = 447 h/s (510 v7)

RX560 2G
"intensity" : 27, "worksize" : 16, "threads" : 2 = 463 h/s (485 v7)

RX560 less hashrate drop than RX550

[RuMiner]

Can you post your string?

which one?

[oxzhor]

Can you post your string?

which one?

start string

[x3t9fi]

2x vega64 from about 4100H/s v7 to about 2900H/s v8 -.-"

[RuMiner]

start string

nothing special here...
%~dp0SRBMiner-CN.exe --config config.txt --pools pools.txt --logfile logs\%date%-%random%.log --apienable --apiport 21555 --apirigname rig09

2x vega64 from about 4100H/s v7 to about 2900H/s v8 -.-"

use 18.6.1, clean cache folder and set intens 124 worksize 16. my 64s dropped only 2%!

[oxzhor]

start string

nothing special here...
%~dp0SRBMiner-CN.exe --config config.txt --pools pools.txt --logfile logs\%date%-%random%.log --apienable --apiport 21555 --apirigname rig09

2x vega64 from about 4100H/s v7 to about 2900H/s v8 -.-"

use 18.6.1, clean cache folder and set intens 124 worksize 16. my 64s dropped only 2%!

Where i need to set intens 124 worksize 16 exaclly?

[RuMiner]

Where i need to set intens 124 worksize 16 exaclly?

in gpu config

Code:

{
"target_temperature" : 47,
"shutdown_temperature" : 85,
"restart_devices_on_startup" : false,
"reboot_script_gpu_watchdog" : "reboot.bat",
"min_rig_speed" : 16000,
"gpu_conf" : 
[ 
	{ "id" : 0, "intensity" : 124, "worksize" : 16, "threads" : 2, "persistent_memory" : false},
	{ "id" : 1, "intensity" : 124, "worksize" : 16, "threads" : 2, "persistent_memory" : false},
	{ "id" : 2, "intensity" : 124, "worksize" : 16, "threads" : 2, "persistent_memory" : false},
	{ "id" : 3, "intensity" : 124, "worksize" : 16, "threads" : 2, "persistent_memory" : false},
	{ "id" : 4, "intensity" : 124, "worksize" : 16, "threads" : 2, "persistent_memory" : false},
	{ "id" : 5, "intensity" : 124, "worksize" : 16, "threads" : 2, "persistent_memory" : false},
	{ "id" : 6, "intensity" : 124, "worksize" : 16, "threads" : 2, "persistent_memory" : false},
	{ "id" : 7, "intensity" : 124, "worksize" : 16, "threads" : 2, "persistent_memory" : false},

]
}

[camchimonline]

What miner for v9?

[BillieCrypt]

After softfork 8.0 on Monero network, this miner became one of the best for AMD cards. Unfortunately, it does not support cards of the R9 280 generation. It would be nice to add support for these old, but rather productive video cards.

[abg00]

some card crash on v8 1.6.8
use cast_xmr. low rate but stable work not crash. only rx470+

[tvukoman]

So people, V8 is here and are pain in the a..  

Vega 56 reference is max affected with change because new algo is max memory intensive and produce even more heat on poor Vega 56 reference design.

See first test of my "rig" with two Vega 56 ref, one Vega 64 Strix, and one RX570 4GB:

Power consumption for rig is 15% more, from 680W to 770W
Hasrate of the rig is down with same configuration: V7: 7170 H/s ; V8: 6870 H/s

But... temperature for Vega 56 is increased from 60 C to 67 C ;-) my Vegas 56 are throttling again and in V7 is super stable...

For now I tried to lower CC from 1442 to 1407 and in configuration of Vega 56 lower intensity from 120 / 16 / 2 to 112 / 16 / 2. Result is lower H/s for about 40 H/s for vega 56, but stability is improved for now:

GPU:4 Threads:8   H/s   CC   MC   °C   RPM Radeon RX Vega Bus:3   1938   1407   1095   63   3459 Radeon RX 580 Series Bus:6   898   1150   1980   51   1739 Radeon RX Vega Bus:11   2032   1442   1100   58   3484 Radeon RX Vega Bus:14   1936   1407   1095   66   3532

So lets all test our Vega-s, post results and we will find best solution for all...

And yes, Rx 570 4GB Nitro, drop for just 4 H/s from 902 to 898, so developers of Monero have older cards and make new algo optimized for Polaris