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

[Markyz23]

Recommended drivers:
Adrenalin 18.5.1
Adrenalin 18.6.1
Adrenalin 2019

So the reccomended driver is also Adrenaline 2019? Which version?

Are there any new improvements for RX 570/560/550?

[1675135793]

1675135793

[1675135793]

1675135793

[1675135793]

1675135793

[doktor83]

i test it with 18.6.1 it's not submit any shear. (uninstall with ddu )
i use with this --disabletweaking but no shear find.

hash rate is fine but after 30m no shear find .  

Try maybe changing the value for fragments.
What does your config looks like?

Try renaming the cache file from v182 that works for you to 186 like :

g1CNV4w168f4hm3_182.srb to g1CNV4w168f4hm3_186.srb

I just tried and 182 kernel works on 186

[Mashy81]

UPX2 is it a smaller scratchpad again, is it lighter than pico?

[akvilon888]

hi dok, i have some rigs w vegas 64&56, and when i use srb i have 10-20% of rejects 0_o , and with the same overclock settings, nothing changed, using trm i have only ~3% as i can see in stats... cant understand this thing...how it can be, what i'm doing wrong?)
and i can see it by profit, cause i'm using nicehash, and i get smaller profit using srb with that too many rejects...

[abg00]

- Added new algo 'Upx2' , the new Uplexa algo

how to test new upx2 algo  before fork?

upd
test on testnet. +10-12% power because v8 algo based

[clousian]

@doktor - Any reason to upgrade from 1.8.3 to this latest version if I'm using RX4 & 5 series cards?

[Blooregardqcazoo]

@doktor83

Tweaking seems still disabled for VII. Tried from both the config file and the command line, with different values for tweak_profile, but it always says it's disabled.

Paste here the first 10 sec of your log

Just realized (from the log) that I had to run the bat file as an administrator to get tweaking enabled. Did that and it seems to work now. Thanks.

[Mashy81]

- Added new algo 'Upx2' , the new Uplexa algo

how to test new upx2 algo  before fork?

upd
test on testnet. +10-12% power because v8 algo based

Use the benchmark alogs feature in the mining software

[primavera8989]

Hey. On one rig srb issues a message "Job timeout detected, reconnecting the pool". This condition can last for hours and the reconnect does not pass. In this case, the cards are under 100% load , srb writes that the shares, but in fact the pool does not receive them.
Windows 10, SRB 1.8.6, moneroocean
https://i.ibb.co/rxFZdKd/timeout.jpg

[doktor83]

Hey. On one rig srb issues a message "Job timeout detected, reconnecting the pool". This condition can last for hours and the reconnect does not pass. In this case, the cards are under 100% load , srb writes that the shares, but in fact the pool does not receive them.
Windows 10, SRB 1.8.6, moneroocean

You are right, after the big networking code rewrite, i forgot to test every option, and this one isn't working as its supposed to.
Going to fix in next release, thanks for reporting.

As a workaround increase "job_timeout" for example to 1 hour (3600 sec)

[rednoW]

Hi, Dok!
Just examined your timings for profile 10.
They are good for hynix HBM2 but for samsung (both low voltage 800-940 mhz vega56 and high voltage 945-1100 mhz vega64 I have much better (more stable and fast) mem timings)
I can share to you if you want ))

[doktor83]

Hi, Dok!
Just examined your timings for profile 10.
They are good for hynix HBM2 but for samsung (both low voltage 800-940 mhz vega56 and high voltage 945-1100 mhz vega64 I have much better (more stable and fast) mem timings)
I can share to you if you want ))

You can share to help the folks here, im not using predefined straps in SRB. They are calculated based on existing settings.

[rednoW]

Hi, Dok!
Just examined your timings for profile 10.
They are good for hynix HBM2 but for samsung (both low voltage 800-940 mhz vega56 and high voltage 945-1100 mhz vega64 I have much better (more stable and fast) mem timings)
I can share to you if you want ))

You can share to help the folks here, im not using predefined straps in SRB. They are calculated based on existing settings.

Ok, this is for Vega 64 LC

This if for reference vega56, stock bios


win10, 19.4.1 driver

[doktor83]

Ok, this is for Vega 64 LC

This if for reference vega56, stock bios

win10, 19.4.1 driver

What is the default value for RFC on the samsung V56 ?

[livada]

Ok, this is for Vega 64 LC

This if for reference vega56, stock bios

win10, 19.4.1 driver

What is the default value for RFC on the samsung V56 ?

vega56 powercolor(samsung ) with 64 bios:
<values cl="20" ras="33" rcdrd="16" rcdwr="10" rcab="47" rcpb="47" rpab="14" rppb="14" rrds="4" rrdl="6" rtp="5" faw="16" cwl="7" wtrs="4" wtrl="9" wr="16" rrefd="8" rdrddd="3" rdrdsd="2" rdrdsc="1" rdrdscl="3" wrwrdd="3" wrwrsd="2" wrwrsc="1" wrwrscl="3" wrrd="1" rdwr="19" ref="3900" mrd="8" mod="15" xs="270" xsmrs="8" pd="8" cksre="10" cksrx="10" rfcpb="160" stag="7" xp="8" cpded="2" cke="8" rddata="19" wrlat="5" rdlat="14" wrdata="1" ckestag="2" rfc="260"></values>

vega64LE(samsung):
<values cl="20" ras="33" rcdrd="16" rcdwr="10" rcab="47" rcpb="47" rpab="14" rppb="14" rrds="4" rrdl="6" rtp="5" faw="16" cwl="7" wtrs="4" wtrl="9" wr="16" rrefd="8" rdrddd="3" rdrdsd="2" rdrdsc="1" rdrdscl="3" wrwrdd="3" wrwrsd="2" wrwrsc="1" wrwrscl="3" wrrd="1" rdwr="19" ref="3900" mrd="8" mod="15" xs="270" xsmrs="8" pd="8" cksre="10" cksrx="10" rfcpb="160" stag="7" xp="8" cpded="2" cke="8" rddata="19" wrlat="5" rdlat="14" wrdata="1" ckestag="2" rfc="260"></values>

vega56 nitro+(hynix)stock bios:
DEFAULT - GPU 5:  Vega 10 XL/XT [Radeon RX Vega 56/64]    Memory state: 0x118 (800MHz)
Timing 1          CL: 18          RAS: 24         RCDRD: 22       RCDWR: 10
Timing 2          RCAb (RC): 37   RCPb (RC): 37   RPAb (RP): 13   RPPb (RP): 13
Timing 3          RRDS: 3         RRDL: 3         RTP: 6
Timing 4          FAW: 13
Timing 5          CWL: 6          WTRS: 4         WTRL: 4
Timing 6          WR: 12
Timing 7          RREFD: 7
Timing 8          RDRDDD: 3       RDRDSD: 2       RDRDSC: 1       RDRDSCL: 3
Timing 9          WRWRDD: 3       WRWRSD: 2       WRWRSC: 1       WRWRSCL: 3
Timing 10         WRRD: 1         RDWR: 18
Timing 12         REF: 3200
Timing 13         MRD: 4          MOD: 12
Timing 14         XS: 216
Timing 16         XSMRS: 216
Timing 17         PD: 7   CKSRE: 8        CKSRX: 8
Timing 20         RFCPB: 96       STAG: 7
Timing 21         XP: 7   CPDED: 2        CKE: 7
Timing 22         RDDATA: 17      WRLAT: 4        RDLAT: 14       WRDATA: 1
Timing 23         CKESTAG: 2
RFC Timing        RFC: 176

[doktor83]

Ok, this is for Vega 64 LC

This if for reference vega56, stock bios

win10, 19.4.1 driver

What is the default value for RFC on the samsung V56 ?

vega56 powercolor(samsung ) with 64 bios:
<values cl="20" ras="33" rcdrd="16" rcdwr="10" rcab="47" rcpb="47" rpab="14" rppb="14" rrds="4" rrdl="6" rtp="5" faw="16" cwl="7" wtrs="4" wtrl="9" wr="16" rrefd="8" rdrddd="3" rdrdsd="2" rdrdsc="1" rdrdscl="3" wrwrdd="3" wrwrsd="2" wrwrsc="1" wrwrscl="3" wrrd="1" rdwr="19" ref="3900" mrd="8" mod="15" xs="270" xsmrs="8" pd="8" cksre="10" cksrx="10" rfcpb="160" stag="7" xp="8" cpded="2" cke="8" rddata="19" wrlat="5" rdlat="14" wrdata="1" ckestag="2" rfc="260"></values>

vega64LE(samsung):
<values cl="20" ras="33" rcdrd="16" rcdwr="10" rcab="47" rcpb="47" rpab="14" rppb="14" rrds="4" rrdl="6" rtp="5" faw="16" cwl="7" wtrs="4" wtrl="9" wr="16" rrefd="8" rdrddd="3" rdrdsd="2" rdrdsc="1" rdrdscl="3" wrwrdd="3" wrwrsd="2" wrwrsc="1" wrwrscl="3" wrrd="1" rdwr="19" ref="3900" mrd="8" mod="15" xs="270" xsmrs="8" pd="8" cksre="10" cksrx="10" rfcpb="160" stag="7" xp="8" cpded="2" cke="8" rddata="19" wrlat="5" rdlat="14" wrdata="1" ckestag="2" rfc="260"></values>

vega56 nitro+(hynix)stock bios:
DEFAULT - GPU 5:  Vega 10 XL/XT [Radeon RX Vega 56/64]    Memory state: 0x118 (800MHz)
Timing 1          CL: 18          RAS: 24         RCDRD: 22       RCDWR: 10
Timing 2          RCAb (RC): 37   RCPb (RC): 37   RPAb (RP): 13   RPPb (RP): 13
Timing 3          RRDS: 3         RRDL: 3         RTP: 6
Timing 4          FAW: 13
Timing 5          CWL: 6          WTRS: 4         WTRL: 4
Timing 6          WR: 12
Timing 7          RREFD: 7
Timing 8          RDRDDD: 3       RDRDSD: 2       RDRDSC: 1       RDRDSCL: 3
Timing 9          WRWRDD: 3       WRWRSD: 2       WRWRSC: 1       WRWRSCL: 3
Timing 10         WRRD: 1         RDWR: 18
Timing 12         REF: 3200
Timing 13         MRD: 4          MOD: 12
Timing 14         XS: 216
Timing 16         XSMRS: 216
Timing 17         PD: 7   CKSRE: 8        CKSRX: 8
Timing 20         RFCPB: 96       STAG: 7
Timing 21         XP: 7   CPDED: 2        CKE: 7
Timing 22         RDDATA: 17      WRLAT: 4        RDLAT: 14       WRDATA: 1
Timing 23         CKESTAG: 2
RFC Timing        RFC: 176

Thanks but i still miss the V56 Samsung default RFC )
Not V56 flashed to V64, because that uses the V64 timings.

[doktor83]

Hi, Dok!
Just examined your timings for profile 10.
They are good for hynix HBM2 but for samsung (both low voltage 800-940 mhz vega56 and high voltage 945-1100 mhz vega64 I have much better (more stable and fast) mem timings)
I can share to you if you want ))

Ok, so now i have examined your V64 samsung timings and i can't see why they would be much better, as you say.

Can you run profile 10 on V64 with samsung mem?
If you can, what's the hashrate with profile 10 vs your timing on the same algo?

[KolordZ]

Hi guys !

I'm really lost with those config.. My rig isn't stable. I got all the time crash, or bug and I don't know why.

I'm using Windows 10 1809 with 18.6.1 drivers and  I got 3 Sapphire RX Vegas 64 Nitro+ and 1 Sapphire Radeon HD 7950

I'm using SRB Miner 1.8.6 with my 3 Vegas and i'm using XMR Stak with my Radeon HD 7950.

This is my config in SRB Miner :

{
"cryptonight_type" : "normalv4",
"intensity" : 0,
"double_threads" : true,
"gpu_conf" :
[
   { "id" : 1, "intensity" : 122, "worksize" : 16, "threads" : 2, "tweak_profile" : 0},
   { "id" : 2, "intensity" : 122, "worksize" : 16, "threads" : 2, "tweak_profile" : 0},
   { "id" : 3, "intensity" : 122, "worksize" : 16, "threads" : 2, "tweak_profile" : 0}
]
}

-----------------------------------------------------------------------------------------------------

SRBMiner-CN.exe --config Config\config-normalv4.txt --pools pools.txt --logfile %LOGTIME%

----------------------------------------------------------------------------------------------------

Sometimes I got GPU crash then SRB try to shutdown my PC and it freezes.

This is my Wattman config :

GPU Clock : -8%
Memory Clock : 1105 Mhz

My Hashrate with tweak profil to 3 :

GPU0: 2149 H/s [T:68c][BUS:10]
GPU1: 2155 H/s [T:68c][BUS:7]
GPU2: 2157 H/s [T:67c][BUS:3]
hashrate: Total: 6461 H/s

That is really low with Vegas 64 Nitro+.. Doktor83, with your 56 you have 2100 Hashrate with each cards with tweak profil 5.. (In your video)

----------------------------------------------------------------------------------------------------

Something to change ?

Doktor83 I have try to use "min_rig_speed" option but there is a bug with this one in SRB 1.8.6. This is not the good average Hashrate. If someone can try and confirm that is a bug..

[livada]

Hi guys !

I'm really lost with those config.. My rig isn't stable. I got all the time crash, or bug and I don't know why.

I'm using Windows 10 1809 with 18.6.1 drivers and  I got 3 Sapphire RX Vegas 64 Nitro+ and 1 Sapphire Radeon HD 7950

I'm using SRB Miner 1.8.6 with my 3 Vegas and i'm using XMR Stak with my Radeon HD 7950.

This is my config in SRB Miner :

{
"cryptonight_type" : "normalv4",
"intensity" : 0,
"double_threads" : true,
"gpu_conf" :
[
   { "id" : 1, "intensity" : 122, "worksize" : 16, "threads" : 2, "tweak_profile" : 0},
   { "id" : 2, "intensity" : 122, "worksize" : 16, "threads" : 2, "tweak_profile" : 0},
   { "id" : 3, "intensity" : 122, "worksize" : 16, "threads" : 2, "tweak_profile" : 0}
]
}

-----------------------------------------------------------------------------------------------------

SRBMiner-CN.exe --config Config\config-normalv4.txt --pools pools.txt --logfile %LOGTIME%

----------------------------------------------------------------------------------------------------

Sometimes I got GPU crash then SRB try to shutdown my PC and it freezes.

This is my Wattman config :

GPU Clock : -8%
Memory Clock : 1105 Mhz

My Hashrate with tweak profil to 3 :

GPU0: 2149 H/s [T:68c][BUS:10]
GPU1: 2155 H/s [T:68c][BUS:7]
GPU2: 2157 H/s [T:67c][BUS:3]
hashrate: Total: 6461 H/s

That is really low with Vegas 64 Nitro+.. Doktor83, with your 56 you have 2100 Hashrate with each cards with tweak profil 5.. (In your video)

----------------------------------------------------------------------------------------------------

Something to change ?

Doktor83 I have try to use "min_rig_speed" option but there is a bug with this one in SRB 1.8.6. This is not the good average Hashrate. If someone can try and confirm that is a bug..

reduce Memory Clock : 1105 Mhz
to 1090mhz and try - 1100 is normal - 1107 is very high(my vega64LE not work normal woith 1102mhz)

[doktor83]

Hi guys !

I'm really lost with those config.. My rig isn't stable. I got all the time crash, or bug and I don't know why.

I'm using Windows 10 1809 with 18.6.1 drivers and  I got 3 Sapphire RX Vegas 64 Nitro+ and 1 Sapphire Radeon HD 7950

I'm using SRB Miner 1.8.6 with my 3 Vegas and i'm using XMR Stak with my Radeon HD 7950.

This is my config in SRB Miner :

{
"cryptonight_type" : "normalv4",
"intensity" : 0,
"double_threads" : true,
"gpu_conf" :
[
   { "id" : 1, "intensity" : 122, "worksize" : 16, "threads" : 2, "tweak_profile" : 0},
   { "id" : 2, "intensity" : 122, "worksize" : 16, "threads" : 2, "tweak_profile" : 0},
   { "id" : 3, "intensity" : 122, "worksize" : 16, "threads" : 2, "tweak_profile" : 0}
]
}

-----------------------------------------------------------------------------------------------------

SRBMiner-CN.exe --config Config\config-normalv4.txt --pools pools.txt --logfile %LOGTIME%

----------------------------------------------------------------------------------------------------

Sometimes I got GPU crash then SRB try to shutdown my PC and it freezes.

This is my Wattman config :

GPU Clock : -8%
Memory Clock : 1105 Mhz

My Hashrate with tweak profil to 3 :

GPU0: 2149 H/s [T:68c][BUS:10]
GPU1: 2155 H/s [T:68c][BUS:7]
GPU2: 2157 H/s [T:67c][BUS:3]
hashrate: Total: 6461 H/s

That is really low with Vegas 64 Nitro+.. Doktor83, with your 56 you have 2100 Hashrate with each cards with tweak profil 5.. (In your video)

----------------------------------------------------------------------------------------------------

Something to change ?

Doktor83 I have try to use "min_rig_speed" option but there is a bug with this one in SRB 1.8.6. This is not the good average Hashrate. If someone can try and confirm that is a bug..

I don't know why you have crashes, but your vega temperatures are too high (thats +10c on the vram).

1.8.6 has 10 tweak profiles, so you have 7 more to try

Also you can disable the watchdog so it wont restart the miner on gpu crash, use --disablegpuwatchdog.

I would first try to lower memory clock, get temps down to at least 60c.

min_rig_speed gets the average hashrate for 'min_rig_speed_duration' time, that is 5 minutes if not set by the user.
So please describe what you think is a bug.