Pages:
Author

Topic: SRBMiner Cryptonight AMD GPU Miner V1.9.3 - native algo switching - page 73. (Read 237261 times)

hero member
Activity: 2548
Merit: 626
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!


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

Feedback on
V. 1.7.6

Hashrates have gone up for normal V8 algo:
280X 455 H/s
270X 400 H/s
290 650H/s

Temperatures seem to be better on the 270X with this version, or maybe now they are read correctly.

Wow such a boost on 270x & 290 awesome!
Thanks again for the report Smiley
newbie
Activity: 22
Merit: 0
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!


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

Feedback on
V. 1.7.6

Hashrates have gone up for normal V8 algo:
280X 455 H/s
270X 400 H/s
290 650H/s

Temperatures seem to be better on the 270X with this version, or maybe now they are read correctly.
hero member
Activity: 2548
Merit: 626
Please help. I have 5 video cards in my rig and I dont have any issues making 4 out of 5 of them work. The 4 that work are RX 480/580 cards. Card 5 which gives me the fit is a R9 290. The secont I fire up srb with the 290 installed I get this problem. I am using 18.6.1 driver but have tried others. Im using SRB 1.7.6 but have had this issue with all versions. My setup is 1 Red Devil RX480, 1 XFX RX580 GTS-xxx, 1 Nitro+ RX580(samsung), 1 Nitro+ RX580(hynix). All 8GB cards. No issues. When I introduce the Sapphire R9 290 Tri-x 4gb card I get this issue. I have virtual memory set at 36000gigs and am running windows 10.

PLEASE HELP!!!!



cache cleared? do you use olcard in config?

Try the --nocache parameter maybe it helps
jr. member
Activity: 288
Merit: 1
Please help. I have 5 video cards in my rig and I dont have any issues making 4 out of 5 of them work. The 4 that work are RX 480/580 cards. Card 5 which gives me the fit is a R9 290. The secont I fire up srb with the 290 installed I get this problem. I am using 18.6.1 driver but have tried others. Im using SRB 1.7.6 but have had this issue with all versions. My setup is 1 Red Devil RX480, 1 XFX RX580 GTS-xxx, 1 Nitro+ RX580(samsung), 1 Nitro+ RX580(hynix). All 8GB cards. No issues. When I introduce the Sapphire R9 290 Tri-x 4gb card I get this issue. I have virtual memory set at 36000gigs and am running windows 10.

PLEASE HELP!!!!



cache cleared? do you use olcard in config?
newbie
Activity: 2
Merit: 0
such a question
Is it possible to view employee statistics from another computer? or just where does srbminer work?
newbie
Activity: 64
Merit: 0
Please help. I have 5 video cards in my rig and I dont have any issues making 4 out of 5 of them work. The 4 that work are RX 480/580 cards. Card 5 which gives me the fit is a R9 290. The secont I fire up srb with the 290 installed I get this problem. I am using 18.6.1 driver but have tried others. Im using SRB 1.7.6 but have had this issue with all versions. My setup is 1 Red Devil RX480, 1 XFX RX580 GTS-xxx, 1 Nitro+ RX580(samsung), 1 Nitro+ RX580(hynix). All 8GB cards. No issues. When I introduce the Sapphire R9 290 Tri-x 4gb card I get this issue. I have virtual memory set at 36000gigs and am running windows 10.

PLEASE HELP!!!!



Initializing OpenCL devices, please wait..

GPU0: Radeon RX 580 Series [ellesmere] [8192 MB][I: 55.0][W: 8][T: 2][F: 1][BUS: 1]
GPU1: Radeon RX 580 Series [ellesmere] [8192 MB][I: 55.0][W: 8][T: 2][F: 1][BUS: 5]
GPU2: Radeon RX 580 Series [ellesmere] [8192 MB][I: 55.0][W: 8][T: 2][F: 1][BUS: 8]
GPU3: Radeon RX 580 Series [hawaii] [4096 MB][I: 28.0][W: 8][T: 2][F: 1][BUS: 6]

---------------------------------------------------------
SRBMiner Cryptonight AMD GPU miner 1.7.6
DevFee ~0.85%
Press 's' to display stats
Press 'h' to display hashrate
Press 'r' to reload pools
Press 'p' to switch to the next pool
Press 'o' to switch to the previous pool
Press 0-9 to disable/enable GPU while mining
---------------------------------------------------------
Total pools: 3

Please wait...

Cryptonight Heavy mode enabled

[2019-01-06 17:32:14] Loading [heavy] kernel for DEVICE BUS_ID[1] ...
[2019-01-06 17:32:14] Loading [heavy] kernel for DEVICE BUS_ID[1] ...
[2019-01-06 17:32:14] Loading [heavy] kernel for DEVICE BUS_ID[5] ...
[2019-01-06 17:32:14] Loading [heavy] kernel for DEVICE BUS_ID[5] ...
[2019-01-06 17:32:15] Loading [heavy] kernel for DEVICE BUS_ID[8] ...
[2019-01-06 17:32:15] Loading [heavy] kernel for DEVICE BUS_ID[8] ...
[2019-01-06 17:32:15] Loading [heavy] kernel for DEVICE BUS_ID[6] ...
[2019-01-06 17:32:15] Error CL_BUILD_PROGRAM_FAILURE when calling clBuildProgram [ctx->Program] for DeviceID 3 (Thread 6)

Stopping miner...Uno momento

Press any key to continue . . .

full member
Activity: 215
Merit: 100
So what is the Loud miner doing in this thread?
newbie
Activity: 31
Merit: 0
What hashrate you have with webchain?

LOUD miner works better. SRB miner not for webchain. Do not use SRB, if you want a stable hashrate.
I have 2000 h/s with my 5xRX550, but after 2 hours rig show 1600 h/s and after 5 hours rig froze.
With LOUD miner my rig works fine for weeks without hanging.
I think the algorithm doesn't work correctly with SRB miner.
hero member
Activity: 2548
Merit: 626
Hey Dok,

Is there a way to figure out the best thread_delay for every card?

What is the math behind it and how can we improve the default 319? I have a few more hash compared to the last version so it is effective.

Thank you!

There is no math really, also its affects algos differently. For example on V8 algo small changes almost make no difference.
IMO the best number should be between 250-450, everything over 500 makes the hashrate lower.
newbie
Activity: 39
Merit: 0
Hey Dok,

Is there a way to figure out the best thread_delay for every card?

What is the math behind it and how can we improve the default 319? I have a few more hash compared to the last version so it is effective.

Thank you!
full member
Activity: 714
Merit: 104
Yes something wrong here, finally i am back to the loud miner
hero member
Activity: 2548
Merit: 626
What hashrate you have with webchain?

it should be a little lower than v7

strange rate on webchain. one rig work fine with rate 80-85% from local.
but if add 1-2 rig have lot of rejected duplicate share and rate 40-60%

[2019-01-06 15:11:54] Pool accepted result 0xC80AD1CD [121ms]
[2019-01-06 15:11:54] Pool sent a new job (ID: 96e16412aab5c1865246e58fd67b6cdc)
[2019-01-06 15:11:56] Pool accepted result 0xD4AEDD68 [92ms][ S]
[2019-01-06 15:11:58] Pool sent a new job (ID: 2825b66be399a209ae1f10f5bf4ef5fb)
[2019-01-06 15:12:01] Pool rejected result 0x64858290 (duplicate share)
[2019-01-06 15:12:11] Pool difficulty: 20000
[2019-01-06 15:12:11] Pool sent a new job (ID: f4b8fcf6c5cec6a16d8d1ea973e90ece)
[2019-01-06 15:12:17] Pool sent a new job (ID: 06ca85cfb801ec7bd6ca43e63a1be417)
[2019-01-06 15:12:25] Pool rejected result 0x98A210FA (high rate of invalid shares)
[2019-01-06 15:12:34] Pool difficulty: 20000
[2019-01-06 15:12:34] Pool sent a new job (ID: 06ca85cfb801ec7bd6ca43e63a1be417)
[2019-01-06 15:12:39] Pool sent a new job (ID: 8c6b120b42492e89f9f546388378ec2b)

you can thank that to the super-duper webchain protocol.
now i don't know why i added it at all, people only have problems and think its miner related.
jr. member
Activity: 288
Merit: 1
What hashrate you have with webchain?

it should be a little lower than v7

strange rate on webchain. one rig work fine with rate 80-85% from local.
but if add 1-2 rig have lot of rejected duplicate share and rate 40-60%

[2019-01-06 15:11:54] Pool accepted result 0xC80AD1CD [121ms]
[2019-01-06 15:11:54] Pool sent a new job (ID: 96e16412aab5c1865246e58fd67b6cdc)
[2019-01-06 15:11:56] Pool accepted result 0xD4AEDD68 [92ms][ S]
[2019-01-06 15:11:58] Pool sent a new job (ID: 2825b66be399a209ae1f10f5bf4ef5fb)
[2019-01-06 15:12:01] Pool rejected result 0x64858290 (duplicate share)
[2019-01-06 15:12:11] Pool difficulty: 20000
[2019-01-06 15:12:11] Pool sent a new job (ID: f4b8fcf6c5cec6a16d8d1ea973e90ece)
[2019-01-06 15:12:17] Pool sent a new job (ID: 06ca85cfb801ec7bd6ca43e63a1be417)
[2019-01-06 15:12:25] Pool rejected result 0x98A210FA (high rate of invalid shares)
[2019-01-06 15:12:34] Pool difficulty: 20000
[2019-01-06 15:12:34] Pool sent a new job (ID: 06ca85cfb801ec7bd6ca43e63a1be417)
[2019-01-06 15:12:39] Pool sent a new job (ID: 8c6b120b42492e89f9f546388378ec2b)
hero member
Activity: 2548
Merit: 626
What hashrate you have with webchain?

it should be a little lower than v7
newbie
Activity: 118
Merit: 0
What hashrate you have with webchain?
hero member
Activity: 2548
Merit: 626
1.7.6
key O P dont work. dont swich to next pool on webchain algo. other algo work

those keys have nothing to do with the used algo. Capslock was on maybe?
jr. member
Activity: 288
Merit: 1
1.7.6
key O P dont work. dont swich to next pool on webchain algo. other algo work
newbie
Activity: 31
Merit: 0
Incredible! The webchain algorithm works. But LOUD Miner work perfectly.
Please, kill this algorithm in SRB miner!
jr. member
Activity: 288
Merit: 1
V1.7.6
- Webchain sending stale shares again

+ Now showing time in ms for shares. This is the time needed for:

test webchain now. speed better. lot of stale shares but lot of good shares too. excellent info time for shares
hero member
Activity: 2548
Merit: 626
V1.7.6
- New tuning parameter 'thread_delay', can be used in cmd (--cgputhreaddelay), config, gpu_conf
- Even faster kernel compilation
- Added HTML statistics page, can be accessed if API is enabled via /stats
- Added 3 new API parameters: --apirigrestarturl, --apirigshutdownurl, --apiminerrestarturl
- Added share 'trip' time from miner->pool->miner on share sending
- Added SSL in interactive configurator
- Removed --runbenchmark, --benchmarkduration parameters
- Fixed algo benchmarking
- Webchain sending stale shares again
- Startup monitor now restarts whole machine on fail, instead of miner
- Fixed a bug with shutdown temperature and false ADL temperature reporting


+ Oh, a new parameter that can impact your hashing speed : thread_delay (or --cgputhreaddelay in start.bat)
It's value goes from 1-1000. It is used only when 2 or more threads are used per SAME GPU.
Simply said, this value controls the timing between thread starting on the same gpu.
The default is 319, this is something i found acceptable across multiple gpu platforms.

+ Added a GUI statistics page, where you can see basic info about your mining session.

Quote
First you must enable API, by using the --apienable parameter in start.bat
Set your rig (computer) name with --apirigname myrigname also in start.bat

After you have started the miner, you can access the stats page in your browser :
http://127.0.0.1:21555/stats

There are also three other parameters that can help you to restart miner, reboot or shutdown your machine remotely :

--apirigrestarturl
This should be a unique string, which accessed in browser results in a computer restart. Miner needs to have admin privileges.

--apirigshutdownurl
This should be a unique string, which accessed in browser results in a computer shutdown. Miner needs to have admin privileges.

--apiminerrestarturl
This should be a unique string, which accessed in browser restart SRBMiner.



Example start.bat:
SRBMiner-CN.exe --config Config\config-normalv8.txt --pools Pools\pools-normalv8.txt --apienable --apirigrestarturl 12345fff --apirigshutdownurl 54321fff --apiminerrestarturl restart_my_srb

Visiting this url restarts your machine:
http://127.0.0.1:21555/12345fff

Visiting this url turns off your machine:
http://127.0.0.1:21555/54321fff

Visiting this url restarts SRBMiner:
http://127.0.0.1:21555/restart_my_srb


Of course if you are going to 'open' this HTTP interface to the world, you better use complex urls so nobody can guess them easilly and turn off your rig Smiley

+ Now showing time in ms for shares. This is the time needed for:

- your miner sent the share to the pool
- pool accepted the share, did some validation and returned a result
- your miner accepted the result

+ Removed the benchmarking per algo, there is the --benchmarkalgos parameter which benches all you got in algos.txt
Pages:
Jump to: