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

[Saphire]

I think you deserve the DevFee but one of my rigs the miner can't connect to DevFee pool and stop working and needed to be restart by menually.

[1714242623]

1714242623

[1714242623]

1714242623

[doktor83]

V1.4.4
- Fixed (hopefully) situation when miner hashed in 'empty' for a long time
- Added 30 minute mining average speed display
- Removed greedybastard switch

ah, im tired. Please all of you having the 'devfee stuck' or 'miner not doing anything' problem, download this version and test.
It should be ok now.

Also removed the greedy switch. This is just taking too much of my free time, and i think the devfee % is not something terrible.

I hope everything will be stable now and i can get myself on the overclocking stuff and the API finally.  

[doktor83]

Hello everyone!

Does this miner support TeraScale/VLIW Radeon GPUs?

it should work, but you wont get speeds like on old claymore.
Which card do you have?

Radeon 5770 1GB GDDR5.

And yeah, hashing speed is only half compared to Claymore v9.7...

download v 1.4.3 and add --usekernel 3 in the bat like this :

%EXE% --config config.txt --pools pools.txt --usekernel 3

this is a test kernel im working on for those oldies

I guess this emulates the Claymore v9.7 kernel?

Thanks for caring about old GPUs (most devs don't), I'll give it a try!

I just tried it and I still only get 80 H/s compared to 176 H/s on Claymore v9.7... intensity is 14 (higher doesn't work), worksize 8, 1 thread.

Another problem with all miners I've tried (except Claymore v9.7) is that there is no "low intensity" option, which means the UI is lagging a lot (if you want to use your PC while mining).

I wish someone could replicate Claymore v9.7 characteristics and I'd gladly pay a mining fee...

can you download 1.4.4 now and try again with --usekernel 3 parameter in bat, but set start intensity to 10 , and use only 1 thread. Then work up the intensity as much as you can.

[ipe4enko]

>> - Removed greedybastard switch

I dont use this options but then i think that i can switch this option i feel freedom )))

[dohfish]

V1.4.4
- Fixed (hopefully) situation when miner hashed in 'empty' for a long time
- Added mining session average speed display
- Removed greedybastard switch

ah, im tired. Please all of you having the 'devfee stuck' or 'miner not doing anything' problem, download this version and test.
It should be ok now.

Also removed the greedy switch. This is just taking too much of my free time, and i think the devfee % is not something terrible.

I hope everything will be stable now and i can get myself on the overclocking stuff and the API finally. 

1.4.4 still crashes instantly when pressing "h" to see hashrate on all of my 15 rigs.

Miner is still rather unstable with restarts every 15-20 minutes.

[doktor83]

V1.4.4
- Fixed (hopefully) situation when miner hashed in 'empty' for a long time
- Added mining session average speed display
- Removed greedybastard switch

ah, im tired. Please all of you having the 'devfee stuck' or 'miner not doing anything' problem, download this version and test.
It should be ok now.

Also removed the greedy switch. This is just taking too much of my free time, and i think the devfee % is not something terrible.

I hope everything will be stable now and i can get myself on the overclocking stuff and the API finally. 

1.4.4 still crashes instantly when pressing "h" to see hashrate on all of my 15 rigs.

Miner is still rather unstable with restarts every 15-20 minutes.

wow so fast you tried the new version on 15 rigs in 5 minutes 

anyways, put --adldisable in .bat and lower your intensity, dont leave it on 0, but set it up yourself, and it wont crash.

[bovineplane]

V1.4.4
- Fixed (hopefully) situation when miner hashed in 'empty' for a long time
- Added mining session average speed display
- Removed greedybastard switch

ah, im tired. Please all of you having the 'devfee stuck' or 'miner not doing anything' problem, download this version and test.
It should be ok now.

Also removed the greedy switch. This is just taking too much of my free time, and i think the devfee % is not something terrible.

I hope everything will be stable now and i can get myself on the overclocking stuff and the API finally. 

Paying the dev fee wasnt an issue with me. It was the freezing that was the issue. I have used an official release of XMR-stak with dev fee since I moved to CN coins. If this new version does not freeze then I will use it with dev fee.

Thanks for the update.

[doktor83]

V1.4.4
- Fixed (hopefully) situation when miner hashed in 'empty' for a long time
- Added mining session average speed display
- Removed greedybastard switch

ah, im tired. Please all of you having the 'devfee stuck' or 'miner not doing anything' problem, download this version and test.
It should be ok now.

Also removed the greedy switch. This is just taking too much of my free time, and i think the devfee % is not something terrible.

I hope everything will be stable now and i can get myself on the overclocking stuff and the API finally. 

Paying the dev fee wasnt an issue with me. It was the freezing that was the issue. I have used an official release of XMR-stak with dev fee since I moved to CN coins. If this new version does not freeze then I will use it with dev fee.

Thanks for the update.

just make sure to find the right intensities for your cards and it wont freeze

[UnclWish]

Doctor, did you fix issue with launching Pitcairn card alone when it 3rd GPU?

[alczaty]

V1.4.4
- Fixed (hopefully) situation when miner hashed in 'empty' for a long time
- Added mining session average speed display
- Removed greedybastard switch

ah, im tired. Please all of you having the 'devfee stuck' or 'miner not doing anything' problem, download this version and test.
It should be ok now.

Also removed the greedy switch. This is just taking too much of my free time, and i think the devfee % is not something terrible.

I hope everything will be stable now and i can get myself on the overclocking stuff and the API finally. 

1.4.4 still crashes instantly when pressing "h" to see hashrate on all of my 15 rigs.

Miner is still rather unstable with restarts every 15-20 minutes.

probably crash because you overclock to much, try to decrease a bit gpu clock

[dohfish]

V1.4.4
- Fixed (hopefully) situation when miner hashed in 'empty' for a long time
- Added mining session average speed display
- Removed greedybastard switch

ah, im tired. Please all of you having the 'devfee stuck' or 'miner not doing anything' problem, download this version and test.
It should be ok now.

Also removed the greedy switch. This is just taking too much of my free time, and i think the devfee % is not something terrible.

I hope everything will be stable now and i can get myself on the overclocking stuff and the API finally. 

1.4.4 still crashes instantly when pressing "h" to see hashrate on all of my 15 rigs.

Miner is still rather unstable with restarts every 15-20 minutes.

probably crash because you overclock to much, try to decrease a bit gpu clock

The machines are stable on all other different software out there - And it has been stable for months.

--adldisable seems to have done the trick for now, going to test it a bit more.

[cryptosize]

Hello everyone!

Does this miner support TeraScale/VLIW Radeon GPUs?

it should work, but you wont get speeds like on old claymore.
Which card do you have?

Radeon 5770 1GB GDDR5.

And yeah, hashing speed is only half compared to Claymore v9.7...

download v 1.4.3 and add --usekernel 3 in the bat like this :

%EXE% --config config.txt --pools pools.txt --usekernel 3

this is a test kernel im working on for those oldies

I guess this emulates the Claymore v9.7 kernel?

Thanks for caring about old GPUs (most devs don't), I'll give it a try!

I just tried it and I still only get 80 H/s compared to 176 H/s on Claymore v9.7... intensity is 14 (higher doesn't work), worksize 8, 1 thread.

Another problem with all miners I've tried (except Claymore v9.7) is that there is no "low intensity" option, which means the UI is lagging a lot (if you want to use your PC while mining).

I wish someone could replicate Claymore v9.7 characteristics and I'd gladly pay a mining fee...

can you download 1.4.4 now and try again with --usekernel 3 parameter in bat, but set start intensity to 10 , and use only 1 thread. Then work up the intensity as much as you can.

I don't mind the devfee, as long as the miner can work as good as Claymore v9.7.

I tried intensity to 10, 67 H/s only, then I increased it to 12 and the entire PC froze. :\

I don't understand why GPU mining it's so hard compared to CPU mining where you just set the available CPU cores...

Radeon 5770 (Juniper) has 800 ALUs (10 Compute Units x 16 SIMD x 5 ALUs).

I don't even understand what intensity is supposed to do ALU-wise... why not allow us to set ALUs manually?

Claymore uses hashcount option and optimal value is 320 hashes (320 ALUs?), 2MB each = 320 x 2 = 640MB of VRAM needed. The GPU has 1024MB of VRAM, so I can still use the PC while mining with no lagging.

Is it so hard to write proper VLIW assembly (from what I've heard at least) and only Claymore has mastered this art of programming? Just wondering...

[Littlegrand]

Has anyone had any luck with 1 thread on rx570s?  Best I can get is 650h/s using 2 threads intensity 29 and workload 8, anyone have any insights?

Sorry 4gb cards on heavy

Edit: got 730 with 1 thread 60 intensity

Thanks,
--J

[ragecage]

write intensities, algorithm and gpu memory. so it is unclear.

[CryptoPlay]

I was mining fine with (6x)Vega 64 on UltraHeavy(the lite one) despite not knowing the config.
Ive set intensity 110, worksize 16 adn threads 2  = 18000h
I did not tried any other config cuz 18k was fine to me. I know it could do more in this algo.

Doctor,
My doubt is, setting threads 2 in the config lines and not changing top from false to true will it work with 2?
Looks you do have the same parameter in two places.


PS. I do not recommend IPBC support pool cuz of the pool config setting too high diff. My client stayed 10min without sending hash with diff 2500000. This do not look a mistake at all...
This one is better: https://cryptoknight.cc/ipbc/


Can someone share a config to Vega 64 that makes it hash 2k in Monero7 algo?

[WhyMe]

I was mining fine with (6x)Vega 64 on UltraHeavy(the lite one) despite not knowing the config.
Ive set intensity 110, worksize 16 adn threads 2  = 18000h
I did not tried any other config cuz 18k was fine to me. I know it could do more in this algo.

Doctor,
My doubt is, setting threads 2 in the config lines and not changing top from false to true will it work with 2?
Looks you do have the same parameter in two places.


PS. I do not recommend IPBC support pool cuz of the pool config setting too high diff. My client stayed 10min without sending hash with diff 2500000. This do not look a mistake at all...
This one is better: https://cryptoknight.cc/ipbc/


Can someone share a config to Vega 64 that makes it hash 2k in Monero7 algo?

Just force diff with wallet.diff

[gazu]

Hello everyone!

Does this miner support TeraScale/VLIW Radeon GPUs?

it should work, but you wont get speeds like on old claymore.
Which card do you have?

Radeon 5770 1GB GDDR5.

And yeah, hashing speed is only half compared to Claymore v9.7...

download v 1.4.3 and add --usekernel 3 in the bat like this :

%EXE% --config config.txt --pools pools.txt --usekernel 3

this is a test kernel im working on for those oldies

I guess this emulates the Claymore v9.7 kernel?

Thanks for caring about old GPUs (most devs don't), I'll give it a try!

I just tried it and I still only get 80 H/s compared to 176 H/s on Claymore v9.7... intensity is 14 (higher doesn't work), worksize 8, 1 thread.

Another problem with all miners I've tried (except Claymore v9.7) is that there is no "low intensity" option, which means the UI is lagging a lot (if you want to use your PC while mining).

I wish someone could replicate Claymore v9.7 characteristics and I'd gladly pay a mining fee...

can you download 1.4.4 now and try again with --usekernel 3 parameter in bat, but set start intensity to 10 , and use only 1 thread. Then work up the intensity as much as you can.

I don't mind the devfee, as long as the miner can work as good as Claymore v9.7.

I tried intensity to 10, 67 H/s only, then I increased it to 12 and the entire PC froze. :\

I don't understand why GPU mining it's so hard compared to CPU mining where you just set the available CPU cores...

Radeon 5770 (Juniper) has 800 ALUs (10 Compute Units x 16 SIMD x 5 ALUs).

I don't even understand what intensity is supposed to do ALU-wise... why not allow us to set ALUs manually?

Claymore uses hashcount option and optimal value is 320 hashes (320 ALUs?), 2MB each = 320 x 2 = 640MB of VRAM needed. The GPU has 1024MB of VRAM, so I can still use the PC while mining with no lagging.

Is it so hard to write proper VLIW assembly (from what I've heard at least) and only Claymore has mastered this art of programming? Just wondering...

hi DOK, I get 320 h/s and with claymore 9.7 360 h/s, both results with OC, excellent work, a little more and you get over it, congratulations. (DISCUSS MY ENGLISH, I USE GOOGLE TRANSLATOR JAJAJAJAJA)
*****    I have a 6950 with 6970 bios and drivers 14.4 *****

[doktor83]

do you get this with kernel 3? you have 512mb 6950?

[doktor83]

V1.4.4
- Fixed (hopefully) situation when miner hashed in 'empty' for a long time
- Added mining session average speed display
- Removed greedybastard switch

ah, im tired. Please all of you having the 'devfee stuck' or 'miner not doing anything' problem, download this version and test.
It should be ok now.

Also removed the greedy switch. This is just taking too much of my free time, and i think the devfee % is not something terrible.

I hope everything will be stable now and i can get myself on the overclocking stuff and the API finally. 

1.4.4 still crashes instantly when pressing "h" to see hashrate on all of my 15 rigs.

Miner is still rather unstable with restarts every 15-20 minutes.

probably crash because you overclock to much, try to decrease a bit gpu clock

The machines are stable on all other different software out there - And it has been stable for months.

--adldisable seems to have done the trick for now, going to test it a bit more.

can you tell on which cards do you have to use the --adldisable ?

[doktor83]

I was mining fine with (6x)Vega 64 on UltraHeavy(the lite one) despite not knowing the config.
Ive set intensity 110, worksize 16 adn threads 2  = 18000h
I did not tried any other config cuz 18k was fine to me. I know it could do more in this algo.

Doctor,
My doubt is, setting threads 2 in the config lines and not changing top from false to true will it work with 2?
Looks you do have the same parameter in two places.


PS. I do not recommend IPBC support pool cuz of the pool config setting too high diff. My client stayed 10min without sending hash with diff 2500000. This do not look a mistake at all...
This one is better: https://cryptoknight.cc/ipbc/


Can someone share a config to Vega 64 that makes it hash 2k in Monero7 algo?

let me quote you this from the first page :

"DON'T comment out or remove intensity and double_threads setting at the top of the config file, just leave it as-is, because when you use a per card setting, those settings will be ignored."