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

[zeef]

Thanks for new version.

When could we see support for a more recent amd drivers?

Thanks

What do you mean?
It works with the latest drivers also

hmm ok nice..

As i seen recommend 18.6.1. did not want to risk.

1 - Can i install it and mine normally without change anything in my config_?

2 - Other question dok, i notice that auto intensity from version 1.7.0 was [56] and now is [60] ,for vega 64, why such increase? is safe to maintain that?

3 - In my rigs i use winver v1709 and disable all updates. I want to use in my desktop pc one vega 64, its ok to install last winver and amd drivers with no lose in performance? in rigs im not updating anything, only for new drives if you say is safe

Many thanks for help  

[1679521368]

1679521368

[1679521368]

1679521368

[Dubai-Industry]

Many thanks to the creator of this great miner!

[justoiurii]

Hi, DOK) Thanks for the upgrade)
UPX algo, RX 480 8G, bc drivers, intens 0 - 3890H/sec (wow!!!). I'm afraid a little) Is it - OK?

[Bllacky]

Hi,

There is an issue with your miner.
The kernel cache created for the
Radeon 280X and Radeon 270X are created for monero with the same name: g9_CNV8_w16_f8_hm1_ha.srb
If you have a system which combines 280X and 270X cards, then you can't use both card, since the kernel generated for one type will fail on the other.

You should generate different kernels for the 280X and 270X. Thank you.

[wgestickley]

Hey Doc,

- will the different Heavy Mode (1-3) effect only 8GB or also 4GB Versions Cards?
- when will Stellite fork approximately?

[doktor83]

Hi,

There is an issue with your miner.
The kernel cache created for the
Radeon 280X and Radeon 270X are created for monero with the same name: g9_CNV8_w16_f8_hm1_ha.srb
If you have a system which combines 280X and 270X cards, then you can't use both card, since the kernel generated for one type will fail on the other.

You should generate different kernels for the 280X and 270X. Thank you.

You are right, the whole GCN1 family is put under the same id (g9), i supposed that a pitcairn binary should work on a tahiti etc.it looks like it's not that way.
I will separate them in the next version.

Can you tell me how are speeds now with the tahiti and pitcairn, compared to other miners?

[doktor83]

Hey Doc,

- will the different Heavy Mode (1-3) effect only 8GB or also 4GB Versions Cards?
- when will Stellite fork approximately?

Why don't you try and see for yourself ?
It should have effect on both cards.

IDK about Stellite fork because they still don't know, and don't have a date yet.

[wgestickley]

Hey Doc,

- will the different Heavy Mode (1-3) effect only 8GB or also 4GB Versions Cards?
- when will Stellite fork approximately?

Why don't you try and see for yourself ?
It should have effect on both cards.

IDK about Stellite fork because they still don't know, and don't have a date yet.

Yeah, of course i will do it!
But i am currently on the road and want to calculate how many of my Cards in my Farm could be effected :-)

Last Time, with Version 1.7.2 i think there was no difference in the 3 Versions or have i done a failure?
Can i put it also on top of the config File where the General stuff for all Cards are or have i to do it in the line for each card where i set intensitiy for each card etc???

[Bllacky]

Hi,

There is an issue with your miner.
The kernel cache created for the
Radeon 280X and Radeon 270X are created for monero with the same name: g9_CNV8_w16_f8_hm1_ha.srb
If you have a system which combines 280X and 270X cards, then you can't use both card, since the kernel generated for one type will fail on the other.

You should generate different kernels for the 280X and 270X. Thank you.

You are right, the whole GCN1 family is put under the same id (g9), i supposed that a pitcairn binary should work on a tahiti etc.it looks like it's not that way.
I will separate them in the next version.

Can you tell me how are speeds now with the Tahiti and Pitcairn, compared to other miners?

It's a bit complicated with the speeds.
I would say they definitely went up for most cards, the 270X has a massive boost, hashing now at 400H/s.
The 280X is now hashing at around 440 H/s which is also an improvement.
The 290 went down to 460H/s initially from 540H/s, but after some tweaking, it went up to 520 H/s.

I only use R9 cards with your miner, the Vegas I have are mining ETH.

There also seems to be an issue with the off-temperature
this is a config example:
{ "id" : 0, "kernel": 1, "intensity" : 60, "worksize" : 16, "threads" : 1, "double_threads" : false, "target_temperature" : 90, "target_fan_speed": 2500, "off_temperature":90, "adl_type": 1},

But the miner still switches-off the video cards at 85C.
Also, I'm not sure if the fan controls work well for Pitcairn cards, most of the time they read 0 rpm or somthing like 10000 rpm and don't seem to spin at all.

[doktor83]

Thanks for all the info!

You example could be written as :

{ "id" : 0, "intensity" : 60, "worksize" : 16, "threads" : 1, "target_temperature" : 90, "target_fan_speed": 2500, "off_temperature":90, "adl_type": 1}

I tested the off_temperature on on a few rigs, it was OK, but going to check again.
Edit : ok found the problem, going to fix in next version

Until then you can set "off_temperature" outside gpu_conf

Also, I'm not sure if the fan controls work well for Pitcairn cards, most of the time they read 0 rpm or somthing like 10000 rpm and don't seem to spin at all.

For those oldies adl_type 2 should be used (in 1.7.3 it is now auto set for GCN1 cards), but to be honest i don't know if the target temperature and fan speed stuff work on those cards, i couldn't test it.

[1.2-GIGAHASH]

Thanks for good software!

[zeef]

Thanks for new version.

When could we see support for a more recent amd drivers?

Thanks

What do you mean?
It works with the latest drivers also

hmm ok nice..

As i seen recommend 18.6.1. did not want to risk.

1 - Can i install it and mine normally without change anything in my config_?

2 - Other question dok, i notice that auto intensity from version 1.7.0 was [56] and now is [60] ,for vega 64, why such increase? is safe to maintain that?

3 - In my rigs i use winver v1709 and disable all updates. I want to use in my desktop pc one vega 64, its ok to install last winver and amd drivers with no lose in performance? Smiley in rigs im not updating anything, only for new drives if you say is safe Smiley Cool

4 - Its possible to put in log the gpu temps?

Many thanks for help   

[tf2addict]

deleted

[wgestickley]

Hey Doc,

i am just playing around with the Heavy Types an see, that the most speed increase is in the RX570ies.

But what is the concret Problem with the Hardware errors?
is every HWR a Share which is not send to the Pool?
I want to calculate the decrease in real Hashrate with it.

When i get with Heavy Mode 3 around 90H/s more in total for a RIG and i got 3 HWR. Is it worth it to use it?

[doktor83]

Thanks for new version.

When could we see support for a more recent amd drivers?

Thanks

What do you mean?
It works with the latest drivers also

hmm ok nice..

As i seen recommend 18.6.1. did not want to risk.

1 - Can i install it and mine normally without change anything in my config_?

2 - Other question dok, i notice that auto intensity from version 1.7.0 was [56] and now is [60] ,for vega 64, why such increase? is safe to maintain that?

3 - In my rigs i use winver v1709 and disable all updates. I want to use in my desktop pc one vega 64, its ok to install last winver and amd drivers with no lose in performance? Smiley in rigs im not updating anything, only for new drives if you say is safe Smiley Cool

4 - Its possible to put in log the gpu temps?

Many thanks for help   

1. Can you change the miner version without config change? Hard to tell,depends which version are you using, if it is 170+ you can swap, config is same.
2. auto intensity is there just as a start point, you should set intensity by hand, at least if you want max performance
3. I think the recommended drivers work best, miner will work with the latest drivers but i think performance will be lower than on recommended.
4. possible, i will add it on the wishlist

[doktor83]

Hey Doc,

i am just playing around with the Heavy Types an see, that the most speed increase is in the RX570ies.

But what is the concret Problem with the Hardware errors?
is every HWR a Share which is not send to the Pool?
I want to calculate the decrease in real Hashrate with it.

When i get with Heavy Mode 3 around 90H/s more in total for a RIG and i got 3 HWR. Is it worth it to use it?

Those are not hw errors, those are compute errors which means the gpu returned a wrong result, it did not pass the validation on the cpu.
those shares are not sent to the pool of course because they would be rejected.

You get 3 compute errors in 1hr, 12hr , 24hr ?
If 24hr then its not much imo.

[wgestickley]

Hey Doc,

i am just playing around with the Heavy Types an see, that the most speed increase is in the RX570ies.

But what is the concret Problem with the Hardware errors?
is every HWR a Share which is not send to the Pool?
I want to calculate the decrease in real Hashrate with it.

When i get with Heavy Mode 3 around 90H/s more in total for a RIG and i got 3 HWR. Is it worth it to use it?

Those are not hw errors, those are compute errors which means the gpu returned a wrong result, it did not pass the validation on the cpu.
those shares are not sent to the pool of course because they would be rejected.

You get 3 compute errors in 1hr, 12hr , 24hr ?
If 24hr then its not much imo.

3 errors in 1 Hours.
Is my Calculation about it right?

When these are 2,75% of the total shares and the increase in Mode 3 is around 1,75% it is not worth to use it, right?

[zeef]

Thanks for new version.

When could we see support for a more recent amd drivers?

Thanks

What do you mean?
It works with the latest drivers also

hmm ok nice..

As i seen recommend 18.6.1. did not want to risk.

1 - Can i install it and mine normally without change anything in my config_?

2 - Other question dok, i notice that auto intensity from version 1.7.0 was [56] and now is [60] ,for vega 64, why such increase? is safe to maintain that?

3 - In my rigs i use winver v1709 and disable all updates. I want to use in my desktop pc one vega 64, its ok to install last winver and amd drivers with no lose in performance? Smiley in rigs im not updating anything, only for new drives if you say is safe Smiley Cool

4 - Its possible to put in log the gpu temps?

Many thanks for help   

1. Can you change the miner version without config change? Hard to tell,depends which version are you using, if it is 170+ you can swap, config is same.
2. auto intensity is there just as a start point, you should set intensity by hand, at least if you want max performance
3. I think the recommended drivers work best, miner will work with the latest drivers but i think performance will be lower than on recommended.
4. possible, i will add it on the wishlist

Thanks for anwser doc

1- what i was asking is if change from recomended drivers for new one require anychange in the config miner file   ?
2- thanks
3- ok, i will stick to that
4- i think is a must have

Regards

[mk111]

Anyone mining UPX with Vega 56's?

I am using SRB V1.7.3 and Radeon 18.8.2.

I get 8700H/s per card but after a few minutes I start getting compute errors

Config:
{ "id" : 0, "intensity" : 230, "double_threads" : true, "old_mode" : true },

CC 1475
MC 1100

When I change the config to "old_mode : false" then I get no more compute errors but hashrate is only 6600H/s.
Even dropping CC, MC and intensity much lower I still get compute errors when using "old_mode" : true

Any advice?

[doktor83]

Anyone mining UPX with Vega 56's?

I am using SRB V1.7.3 and Radeon 18.8.2.

I get 8700H/s per card but after a few minutes I start getting compute errors

Config:
{ "id" : 0, "intensity" : 230, "double_threads" : true, "old_mode" : true },

CC 1475
MC 1100

When I change the config to "old_mode : false" then I get no more compute errors but hashrate is only 6600H/s.
Even dropping CC, MC and intensity much lower I still get compute errors when using "old_mode" : true

Any advice?

Old mode + new drivers are a no go.
Old mode creates kernels well, the old way, which worked OK with drivers up to 18.6.1

+ Added a gpu_conf only parameter 'old_mode'. With the changes i made in 1.7.2, that allow the creation of kernels on drivers newer than 18.6.1, algos that use a scratchpad <=1MB (litev7, mox, dark, upx) suffer from a hashrate drop on Vegas (maybe others too, i didn't test). So if you set this parameter to true, it will create kernel the old way, and the speed on these algos should be back.
Don't forget, if you use old_mode : true, and compile the kernel on a driver newer than 18.6.1, you probably won't be finding any shares!