[Cast XMR] high speed XMR/CryptoNight miner for RX Vega GPUs (2 KHash/s)

[StellarX88]

Is there any free scripts available to restart CastXMR when it hangs or speed drops to 0?

I shared my modified version of JJ Hash Monitor for XMR-Stak awhile back, modified it to work with Cast-XMR.

https://drive.google.com/drive/folders/1TnicWxpB1TtGQCe_cuXND8ha-XsgQsU1

You can go through the changelog and update it to latest Cast-XMR accordingly.

[mk111]

Is there any free scripts available to restart CastXMR when it hangs or speed drops to 0?

I shared my modified version of JJ Hash Monitor for XMR-Stak awhile back, modified it to work with Cast-XMR.

https://drive.google.com/drive/folders/1TnicWxpB1TtGQCe_cuXND8ha-XsgQsU1

You can go through the changelog and update it to latest Cast-XMR accordingly.

I will try it, thank you

[mehmetkorkrek]

Can someone help me. My problem is that, the program closes 20-30 minutes after oflined. Sometimes the internet is interrupted. And it can take 20 to 30 minutes.So the program is shutting itself down. How can I solve this.

[dragonmike]

Can someone help me. My problem is that, the program closes 20-30 minutes after oflined. Sometimes the internet is interrupted. And it can take 20 to 30 minutes.So the program is shutting itself down. How can I solve this.

Use Cast XMR UI. It's been posted in this thread. Neat app, manages your miner with a graphical UI, auto-restarts if average hashrate falls under user defined thresholds etc.

[ayeaye87]

Can someone help me. My problem is that, the program closes 20-30 minutes after oflined. Sometimes the internet is interrupted. And it can take 20 to 30 minutes.So the program is shutting itself down. How can I solve this.

Use Cast XMR UI. It's been posted in this thread. Neat app, manages your miner with a graphical UI, auto-restarts if average hashrate falls under user defined thresholds etc.

I think the topics described at http://vega.miningguides.com/ are also mentioned. I did some of the things described here. I made a mistake on one or two issues. Thank you for your interest.

[Rewqpro]

when fiji

[xpulse]

All,

Can someone tell me why this happened with Vega FE. I did everything as said in video in some post before this one.

I have 3 Vega's 64 and 1 Vega FE.
BlockChain drives, etc... Everything looks good except 1 thing: Vega FE give ONLY 1300-1500 Khash/s all Vega's 64 1980-2010 Kh.
I tried almost everything still FE can't go more than 1500. I'm not using any SoC registry mods. EnableUlps and Crosslink are disabled on all cards.

Any suggestions? Quite a little bit out of idea's what can't be wrong.

CPU pretty powerful Ryzen 5 1600 / 8 Gb RAM + 70GB Swap memory.

Thanks.

[VaynardW]

All,

Can someone tell me why this happened with Vega FE. I did everything as said in video in some post before this one.

I have 3 Vega's 64 and 1 Vega FE.
BlockChain drives, etc... Everything looks good except 1 thing: Vega FE give ONLY 1300-1500 Khash/s all Vega's 64 1980-2010 Kh.
I tried almost everything still FE can't go more than 1500. I'm not using any SoC registry mods. EnableUlps and Crosslink are disabled on all cards.

Any suggestions? Quite a little bit out of idea's what can't be wrong.

CPU pretty powerful Ryzen 5 1600 / 8 Gb RAM + 70GB Swap memory.

Thanks.

Are you using the FE as display output?

[xpulse]

All,

Can someone tell me why this happened with Vega FE. I did everything as said in video in some post before this one.

I have 3 Vega's 64 and 1 Vega FE.
BlockChain drives, etc... Everything looks good except 1 thing: Vega FE give ONLY 1300-1500 Khash/s all Vega's 64 1980-2010 Kh.
I tried almost everything still FE can't go more than 1500. I'm not using any SoC registry mods. EnableUlps and Crosslink are disabled on all cards.

Any suggestions? Quite a little bit out of idea's what can't be wrong.

CPU pretty powerful Ryzen 5 1600 / 8 Gb RAM + 70GB Swap memory.

Thanks.

Are you using the FE as display output?

Nope, Rig is headless

[Rewqpro]

Still have one of my Vegas mining at a good 100h/s slower than the other 5... and no idea why. They are all 56's flashed to 64's, running at 1950h/s except that odd one doing 1850. Same temperature, same clocks, same everything. Is there a logical explanation? Any suggestions?

maybe other memory type on that gpu

Is there a way to find this out by the way?...

Memory type? gpu-z should do it. Hwinfo might.

what should be the memory type to attain the mentioned hash rate and what memory types to avoid , can u tell me.

Vega is the only card that has a new kind of memory called HBM.
All reference design has Samsung HBM chips in it.
It doesn't matter if it is Sapphire,  Asus or 56 and 64.
All reference card was made by AMD itself and it has Samsung in it.

Most of the non reference cards have hynix HBM chips.

Samsung good
Hynix bad

How bad I don't really know. Apparently Hynix memory will do around 900hz . Samsung could do 1200 but usually stays at 1100.

Avoid XFX non reference!!!

That's all

About the problem mentioned.  Is the card in question  an GFX900 instead GFX901?
My guess is, it's not the card. It's something else. If you change the card position,  do it keep the lower hash?
I do have an problematic Sapphire 64, same card as the other 5 in the rig. It is the only that is identified as GFX900 and the only that gave me errors.

Lol, just saw u do have 2 GFX900.  That's your problem. This card is more tricky.  Mine is doing same hash but with 2% error.
No one knows the difference from gfx 900  to gfx 901!
There are 56, 64 and FE as 900 but the majority in the market is 901.

 

actually, you are wrong here. few years back, such kind of GPU as FURY X hit the market.

it also has HBM memory type, not HBM2 like Vega though.

[MickDaster]

are there any improvements being made here? i hope there are

[AlexanderDumas]

When I mine with cast it hangs up after running for a couple of hours. The longest I’ve managed to get it to run is like 8 hours. Anyone else experiencing this problem? Typing S doesn’t do anything I have to exit the application and restart it.

Currently using 2 Vega 56s
Temps stay around 63

Normally tose hangs up are caused by the motherboard BIOS setting of the PCI E slots. some BIOS by default are in AUTO and the best settings should be GEN 1 or GEN 2.
Once I made that change, those hangs up stopped happening. I have been able to run it for days without crash.

Try that one.

[yondy12]

I've been use this tools since 2 days ago running without a problem but sometimes I got "Invalid nonce compute failed" 25 and reject block 25  on "nanopool" after m mining like pass 32 hours , is that safe if m using this command --fastjobswitch --forcecompute --nonicehash ?

[newnonce]

I've been use this tools since 2 days ago running without a problem but sometimes I got "Invalid nonce compute failed" 25 and reject block 25  on "nanopool" after m mining like pass 32 hours , is that safe if m using this command --fastjobswitch --forcecompute --nonicehash ?

As far as I know:

--fastjobswitch :  helps if you are getting a lot of "outdated because of job change" in the stats report
--forcecompute:  use if you are not using Blockchain drivers - takes the place of changing mode from "graphics" to "compute" mode through Wattman settings in AMD driver settings.  I stick with the Blockchain drivers since CAST doc says Blockchain is still better
--nonicehash :  I use this too since the doc mentions something about memory space issues I don't use nicehash but figured it won't hurt

BUT - regarding the invalid NONCE, I get those from having too high of a memory OC.  For instance, I used the VEGAMINING guide put together by CircusDad and the settings for the Vega56's are too high for me.  I have these crappy XFX Dual Vega 56's that don't OC that well so in order to get stable hash without NONCE COMPUTE FAILED errors and lockups, I had to use the 85% SoftPowerPlay settings from his guide and dial back the P2 MEMORY down in OverDriveN from 950MHZ to 935 on one and 925 on another - same manufacturer and model, but the "silicon lottery" was not lucky for me.   That's been my experience so far - Good luck tweaking!

   

[yondy12]

I've been use this tools since 2 days ago running without a problem but sometimes I got "Invalid nonce compute failed" 25 and reject block 25  on "nanopool" after m mining like pass 32 hours , is that safe if m using this command --fastjobswitch --forcecompute --nonicehash ?

As far as I know:

--fastjobswitch :  helps if you are getting a lot of "outdated because of job change" in the stats report
--forcecompute:  use if you are not using Blockchain drivers - takes the place of changing mode from "graphics" to "compute" mode through Wattman settings in AMD driver settings.  I stick with the Blockchain drivers since CAST doc says Blockchain is still better
--nonicehash :  I use this too since the doc mentions something about memory space issues I don't use nicehash but figured it won't hurt

BUT - regarding the invalid NONCE, I get those from having too high of a memory OC.  For instance, I used the VEGAMINING guide put together by CircusDad and the settings for the Vega56's are too high for me.  I have these crappy XFX Dual Vega 56's that don't OC that well so in order to get stable hash without NONCE COMPUTE FAILED errors and lockups, I had to use the 85% SoftPowerPlay settings from his guide and dial back the P2 MEMORY down in OverDriveN from 950MHZ to 935 on one and 925 on another - same manufacturer and model, but the "silicon lottery" was not lucky for me.   That's been my experience so far - Good luck tweaking!

   

I see how abt I put this tuning on my run batch like stak amd :

setx GPU_FORCE_64BIT_PTR 1
setx GPU_MAX_HEAP_SIZE 99
setx GPU_MAX_ALLOC_PERCENT 99
setx GPU_SINGLE_ALLOC_PERCENT 99

is that okay?

[micky123]

Hi All,

I have setup XMR cast and running it, however, it only detects GPU 0 and not GPU1. Tried using the -G 0,1 switch, but no luck! I have checked and confirmed that Crossfire is currently off as well. Can't seem to understand why only 1 card is detected. I am using 2 x RX480 cars from Sapphire. Any ideas?

[RonBurgundy]

I am having an issue with getting the miner to work at all.

I have two vega 56's and I have done circusdad's instructions (among others) and everything runs fine until I connect to a pool. Once it connects my computer locks up. I have tried sooooo many settings and drivers. I'm at my wits end.

Has anyone else run into this? I can not get these things to mine XMR without my computer immediately freezing up.

Thanks guys!

[micky123]

I am having an issue with getting the miner to work at all.

I have two vega 56's and I have done circusdad's instructions (among others) and everything runs fine until I connect to a pool. Once it connects my computer locks up. I have tried sooooo many settings and drivers. I'm at my wits end.

Has anyone else run into this? I can not get these things to mine XMR without my computer immediately freezing up.

Thanks guys!

Have you tried mining other cryptonight coins? If not, that would be a good place to start, just to ensure that the issue is not with the miner. Make sure you have installed the latest AMD drivers for your graphics card, and ensure that they are not OC'd to begin with. Try a different XMR pool if possible, i use dwarf, and have had no issues with them so far. Thermal issues seem to be the likely issue here.

[RonBurgundy]

I am having an issue with getting the miner to work at all.

I have two vega 56's and I have done circusdad's instructions (among others) and everything runs fine until I connect to a pool. Once it connects my computer locks up. I have tried sooooo many settings and drivers. I'm at my wits end.

Has anyone else run into this? I can not get these things to mine XMR without my computer immediately freezing up.

Thanks guys!

Have you tried mining other cryptonight coins? If not, that would be a good place to start, just to ensure that the issue is not with the miner. Make sure you have installed the latest AMD drivers for your graphics card, and ensure that they are not OC'd to begin with. Try a different XMR pool if possible, i use dwarf, and have had no issues with them so far. Thermal issues seem to be the likely issue here.

I have tried different pools and different coins. Some other algo's work but I can't get cryptonight to work at all.

I have stock settings and have tried blockchain, older and latest drivers. I used DDU to uninstall in between too. Sooo frustrating

[jadicomunicaciones]

Hi guys
Could you help me with the compatible pool
the only one that connects me is nanopool.
Thank you.