Topic: [Cast XMR] high speed XMR/CryptoNight miner for RX Vega GPUs (2 KHash/s) - page 19. (Read 206404 times)

I agree -- check your Mem voltage to make sure it wasn't reset lower and you didn't notice.

Hi,

an update is available.

Regards,
glph3k

Cast XMR Version 1.3.0 (2018/6/29)

- Performance improvement for job switches with multiple GPUs
- Support for BitTube (TUBE) V4 Pow Update which introduces the new algo CryptoNightTUBE-Heavy aka CN Saber.
  Stay on --algo=5 to switch automatically from CryptoNightIPBC-Lite to CryptoNightTUBE-Heavy when the update is happening (approx. 2nd July).
  Please note that the hash rate will drop as it is a switch from a Lite to a Heavy variant.

Download Cast XMR 1.3.0 for Windows (64bit)

SHA256 hash of cast_xmr-vega-win64_130.zip:
8d038f45b080a88c75a8dedb0469d52f105521391cc60606ddc34a76e110d0eb

is mineable?

Dear glph3k. My OverDriveNTool. Prev 1.0 run OK but 1.2.5 freeze

https://ibb.co/jTof5T
https://ibb.co/jTof5T

hashrate rx570 4gb vs rx580 4gb?

i know rx580 8gb normaly does 800-850

in Cryptonight heavy

Hi glph3k,

Im using the 110 version of your miner. This is very stable , after this version i have some hangs.

When i start up the miner, it show me an rx560 on gpu9.

I have one connected but is mining using claymore.

Analysing the Hashes is working like an rx580. So is fine i think. What could this be? error in enumerating the devices? i only discover this today, maybe im mining in this way for weeks.

Im using last 18.6.1 drivers, but wiht 18.5.x i got the same problem!

Image:

https://ibb.co/hSS3aT

https://ibb.co/j8Eo9o

Regards

Have you changed clock/voltage settings lately?  My mining window would freeze periodically, sometimes within a couple hours, sometimes a couple of days.  I originally blamed Windows, but after some trial and error, discovered I’d reduced GPU voltage a little too much and after giving my cards more juice, rig has been running solid for over a week now.

I ve been running the older versions very stable on both my rigs.
mining Haven and RYO, now my rigs are getting stuck after some hours.

someone had this problem and have a solution?

thnx

945 Mem clock --> freeze. Why . Try again but not good

Probably related to to high mem clock. 1100MHz is very ambitious and gets more unstable as more work has to be done. Try reducing to 1025 or even 945 and check if it runs stable for  24h first. If yes, it is not a software related problem.

Regards,
glph3k

Hashrate very good but My Rig 8 Vega 64 run on 30 minutes freeze tool.  1100mem --> restart cast xmr --> run OK --> Wait 30 minutes--> freeze (force close) --> rerun --> Not good. Please fix

SRBMiner --> LOw Hashrate but not freeze .

Just wondering what has changed to cause this, i have a dual cpu board and i think only the cards on the CPU1 pcie lanes start anymore.

I think its somehow related to this dual CPU system, as the CPU0 bus id:s are negative, -119, -122 etc.. and SRBMiner for example does not read temperatures from these cards, but it still runs.

But something has changed since 1.1, as that version works flawlessly.

We haven't had much trouble with cards starting, though we have seen that on a few of the FE cards.

Our rigs are one of 2 configs:
Asus B250 Mining mobo w/G4400 CPU, 8GB RAM, 120GB SSD
6 x Vega 64
2 x EVGA 1000W PS

Gigabyte B250 Fintech mobo w/G4400 CPU, 8GB RAM, 120GB SSD
4 x Vega FE
1 x EVGA 1000W PS

Yes, I know we could pack more cards in, but these configs have been incredibly stable, and that's worth some bucks to us.

What we have seen in recent updates is more random hangs, and some of the Vega 56s we were running would not stay stable, so we replaced them with 64s. Since then, no real issues.

Thanks, Gandalph3000 for an amazing piece of code, in spite of all the bitching that goes on here. You'll never please everyone, but our operation is mighty happy with your program!

Has anyone had problems with couple of the latest releases?

From 5 cards, it anymore starts 2. If i start one miner per card, 2 of them will stay up and 3 will close right before the first green message.

Cast XMR 1.1.0 still works perfectly, but the later version do no longer work on my setup.

Any fix for API reporting full hashrates without comma?

This is on Cast XMR mining.  

Individual tests, one card is only doing 200-300 h/s, even though they all have the exact same driver update.  Any solution to this?

GPU0

[16:59:45] Initializing GPU, loading kernel ...
Detected OpenCL Platform: OpenCL 2.1 AMD-APP (2580.6)
Driver Version OK.
GPU0: Radeon RX Vega (gfx900) | 56 Compute Units | Memory (MB): 8176 | Intensity: 7 / 10
[16:59:56] Connecting to 52.168.91.103:5555 ...
[16:59:56] Connected to pool.
[16:59:56] New Job received (CryptoNightV7).
[16:59:56] Difficulty changed. Now: 10000.
[17:00:00] GPU0 | 39°C | Fan 0 RPM | 1671.6 H/s
[17:00:02] GPU0 Found Nonce, submitting...
[17:00:02] Share Accepted -> +1
[17:00:02] Shares: 1 Accepted, 0 Errors | Hash Rate Avg: 1671.6 H/s | Avg Search Time: 0.0 sec
[17:00:04] GPU0 | 41°C | Fan 0 RPM | 1734.8 H/s
[17:00:08] GPU0 | 43°C | Fan 0 RPM | 1734.8 H/s
[17:00:13] GPU0 | 44°C | Fan 0 RPM | 1733.9 H/s
[17:00:17] GPU0 | 46°C | Fan 0 RPM | 1734.3 H/s
[17:00:18] New Job received (CryptoNightV7). Avg Job Time: 22.0 sec
[17:00:18] Difficulty changed. Now: 15000.

GPU3

[16:57:59] Initializing GPU, loading kernel ...
Detected OpenCL Platform: OpenCL 2.1 AMD-APP (2580.6)
Driver Version OK.
GPU3: Radeon RX Vega (gfx900) | 56 Compute Units | Memory (MB): 8176 | Intensity: 7 / 10
[16:58:11] Connecting to 52.168.91.103:5555 ...
[16:58:11] Connected to pool.
[16:58:11] New Job received (CryptoNightV7).
[16:58:11] Difficulty changed. Now: 10000.
[16:58:12] New Job received (CryptoNightV7). Avg Job Time: 1.0 sec
                                                      [16:58:28] GPU3 | 37°C | Fan 0 RPM | 206.3 H/s
[16:58:46] GPU3 | 42°C | Fan 0 RPM | 206.2 H/s
[16:59:03] GPU3 | 47°C | Fan 0 RPM | 205.1 H/s
[16:59:03] GPU3 Found Nonce, submitting...
[16:59:03] Share Accepted -> +1
[16:59:03] Shares: 1 Accepted, 0 Errors | Hash Rate Avg: 205.9 H/s | Avg Search Time: 0.0 sec
[16:59:03] GPU3 Found Nonce, submitting...
[16:59:04] Share Accepted -> +1
[16:59:04] Shares: 2 Accepted, 0 Errors | Hash Rate Avg: 205.9 H/s | Avg Search Time: 26.0 sec
[16:59:18] New Job received (CryptoNightV7). Avg Job Time: 33.5 sec
[16:59:18] Difficulty changed. Now: 15000.
[16:59:21] GPU3 | 51°C | Fan 0 RPM | 203.8 H/s

1) Set up standard PowerPlay tables
2) Restart cards with switch-radeon-gpu
3) Disable HBCC and enable large pages through switch-radeon-gpu
4) Double-check the correct values through OverdriveNTool
5) Run Cast with intensity set to 11

That's all I did and I am sustaining 2100+ hash/s

It seems like I cannot get the remote monitor working.  Remote monitoring is enabled since the dialogue appears during start up and I have made sure cast is allowed by my firewall, however when going to 127.0.0.1:7777 on another device I get a unable to connect error.

Any ideas?

Im only getting 3500+ on intensity 8 (vega 56, cn fast)? any ideas what could be wrong? 1408/906;950/900

Ok, editor is now working for me. But the card fan on water block acting badly. No matter how many rpms I set on fan in ppt, it has always 1000-1100rpm and this mean more than 60degrees.