Claymore's CryptoNote AMD GPU Miner v11.3

[Claymore]

Any idea what could be causing this:

09:29:03:320   f9c   buf: {"id":1,"jsonrpc":"2.0","error":null,"result":{"id":"124601006298325","job":{"blob":"0100beffc9a80519953029eb35686e6603c415a5ce146464a18fc5d4a2ba148c54fce948f82f4e0 0000000e347b57b530720a0174765f22dd11afec72266a4aca2468563cb9692cc5c52ee03","job_id":"148032830562442","target":"7b5e0400"},"status":"OK"}}

09:29:03:320   f9c   parse packet: 303
09:29:03:325   f9c   new buf size: 0
09:29:03:325   f9c   DevFee: Pool Diff 15000
09:29:04:244   f64   GPU 2, GpuMiner k1a failed -4
09:29:04:246   6ec   GPU 3, GpuMiner k1a failed -4
09:29:04:249   f64   GPU memory allocation failed. Try to set at least 16 GB of Virtual Memory: Computer Properties / Advanced System Settings / Performance / Advanced / Virtual Memory

5x 290 rig, I have virtual memory set to 16GB, no overclock, I've actually set memory down to 1250 from stock 1300.

Thanks.

Edit: using 13.12 drivers with 9.2 13.12.

Try to increase size of physical RAM, for 4 290 cards at least 4GB is necessary, 5 cards may require even more. You can also try to increase size of virtual memory.

[1715634886]

1715634886

[1715634886]

1715634886

[1715634886]

1715634886

[1715634886]

1715634886

[MrJay]

Thanks for the reply, I've got 8GB in the system, will try increasing the virtual memory.

[Tiger78]

For 7790(260x) managed to obtain the best values in the config: -h 256-a 4.
Here's what happened:

09:42:03:659   bc8   GPU 4: HashCnt = 256, time = 1102 ms, hashrate = 232 h/s
09:42:03:674   580   GPU 3: HashCnt = 256, time = 1101 ms, hashrate = 232 h/s
09:42:03:694   bc8   Round
09:42:03:704   7ec   GPU 2: HashCnt = 256, time = 1121 ms, hashrate = 228 h/s
09:42:03:709   580   Round
09:42:03:719   d10   GPU 1: HashCnt = 256, time = 1122 ms, hashrate = 228 h/s
09:42:03:744   d10   Round
09:42:03:754   730   GPU 0: HashCnt = 256, time = 1117 ms, hashrate = 229 h/s
09:42:03:764   730   Round
09:42:04:769   bcc   GPU 4: HashCnt = 256, time = 1075 ms, hashrate = 238 h/s
09:42:04:779   bcc   Round
09:42:04:804   d54   GPU 3: HashCnt = 256, time = 1079 ms, hashrate = 237 h/s
09:42:04:809   d54   Round
09:42:04:814   c90   GPU 2: HashCnt = 256, time = 1074 ms, hashrate = 238 h/s
09:42:04:824   c90   Round
09:42:04:839   398   GPU 1: HashCnt = 256, time = 1077 ms, hashrate = 237 h/s
09:42:04:849   398   Round
09:42:04:854   aa4   GPU 0: HashCnt = 256, time = 1074 ms, hashrate = 238 h/s
09:42:04:864   aa4   Round
09:42:05:869   bc8   GPU 4: HashCnt = 256, time = 1072 ms, hashrate = 238 h/s
09:42:05:894   bc8   Round
09:42:05:909   580   GPU 3: HashCnt = 256, time = 1079 ms, hashrate = 237 h/s
09:42:05:919   7ec   GPU 2: HashCnt = 256, time = 1075 ms, hashrate = 238 h/s

[MrJay]

Try to increase size of physical RAM, for 4 290 cards at least 4GB is necessary, 5 cards may require even more. You can also try to increase size of virtual memory.

I came home to the rig only showing 4 cards and the 5th's fan spinning at 100%.  moved the 5th card in to a different pci slot, increased virtual mem to 18GB:

5x Sapphire R9 290 Tri X @ 1000/1250

Speed: 3731 h/s, TotalHashes: 503585K, DevHashes: 12588K Mining time: 38:14

10:07:25:440   99c   GPU 0: HashCnt = 1280, time = 1698 ms, hashrate = 753 h/s
10:07:25:480   99c   Round
10:07:25:880   594   GPU 4: HashCnt = 1280, time = 1702 ms, hashrate = 752 h/s
10:07:25:905   594   Round
10:07:26:035   eb0   GPU 1: HashCnt = 1280, time = 1697 ms, hashrate = 754 h/s
10:07:26:070   eb0   Round
10:07:26:100   890   GPU 3: HashCnt = 1280, time = 1696 ms, hashrate = 754 h/s
10:07:26:115   890   Round
10:07:26:705   77c   GPU 2: HashCnt = 1280, time = 1746 ms, hashrate = 733 h/s
10:07:26:740   77c   Round

Looking good.

[gaalx]

Are mined now Monero on Claymore GPU Miner 9.2 beta. So it is not clear how to set the frequency. One rig working 947-1250, and the second 1100-1250 (based on CPU info). Accordingly, various profits. Which team is responsible for setting the frequency? -gpu-engine 1000 -gpu-memclock 1500 - the program gives that does not know the command.

[Claymore]

Are mined now Monero on Claymore GPU Miner 9.2 beta. So it is not clear how to set the frequency. One rig working 947-1250, and the second 1100-1250 (based on CPU info). Accordingly, various profits. Which team is responsible for setting the frequency? -gpu-engine 1000 -gpu-memclock 1500 - the program gives that does not know the command.

Miner does not support overclocking. You can use MSI AfterBurner for that. Or "GPUs Control Utility v1.1" that you can download by link in OP of this topic:  https://bitcointalk.org/index.php?topic=693680.0

[gaalx]

Thanks for the reply. There are 8 pieces 290 GPU. Settings - 1030/1350. Config file whose name ends with options -a -r 0. 1 in the same farm - three 790 and one 680. The second three 680 and one 790. It depends on the type of memory?

[anatolikostis]

Thanks for the reply. There are 8 pieces 290 GPU. Settings - 1030/1350. Config file whose name ends with options -a -r 0. 1 in the same farm - three 790 and one 680. The second three 680 and one 790. It depends on the type of memory?

It does: 290 with Hynix = 290x with Elpida.

[gaalx]

initialization 290 GPU program only sees 3GB. actual card with 4 GB. So it should be or is it some kind of problem?

[Claymore]

initialization 290 GPU program only sees 3GB. actual card with 4 GB. So it should be or is it some kind of problem?

It's ok.

[capler]

Who mine on linux? How much hash on 280x?

GPU 1: HashCnt = 896, time = 1774 ms, hashrate = 505 h/s
GPU 0: HashCnt = 832, time = 1535 ms, hashrate = 542 h/s
GPU 3: HashCnt = 672, time = 1281 ms, hashrate = 524 h/s
GPU 2: HashCnt = 512, time = 1071 ms, hashrate = 478 h/s
04/02/15-00:23:24 - New job received from xmr-eu.dwarfpool.com:8050
Speed: 1987 h/s, TotalHashes: 3799K, DevHashes: 94K Mining time: 00:31

gpu: 0-7970, 1-7950, 2-7870 , 3-280x

[mik_druid]

I've downloaded the latest drivers 5.3 beta, it the miner seems no to work with the - a-1 and a-2 options. It works fine with a-3 ( I 've r290s ) Do you know what may cause that?

[kickstarter]

Most of the times I get 1180 and on rare occasions it hashes at 1330. 3 x Gigabyte 270 non x.. Any thought on what gives it that extra boost?

Overclocking via AMD Overdrive doesn't seem to make much difference. Is there a way to pass clock settings to the miner in the bat file?

Thanks

[Claymore]

I would hazard a guess that internally, it uses one of four OpenCL implementations based on the number you pass to the a switch. Since the AMD OpenCL compiler seems to be getting (slowly) less stupid, it's probable that Claymore has code in the OpenCL kernels specific to bugs in older drivers, that cause it not to work with 15.3.

Yes miner uses several kernels that are optimized for different cards. The problem is that I use some optimizations that are not allowed directly in OpenCL. They work properly on some driver versions and don't work on others.
I will release an update soon that supports overclocking and then will check 15.x drivers.

[Grim]

04/09/15-15:43:18 - SHARE FOUND (target 5000) - (GPU 0 of 1)

WATCHDOG: Thread 1 hangs in OpenCL call, you need to restart miner
WATCHDOG: GPU memory error, you need to restart miner
Restarting OK, exit...


I get that error on Thread 0 or 1 on a single 7950 every few minutes.
Driver 13.12

Can anyone help to solve this?

TIA

[Oscilson]

04/09/15-15:43:18 - SHARE FOUND (target 5000) - (GPU 0 of 1)

WATCHDOG: Thread 1 hangs in OpenCL call, you need to restart miner
WATCHDOG: GPU memory error, you need to restart miner
Restarting OK, exit...


I get that error on Thread 0 or 1 on a single 7950 every few minutes.
Driver 13.12

Can anyone help to solve this?

TIA

You need to reduce the clock frequency.

[Grim]

04/09/15-15:43:18 - SHARE FOUND (target 5000) - (GPU 0 of 1)

WATCHDOG: Thread 1 hangs in OpenCL call, you need to restart miner
WATCHDOG: GPU memory error, you need to restart miner
Restarting OK, exit...


I get that error on Thread 0 or 1 on a single 7950 every few minutes.
Driver 13.12

Can anyone help to solve this?

TIA

You need to reduce the clock frequency.

everything stock

[Claymore]

04/09/15-15:43:18 - SHARE FOUND (target 5000) - (GPU 0 of 1)

WATCHDOG: Thread 1 hangs in OpenCL call, you need to restart miner
WATCHDOG: GPU memory error, you need to restart miner
Restarting OK, exit...


I get that error on Thread 0 or 1 on a single 7950 every few minutes.
Driver 13.12

Can anyone help to solve this?

TIA

TROUBLESHOOTING

1. Install Catalyst v13.12.
2. Disable overclocking.
3. Set GPU_MAX_ALLOC_PERCENT 100.
4. Set Virtual Memory 16 GB.
5. Reboot computer.

If it does not help, the only option is to add more RAM.

[Jeremycoin]

Where is the 32bit??? I just downloaded the 64bit but my computer is not support it

[Claymore]

Where is the 32bit??? I just downloaded the 64bit but my computer is not support it

From "Readme!!!.txt":

...
This version is for Windows x64 and Linux x64 only. No 32-bit support.
...