Claymore's CryptoNote AMD GPU Miner v11.3

[Claymore]

Hi,

Thanks a lot for the linux support miner is awesome,,

The first machine of test,, i use centos 6.5 X64 and have a hd 5770 http://www.sapphiretech.com/presentation/product/product_index.aspx?pid=305&lid=1

2 gb of memory, in the linux my hash is only 86 hs also in windows was 149hs when overclocked. i see too, when program starting, it only recognize 512MB of my graphic card,,, whnm it have 1024mb; see the output.

catalyst 14.6

GPU #0: name: Juniper, 512 MB available, 9 units

Now i try with catalyst 13.12 has the same.

i tried to help with something

BB thanks.

Try to make OpenCL see all memory of your card. For Windows use "GPU_MAX_ALLOC_PERCENT", though I don't know if it works in Linux. Memory size is very important thing for good mining speed
.

Thanks for reply,

im using centos x64 and use "export GPU_MAX_ALLOC_PERCENT=100" Before exec the comand of the miner. Ati Catalyst show 1024MB in the info.

1. How many GPUs in the system?
2. How much RAM installed in the system?

[1713254712]

1713254712

[1713254712]

1713254712

[uray]

Here is the full log, when I start the miner:
-----------

Code:

./miner -u 47NHu3QopHC7sBrqraNJwqD9xXAWErAsDiEr3QKQcX3rFyKfGa2CPJcXsG9qEiYHNDT2nf6BwZPFMW3rxtES1MPwN1bCMTm -p x -a 0

имммммммммммммммммммммммммммммммммммммммммммммммммммммммммммммммм╩
╨            Claymore CryptoNote GPU Miner  v7.0 Beta            ╨
xмммммммммммммммммммммммммммммммммммммммммммммммммммммммммммммммм╪

Cards available: 6
GPU #0: name: Tahiti, 2861 MB available, 28 units
GPU #1: name: Tahiti, 2869 MB available, 28 units
GPU #2: name: Tahiti, 2869 MB available, 28 units
GPU #3: name: Tahiti, 2869 MB available, 28 units
GPU #4: name: Tahiti, 2869 MB available, 28 units
GPU #5: name: Tahiti, 2869 MB available, 28 units
Total cards: 6

Catalyst 13.12 is REQUIRED for best performance and compatibility
At least 16 GB of Virtual Memory is required for multi-GPU systems
Make sure you defined GPU_MAX_ALLOC_PERCENT 100
Be careful with overclocking, use default clocks for first tests
Initializing...

POOL version
Internal error: Input OpenCL binary is not for the target!
Internal error: Input OpenCL binary is not for the target!
GPU 0: -a 1 mode selected
GPU 1: -a 1 mode selected
GPU 2: -a 1 mode selected
GPU 3: -a 1 mode selected
GPU 4: -a 1 mode selected
GPU 5: -a 1 mode selected
4 pools specified.
Stratum - connecting to 'mine.moneropool.org' <104.131.59.214> port 80
Stratum - connecting to 'mine.moneropool.org' <104.131.59.214> port 80
Stratum - Connected
Stratum - Connected
Pool Diff 10000
DevFee: Pool Diff 10000
GPU 0, GpuMiner k1a failed -48
GPU memory allocation failed. Try to set at least 16 GB of Virtual Memory: Computer Properties / Advanced System Settings / Performance / Advanced / Virtual Memory
GPU 1, GpuMiner k1a failed -48
GPU 3, GpuMiner k1a failed -48
GPU memory allocation failed. Try to set at least 16 GB of Virtual Memory: Computer Properties / Advanced System Settings / Performance / Advanced / Virtual Memory
GPU 5, GpuMiner k1a failed -48
GPU memory allocation failed. Try to set at least 16 GB of Virtual Memory: Computer Properties / Advanced System Settings / Performance / Advanced / Virtual Memory
GPU 2, GpuMiner k1a failed -48
GPU 4, GpuMiner k1a failed -48
GPU memory allocation failed. Try to set at least 16 GB of Virtual Memory: Computer Properties / Advanced System Settings / Performance / Advanced / Virtual Memory
GPU memory allocation failed. Try to set at least 16 GB of Virtual Memory: Computer Properties / Advanced System Settings / Performance / Advanced / Virtual Memory
GPU memory allocation failed. Try to set at least 16 GB of Virtual Memory: Computer Properties / Advanced System Settings / Performance / Advanced / Virtual Memory
09/18/11-15:23:53 - New job received from mine.moneropool.org:80
Speed: 0 h/s, TotalHashes: 0K, DevHashes: 0K Mining time: 00:00

The other thing that is bugging me, that in the both archives uploaded, have identical sizes, and both give me the warning "Catalyst 13.12 is REQUIRED for best performance and compatibility", although the files in the archives are different.

I use driver 14.4, set swap to 20GB, system memory is 2GB, 6 x 7950 cards. Operating system is Ubunty Server 14.04 LTS. Tried both versions, 13.12 and 14.6, both are giving me the error above.

+1 does not work on multi gpu, i have 16GB of RAM, 4GB of swap space, 2 GPU 280x, Linux 64

the symptom is it work if i use -di for single gpu, for multi it seems the miner keep allocating on wrong gpu index, i notice that if i run miner on single gpu, then run another instance to use second gpu only, miner says gpu free memory is down to 900MB, yes it should down from 3GB to 0.9GB for first gpu, but not for second gpu, so when the second instance want to allocate memory for second gpu, somehow it allocate to wrong gpu that is currently used hence the memory allocation failure

[pgxz]

Hi,

Thanks a lot for the linux support miner is awesome,,

The first machine of test,, i use centos 6.5 X64 and have a hd 5770 http://www.sapphiretech.com/presentation/product/product_index.aspx?pid=305&lid=1

2 gb of memory, in the linux my hash is only 86 hs also in windows was 149hs when overclocked. i see too, when program starting, it only recognize 512MB of my graphic card,,, whnm it have 1024mb; see the output.

catalyst 14.6

GPU #0: name: Juniper, 512 MB available, 9 units

Now i try with catalyst 13.12 has the same.

i tried to help with something

BB thanks.

Try to make OpenCL see all memory of your card. For Windows use "GPU_MAX_ALLOC_PERCENT", though I don't know if it works in Linux. Memory size is very important thing for good mining speed
.

Thanks for reply,

im using centos x64 and use "export GPU_MAX_ALLOC_PERCENT=100" Before exec the comand of the miner. Ati Catalyst show 1024MB in the info.

1. How many GPUs in the system?
2. How much RAM installed in the system?

Have 2GB RAM 1 GPU.. i say that in first post.

Searching i found that!  http://devgurus.amd.com/thread/158397
Hey Sucessul here

in linux you need to put this 2 lines,, the first line solve my problem. thanks for the support.

export GPU_MAX_HEAP_SIZE=100
export GPU_MAX_ALLOC_PERCENT=100

[Claymore]

Hi,

Thanks a lot for the linux support miner is awesome,,

The first machine of test,, i use centos 6.5 X64 and have a hd 5770 http://www.sapphiretech.com/presentation/product/product_index.aspx?pid=305&lid=1

2 gb of memory, in the linux my hash is only 86 hs also in windows was 149hs when overclocked. i see too, when program starting, it only recognize 512MB of my graphic card,,, whnm it have 1024mb; see the output.

catalyst 14.6

GPU #0: name: Juniper, 512 MB available, 9 units

Now i try with catalyst 13.12 has the same.

i tried to help with something

BB thanks.

Try to make OpenCL see all memory of your card. For Windows use "GPU_MAX_ALLOC_PERCENT", though I don't know if it works in Linux. Memory size is very important thing for good mining speed
.

Thanks for reply,

im using centos x64 and use "export GPU_MAX_ALLOC_PERCENT=100" Before exec the comand of the miner. Ati Catalyst show 1024MB in the info.

1. How many GPUs in the system?
2. How much RAM installed in the system?

Have 2GB RAM 1 GPU.. i say that in first post.

Well, I don't have any ideas. For Windows such issues can be due to a lot of GPUs and/or small RAM size. For Linux I don't know, it must be something related to Linux+Catalyst.

[pgxz]

Hi,

Thanks a lot for the linux support miner is awesome,,

The first machine of test,, i use centos 6.5 X64 and have a hd 5770 http://www.sapphiretech.com/presentation/product/product_index.aspx?pid=305&lid=1

2 gb of memory, in the linux my hash is only 86 hs also in windows was 149hs when overclocked. i see too, when program starting, it only recognize 512MB of my graphic card,,, whnm it have 1024mb; see the output.

catalyst 14.6

GPU #0: name: Juniper, 512 MB available, 9 units

Now i try with catalyst 13.12 has the same.

i tried to help with something

BB thanks.

Try to make OpenCL see all memory of your card. For Windows use "GPU_MAX_ALLOC_PERCENT", though I don't know if it works in Linux. Memory size is very important thing for good mining speed
.

Thanks for reply,

im using centos x64 and use "export GPU_MAX_ALLOC_PERCENT=100" Before exec the comand of the miner. Ati Catalyst show 1024MB in the info.

1. How many GPUs in the system?
2. How much RAM installed in the system?

Have 2GB RAM 1 GPU.. i say that in first post.

Well, I don't have any ideas. For Windows such issues can be due to a lot of GPUs and/or small RAM size. For Linux I don't know, it must be something related to Linux+Catalyst.

SOLVED:

Have 2GB RAM 1 GPU.. i say that in first post.

Searching i found that!  http://devgurus.amd.com/thread/158397
Hey Sucessul here

in linux you need to put this 2 lines,, the first line solve my problem. thanks for the support.

export GPU_MAX_HEAP_SIZE=100
export GPU_MAX_ALLOC_PERCENT=100

If i help plz donate: 4A3agozMdgJdKojgeSrWkJLLjR7PdTXYw12rzdJXVscv5SaBwYCg8hBAyPgN6xGibmKNPq6q9DyLfJw zzy2FMQjTSzzgptf (XMR) (monero)

[Claymore]

ANN: I will release an update soon, most speedup (about 5%) is for 280x cards and for old cards.

[OrientA]

ANN: I will release an update soon, most speedup (about 5%) is for 280x cards and for old cards.

Does it consume more power?

[pgxz]

ANN: I will release an update soon, most speedup (about 5%) is for 280x cards and for old cards.

ME too .. UP

[bensam123]

Make that 25% now. XMR is reaching x11 levels of profitability, which isn't profitable -_-

Neat, so no one should ever release Nix miners if they want mining to stay in the hands of the average joe.

LOL - so force people to use Windows, because *nix users aren't the average joe. The sheer amount of stupid in that comment is staggering.

Or the average data center runs nix along with giant server farms full of miners because its much easier to manage on a wide scale. Perhaps elitist shit like the comment above is part of what keeps your everyday user away from nix too.

Don't blame me that you happen to be lumped in with the chinese farmer demograph. If you have 1-2 computers, you could easily run windows on them and manage them (and I'm sure you dual boot already because you can't run half the programs you use everyday on nix). But chances are if you're running nix you have shit ton miners. Nix isn't a everyday driver, it has very specialized uses, which it does very well, which is why people use it for very specific certain tasks... like databases, servers, render farms... mining farms.


Is elpedia memory going to be worked on in the new release?

[omnik]

ANN: I will release an update soon, most speedup (about 5%) is for 280x cards and for old cards.

Any news when we can expect new updated miner?
Tnx

[Claymore]

ANN: I will release an update soon, most speedup (about 5%) is for 280x cards and for old cards.

Any news when we can expect new updated miner?
Tnx

1-2 days.

[MTJ151]

Looking forward to it

[Prelude]

ANN: I will release an update soon, most speedup (about 5%) is for 280x cards and for old cards.

Any news when we can expect new updated miner?
Tnx

1-2 days.

Will the increase in speed cause an increase in power consumption?

[OrientA]

ANN: I will release an update soon, most speedup (about 5%) is for 280x cards and for old cards.

Any news when we can expect new updated miner?
Tnx

1-2 days.

Will the increase in speed cause an increase in power consumption?

If that is the case, there is no point of the new miner. We can overclock ourselves. We will see.

[Claymore]

Will the increase in speed cause an increase in power consumption?

It depends on "-a" mode. Right now "-a 1" means max hashrate but more power consumption. "-a 2" means less power consumption but a bit less hashrate. "-a 3" mode is for 6xxx cards.
To get max hashrate for different cards I had to add "-a 4" mode and also add "-h" option that sets number of hashes that GPU handles during one round. So now it depends on GPU what "-a" mode and "-h" option are the best. In some cases you will get more hashrate at same power consumption, but sometimes it will increase power consumption. In v8.0 you have to spend some time to get best speed, but for some cases speedup is very good, I got the following results after some tuning (compared to v7.0):

270x card: 19% speedup.
280x card: 7% speedup.
290x card: 1-2% speedup.
6950 2GB card: 3% speedup.

I need to finish a special "calibration" mode that helps to find best "-h" option value, and finish some other things, so it will take a couple of days to release new version.

[ZXSpectrum]

Will the increase in speed cause an increase in power consumption?

It depends on "-a" mode. Right now "-a 1" means max hashrate but more power consumption. "-a 2" means less power consumption but a bit less hashrate. "-a 3" mode is for 6xxx cards.
To get max hashrate for different cards I had to add "-a 4" mode and also add "-h" option that sets number of hashes that GPU handles during one round. So now it depends on GPU what "-a" mode and "-h" option are the best. In some cases you will get more hashrate at same power consumption, but sometimes it will increase power consumption. In v8.0 you have to spend some time to get best speed, but for some cases speedup is very good, I got the following results after some tuning (compared to v7.0):

270x card: 19% speedup.
280x card: 7% speedup.
290x card: 1-2% speedup.
6950 2GB card: 3% speedup.

I need to finish a special "calibration" mode that helps to find best "-h" option value, and finish some other things, so it will take a couple of days to release new version.

Is the Linux version faster than the Windows version? Any stats on that?

Thanks.

[Claymore]

Is the Linux version faster than the Windows version? Any stats on that?

Thanks.

Linux version has the same speed.

[Prelude]

Will the increase in speed cause an increase in power consumption?

It depends on "-a" mode. Right now "-a 1" means max hashrate but more power consumption. "-a 2" means less power consumption but a bit less hashrate. "-a 3" mode is for 6xxx cards.
To get max hashrate for different cards I had to add "-a 4" mode and also add "-h" option that sets number of hashes that GPU handles during one round. So now it depends on GPU what "-a" mode and "-h" option are the best. In some cases you will get more hashrate at same power consumption, but sometimes it will increase power consumption. In v8.0 you have to spend some time to get best speed, but for some cases speedup is very good, I got the following results after some tuning (compared to v7.0):

270x card: 19% speedup.
280x card: 7% speedup.
290x card: 1-2% speedup.
6950 2GB card: 3% speedup.

I need to finish a special "calibration" mode that helps to find best "-h" option value, and finish some other things, so it will take a couple of days to release new version.

Good stuff!

[huhliki.k]

Help, often there is an error:
 
Error in server response
: {"id":1,"jsonrpc":"2.0","error":{"code":-1,"message":"Unauthenticated"}}

What could be the problem?

[Claymore]

Help, often there is an error:
 
Error in server response
: {"id":1,"jsonrpc":"2.0","error":{"code":-1,"message":"Unauthenticated"}}

What could be the problem?

What pool? You can set "-dbg 2" option and send me the log file.