CCminer(SP-MOD) Modded GPU kernels.

[sp_]

And the speed? Any faster?

I compiled for windows with cuda 6.5. There is a windows bug in it, my 750Ti is only at 2.2+MHs instead of the possible 9+MHs.

Did you build with debug? then build release. I think the kernal need some tuning. Might work on it later. I see work has been done by tvprovut in his fork.

[1714695557]

1714695557

[1714695557]

1714695557

[1714695557]

1714695557

[skunk]

There is no projectfile for visual studio 2013. and the cuda 7.0 compiler is slow. Perhaps 30% faster with cuda 6.5.

You can tip the crypto mining blog, they have compiled a windows version before..

Ask them to build for cuda 6.5

doesn't make much of a difference actually, everything is slowed down my mem load/store. (moving to 32bit would probably help a little as well but get some problem with some dll)
actually can't get my 780ti/750ti to work properly because of the high mem allocation (which actually isn't so high... but it doesn't like that)
 

under linux ethminer allocates 1087mb on each gpu and almost 800mb on the system, i wonder if it's moving hashes between cpu and gpu's memory...

[ryen123]

And the speed? Any faster?

I compiled for windows with cuda 6.5. There is a windows bug in it, my 750Ti is only at 2.2+MHs instead of the possible 9+MHs.

Did you build with debug? then build release. I think the kernal need some tuning. Might work on it later. I see work has been done by tvprovut in his fork.

I always build release. From what I've read on the other forum, the same code works correctly in linux. Others that run in windows get bad hashrates. Whatever the problem is, is beyond my understanding.

My system: Win10, Nvidia driver 355.60, CUDA 6.5

[skunk]

And the speed? Any faster?

I compiled for windows with cuda 6.5. There is a windows bug in it, my 750Ti is only at 2.2+MHs instead of the possible 9+MHs.

Did you build with debug? then build release. I think the kernal need some tuning. Might work on it later. I see work has been done by tvprovut in his fork.

I always build release. From what I've read on the other forum, the same code works correctly in linux. Others that run in windows get bad hashrates. Whatever the problem is, is beyond my understanding.

My system: Win10, Nvidia driver 355.60, CUDA 6.5

windows 10 is the problem, it sends the missing hashes to microsoft's pool 

[ryen123]

And the speed? Any faster?

I compiled for windows with cuda 6.5. There is a windows bug in it, my 750Ti is only at 2.2+MHs instead of the possible 9+MHs.

Did you build with debug? then build release. I think the kernal need some tuning. Might work on it later. I see work has been done by tvprovut in his fork.

I always build release. From what I've read on the other forum, the same code works correctly in linux. Others that run in windows get bad hashrates. Whatever the problem is, is beyond my understanding.

My system: Win10, Nvidia driver 355.60, CUDA 6.5

windows 10 is the problem, it sends the missing hashes to microsoft's pool 

  haha...

[djm34]

Nvidia fixed a memory allocation bug in one of the latest drivers. 353.30 or newer.you should upgrade

using 353.62... my 750ti is running at a few kh/s

[sp_]

There is no projectfile for visual studio 2013. and the cuda 7.0 compiler is slow. Perhaps 30% faster with cuda 6.5.

You can tip the crypto mining blog, they have compiled a windows version before..

Ask them to build for cuda 6.5

doesn't make much of a difference actually, everything is slowed down my mem load/store. (moving to 32bit would probably help a little as well but get some problem with some dll)
actually can't get my 780ti/750ti to work properly because of the high mem allocation (which actually isn't so high... but it doesn't like that)
 

under linux ethminer allocates 1087mb on each gpu and almost 800mb on the system, i wonder if it's moving hashes between cpu and gpu's memory...

800meg is alot. Whatbis your hashrate?

[djm34]

There is no projectfile for visual studio 2013. and the cuda 7.0 compiler is slow. Perhaps 30% faster with cuda 6.5.

You can tip the crypto mining blog, they have compiled a windows version before..

Ask them to build for cuda 6.5

doesn't make much of a difference actually, everything is slowed down my mem load/store. (moving to 32bit would probably help a little as well but get some problem with some dll)
actually can't get my 780ti/750ti to work properly because of the high mem allocation (which actually isn't so high... but it doesn't like that)
 

under linux ethminer allocates 1087mb on each gpu and almost 800mb on the system, i wonder if it's moving hashes between cpu and gpu's memory...

ha ok... I get 1159MB allocated, but the problem it has problem to access it efficiently... power usage for the 750ti gets around 16%
I have the same problem with cryptonight actually and have to decrease mem allocation to below 1GB

[skunk]

There is no projectfile for visual studio 2013. and the cuda 7.0 compiler is slow. Perhaps 30% faster with cuda 6.5.

You can tip the crypto mining blog, they have compiled a windows version before..

Ask them to build for cuda 6.5

doesn't make much of a difference actually, everything is slowed down my mem load/store. (moving to 32bit would probably help a little as well but get some problem with some dll)
actually can't get my 780ti/750ti to work properly because of the high mem allocation (which actually isn't so high... but it doesn't like that)
 

under linux ethminer allocates 1087mb on each gpu and almost 800mb on the system, i wonder if it's moving hashes between cpu and gpu's memory...

800meg is alot. Whatbis your hashrate?

~52 mhs, so ~8.7 mhs per card

[sambiohazard]

Today's share sent

djm34: 047171eabd9db41dea5d4f8aece3647bd6990118c09c4ef0c056933e6920f5a7
pallas: ff450e633bec606b921855a307a4fbad1f37c1afa23e69ba426184fe7b2cae3f
sp_: 13acb0a16bb2c0252f04a864c24e59ba8e92cbc0080809476c6c15ad6bc2bfaf

Thanks. Keep going

[tbearhere]

@djm34  ......Sent you some beers for your Lyra2v2.v4....or is it a 6 pack.      .05btc

TrxID: cdd1ad95ace6de79024b50863a51743c0cead3125c09cda59ea09802176ea90c

4.6 mh 750ti oc 1350.

[blue-yu]

Guys, ccminer is reporting 28.965 kH/s, but Nicehash is reporting only around 15.000 kH/s! What should be wrong?

[hashbrown9000]

@blue_yu, try mining on give me coins pool with no diff setting.  that's the only way i could get any vtc pool to agree with my local total hashrate.

[Genoil]

There is no projectfile for visual studio 2013. and the cuda 7.0 compiler is slow. Perhaps 30% faster with cuda 6.5.

You can tip the crypto mining blog, they have compiled a windows version before..

Ask them to build for cuda 6.5

doesn't make much of a difference actually, everything is slowed down my mem load/store. (moving to 32bit would probably help a little as well but get some problem with some dll)
actually can't get my 780ti/750ti to work properly because of the high mem allocation (which actually isn't so high... but it doesn't like that)
 

I've spent a lot of time already trying to figure out why the algo doesn't perform well on some windows/nvidia combo's. this gtx750ti i have at home gets only 500KH, but when i pull it out and place it along a gtx780 i run at work, it pulls 8MH/s (as it should, and most linux people report).

The 1GB RAM allocation (both on RAM and per GPU) is for the "dagger" file. This is used in the "middle" part of the ethash algo, where a random 64 node path is walked through that file. This takes up most of the time, about 85% compared to the keccak stuff. The dagger file updates each 30,000 blocks. This all with the goal of fighting off ASIC development.

The code as I have in git now is a bit of a mess, with two approaches of running the dagger loop dumped into a single file. One approach uses shared memory (compute_hash), whereas the other one warp shuffles instead (compute_hash_shuffle). Turned out to hardly make a difference. By default shuffle is now used.

Currently looking into splitting the whole thing back up into 3 parts (keccak/dagger/keccak) and see if i can run these kernels concurrently. there's a branch for that, cudaminer-staged.   

[djm34]

There is no projectfile for visual studio 2013. and the cuda 7.0 compiler is slow. Perhaps 30% faster with cuda 6.5.

You can tip the crypto mining blog, they have compiled a windows version before..

Ask them to build for cuda 6.5

doesn't make much of a difference actually, everything is slowed down my mem load/store. (moving to 32bit would probably help a little as well but get some problem with some dll)
actually can't get my 780ti/750ti to work properly because of the high mem allocation (which actually isn't so high... but it doesn't like that)
 

I've spent a lot of time already trying to figure out why the algo doesn't perform well on some windows/nvidia combo's. this gtx750ti i have at home gets only 500KH, but when i pull it out and place it along a gtx780 i run at work, it pulls 8MH/s (as it should, and most linux people report).

The 1GB RAM allocation (both on RAM and per GPU) is for the "dagger" file. This is used in the "middle" part of the ethash algo, where a random 64 node path is walked through that file. This takes up most of the time, about 85% compared to the keccak stuff. The dagger file updates each 30,000 blocks. This all with the goal of fighting off ASIC development.

The code as I have in git now is a bit of a mess, with two approaches of running the dagger loop dumped into a single file. One approach uses shared memory (compute_hash), whereas the other one warp shuffles instead (compute_hash_shuffle). Turned out to hardly make a difference. By default shuffle is now used.

Currently looking into splitting the whole thing back up into 3 parts (keccak/dagger/keccak) and see if i can run these kernels concurrently. there's a branch for that, cudaminer-staged.  

thanks, that answers the question I forgot to ask (about the staged branch).  

actually I have 2 980 which runs ok, but the 2 other card 780ti + 750ti are doing about 1.5MH/s together... (can be due to the fact they are on riser...)

edit: I get "illegal memory access" with the shared memory method

[scryptr]

ETHMINER and LINUX--

Somebody shared a pre-compiled CUDA Compute 50 Linux binary on Genoil's blog thread, and I toyed with it.  Just as in CCminer, there are shared library files.  The file will not run unless you have the Ethereum system installed properly.  I don't think I will install it on my 750ti rig.

If I rebuild my AMD 280x rig, I will install it, and run ethminer instead of sgminer.  It all depends on ETH and its value on the coin market.  I do suspect that Ethminer would run smoothly, and with less energy usage and the headaches associated with sgminer.       --scryptr

[Genoil]

edit: I get "illegal memory access" with the shared memory method

hm that probably got borked in all the mess...it used to be fine.

i'll see if i can fix that tomorrow.

[hashbrown9000]

Somebody shared a pre-compiled CUDA Compute 50 Linux binary on Genoil's blog thread, and I toyed with it.

linky?

[scryptr]

Somebody shared a pre-compiled CUDA Compute 50 Linux binary on Genoil's blog thread, and I toyed with it.

linky?

LINKY--

Here is a link to page 6 of Genoil's thread:  "https://forum.ethereum.org/discussion/2227/cuda-miner/p6"

Look for poster SuchFakeAccount near the bottom of the page.  He kindly said he would compile other binaries if asked.  Pages 6 and 7 have a few binaries on them, just read the thread.

The file has some dependencies, and I don't have them.  If I could place the dependencies in the directory like Windows *.dll files, I would.  I just don't want to install Ethereum on my good rig and then have complications with CCminer compilation.

--scryptr

[zTheWolfz]

What sort of luck are you guys having with mining Ethereum? I've been mining for about 24 hours and haven't found a block. Everything is so glitchy and buggy I can't tell if it's working right even though it should be.

Since there is no pool it seems as though this might be sorta pointless too. Pools are there because mining alone doesn't net the same results, even though it's supposed to be averaged over time it usually doesn't work that way. I may switch back to VTC till this ends up a bit more mature.

I HIT ONE BLOCK--

I have 5 ETH coins after mining solo since sometime Friday.  The prediction is about 10 coins a week for 25Mh/s, my current speed.  Without a GUI wallet, I have to check my balance every few hours or so.  Mining works.  If optimized in CUDA code, I could earn a lot more.       --scryptr

Try to see if 2>> ETHlog.txt flag will work with the miner software, if it does it will place a log file within the folder where the miner resides.
the 2>> name-can-be-what-ever-you-want-log.txt You can then open that log with notepad and use the search for yay! to find blocks you found
or boo! to find bad blocks within the log file. Asumming it uses yay and boo.