ethminer-0.9.41-genoil-1.1

[Amph]

so i guess there will be no more 1.0.9 and pascal support will be added in the 1.2+ version

btw i still have that heavy cpu utilization, but only when mining in stratum mode on some pool like cointron

[1713520370]

1713520370

[scryptr]

HOW DO I CLONE VERSION 110? --

I must have cloned the current version 1.0.8 from GIT.  How do I clone the v110?  I have researched, but am not familiar with "checkout" commands. Working in Linux, here.       --scryptr

cd cpp-ethereum
git checkout 110

I think, I tend to forget these things. If it doesn't work because you cloned the repo earlier, git pull first.

GIT CHECKOUT 110 WORKED--

And, I was able to correctly build pre-release v1.1.  The generation of the DAG file directly on the GPUs took about 5 seconds, rather than minutes for on-disk generation.  No more DAG directory to clean up!  The version label was correct, and in place at program launch.

I did notice that the percentage completion display during DAG file generation scrolled down the screen instead of remaining in one spot.  It went by with speed, however.    Great job!  Thank you...       --scryptr

[Pustul]

I just want to say thank you so much Genoil for the continued support and the great update

Nice work on the DAG creation!

[moaddip]

btw i still have that heavy cpu utilization, but only when mining in stratum mode on some pool like cointron

same problem

[jstefanop]

I just ran a little test on both Nvidia (CUDA) and AMD (OpenCL) where I switched between current and next DAG every 5 blocks. No problems whatsoever. I don't have a multi-card rig, but I don't see how that should be any different, other than maybe a bit of clutter in the log output while changing DAGs.

Unfortunately DAGs are created so fast that the custom pcie bus we are running all these GPUs on is getting overloaded and the OpenCL calls hang. I guess I got lucky and it started the first time but quickly crashed.

The only way to solve this issue is to implement the same serialization fix claymour did on his miner (i.e. generate DAGs one by one serially, and any OpenCL calls have to have a delay in them (especially when new work is pushed to all the GPUs). Unfortunately I don't think thats an easy fix with ethminer and how threaded it is.

FYI this wont effect most users running less than 6 GPUs

This on 1.1 i assume. There's not so much going over the bus, just about 50-100MB of DAG cache. When did it actually crash? During DAG generation or during mining? Because when mining, it;s not really that much different from 1.0.8, other than some code commented out. It must be something else then this, perhaps removing all this DAG crap opened up another weird ethminer bug.

no i have debugged the issues extensively and confirmed its a bug with how fglrx makes openCL calls with multiple GPUs (and its made worse with bridges we use). Here is an example of where it usually crashes

Code:

[OPENCL]:Printing program log
[OPENCL]:
[OPENCL]:Creating cache buffer
[OPENCL]:Creating DAG buffer
[OPENCL]:Loading kernels
[OPENCL]:Writing cache buffer
0%
67%
[OPENCL]:Creating buffer for header.
[OPENCL]:Creating mining buffer 0
[OPENCL]:Creating mining buffer 1
[OPENCL]:Generating DAG data
8%
42%
25%
50%
33%
67%
8%
75%
17%
0%
50%
33%
58%
42%
75%
17%
83%
25%
8%
92%
58%
42%
50%
67%
83%
25%
92%
33%
17%
67%
50%
58%
75%
33%
42%
25%
75%
58%
67%
83%
92%
42%
50%
33%
83%
  ℹ  15:06:03|gpuminer0  set work to: #6a82dd23, target  #0000000112e0be82
67%
92%
75%
50%
58%
  ℹ  15:06:03|gpuminer1  set work to: #6a82dd23, target  #0000000112e0be82
42%
75%
83%
92%
58%
67%
50%
  ℹ  15:06:03|gpuminer2  set work to: #6a82dd23, target  #0000000112e0be82
83%
  ℹ  15:06:04|gpuminer3  set work to: #6a82dd23, target  #0000000112e0be82
92%
67%
75%
58%
75%
83%
67%
92%
83%
75%
92%
83%
92%

As you can see there is way to much I/O going on between 9 GPUs in short amount of time, which causes the openCL call hangs. The only fix until AMD decides to revamp the AMDGPU drivers is to serialize the OpenCL calls and make sure they don't overlap between GPUs.

[Genoil]

found a bug in 1.1 cauisng it to hang after a random amount of blocks.. it's not production ready yet, stay tuned  

@jstefanop: during DAG generation ok. serialisation seems the bvest way forward then.

[greaterninja]

Hello Genoil,


Thank you for your great work!

I'm using  version 1.10   x 6 AMD GPU setup on ubuntu 14.04.

Should I rollback my version of ethminer?

[Genoil]

Hello Genoil,


Thank you for your great work!

I'm using  version 1.10   x 6 AMD GPU setup on ubuntu 14.04.

Should I rollback my version of ethminer?

It has a chance of deadlock in stratum mode. If it loads the GPUs and you use getwork, it should be fine

[moaddip]

support stratum  from dwarf?

[Genoil]

support stratum  from dwarf?

nope.

RE: CPU usage: I'm considering rewriting the stratum client to be single threaded. I think that will reduce CPU load on simple CPUs.

RE: 1.0.9. I figured getting rid of the DAG would be a nice milestone to push the minor version number a notch  . 1.0.8 already runs on Pascal by the way. Just cmake with -DCOMPUTE=61 (CUDA 8 required) or use OpenCL. And use Linux for now to get a bit of hashrate.

RE: deadlock: have pushed new src/binary that may resolve the deadlock. Still running a test.

[tolazy]

for some reason i only get 25mhs with 1.1
i'm getting 29mhs with 1.08
and with claymore i get 31mhs

[Monzsta]

No more DAG?  Does this mean I can throw the old 5850 back in there?

[Xdragon]

^ No

[pehoko]

No more DAG?  Does this mean I can throw the old 5850 back in there?

Yes if you expand gpu memory to 2gb.

[Pustul]

No more DAG?  Does this mean I can throw the old 5850 back in there?

There is still a DAG, but it is generated directly on the GPU instead of on your hard drive and then used by the GPU. So no...

[Genoil]

Genoil - I KNOW this is a nitpick - but I noticed as I was going through and rewriting a lot of the Ethash OpenCL... this one just bugs the shit outta me:

Code:

bool update_share = thread_id == ((a >> 2) & (THREADS_PER_HASH - 1));

Were it performance critical, that would be ouch. As is, on AMD, it just is a little eww, when you could:

Code:

bool update_share = thread_id == amd_bfe(a, 2U, 3U)

Thanks I'll have a look at it. Up until now it's been highly demotivating to try and optimize the OpenCL kernel, because nothing really made a difference. CUDA has been much more willing to give me a few % improvement over the baseline.

[Tmdz]

The miner works great, really stable but I have a question.

On my rig that I just put together the miner works fine running it but if I go to close the miner it locks the computer up and I have to force a shutdown.  I think it is a graphical issue of some sort.  I have 2 R9 290 reference card hooked up to pci-e x1 powered risers and it is older lga 775 hardware with win 8.1.  On my other 775 machine with 2x hd 7950's I do not have the issue and works flawlessly.

I have the machine using video from one of the cards, I think the issue goes away if I force video through integrated hardware but I have no control over the cards, so I cant see temps, change fan speed, or undervolt in msi afterburner.

Edit: posting on the main forum

[dayverr]

Hello every one. I am new here. How can i mine with amd and nvdia cards in one rig (gf970,gf960,r370,r380). I use win7 x64, 15.12 amd driver and  347.52 nvdia driver.
ethminer-0.9.41-genoil-1.0.8 and eth-proxy (stratum i think)

Miner use only amd cards.
I hear something about -R option, but i have no idea in what version it works, and where and how i must to use it for good result.

this is my bat file :

setx GPU_FORCE_64BIT_PTR 0
setx GPU_MAX_HEAP_SIZE 100
setx GPU_USE_SYNC_OBJECTS 1
setx GPU_MAX_ALLOC_PERCENT 100
setx GPU_SINGLE_ALLOC_PERCENT 100
ethminer.exe --farm-recheck 200 -G -F http://127.0.0.1:8080/mg970
pause


what  i must to edit?

ps. sorry for my english, i try to study it time to time.

[Genoil]

Hello every one. I am new here. How can i mine with amd and nvdia cards in one rig (gf970,gf960,r370,r380). I use win7 x64, 15.12 amd driver and  347.52 nvdia driver.
ethminer-0.9.41-genoil-1.0.8 and eth-proxy (stratum i think)

Miner use only amd cards.
I hear something about -R option, but i have no idea in what version it works, and where and how i must to use it for good result.

this is my bat file :

setx GPU_FORCE_64BIT_PTR 0
setx GPU_MAX_HEAP_SIZE 100
setx GPU_USE_SYNC_OBJECTS 1
setx GPU_MAX_ALLOC_PERCENT 100
setx GPU_SINGLE_ALLOC_PERCENT 100
ethminer.exe --farm-recheck 200 -G -F http://127.0.0.1:8080/mg970
pause


what  i must to edit?

ps. sorry for my english, i try to study it time to time.

Launch another ethminer, replace -G with -U. If 1.0.8, add -R path/to/directory. If new 1.1, you no longer need -R.

[dayverr]

Thanks a lot.