|
Pckomp, you should lower the mem clock as much as possible in order to save power. You'll not lose hashes.
|
|
|
|
I can connect 2 cards in the same rig to 2 different pools?
The contemporaneity is for the 2 cards in the same rig, not the third one.
The third card (in a different rig) is pluggued with a 7950 that remain stable.
I thought they where half in a rig and half in the other. Well, actually you could run two instances of cgminer selecting the devices: --device|-d <arg> Select device to use, one value, range and/or comma separated (e.g. 0-2,4) default: all |
|
|
|
Here is the fun part:
I have3 R9 290 in 2 different rig that always loos 0.2 MHs after 10 minutes, no matter what overclocking I gave to them.
BUT: the cards looses their MHS at the exact same second...I tried several reboot and look at my screen during a few minutes: It's stable for a while, then after a few minutes, at the exact same seconds, my 2 cards loose 0.2Mhs.
Are they connected to the same pool? Try two different ones and see if you still have the contemporaneity. |
|
|
|
I'll give it a try. By the way can you tell me wich excact Linux version you used with your 290?
Should I do a git clone ./autogen.sh ./configure make?
It said download and extract...
Once I got that installed, I launched it using od6config --list for example, no funky move of chmod of any sort? (I'm a total noob in Linux)
I'm on ubuntu 13.10. It doesn't seem to need superuser privileges. I did autogen configure make and run it from there without issues. Maybe: sudo aticonfig --od-enable --adapter=all before. |
|
|
|
I did some experiments with the opencl code in cgminer_keccak. I managed to get a little hashrate increase, nothing to write home about (0.5-2% depending on the card and intensity). If you want to try it out, get this file: https://www.dropbox.com/s/jc0s0gg12dpbjqu/keccak130718.clreplace the original one in the main cgminer folder, remove all the .bin files and run the miner as usual. |
|
|
|
I still need advice for R9 290. My cards looses 9% hashrate after 10 minutes on 2 rigs.
If it did not, please tell me what system you use and I gave it a test (bamt1.2, 1.3, 1.5...)
If you have some ideas or some tests I could do, you're welcome too.
you said you are on linux... did you try this tool: http://epixoip.github.io/od6config/did wonders to my 290, it can stabilize the clock to where I want it. |
|
|
|
I need some help with 290 cards.
I always started well with 2.4M hash, then after a few minutes, it starts decreasing slowly, all the way tl 2.1M.
It did that on all my 290, and just on my 290s: on my 280 I have some stable 2.1M from the beginning to the end.
I first suspected it was some sort of bios issue, but I got the same with a custom bios.
Any-one have a stable mining with a 290 that last a few hours above 2.4M?
I'm in a much worse situation: my 290 does a maximum of 500K, while an old 7950 does almost 1700K. |
|
|
|
I did not put the whole address here (but I can as it seems to be safe)
export DISPLAY=:0
export GPU_MAX_ALLOC_PERCENT=100
export GPU_USE_SYNC_OBJECTS=1
cgminer --keccak --no-submit-stale -o stratum+tcp://erebor.dwarfpool.com:3339 -u 8192mXkp7Xkg8imomY5Ka93dpEdTtGNYurjPhR -p 1
--failover-only -o stratum+tcp://moria.dwarfpool.com:3339 -u 8192mXkp7Xkg8imomY5Ka93dpEdTtGNYurjPhR -p 1
Anyway, I was thinking that even with a wrong address you will mine on the pool (but don't get any coins)
that 8192 at the front of the address looks like a copy paste error, maybe it was a --thread-concurrency leftover? |
|
|
|
Thanks
I think I use it: the 8192mXkp7Xkg8imomY5Ka just after the -u is my Wallet ID (It's a Wallet from Bter)
it doesn't look like a valid maxcoin address. |
|
|
|
Holy cow, this pool needs to find a way to decrease reject rates... around 15% here... really want to try this but reject rates are way too high. Please put up some server at US EAST, until then i will stay on middlecoin, getting only 1,5% rejects there.
payout is what counts ;-) Yeah well if you can make 10% more b/c 10% your hash is wasted. So..who wins there? Terk. this is not the point: if I get more btc on clevermining with 10% rejects than on middlecoin with 1% rejects, where should I mine? also high rejects may not always be fixable: it may depend on the coin mined. |
|
|
|
BTW. would anybody be interested in my patch to cpuminer-1gh, optimizing the keccak (3400 to 8100 khash/s on Core i5-2600K)? I guess nobody would care (40 W for 8 Mhash/s), but I was just testing for fun
Sure, I'd be interested. sent patch to Atrides, hopefully gets into 1gh cpuminer. kecasm.patch SHA3-512 checksum d1a2a36620e0b6b392ee6400276a8ede9b08a9ab23e2607ab6e16aeeef50d3ca679ffbcfcb19106 1f1247e1f56d1a6574818b50d7181cda739ff7d0d8aa35963 Thank you. Maybe you could commit it to https://github.com/Max-Coin/cpuminer so it's for everybody. |
|
|
|
Holy cow, this pool needs to find a way to decrease reject rates... around 15% here... really want to try this but reject rates are way too high. Please put up some server at US EAST, until then i will stay on middlecoin, getting only 1,5% rejects there.
payout is what counts ;-) |
|
|
|
#!/bin/sh
export DISPLAY=:0
export GPU_MAX_ALLOC_PERCENT=100
export GPU_USE_SYNC_OBJECTS=1
cd /home/zero/cgminer
./cgminer --scrypt -o stratum+tcp://us.clevermining.com:3333 -u 1HtQddmduuhgd9j96afd26qx1J3GLsHRJu -p one -I 14 -g 1 --lookup-gap 2 -w 256 --thread-concurrency 10048 --gpu-powertune 20 --temp-cutoff 90 --temp-overheat 85 --temp-target 70 --gpu-fan 30-90 --auto-fan --gpu-engine 1200 --gpu-memclock 1250 --expiry 1 --scan-time 1 --queue 1 --no-submit-stale --api-port 3000 --api-listen --api-network
with these settings i have 100% rejects, how can i fix them ? thanks
1 - which card(s) 2 - you should submit stales 3 - I believe queue is better at 0, scantime and expire very high (es. 9999) |
|
|
|
I can't get a decent darkcoin hashrate with my sapphire r9 290, while the other cards are ok.
It looks like the memory clock stays at 150 no matter what I do.
Any advice?
I did some other tests today, using the quark algorythm instead, forcing the clock with od6config: getting 1/5 the hashrate :-( |
|
|
|
Found a bug in maxcoin-qt g047adead I have never experienced before: when I change desktop away and back to where maxcoin-qt was, it has disappeared from view!
But it's still running and writing happily to log files. Does not show up in alt-tab list, either.
Fedora 20
qt-x11-4.8.5-15.fc20.x86_64
openbox-3.5.2-2.fc20
xorg-x11-server-Xorg-1.14.4-5.fc20.x86_64
BTW. would anybody be interested in my patch to cpuminer-1gh, optimizing the keccak (3400 to 8100 khash/s on Core i5-2600K)? I guess nobody would care (40 W for 8 Mhash/s), but I was just testing for fun
I'd test it, for fun  Maybe you'd like to have a look at keccak.cl as well?  |
|
|
|
anything that could be improved?
a bigger psu? 300W per card = 1200W, you are right at the limit. an r9 290 can make 450 with a bit of overclock, but not in your case, cause of not enough current. |
|
|
|
second night in a row that i find out that my rig is shut down, 10 hours wasted, that is it from me
hmmm are you overusing the power supply? that's a good shutdown reason. check it with a power meter. |
|
|
|
Hello, I'm crossposting here from the sph sgminer thread, I know I shouldn't but nobody answers me there:
I have a sapphire r9 290 which doesn't want to move the memory clock from 150 when mining darkcoins, leading to 1/10 of the available hashrate. All other cards are fine, and the card mines fine with other miners (engine clock can be changed instead).
Please help!
try editing the sgminer default .conf file to force values.... also, make sure you compiled with the right AMD sdk + AMD APP I run it like this, but I've tried any other possible setting: -I 19 -w 128 --gpu-engine 1100 --gpu-memclock 1250 --gpu-powertune 20 I'm on ubuntu with fglrx 2:13.200~beta-0ubuntu1~xedgers~saucy2. Compiled many flavors of cgminer and sgminer without issues in the past. my r 290x will not run stable above engine 1060 I tried rebooting and not setting any clock at all, same problem... 150 and no more :-( |
|
|
|
I can't get a decent darkcoin hashrate with my sapphire r9 290, while the other cards are ok.
It looks like the memory clock stays at 150 no matter what I do.
Any advice?
EDIT: if I run another miner after this one, the problem persists even there, until I reboot.
if your mem clock is at 150 it means the drivers stopped responding usually because you tried to over clock either the gpu or memory too far. This requires a reboot to fix. try different settings till it stops crashing unfortunately reboot doesn't fix it. actually it's not crashing, just auto-throttling the memory clock to 150. i tried not setting any clock at all, same problem... |
|
|
|
Hello, I'm crossposting here from the sph sgminer thread, I know I shouldn't but nobody answers me there:
I have a sapphire r9 290 which doesn't want to move the memory clock from 150 when mining darkcoins, leading to 1/10 of the available hashrate. All other cards are fine, and the card mines fine with other miners (engine clock can be changed instead).
Please help!
try editing the sgminer default .conf file to force values.... also, make sure you compiled with the right AMD sdk + AMD APP I run it like this, but I've tried any other possible setting: -I 19 -w 128 --gpu-engine 1100 --gpu-memclock 1250 --gpu-powertune 20 I'm on ubuntu with fglrx 2:13.200~beta-0ubuntu1~xedgers~saucy2. Compiled many flavors of cgminer and sgminer without issues in the past. |
|
|
|
|
Show Posts
