[ANN] [ASIC-RESISTANT] UltraCoin (UTC) - Ultrafast 6 second transactions!!

[Demontager]

lol keep daydreaming...

Try this, download yacminer 3.5 and use the following settings:

--scrypt-chacha --nfmin 4 --nfmax 30 --starttime 1388361600 --no-submit-stale --expiry 10 --scan-time 1 --queue 0 --gpu-engine 1120 --gpu-memclock 1200 --gpu-fan 0-100 --temp-target 75 --temp-cutoff 100 --temp-overheat 90 --auto-gpu --auto-fan --rawintensity 5120 --buffer-size 1698 -g 1 -w 256 --lookup-gap 4

Same setting as yours, but with lowered "buffer-size" to  "1624", because 1698 unable to allocate

Code:

 [2014-04-06 21:35:09] GPU 0: bufsize for thread @ 1698MB based on buffer-size
 [2014-04-06 21:35:09] GPU 1: bufsize for thread @ 1698MB based on buffer-size
 [2014-04-06 21:35:09] Network diff set to 77.2K
 [2014-04-06 21:35:09] Error -4: Enqueueing kernel onto command queue. (clEnqueueNDRangeKernel)
 [2014-04-06 21:35:09] GPU 0 failure, disabling!
 [2014-04-06 21:35:09] Thread 0 being disabled
 [2014-04-06 21:35:09] Error -4: Enqueueing kernel onto command queue. (clEnqueueNDRangeKernel)

anyway 31 khas/s with no HW looks reasonable. Thank you

[1715201014]

1715201014

[1715201014]

1715201014

[1715201014]

1715201014

[1715201014]

1715201014

[shpakov]

Guys, any config for r9 280x ? I still got HW

I got 60kh with yacminer and no hw, but its hard to reach if you have more than 1 card
need some woodoo magic, deleting binaries, etc

Я пoлyчaю 64кH нa 4*280x бeз вcякoй мaгии. Читaйтe мoи cooбщeния paнee, я дaвaл нacтpoйки и для 4*280x и для 4*270x (39кH)

[jakiman]

I have asked a community leader from another coin community to provide us with a capable coder to assist in giving the client a much needed overhaul. Hopefully we'll get an answer today. Yes Nitro is a problem at the moment, as well as a solution. Nitro does run faster than the network, by about 300%
This does not give other pools a level playing field not even Nitro2. When the client update comes, We will take drastic action to spread the hash. This step at the moment will be fruitless. I am well aware of the danger of having one pool. I am hoping that in a week time with a good coder we can have a new client and we will update and run this new client without any alteration in the process of how it runs. That should make all the pools equal in speed and performance. We detected this bug in the first week, and reported it, and I did mention, it will get worse with every Nfactor change. We are working on some optimizations on Nitro2 at the moment to bring it up to speed. Just bare with us one more week, till we get the client fixed. This bug btw does not affect the average user, but some users might notice the client is not starting up as quick, or going into a coma, before it wakes up 10 to 15 minutes later, just imagine this happening on a client with 1000+ workers. I know some of you will not believe what I say regardless of how many times I say it, but I have been honest and truthful about it. At one point our client was doing that, we did not tamper with the code to fix it, as that will make us responsible for any screwups. We just applied an expensive fix. Just give us one more week to work on a solution that will benefit both pools and community. Also there is some other issues that need to be addressed as well, PoS, Coin Control, Block reward, and once done it will make UTC stronger and more trustworthy. It is solutions like these that will make UTC a better coin for the future. I hope once the coder/lead developer is present he might be able to describe to you the problem in more details, and apply the fix in a short time.

Thanks for the great news and support!

I noticed you removed invitations from Nitro. That's a good start also.
As long as Nitro1 doesn't grow larger, it should force others to grow.

[Demontager]

Guys, any config for r9 280x ? I still got HW

I got 60kh with yacminer and no hw, but its hard to reach if you have more than 1 card
need some woodoo magic, deleting binaries, etc

Я пoлyчaю 64кH нa 4*280x бeз вcякoй мaгии. Читaйтe мoи cooбщeния paнee, я дaвaл нacтpoйки и для 4*280x и для 4*270x (39кH)

Stop posting in russian, respect others! (Пoжaлyйcтa, пишитe нa aнглийcкoм, yвaжaйтe дpyгиx coбeceдникoв!)

[xwebnetwork]

Thanks to everyone for the yacminer tips. Got it nicely hashing along now on two of my rigs. I had the same issue as others with 4 x 280x, needed to drop the buffer to 1760 and drop R until no hardware errors. Now the rig is running @ 200 kh/s  

So you have 4 x 280X's running at 50KHs per with 8GB RAM? May I have your final settings? Thanks!

Here you go, I was the first to post it which now everyone is using AFAIK.  
https://bitcointalk.org/index.php?topic=413978.msg6093996#msg6093996

280x(7970) @ 60Kh/s

55Kh/s at 1025/1500 @ 1.075v
60Kh/s at 1110/1500 @ 1.150v

1. Use latest YACMiner. https://github.com/Thirtybird/YACMiner/releases
2. Use following core settings: (Add other settings, more pools as needed)

yacminer.exe --scrypt-chacha -o <pool> -u <id.worker> -p <password> -R 8000 -g 1 -w 128 --buffer-size 2656 --lookup-gap 3 --temp-cutoff 85 --nfmin 4 --nfmax 30 --starttime 1388361600

Note: This may ONLY work on a PC with 8GB "physical" RAM regardless of VM size. (I tried up to 32GB)
If you have 4GB RAM machine and it fails, try using a lower buffer-size of 1760 and -R of 6144 or 5888.
As the rawintensity is lower, you will end up with a lower hash rate of about 45Kh/s unfortunately.

If this helped: UXLaLvo8L4eV9uC355WRzTX88GA7iFja2g  

Some tutorial on YACMiner for everyone

YACMiner allows finer control of the intensity value so you can squeeze every bit of its hash rate which ultracoin-miner cannot.

--buffer-size cannot go over the size of your video card's RAM amount. It's usually around 10% less value.
-R is basically Intensity equivalent in cgminer. YACMiner allows for more granular in-between Intensity values. (e.g. 12 = 4096 & 13 = 8192)

Goal is to try and squeeze in as many threads (intensity) in to the amount of buffer without it going into dynamic memory.
So first try to find the largest max buffer size your system will work with. Go up in increments of 128 until it fails to start.
Depending on system RAM size, you may need to start as low as 50% of your single video card's RAM amount. (e.g. 3GB video card  1536)
Once you know the max buffer size that works, set it as that and leave it alone.

Now just adjust the -R value in increments of 128. Good to start from something low such as 4096 and work your way up.
If you set it too high, you will either get HW errors and much less WU or it'll just fail/lockup/crash etc. Dial it down if so.

Lastly, check your WU over at least 15-20minutes. Clocking higher might givss higher hash rate but WU might not go up.
You might end up using 5-10% more electricity but end up not actually doing any more real work. So keep an eye on WU.

Awesome. Worked like a charm. Slightly modified the clocks to 1000/1250 and still pulling 55KHs per.

Code:

--scrypt-chacha -o <pool> -u <id.worker> -p <password> -R 8000 -g 1 -w 128 --buffer-size 2656 --lookup-gap 3 --temp-cutoff 85 --nfmin 4 --nfmax 30 --starttime 1388361600 --gpu-engine 1000 --gpu-memclock 1250 --auto-fan --temp-target 75 --temp-overheat 80 --temp-cutoff 85

Also, YACMiner does work with CGWatcher. Just need to rename the EXE to "Cgminer.exe". Furthermore, on my one Windows 8 Rig, one of the cards doesn't seem to play well with YAC. Works fine with CGMiner. Have rebooted but will keep on tweaking to figure it out. Thank you again!

[shpakov]

Stop posting in russian, respect others! (Пoжaлyйcтa, пишитe нa aнглийcкoм, yвaжaйтe дpyгиx coбeceдникoв!)

Well, if you're too lazy to use the built-in browser translator in order to read my posts, as I do for what to read this forum, just nothing else will be writing;)
Speak their native language is not no respect for others.

(Xopoшo,ecли Baм лeнь вocпoльзoвaтьcя вcтpoeнным в бpayзep пepeвoдчикoм, чтo бы пpoчитaть мoи cooбщeния,кaк этo дeлaю я для тoгo,чтo бы читaть дaнный фopyм, тo  пpocтo бoльшe ничeгo нe бyдy пиcaть
Paзгoвapивaть нa poднoм языкe нe являeтcя HE yвaжeниeм к дpyгим.)

[tungz10]

Thanks to everyone for the yacminer tips. Got it nicely hashing along now on two of my rigs. I had the same issue as others with 4 x 280x, needed to drop the buffer to 1760 and drop R until no hardware errors. Now the rig is running @ 200 kh/s  

So you have 4 x 280X's running at 50KHs per with 8GB RAM? May I have your final settings? Thanks!

Here you go, I was the first to post it which now everyone is using AFAIK.  
https://bitcointalk.org/index.php?topic=413978.msg6093996#msg6093996

280x(7970) @ 60Kh/s

55Kh/s at 1025/1500 @ 1.075v
60Kh/s at 1110/1500 @ 1.150v

1. Use latest YACMiner. https://github.com/Thirtybird/YACMiner/releases
2. Use following core settings: (Add other settings, more pools as needed)

yacminer.exe --scrypt-chacha -o <pool> -u <id.worker> -p <password> -R 8000 -g 1 -w 128 --buffer-size 2656 --lookup-gap 3 --temp-cutoff 85 --nfmin 4 --nfmax 30 --starttime 1388361600

Note: This may ONLY work on a PC with 8GB "physical" RAM regardless of VM size. (I tried up to 32GB)
If you have 4GB RAM machine and it fails, try using a lower buffer-size of 1760 and -R of 6144 or 5888.
As the rawintensity is lower, you will end up with a lower hash rate of about 45Kh/s unfortunately.

If this helped: UXLaLvo8L4eV9uC355WRzTX88GA7iFja2g  

Some tutorial on YACMiner for everyone

YACMiner allows finer control of the intensity value so you can squeeze every bit of its hash rate which ultracoin-miner cannot.

--buffer-size cannot go over the size of your video card's RAM amount. It's usually around 10% less value.
-R is basically Intensity equivalent in cgminer. YACMiner allows for more granular in-between Intensity values. (e.g. 12 = 4096 & 13 = 8192)

Goal is to try and squeeze in as many threads (intensity) in to the amount of buffer without it going into dynamic memory.
So first try to find the largest max buffer size your system will work with. Go up in increments of 128 until it fails to start.
Depending on system RAM size, you may need to start as low as 50% of your single video card's RAM amount. (e.g. 3GB video card  1536)
Once you know the max buffer size that works, set it as that and leave it alone.

Now just adjust the -R value in increments of 128. Good to start from something low such as 4096 and work your way up.
If you set it too high, you will either get HW errors and much less WU or it'll just fail/lockup/crash etc. Dial it down if so.

Lastly, check your WU over at least 15-20minutes. Clocking higher might givss higher hash rate but WU might not go up.
You might end up using 5-10% more electricity but end up not actually doing any more real work. So keep an eye on WU.

Awesome. Worked like a charm. Slightly modified the clocks to 1000/1250 and still pulling 55KHs per.

Code:

--scrypt-chacha -o <pool> -u <id.worker> -p <password> -R 8000 -g 1 -w 128 --buffer-size 2656 --lookup-gap 3 --temp-cutoff 85 --nfmin 4 --nfmax 30 --starttime 1388361600 --gpu-engine 1000 --gpu-memclock 1250 --auto-fan --temp-target 75 --temp-overheat 80 --temp-cutoff 85

Also, YACMiner does work with CGWatcher. Just need to rename the EXE to "Cgminer.exe". Furthermore, on my one Windows 8 Rig, one of the cards doesn't seem to play well with YAC. Works fine with CGMiner. Have rebooted but will keep on tweaking to figure it out. Thank you again!

same with me out of 6 cards 1 cards posted 0.000k/0.000kh/s...

[jakiman]

Snip: https://bitcointalk.org/index.php?topic=413978.msg6101155#msg6101155

Awesome. Worked like a charm. Slightly modified the clocks to 1000/1250 and still pulling 55KHs per.

Code:

--scrypt-chacha -o <pool> -u <id.worker> -p <password> -R 8000 -g 1 -w 128 --buffer-size 2656 --lookup-gap 3 --temp-cutoff 85 --nfmin 4 --nfmax 30 --starttime 1388361600 --gpu-engine 1000 --gpu-memclock 1250 --auto-fan --temp-target 75 --temp-overheat 80 --temp-cutoff 85

Also, YACMiner does work with CGWatcher. Just need to rename the EXE to "Cgminer.exe". Furthermore, on my one Windows 8 Rig, one of the cards doesn't seem to play well with YAC. Works fine with CGMiner. Have rebooted but will keep on tweaking to figure it out. Thank you again!

I'm glad it worked!

Do you have 8GB system RAM? Try increasing virtual memory to at least 16GB also to see if it makes a difference.
Maybe your system cannot address enough RAM to initialize. (try closing some apps etc. =p)
Other than that, it's all about finding the sweet spot of -R and --buffer-size really.

[ivanlabrie]

Stop posting in russian, respect others! (Пoжaлyйcтa, пишитe нa aнглийcкoм, yвaжaйтe дpyгиx coбeceдникoв!)

Well, if you're too lazy to use the built-in browser translator in order to read my posts, as I do for what to read this forum, just nothing else will be writing;)
Speak their native language is not no respect for others.

(Xopoшo,ecли Baм лeнь вocпoльзoвaтьcя вcтpoeнным в бpayзep пepeвoдчикoм, чтo бы пpoчитaть мoи cooбщeния,кaк этo дeлaю я для тoгo,чтo бы читaть дaнный фopyм, тo  пpocтo бoльшe ничeгo нe бyдy пиcaть
Paзгoвapивaть нa poднoм языкe нe являeтcя HE yвaжeниeм к дpyгим.)

操你妈的屄

+1

Somehow google translate isn't translating stuff on the go cause the site is in "English".
Podria postear en español y tampoco sería apropiado...most people will understand English (despite how sad that makes you or me lol)

[xwebnetwork]

Snip: https://bitcointalk.org/index.php?topic=413978.msg6101155#msg6101155

Awesome. Worked like a charm. Slightly modified the clocks to 1000/1250 and still pulling 55KHs per.

Code:

--scrypt-chacha -o <pool> -u <id.worker> -p <password> -R 8000 -g 1 -w 128 --buffer-size 2656 --lookup-gap 3 --temp-cutoff 85 --nfmin 4 --nfmax 30 --starttime 1388361600 --gpu-engine 1000 --gpu-memclock 1250 --auto-fan --temp-target 75 --temp-overheat 80 --temp-cutoff 85

Also, YACMiner does work with CGWatcher. Just need to rename the EXE to "Cgminer.exe". Furthermore, on my one Windows 8 Rig, one of the cards doesn't seem to play well with YAC. Works fine with CGMiner. Have rebooted but will keep on tweaking to figure it out. Thank you again!

I'm glad it worked!

Do you have 8GB system RAM? Try increasing virtual memory to at least 16GB also to see if it makes a difference.
Maybe your system cannot address enough RAM to initialize. (try closing some apps etc. =p)
Other than that, it's all about finding the sweet spot of -R and --buffer-size really.

Yeah, 8GB. Its the exact same setup as the other rigs, just Win 8. Still working on it. Will keep you guys posted.

[MasterCATZ]

Just tweaking my yacminer now  for R9 290's
The MAX I could get
bufsize for thread @ 3624MB based on buffer-size

a few cards would not take it so its + / - a couple of Meg 

Code:

yacminer version 3.5.0-yac2 - Started: [2014-04-07 07:00:54]
--------------------------------------------------------------------------------
 (5s):55.30K (avg):61.12Kh/s | A:80  R:0  HW:0  U:53.29/m  WU:61.28/m  FB:0
 ST: 1  SS: 0  NB: 3  LW: 29  GF: 0  RF: 0
 Connected to pickaxe.pool.pm diff 16 NF 12 with stratum as user MC.R9x6pack
 Block: c3624e7be9699175...  Diff:88.9K  Started: [07:01:31]  Best share: 51
--------------------------------------------------------------------------------
 [P]ool management [G]PU management [S]ettings [D]isplay options [Q]uit
 GPU 0:  58.0C 5012RPM | 64.35K/61.12Kh/s | A:80 R:0 HW:0 U:53.29/m T:1 rI:262144
--------------------------------------------------------------------------------

 [2014-04-07 07:00:53] Started yacminer 3.5.0-yac2
 [2014-04-07 07:00:53] Loaded configuration file /opt/miners/YACMiner/UTCp2p.conf
 [2014-04-07 07:00:53] Probing for an alive pool
 [2014-04-07 07:00:54] Pool 1 difficulty changed to 64
 [2014-04-07 07:00:54] Switching to pool 1 stratum+tcp://ultra.nitro.org:3337 - first alive pool
 [2014-04-07 07:00:54] Pool 2 difficulty changed to 16
 [2014-04-07 07:00:54] GPU 0: bufsize for thread @ 3624MB based on buffer-size
 [2014-04-07 07:00:54] Network diff set to 86.1K
 [2014-04-07 07:00:54] Pool 0 difficulty changed to 16
 [2014-04-07 07:00:55] Pool 0 stratum+tcp://pickaxe.pool.pm:3306 alive
 [2014-04-07 07:00:55] Switching to pool 0 stratum+tcp://pickaxe.pool.pm:3306
 [2014-04-07 07:00:59] API running in local read access mode on port 4028 (8)
 [2014-04-07 07:01:16] Pool 0 stale share detected, submitting as user requested
 [2014-04-07 07:01:16] Accepted 998d4eed Diff 37/16 GPU 0 pool 0
 [2014-04-07 07:01:16] Stratum from pool 0 requested work restart
 [2014-04-07 07:01:18] Network diff set to 86.8K
 [2014-04-07 07:01:18] Stratum from pool 0 detected new block
 [2014-04-07 07:01:24] Accepted f09a0adb Diff 22/16 GPU 0 pool 0
 [2014-04-07 07:01:31] Network diff set to 88.9K
 [2014-04-07 07:01:31] Stratum from pool 0 detected new block
 [2014-04-07 07:01:57] Accepted b740962a Diff 39/16 GPU 0 pool 0
 [2014-04-07 07:02:05] Pool 0 stale share detected, submitting as user requested
 [2014-04-07 07:02:05] Accepted 2656a1fd Diff 51/16 GPU 0 pool 0
 [2014-04-07 07:02:17] Accepted 4f3a213f Diff 23/16 GPU 0 pool 0

Code:

[code]
scrypt-chacha" : true,
"worksize" : "64",
"shaders" : "2560",
"lookup-gap" : "3",
"buffer-size" : "3624",
"rawintensity" : "262144",
"gpu-engine" : "300-940",
"gpu-fan" : "40-85",
"temp-cutoff" : "80",
"temp-overheat" : "75",
"temp-target" : "73",
"api-mcast-port" : "4028",
"api-port" : "4028",
"api-listen" : true,
"expiry" : "10",
"gpu-dyninterval" : "7",
"gpu-platform" : "0",
"gpu-threads" : "1",
"log" : "5",
"no-pool-disable" : true,
"queue" : "0",
"scan-time" : "3",
"temp-hysteresis" : "3",
"kernel-path" : "/usr/local/bin"

[/code]









passed I 17 @ 65536
passed I 18 @ 131072
passed I 19 @ 262144
passed I 20 @ 524288
passed I 21 @ 1048576

passed I 22 @ 2097152 ( rejects 50% no HW errors )
it should not be still working !!!
no increase in khs / WU showing only slight temp rise

switching to X
passed 32
passed 128
passed 256
Am I just stuck at 50 khs 30 WU no settings seem to effect it ( tho R rate does go up )  
apart from changing GPU core which drops speed in half
I was mining with 3 threads on Last N factor at 170 WU with TC 65536


trying another GPU thread
failed
bufsize for thread @ 1812MB based on buffer-size
bufsize for thread @ 1796MB based on buffer-size

testing  
bufsize for thread @ 1706MB based on buffer-size
half hash speed / getting HW -R 524288

MAX
bufsize for thread @ 1710MB based on buffer-size
half hash / Wu speeds no HW -R 524288

back to single thread

Code:

yacminer version 3.5.0-yac2 - Started: [2014-04-07 05:42:02]
--------------------------------------------------------------------------------
 (5s):50.57K (avg):44.11Kh/s | A:480  R:2  HW:0  U:33.86/m  WU:35.98/m  FB:0
 ST: 1  SS: 0  NB: 25  LW: 260  GF: 0  RF: 0
 Connected to pickaxe.pool.pm diff 16 NF 12 with stratum as user MC.R9x6pack
 Block: b7dc23c6825cb0d5...  Diff:92.6K  Started: [05:56:03]  Best share: 721
--------------------------------------------------------------------------------
 [P]ool management [G]PU management [S]ettings [D]isplay options [Q]uit
 GPU 0:  69.0C 3116RPM | 50.55K/44.11Kh/s | A:480 R:2 HW:0 U:33.86/m T:1 rI:131072
--------------------------------------------------------------------------------

GPU 0: 50.6 / 44.0 Kh/s | A:30  R:1  HW:0  U:34.28/m  I:0 xI:0  rI:131072
69.0 C  F: 52% (3115 RPM)  E: 940 MHz  M: 1250 Mhz  V: 0.000V  A: 100%  P: 20%
Last initialised: [2014-04-07 05:42:02]
Thread 0: 50.6 Kh/s Enabled ALIVE

[E]nable [D]isable [I]ntensity [x]Intensity R[a]w Intensity [R]estart GPU [C]hange settings
Or press any other key to continue

you have me stumped yet again thirtybird ?

rI:131072

and you think ?
-R = 10752

tho I must admit its default LG 4 atm


Thirty Bird I think my cards are broken I can not seem to get HW:0 to go up !!!
one more level and I am rebooting

Code:

yacminer version 3.5.0-yac2 - Started: [2014-04-07 05:42:02]
--------------------------------------------------------------------------------
 (5s):26.62K (avg):46.72Kh/s | A:960  R:6  HW:0  U:37.86/m  WU:40.94/m  FB:0
 ST: 1  SS: 0  NB: 45  LW: 432  GF: 0  RF: 0
 Connected to pickaxe.pool.pm diff 16 NF 12 with stratum as user MC.R9x6pack
 Block: 6d22e672bc288edf...  Diff:97.6K  Started: [06:06:32]  Best share: 1.39K
--------------------------------------------------------------------------------
 [P]ool management [G]PU management [S]ettings [D]isplay options [Q]uit
 GPU 0:  68.0C 3068RPM | 50.35K/46.72Kh/s | A:960 R:6 HW:0 U:37.86/m T:1 rI:524288
--------------------------------------------------------------------------------

 [2014-04-07 06:06:03] Pool 0 stale share detected, submitting as user requested
 [2014-04-07 06:06:04] Accepted 2cc2fad1 Diff 16/16 GPU 0 pool 0
 [2014-04-07 06:06:14] Pool 0 stale share detected, submitting as user requested
 [2014-04-07 06:06:14] Accepted 5d2d6ae8 Diff 74/16 GPU 0 pool 0
 [2014-04-07 06:06:32] Network diff set to 97.6K
 [2014-04-07 06:06:32] Stratum from pool 0 detected new block
 [2014-04-07 06:06:35] Pool 0 stale share detected, submitting as user requested
 [2014-04-07 06:06:35] Rejected 05814826 Diff 45/16 GPU 0 pool 0 (Job '24b7' not found)
 [2014-04-07 06:06:45] Pool 0 stale share detected, submitting as user requested
 [2014-04-07 06:06:45] Pool 0 stale share detected, submitting as user requested
 [2014-04-07 06:06:45] Accepted 75ad8769 Diff 26/16 GPU 0 pool 0
 [2014-04-07 06:06:45] Accepted 0534b8a6 Diff 34/16 GPU 0 pool 0
 [2014-04-07 06:06:55] Pool 0 stale share detected, submitting as user requested
 [2014-04-07 06:06:55] Pool 0 stale share detected, submitting as user requested
 [2014-04-07 06:06:56] Accepted d291f214 Diff 50/16 GPU 0 pool 0
 [2014-04-07 06:06:56] Accepted 1fd8be0f Diff 73/16 GPU 0 pool 0

Code:

scrypt-chacha" : true,
"worksize" : "64",
"shaders" : "2560",
"lookup-gap" : "3",
"buffer-size" : "3624",
"rawintensity" : "10752",
"gpu-engine" : "300-940",
"gpu-fan" : "40-85",
"temp-cutoff" : "80",
"temp-overheat" : "75",
"temp-target" : "73",
"api-mcast-port" : "4028",
"api-port" : "4028",
"api-listen" : true,
"expiry" : "10",
"gpu-dyninterval" : "7",
"gpu-platform" : "0",
"gpu-threads" : "1",
"log" : "5",
"no-pool-disable" : true,
"queue" : "0",
"scan-time" : "3",
"temp-hysteresis" : "3",
"kernel-path" : "/usr/local/bin"

LG 1 gives HW errors
LG 2 gives HW errors
LG 3/4 fine

[Thirtybird]

just tweaking my yacminer now  for R9 290's

bufsize for thread @ 3624MB based on buffer-size

LG=3
-R = 10752

or

LG =4
-X = 32 or 128 or whatever you please

I think LG =3 will be slightly faster.  let me know if I am right

[MasterCATZ]

just tweaking my yacminer now  for R9 290's

bufsize for thread @ 3624MB based on buffer-size

LG=3
-R = 10752

or

LG =4
-X = 32 or 128 or whatever you please

I think LG =3 will be slightly faster.  let me know if I am right

70 WU

hey I just sent the new settings to the farm , some GPU's will not take the buffer size ?? any idea's ?? all the same cards

*edit *


ok then all exact rigs

o well you win some you loose some

I am trying scattered -R settings to see what numbers works best
( got 100 X GPU's to trial on )

so far 60 WU 70 WU seems to be the highest @ 940 core might be time to start OC'ing these cards their is hardly any heat coming out of them compared to other N factors ,
also noticing PSU's are pulsing ( dedicated GPU PSU's the lights are barely flicking )

[jakiman]

hey I just sent the new settings to the farm , some GPU's will not take the buffer size ?? any idea's ?? all the same cards

It's not the GPU that can't take it AFAIK.
I think it just cannot address that much memory into system RAM so fails on some cards.

[mackbells]

Nice COIn.

[jakiman]

290 @ 71Kh/s

I helped my friend tweak his pair of 290 to 71Kh/s @ 1070Mhz core.
Note: It's hasn't been tested for extended period yet. But so far OK.

yacminer.exe --scrypt-chacha -o <pool> -u <id.worker> -p <password> -R 9920 -g 1 -w 128 --buffer-size 3624 --lookup-gap 3 --temp-cutoff 85 --nfmin 4 --nfmax 30 --starttime 1388361600

[MasterCATZ]

290 @ 71Kh/s

I helped my friend tweak his pair of 290 to 71Kh/s @ 1070Mhz core.
Note: It's hasn't been tested for extended period yet. But so far OK.

yacminer.exe --scrypt-chacha -o <pool> -u <id.worker> -p <password> -R 9920 -g 1 -w 128 --buffer-size 3624 --lookup-gap 3 --temp-cutoff 85 --nfmin 4 --nfmax 30 --starttime 1388361600

good going ....
MBA_C6711101-100_K2_AGR_STILTMP_V48
I only get 27k at those settings

Code:

yacminer version 3.5.0-yac2 - Started: [2014-04-07 07:59:52]
--------------------------------------------------------------------------------
 (5s):25.73K (avg):26.47Kh/s | A:16  R:1  HW:0  U:11.70/m  WU:21.93/m  FB:0
 ST: 1  SS: 0  NB: 3  LW: 29  GF: 0  RF: 0
 Connected to pickaxe.pool.pm diff 16 NF 12 with stratum as user MC.R9x6pack
 Block: 037dbed8136d4042...  Diff:85.4K  Started: [08:01:17]  Best share: 78
--------------------------------------------------------------------------------
 [P]ool management [G]PU management [S]ettings [D]isplay options [Q]uit
 GPU 0:  50.0C 5033RPM | 27.05K/27.64Kh/s | A:16 R:1 HW:0 U:11.70/m T:1 rI:10752
--------------------------------------------------------------------------------

GPU 0: 27.1 / 27.4 Kh/s | A:1  R:0  HW:0  U:15.89/m  I:0 xI:0  rI:10752
51.0 C  F: 85% (5004 RPM)  E: 938 MHz  M: 1250 Mhz  V: 0.000V  A: 100%  P: 20%
Last initialised: [2014-04-07 07:59:53]
Thread 0: 27.1 Kh/s Enabled ALIVE

[E]nable [D]isable [I]ntensity .1[x]Intensity R[a]w Intensity [R]estart GPU [C]hange settings
Or press any other key to continue
Temp: 51.0 C
Fan Speed: 85% (5017 RPM)
Engine Clock: 937 MHz
Memory Clock: 1250 Mhz
Vddc: 0.000 V
Activity: 100%
Powertune: 20%
Fan autotune is disabled (40-85)
GPU engine clock autotune is disabled (300-940)
Change [A]utomatic [E]ngine [F]an [M]emory [V]oltage [P]owertune
Or press any other key to continue
Enter GPU engine clock speed (300 - 1500 Mhz):
1070
Driver reports success but check values below
Temp: 50.0 C
Fan Speed: 85% (5028 RPM)
Engine Clock: 1070 MHz
Memory Clock: 150 Mhz
Vddc: 0.000 V.1
Activity: 100%
Powertune: 20%
Fan autotune is disabled (40-85)
GPU engine clock autotune is disabled (300-1070)
Change [A]utomatic [E]ngine [F]an [M]emory [V]oltage [P]owertune
Or press any other key to continue

for some reason the memory speed drops to 150 mhz with low intensities ...

*edit*

crap forgotten I am going to have to BIOS flash them all again to an higher voltage for OC'ing :S
they are all 1.1V MBA_C6711101-100_K2_AGR_STILTMP_V48
guess 1.2V if I want to OC them to the max   or would an bump to 1.15V do ?
I'll have to calculate if the power consumptions going to be worth the extra WU now :S


cards are ranging from 40 KHS to 70 KHS
all same settings apart from  -R -X settings

just a matter of ruling out the crap numbers


few workers got banned from high reject counts

[jakiman]

for some reason the memory speed drops to 150 mhz with low intensities ...

Are you running Lookup-gap 3?

You should be able to get about 20-25% higher than my 280x at stock 290 clocks.
My 280x underclocked/undervolted does 55Kh/s. At stock voltage / clock, it does 60Kh/s.
Your memory going to 150Mhz surely sounds related also. That doesn't look right.

My friend is now running -R 10240 and is stable at 71.2 Kh/s on his 290. (clocked at 1040Mhz now)

[fredeq]

Thanks to everyone for the yacminer tips. Got it nicely hashing along now on two of my rigs. I had the same issue as others with 4 x 280x, needed to drop the buffer to 1760 and drop R until no hardware errors. Now the rig is running @ 200 kh/s  

So you have 4 x 280X's running at 50KHs per with 8GB RAM? May I have your final settings? Thanks!

Here you go, I was the first to post it which now everyone is using AFAIK.  
https://bitcointalk.org/index.php?topic=413978.msg6093996#msg6093996

280x(7970) @ 60Kh/s

55Kh/s at 1025/1500 @ 1.075v
60Kh/s at 1110/1500 @ 1.150v

1. Use latest YACMiner. https://github.com/Thirtybird/YACMiner/releases
2. Use following core settings: (Add other settings, more pools as needed)

yacminer.exe --scrypt-chacha -o <pool> -u <id.worker> -p <password> -R 8000 -g 1 -w 128 --buffer-size 2656 --lookup-gap 3 --temp-cutoff 85 --nfmin 4 --nfmax 30 --starttime 1388361600

Note: This may ONLY work on a PC with 8GB "physical" RAM regardless of VM size. (I tried up to 32GB)
If you have 4GB RAM machine and it fails, try using a lower buffer-size of 1760 and -R of 6144 or 5888.
As the rawintensity is lower, you will end up with a lower hash rate of about 45Kh/s unfortunately.

If this helped: UXLaLvo8L4eV9uC355WRzTX88GA7iFja2g  

Some tutorial on YACMiner for everyone

YACMiner allows finer control of the intensity value so you can squeeze every bit of its hash rate which ultracoin-miner cannot.

--buffer-size cannot go over the size of your video card's RAM amount. It's usually around 10% less value.
-R is basically Intensity equivalent in cgminer. YACMiner allows for more granular in-between Intensity values. (e.g. 12 = 4096 & 13 = 8192)

Goal is to try and squeeze in as many threads (intensity) in to the amount of buffer without it going into dynamic memory.
So first try to find the largest max buffer size your system will work with. Go up in increments of 128 until it fails to start.
Depending on system RAM size, you may need to start as low as 50% of your single video card's RAM amount. (e.g. 3GB video card  1536)
Once you know the max buffer size that works, set it as that and leave it alone.

Now just adjust the -R value in increments of 128. Good to start from something low such as 4096 and work your way up.
If you set it too high, you will either get HW errors and much less WU or it'll just fail/lockup/crash etc. Dial it down if so.

Lastly, check your WU over at least 15-20minutes. Clocking higher might givss higher hash rate but WU might not go up.
You might end up using 5-10% more electricity but end up not actually doing any more real work. So keep an eye on WU.

Thanks for this quick tutorial jakiman. Buffer size did not work on my main card, so I had to lower it to 2432 though. Some utc love coming your way!

Actually I had a hunch that if it is my main card, it is probably out of memory, due to some visual stuff. I tuned down all the visual things in windows 7 and 2560 is now a go!

[reflexmk]

@jakiman,

your settings should work on 280 dualx, right? thanks