The big VTC Vertcoin settings thread

[mmihai1978]

FINALLY tweaked my 280x issue:

if your card is not working with -g 2 then use setx GPU_MAX_ALLOC_PERCENT 200 before starting cgminer / vertminer.

now my hashrate is up 10% from 300 to 330.

[1713402921]

1713402921

[Galgenfrist]

Win 7 64 Rig with 3 x Gigabyte r9 290 (totally failed to get Linux to work, just haven't got the skills )

http://farm3.staticflickr.com/2856/12172813925_1003f5987e_o.jpg

Settings for cgminer (so far stable), temps are very stable, have a blower fan just moving the hot air away from the rig, without it they peak at 83c, with the window open (it's a cold day) they peak at 75c

setx GPU_MAX_ALLOC_PERCENT 100
setx GPU_USE_SYNC_OBJECTS 1
cgminer --scrypt  -o -u -p -I 19 -g 1 -w 256 --lookup-gap 2 --thread-concurrency 24550 --gpu-engine 1000 --gpu-memclock 1500 --gpu-powertune 10 -v 1 --auto-fan --temp-cutoff 85 --temp-overheat 83 --temp-target 78

Adding another rig tomorrow with the same spec.

Happy mining!

 

This works fine, thanks a lot! I'm getting 450kh/s on my 290 Tri-X.

[snogcel]

Pulled the latest source from github and saw a ~3% jump in hashrate - just a heads up (https://github.com/Bufius/vertminer-gpu/commits/master)

[psybase64]

Pulled the latest source from github and saw a ~3% jump in hashrate - just a heads up (https://github.com/Bufius/vertminer-gpu/commits/master)

I just copied the latest https://github.com/Bufius/vertminer-gpu/blob/master/scrypt140124.cl file from that github link and overwrote the scrypt140121.cl in the vertminer.0.5.2 folder.

Deleted the matching .bin file so the miner compiles a new bin and now getting 367khs now on my R9280xs, up from the 360khs using the scrypt140121.cl that came with Vertminer 0.5.2

Can someone compile a new Windows version from that github so we can use vertminer 0.5.3?

[financial_analyst]

I get this error:

Code:

[2014-02-01 16:12:46] Started cgminer 3.7.2
 [2014-02-01 16:12:46] Probing for an alive pool
 [2014-02-01 16:12:47] Pool 0 difficulty changed to 32
 [2014-02-01 16:12:47] Network diff set to 11
 [2014-02-01 16:12:47] Error -4: Enqueueing kernel onto command queue. (clEnqueueNDRangeKernel)
 [2014-02-01 16:12:47] Error -4: Enqueueing kernel onto command queue. (clEnqueueNDRangeKernel)
 [2014-02-01 16:12:47] GPU 0 failure, disabling!
 [2014-02-01 16:12:47] GPU 1 failure, disabling!

with the following code:

Code:

setx GPU_MAX_ALLOC_PERCENT 100
setx GPU_USE_SYNC_OBJECTS 1
cgminer.exe  --scrypt -o stratum+tcp://s.la.vertcoin.org:3333 -u XXX -p XXX 2>logfile_VERT.txt
-thread-concurrency 7680
-gpu-engine 900-1025
-temp-overheat 95
-temp-target 80
-v 1
-w 256
-Q 1
-G 1
-temp-cutoff 99
-temp-overheat 95
-temp-target 78
-gpu-vddc 1.100
-gpu-fan 45-100
-gpu-powertune 10
-auto-gpu
-auto-fan

Any ideas? I've got 1 280x and 1 7950. System RAM is 4GB, have tried a bunch of different thread concurrency/intensity combinations but I still get this error. I am using the modified cgminer.

[goin2mars]

I don't know yet, I am waiting for our power meter to arrive to do some calculations. We are running a 1350W Enermax unit, which supposedly is capable of 4 x 290's (though I've not tried it yet). So by that rationale, we are drawing somewhere between 900-1000w of juice.

When our power meter arrives I will post up the numbers and work out what in real terms our power costs are.

cheers mate

depending on what your numbers are it might help me get to the next level.

watching this with keen eyes

good luck guys/gals

ua

Power meters finally ordered, finding them in stock was a pain. They'll be arriving Monday

Hi,

I'm running 5x 290 non x's, 2 are elpdia on a different comp and 3 are hynix on another comp.

The hynix I've gotten @ 472k# for a few days stable before I switched back to multipool, but the elpdia I keep running into stability problems after a day. The wattage on the hynix's I remember being a little bit lower (NMT 10-15 watts) than I usually run regular cgminer. .which runs at 1110w with no undervolting (hynix cards running at 943k#), 92.5% efficient at that load, and 80w on the wattmeter side used by the system with no cards (cpu, mobo, peripherals). So after converting i'm seeing a usage of about 317.58w per hynix card. Using killawatt to measure wattages btw.

Decided to run vertminer while I was responding, so when running it on my hynix 290 non x rig @473k#, I'm seeing 1100 +/- 6watts . . so about 314.5w/hynix 290 on the dc side of the psu. . .about 1.501 k#/watt

Do you know if the cards you're running were elpdia or hynix? I'd like to switch the elpdia rig over to vertcoin but don't want to be tempted into an hour or two of messing w/ configs again and restarting the rig 20-30 times.

[normandy]

For some reason I can't push my intensity or thread concurrency over 13 and 8192 respectively without generating a mess of hardware errors and lots of rejects. Running 4x Sapphire R9 290 on Win 8 x64. Get over 820K per card when mining LTC. Any ideas? Currently I am getting around 272K per card mining VTC.

My friend, looks like we are at the same boat.

me using Win7 x64 and Sapphire R9 290, hashing 840K for LTC with TC31992 I-20 now hashing 320K for VTC with TC 8092 I-13. I tried the "optimal" 24450tc and i-20 and rig went freezed then crashed. Now it seems the optimal setting for my 290s is tc8192 and i-13 but it's way below even 350K compare to others hashing 400K.

I am just wondering how to manage the cards to get up for both TC and Intensity.

Thanks

[wezra]

I don't know yet, I am waiting for our power meter to arrive to do some calculations. We are running a 1350W Enermax unit, which supposedly is capable of 4 x 290's (though I've not tried it yet). So by that rationale, we are drawing somewhere between 900-1000w of juice.

When our power meter arrives I will post up the numbers and work out what in real terms our power costs are.

cheers mate

depending on what your numbers are it might help me get to the next level.

watching this with keen eyes

good luck guys/gals

ua

Power meters finally ordered, finding them in stock was a pain. They'll be arriving Monday

Hi,

I'm running 5x 290 non x's, 2 are elpdia on a different comp and 3 are hynix on another comp.

The hynix I've gotten @ 472k# for a few days stable before I switched back to multipool, but the elpdia I keep running into stability problems after a day. The wattage on the hynix's I remember being a little bit lower (NMT 10-15 watts) than I usually run regular cgminer. .which runs at 1110w with no undervolting (hynix cards running at 943k#), 92.5% efficient at that load, and 80w on the wattmeter side used by the system with no cards (cpu, mobo, peripherals). So after converting i'm seeing a usage of about 317.58w per hynix card. Using killawatt to measure wattages btw.

Decided to run vertminer while I was responding, so when running it on my hynix 290 non x rig @473k#, I'm seeing 1100 +/- 6watts . . so about 314.5w/hynix 290 on the dc side of the psu. . .about 1.501 k#/watt

Do you know if the cards you're running were elpdia or hynix? I'd like to switch the elpdia rig over to vertcoin but don't want to be tempted into an hour or two of messing w/ configs again and restarting the rig 20-30 times.

I think the first batch of cards we had must be elpida, because the 2nd batch of three GB 290's (exactly the same model purchased a week apart, on an identically specced rig) are getting 450 kh/s which some say makes them likely to be hynix cards. I've not looked into to much becuase both rigs are running so reliably.

[goin2mars]

For some reason I can't push my intensity or thread concurrency over 13 and 8192 respectively without generating a mess of hardware errors and lots of rejects. Running 4x Sapphire R9 290 on Win 8 x64. Get over 820K per card when mining LTC. Any ideas? Currently I am getting around 272K per card mining VTC.

My friend, looks like we are at the same boat.

me using Win7 x64 and Sapphire R9 290, hashing 840K for LTC with TC31992 I-20 now hashing 320K for VTC with TC 8092 I-13. I tried the "optimal" 24450tc and i-20 and rig went freezed then crashed. Now it seems the optimal setting for my 290s is tc8192 and i-13 but it's way below even 350K compare to others hashing 400K.

I am just wondering how to manage the cards to get up for both TC and Intensity.

Thanks

I run win 7 64 also, I shut off windows aero, and windows search, and set the computer to optimize for best performance, and favor programs over background processes. Not sure if any of that would make a difference for you but it's something to try. Also, you need to have more that 8GB mobo ram otherwise you can't go over the 8192 for these cards. I'm guessing you have your gpu threads set to one. Only other thing I can think of short of a fresh OS/driver install is running the bag file allocating your max gpu percentage and sync objects. Do you have all of this covered?

[goin2mars]

I don't know yet, I am waiting for our power meter to arrive to do some calculations. We are running a 1350W Enermax unit, which supposedly is capable of 4 x 290's (though I've not tried it yet). So by that rationale, we are drawing somewhere between 900-1000w of juice.

When our power meter arrives I will post up the numbers and work out what in real terms our power costs are.

cheers mate

depending on what your numbers are it might help me get to the next level.

watching this with keen eyes

good luck guys/gals

ua

Power meters finally ordered, finding them in stock was a pain. They'll be arriving Monday

Hi,

I'm running 5x 290 non x's, 2 are elpdia on a different comp and 3 are hynix on another comp.

The hynix I've gotten @ 472k# for a few days stable before I switched back to multipool, but the elpdia I keep running into stability problems after a day. The wattage on the hynix's I remember being a little bit lower (NMT 10-15 watts) than I usually run regular cgminer. .which runs at 1110w with no undervolting (hynix cards running at 943k#), 92.5% efficient at that load, and 80w on the wattmeter side used by the system with no cards (cpu, mobo, peripherals). So after converting i'm seeing a usage of about 317.58w per hynix card. Using killawatt to measure wattages btw.

Decided to run vertminer while I was responding, so when running it on my hynix 290 non x rig @473k#, I'm seeing 1100 +/- 6watts . . so about 314.5w/hynix 290 on the dc side of the psu. . .about 1.501 k#/watt

Do you know if the cards you're running were elpdia or hynix? I'd like to switch the elpdia rig over to vertcoin but don't want to be tempted into an hour or two of messing w/ configs again and restarting the rig 20-30 times.

I think the first batch of cards we had must be elpida, because the 2nd batch of three GB 290's (exactly the same model purchased a week apart, on an identically specced rig) are getting 450 kh/s which some say makes them likely to be hynix cards. I've not looked into to much becuase both rigs are running so reliably.

Yeah I keep running into alternating ram here. Every card I seem to buy it switches from Elpdia to Hynix or Hynix to Elpdia. It's like they know.

Anyways. . would you mind checking at least the card you're running? Here is the link for memoryinfo:

http://www.techpowerup.com/downloads/2326/asus-radeon-memoryinfo-1-005/

It will give you the memory info for your current display card. To check all mine I just hooked up the monitor to the one I wanted to check and took the power off the other cards. But to just check the one you're using currently all you have to do is run that.

I'm gonna put your config in my Elpdia rig and check it out at least, thanks for posting it!

[anonuser777]

Just for shits and giggles I tried mining vertcoins with an old 512MB Radeon HD 4670.  

http://www.gpureview.com/radeon-hd-4670-pci-e-card-579.html

Getting just over 20.3 kh/s with Vertcoin-CGMinerAleksNR3, using:

GPU_MAX_ALLOC_PERCENT 100
--scrypt --shaders 320 -w 64 -I 14 --thread concurrency 512 -g 2

Was running at 19.7-ish kh/s until I overclocked the GPU at 770, memclock at 1075. Still some room for optimization there, but not worth the effort really. Not too bad for a 5 year old mid-range card!

PS: If anyone in the UK wants to PM me an offer for the card feel free! Max Power Draw:   59 W  

[themike5000]

Has anyone managed to find a setup for the 7850 that will not give HW Errors above -I 10 ?

I've tried a number of the setup's on here for the 7850/r9 270 and they all come out with the same results, anything I 11 or above gives hw errors, i10 and under doesn't, although at i10 it's hash rate is at 70kh, which is 10kh slower then my old 6770!

My Powercooler 7850 is getting 170 kh/s at only 950 clock and 750 mem.
cgminer.exe --scrypt -o xxx -u xxx -p xxx -I 12 -w 256 -g 2 -v 1 --thread-concurrency 4000 --lookup-gap 2

[financial_analyst]

Is it possible to mine VRT with 4 GB of system RAM?

[mmihai1978]

Pulled the latest source from github and saw a ~3% jump in hashrate - just a heads up (https://github.com/Bufius/vertminer-gpu/commits/master)

I just copied the latest https://github.com/Bufius/vertminer-gpu/blob/master/scrypt140124.cl file from that github link and overwrote the scrypt140121.cl in the vertminer.0.5.2 folder.

Deleted the matching .bin file so the miner compiles a new bin and now getting 367khs now on my R9280xs, up from the 360khs using the scrypt140121.cl that came with Vertminer 0.5.2

Can someone compile a new Windows version from that github so we can use vertminer 0.5.3?

on my 280x hashrate decreases from 340 to 290 with the new .cl file from github.

[ieatm3s]

my R9 290 Hynix(475 khash) & Elpida(450 khash) settings, I have 8gb of system memory. 13.12WHQL drivers with newest AMD SDK, Windows 8.1 x64 pro.

--Thread-concurrency 24550
-I 20
--gpu-engine 1040/1000
--gpu-memory 1500/1500
 
Powertune: 50% (in AB)
undrevolt -12 mV (although shouldnt have anything to do with hashrate.)


Vkeev6vHsYm5HtZV4EXBq7d7oiagPUMSAk

cheers

[psybase64]

Pulled the latest source from github and saw a ~3% jump in hashrate - just a heads up (https://github.com/Bufius/vertminer-gpu/commits/master)

I just copied the latest https://github.com/Bufius/vertminer-gpu/blob/master/scrypt140124.cl file from that github link and overwrote the scrypt140121.cl in the vertminer.0.5.2 folder.

Deleted the matching .bin file so the miner compiles a new bin and now getting 367khs now on my R9280xs, up from the 360khs using the scrypt140121.cl that came with Vertminer 0.5.2

Can someone compile a new Windows version from that github so we can use vertminer 0.5.3?

on my 280x hashrate decreases from 340 to 290 with the new .cl file from github.

Strange... both my 4 x R9280x rigs are up from360 to 367-368 after updating that cl file.

Anyone else tried it and any success?

[skleb]

3x Sapphire 7870 GHz edition 205KH/s each total of ~620KH/s (just about WU of 500KH/s).

./vertminer [] --worksize 256 --thread-concurrency 16192 -g 1 --intensity 19  --temp-target 75 --gpu-engine 940 --gpu-memclock 1250

I live in New England and have this rig in the basement with a small fan pointed towards it. Things stay right around 55C at 40% fan.

[Kisu]

Anyone running vertminer 0.5.3 on BAMT?

I got the latest version from GitHub and installed it with these instructions:
https://bitcointalk.org/index.php?topic=404364.msg4668635#msg4668635

It works, but seems I can't change GPU settings. Not from thecgminer.conf, not from the command line, not from the vertminer. Also, I can't see the card temps at all. The web monitor shows the temps just fine though on http://myip/cgi-bin/status.pl

Even the ./vertminer --help shows only these gpu options, so I suppose the rest are actually missing?
--disable-gpu|-G    Disable GPU mining even if suitable devices exist
--gpu-dyninterval <arg> Set the refresh interval in ms for GPUs using dynamic intensity (default: 7)
--gpu-platform <arg> Select OpenCL platform ID to use for GPU mining (default: -1)
--gpu-threads|-g <arg> Number of threads per GPU (1 - 10) (default: -1)
--lookup-gap <arg>  Set GPU lookup gap for scrypt mining, comma separated
--intensity|-I <arg> Intensity of GPU scanning (d or -10 -> 21,default: d to maintain desktop interactivity)

The only way to change core/mem is to set new defaults in gpu bios and flash the card with that. Which is a bit annoying when searching for the best values..

[snogcel]

Anyone running vertminer 0.5.3 on BAMT?

I got the latest version from GitHub and installed it with these instructions:
https://bitcointalk.org/index.php?topic=404364.msg4668635#msg4668635

It works, but seems I can't change GPU settings. Not from thecgminer.conf, not from the command line, not from the vertminer. Also, I can't see the card temps at all. The web monitor shows the temps just fine though on http://myip/cgi-bin/status.pl

Even the ./vertminer --help shows only these gpu options, so I suppose the rest are actually missing?
--disable-gpu|-G    Disable GPU mining even if suitable devices exist
--gpu-dyninterval <arg> Set the refresh interval in ms for GPUs using dynamic intensity (default: 7)
--gpu-platform <arg> Select OpenCL platform ID to use for GPU mining (default: -1)
--gpu-threads|-g <arg> Number of threads per GPU (1 - 10) (default: -1)
--lookup-gap <arg>  Set GPU lookup gap for scrypt mining, comma separated
--intensity|-I <arg> Intensity of GPU scanning (d or -10 -> 21,default: d to maintain desktop interactivity)

The only way to change core/mem is to set new defaults in gpu bios and flash the card with that. Which is a bit annoying when searching for the best values..

Had this same problem two different times (two weeks apart so I forgot the solution) - be sure to run ./buildit.sh and NOT sudo ./buildit.sh. running as superuser seems to break the Open-GL / ADL stuff from what I can tell (assuming you copied the ADL_SDK .h files into the appropriate folder?)

[mik_druid]

Is it possible to mine VRT with 4 GB of system RAM?

yes I ve got 4gb works fine at 170kh on r9 270
considering buying more ram