Pages:
Author

Topic: The big VTC Vertcoin settings thread - page 4. (Read 264278 times)

newbie
Activity: 1
Merit: 0
April 29, 2014, 02:25:40 PM
Good afternoon;

That's my first time I try to mine, I wrote some of your configurations and the "best"  Roll Eyes result obtained is:

http://s11.postimg.org/6lj9c634z/error.jpg

I've got a Intel Core 2Quad Q8300 with 6 GB of Ram with an Ati Radeon X1950GT.

This is my configuration:

"@echo off
setx GPU_MAX_ALLOC_PERCENT 100
setx GPU_USE_SYNC_OBJECTS 1
COLOR 0A
vertminer.exe --scrypt-vert -o  stratum+tcp://stratum.spaincoinpool.com:3333 -u "User" -p "Password"  --thread-concurrency 8192 -I 13 --shaders 1792 --gpu-engine 1100 --gpu-memclock 1450 --gpu-powertune 20 -w 256 -g 2 --lookup-gap 2 --temp-overheat 85 --temp-cutoff 90 "



-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

For Vertminter configuration:

{
"pools" : [
   {
      "url" : "stratum+tcp://stratum.spaincoinpool.com:3333 -u "User"  -p "Password" ",
      "user" : "User",
      "pass" : "Password"
   }
]
,

"intensity" : "18",
"vectors" : "1",
"worksize" : "256",
"lookup-gap" : "2",
"thread-concurrency" : "22400",
"shaders" : "1792",
"gpu-engine" : "1000",
"gpu-fan" : "60",
"gpu-memclock" : "1400",
"gpu-memdiff" : "0",
"gpu-powertune" : "0",
"gpu-vddc" : "0.000",
"temp-cutoff" : "95",
"temp-overheat" : "85",
"temp-target" : "75",
"gpu-threads" : "1",
"hotplug" : "5",
"log" : "5",
"no-pool-disable" : true,
"queue" : "1",
"scan-time" : "10",
"scrypt" : true,
"temp-hysteresis" : "3",
"shares" : "0",
"kernel-path" : "/usr/local/bin"
}


Obviously, I've changed User and Password for my real name of worker.

Some help, please?
legendary
Activity: 1109
Merit: 1000
April 28, 2014, 02:14:27 PM
Ok, I have the latest version of vertminer and I have 4x 7850s which work well under cgminer (Scrypt). I get about 420KHs (Scrypt) for each card.
I tried to switch to vertminer. I looked throuh this thread and found recomendations for 7850s and tried some of them. Here's the problem.
Every time I try to start vertminer using any of the recommended settings, all Iget is "failed GPU"s when are OFFLINE from the start.
Sometimes, I get one or two of the total of 4 to work (seems random) but KHs is all very low, around 10-12 KHs.
I have 2 other miners running R9-280x cards just fine although the hash raes are about 10-20% below what others in this thread are getting, but they do work and are not failing or throwing any HW errors.
It's my effort to get the rig with the 7850s which is totally blowing up.
Again, ive cut-and-pasted several configs for the 7850 from this thread, but nothing seems to be working. The GPU's are failed right at the start and I can't bring them online.
Looking for any suggestions on how to proceed.
I mean, if I could at least get them online, then I could tune them, but when they report as failed immediately, there's not much tuning I know how to do to fix that.
newbie
Activity: 5
Merit: 0
April 24, 2014, 06:34:38 PM
I've run into a problem with my graphics card settings using Vertminer, as I have 1 simple intel graphics card and another stronger Nividia one. Whenever I start vertminer it automatically uses the intel one over the Nvidia one, is there any way to specify that I use the Nividia card in the vertminer.conf? I'm new to mining so I have no idea Tongue
legendary
Activity: 1540
Merit: 1011
FUD Philanthropist™
April 21, 2014, 12:01:05 AM
Hi to all.

Please, give me some advice about tuning 7970 card. I have an MSI Lightning BE (7970 with stock Core/Mem 1150/1500). All my experiments have resulted in 349-351Kh/s with some HW errors Sad

My current settings is:
Thread concurrency 6136 (experimentally, this gives the best hashrate)
Core 1025
Mem 1495
Intensity 13 (12 gives no HW errors, but hashrate is about 220Kh/s)
Vertminer 0.5.3 (on 0.5.2 hashrate is about 290-300KH/s)
Added setx GPU_MAX_ALLOC_PERCENT 200 and setx GPU_USE_SYNC_OBJECTS 1 at system startup.
The HW error vs Accepted is nearly 50% Sad

Does anybody have experience with these cards? What core/mem speeds should be? And what is the general ratio of core/mem for vertminer?

Thanks in advance.

GPU_MAX_ALLOC_PERCENT 200

should be 100 ..as in 100 percent Wink

i don't run with any errors ever.. it's 24/7 at 0 or i fix the problem Smiley
legendary
Activity: 1540
Merit: 1011
FUD Philanthropist™
April 20, 2014, 11:58:07 PM
the typical and classic max is 18 some *newer cgminer mod versions use a RAW method combined with a max of 13.
so i don't think 20 or 21 or higher would be valid and it would default to some other value.. i'm guessing


20 was the old max in cgminer.  Up to 23 is valid in the newer branches of cgminer such as sgminer.  And any scrypt or nscrypt miner worth his salt should be using sgminer or some other maintained branch by now.

Classic intensity values (when you aren't using RawIntensity or shader-based XIntensity) launch 2^I threads. e.g.
I:13 = 8192 threads
I:14 = 16384 threads
I:15 = 32768 threads
I:16 = 65536 threads
I:17 = 131072 threads
I:18 = 262144 threads
I:19 = 524288 threads
I:20 = 1048576 threads
I:21 = 2,097,152 threads
I:22 = 4,194,304 threads
I:23 = 8,388,608 threads

As you can see, the number of threads increases exponentially with higher values on I.  This is why using Xintensity for tuning is so much more effective.  Not to mention these values rarely match up with the number of shaders on the card, and the optimal number of threads to launch is a multiple of the number of shaders on the card.  This is why I=13 is the most effective on many devices - because they have 8192/N shaders.

---

The sgminer git tree even has pool-specific algorithm/nfactor support, and preliminary kernel hotswapping!

interesting thanks for sharing Smiley
hero member
Activity: 518
Merit: 500
April 20, 2014, 11:23:39 PM
I've noticed that most cards in general, especially VLIW 4/5 cards, run good with a TC= 4x shaders count. That GCN cards with 514-bit memory, ie. R9 290/290X, can do TC=8x shaders count. Other GCN cards, 256-bit memory, can do from anywhere from TC=4x~8x shaders count with the exception of the hd7970/R9 280X (384-bit memory) which does 4x shaders count for the TC. Running default clocks sometimes helps too.



what does this mean? how can I fix it?

Your TC is too high for the device 1 card/gpu 1.

Is gpu 1 a 7970/280x? If yes, then a TC of 8192 or 8193 should work.
member
Activity: 135
Merit: 10
April 16, 2014, 11:58:03 AM
The ratio you can get varies depending on a lot of factors.

Raise TC at I=13, until the HW errors go away (well, a few a day is acceptable). It can be hard to give up the hashrate, I know! But it is an illusion - WU will confirm this, although it takes a really long time for it to stabilize.

Make sure you have 2 gpu threads. You can also experiment with work size in multiples of 64. Your memclock should be a pretty hard upper bound (beyond which weird shit starts to happen), but keep reoptimizing that engine clock with changes to the other factors. Don't push engine too hard though, only to the "bend" where it starts to hit diminishing returns on hash rate and rejects go up. If you don't see this then you probably can increment the intensity and find higher hash rates. TC has to go up with intensity, though (often times the best settings have raw intensity, which is 2^I, equal to TC).
member
Activity: 130
Merit: 10
April 16, 2014, 10:33:55 AM
Hi to all.

Please, give me some advice about tuning 7970 card. I have an MSI Lightning BE (7970 with stock Core/Mem 1150/1500). All my experiments have resulted in 349-351Kh/s with some HW errors Sad

My current settings is:
Thread concurrency 6136 (experimentally, this gives the best hashrate)
Core 1025
Mem 1495
Intensity 13 (12 gives no HW errors, but hashrate is about 220Kh/s)
Vertminer 0.5.3 (on 0.5.2 hashrate is about 290-300KH/s)
Added setx GPU_MAX_ALLOC_PERCENT 200 and setx GPU_USE_SYNC_OBJECTS 1 at system startup.
The HW error vs Accepted is nearly 50% Sad

Does anybody have experience with these cards? What core/mem speeds should be? And what is the general ratio of core/mem for vertminer?

Thanks in advance.
full member
Activity: 210
Merit: 100
April 14, 2014, 10:29:25 AM
vertminer config for sapphire 290 tri-x plzzz ? thanks
sr. member
Activity: 420
Merit: 250
April 13, 2014, 05:42:29 PM
setx GPU_MAX_ALLOC_PERCENT 100
setx GPU_USE_SYNC_OBJECTS 1
COLOR 0A
vertminer -o stratum+tcp://useast.coinsolver.com:3334 -u unknown -p xx -w 512 -I 14,19 -g 1 --lookup-gap 2 --auto-fan --thread-concurrency 24550,12408 --gpu-reorder --gpu-engine 1115,1000 --gpu-memclock 1425,1250 --gpu-vddc 1.25
pause

can anyone tell me what is wrong with this config?

1st slot is 270 ( recently is 290 non x) and 2nd slot is 290 ( recently is 270) I have to use "--gpu-reorder" to work with it.

270 got at least 130 kh/s with less than 100 WU/s  

290 got 370-380 kh/s with 350 WU/s.

how can I make the 270 gpu faster than normal ?
sr. member
Activity: 420
Merit: 250
April 13, 2014, 12:07:15 PM


what does this mean? how can I fix it?
hero member
Activity: 700
Merit: 500
April 11, 2014, 10:16:36 PM
the typical and classic max is 18 some *newer cgminer mod versions use a RAW method combined with a max of 13.
so i don't think 20 or 21 or higher would be valid and it would default to some other value.. i'm guessing


20 was the old max in cgminer.  Up to 23 is valid in the newer branches of cgminer such as sgminer.  And any scrypt or nscrypt miner worth his salt should be using sgminer or some other maintained branch by now.

Classic intensity values (when you aren't using RawIntensity or shader-based XIntensity) launch 2^I threads. e.g.
I:13 = 8192 threads
I:14 = 16384 threads
I:15 = 32768 threads
I:16 = 65536 threads
I:17 = 131072 threads
I:18 = 262144 threads
I:19 = 524288 threads
I:20 = 1048576 threads
I:21 = 2,097,152 threads
I:22 = 4,194,304 threads
I:23 = 8,388,608 threads

As you can see, the number of threads increases exponentially with higher values on I.  This is why using Xintensity for tuning is so much more effective.  Not to mention these values rarely match up with the number of shaders on the card, and the optimal number of threads to launch is a multiple of the number of shaders on the card.  This is why I=13 is the most effective on many devices - because they have 8192/N shaders.

---

The sgminer git tree even has pool-specific algorithm/nfactor support, and preliminary kernel hotswapping!
legendary
Activity: 1540
Merit: 1011
FUD Philanthropist™
April 11, 2014, 10:00:28 PM
is it normal that WU less than kh/s on 290 non x? I got 420 kh/s and WU 319 m/s.

is it normal that WU less than kh/s  on 270x? I got 186 kh/s and WU 140-150 m/s.

Try reduce the Intensity?  I guess your 290 non X is -I 21.  Try -I 20.

the typical and classic max is 18 some *newer cgminer mod versions use a RAW method combined with a max of 13.
so i don't think 20 or 21 or higher would be valid and it would default to some other value.. i'm guessing

for example vertminer 0.5.2 max is 18
for example vertminer 0.5.3 max is 13

edit:
i read the readme and i guess i'm wrong on that 20 may be the old maximum not 18.. oops

this is from the SCRYPT_README file..

Quote
--intensity XX (-I XX)

Just like in bitcoin mining, scrypt mining takes an intensity, however the
scale goes from 0 to 20 to mimic the "Aggression" used in mtrlt's reaper. The
reason this is crucial is that too high an intensity can actually be
disastrous with scrypt because it CAN run out of ram. High intensities
start writing over the same ram and it is highly dependent on the GPU, but they
can start actually DECREASING your hashrate, or even worse, start producing
garbage with HW errors skyrocketing. Note that if you do NOT specify an
intensity, cgminer uses dynamic mode which is designed to minimise the harm
to a running desktop and performance WILL be poor. The lower limit to intensity
with scrypt is usually 8 and cgminer will prevent it going too low.
SUMMARY: Setting this for reasonable hashrates is mandatory.
member
Activity: 90
Merit: 10
April 11, 2014, 07:42:36 AM
is it normal that WU less than kh/s on 290 non x? I got 420 kh/s and WU 319 m/s.

is it normal that WU less than kh/s  on 270x? I got 186 kh/s and WU 140-150 m/s.

Try reduce the Intensity?  I guess your 290 non X is -I 21.  Try -I 20.
legendary
Activity: 1540
Merit: 1011
FUD Philanthropist™
April 09, 2014, 09:59:26 PM
Is it normal if at the moment i use another programm in windows while mining with vertminer hashrate goes heavily down ?


If i open a wallet for example, hashrate drops down to 15% of total. It takes minimum one minute to come back again and does
not reach 100 % of old value.

I tryed to mine DRK with sgminer and it doens`t make any difference in hashrate if i use other programms or not.



yeah that is normal.. vertminer cgminer etc all require a bit of cpu power available at all times to work mining..
so if you all a of a sudden max out your cpu you should see a drop in hash rate until you free up the cpu again.
the cpu has to orchestrate the work like a conductor standing in front of a band.. he is not doing much but if you screw with him the band goes to hell lol
like the cpu is what is feeding the work to the gpu so if it has any delay in sending work then work speed slows down. make sense ?

i use vertminer myself with 1 gpu.. a Powercolor 7950 and i occasionally mine with my cpu on and off too or do video encoding etc
and i have different computers with different cpu's and it doesn't matter which one i use they all follow the same rules or problems etc.
my main machine is an i7 with 4 cores and 8 threads so if i use the cpu i make sure to leave 2 threads free for the GPU miner.
(so i would run the cpu vertminer with 6 threads via commandline) while gpu mining at the same time.

On another computer i have.. an nvidia card mining VTC with cudaminer with a Core 2 Duo e7500 and that is only 2 cores / 2 threads
and if use up the the cores on that cpu it slows my nvidia miner down too. I rarely ever cpu mine on that one because
it's not worth slowing down the faster nvidia gpu miner.. so yeah it's normal Wink
Thanks for your detailed answer. I have noticed that the wallets only use much CPU power while they are not synched. After the are synched
and still open and stay synched it does not effect my mining power.

Oh yeah wallets are notorious at sapping CPU power maxing it out (until they are synced)
It's normal and some are way worse than others depending on the coin.. some people may not know that.

Still mining away using cgwatcher and Vertminer 0.5.2

Here is my whole bat file i use for my Powercolor 7950 in case people want to see it ?
Notice work size 256 ? i think the miner is hard coded at that value so if you try 512 or 128 it won't work..
You can see this by looking at the *.bin file name.. maybe some other miner version can let you change that i don't know.
And the second pool i have listed is a fall back if the main one goes down and that is pointed at my VTC wallet with a conf i made for the wallet for solo mining.
obviously i would have to have the wallet on and synced at all times to do solo mining with Vertminer though (if the main pool goes down)
Also notice a key setting to experiment with here is GPU Threads Wink

Code:
{
"pools" : [
{
"url" : "stratum+tcp://stratum.vertco.in:8080",
"user" : "secretname.2",
"pass" : "x"
},
{
"url" : "http://127.0.0.1:20092",
"user" : "user",
"pass" : "pass"
}
]
,
"intensity" : "18",
"vectors" : "1",
"worksize" : "256",
"lookup-gap" : "2",
"thread-concurrency" : "22400",
"shaders" : "1792",
"gpu-engine" : "1000",
"gpu-fan" : "60",
"gpu-memclock" : "1400",
"gpu-memdiff" : "0",
"gpu-powertune" : "0",
"gpu-vddc" : "0.000",
"temp-cutoff" : "95",
"temp-overheat" : "85",
"temp-target" : "75",
"api-mcast-port" : "4028",
"api-port" : "4028",
"no-submit-stale" : true,
"expiry" : "120",
"gpu-dyninterval" : "7",
"gpu-platform" : "0",
"gpu-threads" : "1",
"hotplug" : "5",
"log" : "5",
"no-pool-disable" : true,
"queue" : "1",
"scan-time" : "10",
"scrypt" : true,
"temp-hysteresis" : "3",
"shares" : "0",
"kernel-path" : "/usr/local/bin"
}


Also here is the contents of my bat file showing how to specify a .Conf config file..
I have no idea if i used the GPU map feature correctly though, i want to skip my integrated Intel 4770k Haswell graphics chip and only mine with my Radeon.
Note: you do not have to specify the full path like below.. cgwatcher changed my bat file and did that.

Code:
C:\Miners\vertminer-0.5.2\cgminer.exe --gpu-map 2:1,1:2 --config C:\Miners\vertminer-0.5.2\cgminer.conf
newbie
Activity: 1
Merit: 0
April 09, 2014, 05:51:44 PM
Hi!

I have a couple of questions.
I won't be doing any serious mining right now, I just want to learn how things work before I buy a more powerfull computer.
I mine with my laptop, which has a Nvidia 750m (and integrated Intel graphic chip).
When I start vertminer it automatically chooses my crappy intel chip and not my Nvidia card, can I somehow fix this? Shouldn't it recognize and automatically use my Nvidia card?
Or should I use Cudaminer? I can mine vertcoin at about 44 Kh/s with Cudaminer (as I said, crappy card), but I'm wondering if Vertminer is more optimised or better?

I've heard people saying that you need lots of RAM, but Cudaminer only uses about 30MB ram when mining Vertcoin, how is this possible?

And also, how much can you "tweak" your program/miner for more performance?
This is my complete .bat, which I just double click to start mine:

"cudaminer -o stratum+tcp://eu.bitcrush.info:3444 -u xxx.xxx -p xxx --algo=scrypt:2048"


As I said, I'm completely new to this, but want to learn how to do it correctly so to speak  Smiley
newbie
Activity: 66
Merit: 0
April 08, 2014, 04:00:10 AM
What is the maximum Kh/s for 7870?

I have three on my rig and I get:
185 Kh/s VTC

on LTC, DOGE I get around 405 Kh/s.

I read that someone get around 200? But I don't believed because I tried everything Smiley
newbie
Activity: 13
Merit: 0
April 08, 2014, 03:21:53 AM
djaychela:  Awesome! Anytime man. I mined 3.7VTC in about 9 hours... Grin
newbie
Activity: 35
Merit: 0
April 08, 2014, 03:12:19 AM
nIcKeLbAcK: Top man!  Downloaded it, tweaked settings for 5 minutes and already got the performance I had with the two separate runners, but with one card slightly above, one slightly below (130 each vs 110, 150).  Will tweak more later, just going to let it run and see if it's happy.  The Raw Intensity setting seems to be a useful thing - I had no idea until reading that thread of the exponential increase that Intensity led to, no wonder I:12 was fine and I:13 was instant HW failure runs!  THanks again.
newbie
Activity: 13
Merit: 0
April 08, 2014, 02:21:38 AM
djaychela: Finally figured it out. Ok, Using the following conf file in conjunction with VertiKev https://bitcointalksearch.org/topic/ann-new-thekev-vertminer-ultimate-vertminer-guide-for-windows-466867  .......  I have found success. My dual HD6970's are pushing 420kh/s. Now, I launch w RI of 10240, but push it to 20480. GPU0 69C and GPU1 74C (this card always runs hot, dam XFX crap. Hope this helps. The key was ThreadConcurrency. Halve your normal TC, delete shaders. If you need shaders in then match it to your TC +or- 1
Hope this helps.
...and remember to delete your *.bin file(s) before launching the new *.bat or conf file.....cheers

Ed4252    TC is probably the issue you are having as well. So halve it.
,
"rawintensity" : "10240",
"scrypt-vert" : true,
"vectors" : "1",
"worksize" : "256",
"lookup-gap" : "2",
"gpu-engine" : "950",
"gpu-memclock" : "1365",
"thread-concurrency" : "7520",
"gpu-fan" : "30-85, 30-100",
"temp-target" : "70,75",
"temp-cutoff" : "85",
"auto-fan" : true,
"api-listen" : true,
"api-port" : "4028",
"expiry" : "15",
"gpu-dyninterval" : "7",
"gpu-threads" : "1",
"hotplug" : "0",
"log" : "5",
"no-pool-disable" : true,
"failover-only" : true,
"queue" : "2",
"scan-time" : "5",
"temp-hysteresis" : "3",
"device" : "0-1"
}
Pages:
Jump to: