Topic: The big VTC Vertcoin settings thread - page 31. (Read 264260 times)

Hi guys, I've been messing with stability issues as well with my reference asus 290 non x's. Anything over I17 would cause freezes or crashes or bluescreens after anywhere from a few minutes to a day.

I have been searching for a stable config that I can run and forget about for an extended period of time, and think I have produced one. I've been running for over 24 hours now with no problems, engine throttling, or light pulsing.

The hashes sit at 417, but I'm willing to let that slide due to lower temps and happy cards. Here is what I run.

"intensity" : "17",
"vectors" : "1",
"worksize" : "256",
"kernel" : "scrypt",
"lookup-gap" : "2",
"thread-concurrency" : "24550",
"shaders" : "0,0",
"gpu-engine" : "947",
"gpu-fan" : "80",
"gpu-memclock" : "1250",
"gpu-memdiff" : "0",
"gpu-powertune" : "20",
"gpu-vddc" : "0.000",
"temp-cutoff" : "99",
"temp-overheat" : "95",
"temp-target" : "90",
"api-mcast-port" : "4028",
"api-port" : "4028",
"expiry" : "120",
"gpu-dyninterval" : "7",
"gpu-platform" : "0",
"gpu-threads" : "1",
"hotplug" : "5",
"log" : "5",
"no-pool-disable" : true,
"no-submit-stale" : true,
"queue" : "1",
"scan-time" : "60",
"scrypt" : true,
"temp-hysteresis" : "3",
"shares" : "0",
"kernel-path" : "/usr/local/bin"
}


I have taken out the gpu-dyninterval setting with no problems as well. My setup is 8Gb ram, 3 cards on one rig 2 on another. Have not tested / 4 cards on one rig as of yet, and won't be able to until 2 more come in the mail next week. I'm also running windows 7 64 bit.

Other clock pairs I want to test include 1000/1500, 1050/1450 and 1075/1470 as I have excellent results from those using regular cgminer. I also sometimes run TC 25000.

I realize there are faster speeds these cards can get @i20 and 1075/1450 got me around 470+...but I'm looking for stability as well as manageable temps and reasonable power usage so if you're looking for speed then use a different config.

@ Adaseb.

"I get the exact same results with slightly different settings.
Seems like the max for the 270X"

Yeah, I tried overclock. Nothing better. So, running stock and cool.

I get the exact same results with slightly different settings.

Seems like the max for the 270X

Mine just had this problem. All of a sudden one of the fans went crazy. I checked and its 511 C. Computer was frozen.

Restarted it and everything seems fine. But I wonder if this vertminer is killing our GPUs

vertminer 0.5.3. with one 2 GB R9 270X. 8 hour run.

I have the same issue with my 290 Tri-X, 1000/1500 with intensity higher than 13 just makes the display driver crash after 2 minutes.

Context :
Sapphire R9 290 Tri-X
8GB RAM
860W PSU
Windows 7 64-bit
Catalyst 13.12
SDK 2.9
vertminer 0.5.3
GPU_MAX_ALLOC_PERCENT 200

Now two settings that work for me and run stable 24/7 with 0 HW errors :

285 kh/s
"intensity": "13",
"worksize": "256",
"lookup-gap" : "2",
"thread-concurrency": "8192",
"gpu-engine": "1000",
"gpu-memclock": "1250",
"gpu-powertune" : "20",
"gpu-threads" : "2"


397 kh/s
"intensity": "17",
"worksize": "256",
"lookup-gap" : "2",
"thread-concurrency": "24550",
"gpu-engine": "900",
"gpu-memclock": "1350",
"gpu-powertune" : "20",
"gpu-threads" : "1"

For the second one, -I 18 crashes instantly.
Hope it helps.

nope, sorry. tons of HW errors.

Hi, there was no shortage of these settings when using single. I wear two cards together in the GPU1 at hw error occurs. Does anyone have any idea?

My Cards: 2 * Asus DCIIT

setx GPU_MAX_ALLOC_PERCENT 200
setx GPU_USE_SYNC_OBJECTS 1
vertminer.exe --scrypt-vert -o stratum+tcp://stratum.ecurie.io:3333 -u user -p pass -w 256 -I 13 -g 2 --lookup-gap 2 --gpu-fan 85,85 --gpu-engine 1030,1030 --gpu-memclock 1500,1500 --thread-concurrency 8192,8192 --gpu-powertune -20,-20 --expiry 1 --scan-time 1

is this the 270 or the 270X ?

A word of warning. I don't know if it's really related, but after mining vertcoin for a few days, 2 of my rigs exhibited a problem. Some of the cards showed 511 C temp and 0 RPM, and then the computer froze. This then happened all the time even after many reboots, after I changed the risers around and changed riser power, it started working again.
2 of my 7950 with this problem, after this, did not hash over 580 KH/s on normal cgminer anymore (previously 620 KH/s steady). One of my 280X now seems to have stability issues and for some reason lower temperature, and auto-fan speed for this card is consistently lower than on all my other cards.

Try increase temp-target to 95 and temp-overheat to 100 may get 440kh/s.  Hynix should be 500kh/s.  But the GPU temp should stabilize at room temp + 50 degree Celsius.

can you share your setting?

So I am trying to get my gigabyte 280x's going and I am using the following config

-v 1 -w 256 -I 12 -g 2 --lookup-gap 2 --gpu-powertune -20 --temp-target 73 --temp-overheat 85 --temp-cutoff 90 --gpu-memclock 1500 --gpu-engine 1000 --thread-concurrency 8191

GPU 0 runs is basically not hashing



No clue whats going on.

Reboot fixed it...

I have no idea how you guys are getting 50% of your regular rate. The best I can do with mine is about 33%

Regular hash rate for me is ~3200Khash/s
with Vert I am getting ~1100KHash/s


still worth it in terms of what BTC I am earning daily as compared to mining Doge.

270X

-w 256 -I 12 -g 2 --thread-concurrency 4448 --shaders 1280 --gpu-engine 1111 --gpu-memclock 1500 --lookup-gap 2 --remove-disabled --temp-cutoff 95

Around 150Khash and WU = 150

270

-w 256 -I 12,12,12 -g 2 --thread-concurrency 4448 --shaders 1280,1280,1280 --gpu-engine 950 --gpu-memclock 1400 --lookup-gap 2 --remove-disabled


Around 165Khash/s and WU = 160

So the 270 makes more WU than the 270X.

The 270 is on Windows 7 with 8GB RAM

the 270X is on Xubuntu with 2GB RAM.

Hi,

thanks for the settings posted in reply to my questions regarding ASUS DC2 Top 280x.

I had been trying all of those myself already. The point seems to be I simply can't use I=13 or TC higher then 6000ish. I 13 give me many HW errors, the usual TCs 8191,8192 and 8193 make my cards run out of memory buffer and disables them (as I already posted in Original post).
The Rig now sports 8 Gigs of RAM but nothing changes.

Could it be related to me running my mining rigs without AMD SDK and with standard Win 8.1 drivers?

For normal scrypt this works fine, actually it works really really well (> 800 kh/s with ASUS, 750-770 with MSI and Sapphire 280x).

I'd really love to mine this coin and support the network (and enjoy the nice profitability) but somehow it doesn't feel right with such poorly optimized results.
Half the normal scrypt hashrate would be fine (400 on Asus and 370 on MSI/Sapphire) but I only get 320 on the Asus and less then 300 on the others...bleargh!

Anyone successfully have r280xs running on guiminer with vert at present?

http://i.imgur.com/ys4Ve1e.png

My conf is:


{
"pools" : [
      {
                "url" : "stratum+tcp://mine1.coinmine.pl:6350",
                "user" : "user.user",
                "pass" : "pw"
        }


]
,
"scrypt" : true
}


and my bat is

setx GPU_MAX_ALLOC_PERCENT 100
setx GPU_USE_SYNC_OBJECTS 1
cgminer --scrypt --intensity 12 -g 1 -w 64 -o stratum+tcp://mine1.coinmine.pl:6350 -u user.user -p pw



I'm using a 6870 but shouldn't I be getting somewhere around 150 k/hs? How do I fix this?

TC 8192. Thanks.

When I had that error (GPU disabling), it was a thread-concurrency that was too high.


Anyone notice that the temp-target and gpu-fan settings don't seem to work in vertminer? I have a temp-target set way above my current temps and a gpu-fan setting of "0-50" same as what i ran with cgminer, and vertminer doesn't seem to care. My target is 76C, my cards are running at 71C but the fans are going at 61% instead of topping out at 50%. Is this a bug or a feature yet to be implemented?

Also, in a future version, can we have the in-program gpu setting changes for things like gpu engine, memory, fan, automation, etc.

What's the problems? 2x7970
.bat