7970 + cgminer + Litecoin

[Waramp22]

Anyone getting good hashrates running a 7970 + cgminer + Litecoin?

Ive been having nothing but trouble trying to get consistant hashing over 550kh/s. I can get 650-700 for 20 minutes and then i get drivers crashing. Windows 7 64bit. Temps are below 80C on the GPU and below 90C on the VRM. Ive run furmark for an hour straight to rule out it being a bad card. Ive tried many different driver setups, 2 fresh installs.

This might have something to do with it...
http://support.amd.com/us/kbarticles/Pages/737-27116RadeonSeries-ATIKMDAGhasstoppedrespondingerrormessages.aspx

Please list the catalyst version you are running, amd app sdk version, and your command line for cgminer + hashrate.

Im willing to pay in litecoin for a consistant 700kh/s.

[Steiner11]

Watching this with much interest as on my Gigabyte 7970 clocked to 1Ghz engine and 1,375Mhz on the memory I'm only getting in the mid 400's.

cgminer as follows:

"intensity" : "13",
"vectors" : "1",
"worksize" : "256",
"kernel" : "scrypt",
"lookup-gap" : "0",
"shaders" : "2048",
"thread-concurrency" : "24576",
"gpu-engine" : "0-0",
"gpu-fan" : "0-85",
"gpu-memclock" : "0",
"gpu-memdiff" : "0",
"gpu-powertune" : "20",
"gpu-vddc" : "0.000",
"temp-cutoff" : "95",
"temp-overheat" : "85",
"temp-target" : "75",
"api-port" : "4028",
"auto-fan" : true,
"expiry" : "120",
"gpu-dyninterval" : "7",
"gpu-platform" : "0",
"gpu-threads" : "1",
"hotplug" : "5",
"log" : "5",
"no-pool-disable" : true,
"queue" : "1",
"scan-time" : "60",
"scrypt" : true,
"temp-hysteresis" : "3",
"shares" : "0",
"kernel-path" : "/usr/local/bin"

[elchorizo]

Unfortunately the 7970's just don't mine Litecoin as well as they do Bitcoin, at least in my experience. First of all make sure you get the latest version of cgminer (3.X) as it has some tweaks for the 7970 and the readme has been updated with more recommended settings for it that most people were not using. Once you update and use the new suggestion settings you should be able to get 600 pretty consistent -- but I don't think you'll ever reach 700.

Here is the relevant info from his updated github README...

TUNING AN AMD RADEON 7970
Example tuning a 7970 for Scrypt mining:

On linux run this command:
export GPU_MAX_ALLOC_PERCENT=100
or on windows this:
setx GPU_MAX_ALLOC_PERCENT 100
in the same console/bash/dos prompt/bat file/whatever you want to call it,
before running cgminer.

First, find the highest thread concurrency that you can start it at. They should
all start at 8192 but some will go up to 3 times that. Don't go too high on the
intensity while testing and don't change gpu threads. If you cannot go above
8192, don't fret as you can still get a high hashrate.

Delete any .bin files so you're starting from scratch and see what bins get
generated.

First try without any thread concurrency or even shaders, as cgminer will try to
find an optimal value
cgminer -I 13

If that starts mining, see what bin was generated, it is likely the largest
meaningful TC you can set.
Starting it on mine I get:
scrypt130302Tahitiglg2tc22392w64l8.bin

See tc22392 that's telling you what thread concurrency it was. It should start
without TC parameters, but you never know. So if it doesn't, start with
--thread-concurrency 8192 and add 2048 to it at a time till you find the highest
value it will start successfully at.

Then start overclocking the eyeballs off your memory, as 7970s are exquisitely
sensitive to memory speed and amazingly overclockable but please make sure it
keeps adequately cooled with --auto-fan! Do it while it's running from the GPU
menu. Go up by 25 at a time every 30 seconds or so until your GPU crashes. Then
reboot and start it 25 lower as a rough start. Mine runs stable at 1900 memory
without overvolting. Overvolting is the only thing that can actually damage your
GPU so I wouldn't recommend it at all.

Then once you find the maximum memory clock speed, you need to find the sweet
spot engine clock speed that matches it. It's a fine line where one more MHz
will make the hashrate drop by 20%. It's somewhere in the .57 - 0.6 ratio range.
Start your engine clock speed at half your memory clock speed and then increase
it by 5 at a time. The hashrate should climb a little each rise in engine speed
and then suddenly drop above a certain value. Decrease it by 1 then until you
find it climbs dramatically. If your engine clock speed cannot get that high
without crashing the GPU, you will have to use a lower memclock.

Then, and only then, bother trying to increase intensity further.

My final settings were:
--gpu-engine 1141  --gpu-memclock 1875 -I 20
for a hashrate of 745kH.

Note I did not bother setting a thread concurrency. Once you have the magic
endpoint, look at what tc was chosen by the bin file generated and then hard
code that in next time (eg --thread-concurrency 22392) as slight changes in
thread concurrency will happen every time if you don't specify one, and the tc
to clock ratios are critical!

Good luck, and if this doesn't work for you, well same old magic discussion
applies, I cannot debug every hardware combo out there.

Your numbers will be your numbers depending on your hardware combination and OS,
so don't expect to get exactly the same results!

You'll note that he is able to get high rates but I've not had as much luck nor has my friend.

[Marrs]

Anyone getting good hashrates running a 7970 + cgminer + Litecoin?

Ive been having nothing but trouble trying to get consistant hashing over 550kh/s. I can get 650-700 for 20 minutes and then i get drivers crashing. Windows 7 64bit. Temps are below 80C on the GPU and below 90C on the VRM. Ive run furmark for an hour straight to rule out it being a bad card. Ive tried many different driver setups, 2 fresh installs.

This might have something to do with it...
http://support.amd.com/us/kbarticles/Pages/737-27116RadeonSeries-ATIKMDAGhasstoppedrespondingerrormessages.aspx

Please list the catalyst version you are running, amd app sdk version, and your command line for cgminer + hashrate.

Im willing to pay in litecoin for a consistant 700kh/s.

Code:

--auto-fan --gpu-fan 50-100 --gpu-powetune 20 --gpu-engine 1023 --gpu-memclock 1690 -I 13 -w 256 --thread-concurrency 8192 -g 2

Set your memory clock as high as you can get it without crashing, set your core clock at (memory clock * 0.605).

[vonross2012]

This is what im getting without a lot of tweeking.

[vonross2012]

How do I do this using Guiminer-script?

TUNING AN AMD RADEON 7970
Example tuning a 7970 for Scrypt mining:

On linux run this command:
export GPU_MAX_ALLOC_PERCENT=100
or on windows this:
setx GPU_MAX_ALLOC_PERCENT 100
in the same console/bash/dos prompt/bat file/whatever you want to call it,
before running cgminer.

[Steiner11]

Set your memory clock as high as you can get it without crashing, set your core clock at (memory clock * 0.605).

I tried setting this, 1375 *.605 = 831Mhz on the engine and my hashes dropped to about 420khash/s, re-upped to 1Ghz and it's hashing at somewhere between 480-500kh/s. 

[Waramp22]

Set your memory clock as high as you can get it without crashing, set your core clock at (memory clock * 0.605).

I tried setting this, 1375 *.605 = 831Mhz on the engine and my hashes dropped to about 420khash/s, re-upped to 1Ghz and it's hashing at somewhere between 480-500kh/s. 

Are you running a batch file? if so, post it i might have something that will get you some more speed.

[Marrs]

Set your memory clock as high as you can get it without crashing, set your core clock at (memory clock * 0.605).

I tried setting this, 1375 *.605 = 831Mhz on the engine and my hashes dropped to about 420khash/s, re-upped to 1Ghz and it's hashing at somewhere between 480-500kh/s. 

1375 is the stock memory clock. You need to raise it as high as it will go if you want good hash rates.

Some people are able to get up to 1800, but mine only goes to 1690 or so before it crashes.

Once you have stable overclocked memory, set the core clock with the .605 ratio.

These are the other settings that work well for me with that ratio: -I 13 -w 256 --thread-concurrency 8192 -g 2

[Waramp22]

Try this guys. Getting 615-650 with it. Powercolor 7970 + Catalyst 13.5 Beta drivers.

setx GPU_MAX_ALLOC_PERCENT=100
cgminer.exe --scrypt -o stratum+tcp//coinotron.com:3334 -u Waramp22.Desktop -p XXXXXXX --intensity 13 -g 2 --thread-concurrency 8192 --lookup-gap 2 --gpu-powertune 20 --gpu-fan 65 --auto-gpu --no-submit-stale

Donate if it works!   LRcHY7HoCUXfEeDdSjeeMPrV7esmf6ETa1


EDIT: It auto adjusts my gpu and ends up with 925 core, 1730 mem clocks.

[Steiner11]

I've always heard that hitting good numbers mining LTC was a lot of playing around and now I know everybody was right. 

I think I may be able to do better but following everyone's advice I played around with mem clocks and settled on 1.6Ghz, engine running at 1Ghz and I was getting around 540khash/s.  I did the calculation on that and then set my engine down to 965 in CCC and my hashes instantly popped up to 634khash/s average.  No idea... I have no idea why the engine running slower would make it hash faster, I'm sure google would turn up something if I really wanted to know.

I'm going to continue playing around to see if I can tweak any more speed out of it.  Thanks guys!!

[crazyates]

I've always heard that hitting good numbers mining LTC was a lot of playing around and now I know everybody was right. 

I think I may be able to do better but following everyone's advice I played around with mem clocks and settled on 1.6Ghz, engine running at 1Ghz and I was getting around 540khash/s.  I did the calculation on that and then set my engine down to 965 in CCC and my hashes instantly popped up to 634khash/s average.  No idea... I have no idea why the engine running slower would make it hash faster, I'm sure google would turn up something if I really wanted to know.

I'm going to continue playing around to see if I can tweak any more speed out of it.  Thanks guys!!

Ckolivas said that LTC is very timing sensitive, so a lower clock might bring a lot of the internal latencies into sync for a more efficient flow, where a high clock end up causing those minute latencies to stack up on top of each other, slowing everything down.

[Aggrophobia]

delete all bin files and start again

--scrypt -o url -u worker -p pass -I 13 -w 256 -g 2 --gpu-engine x  --gpu-memclock x --lookup-gap 2 --shaders 2048

should be in your bachfile

try to find your maximum memory clock (phps you can adjust your memory voltage up to 1.7V with unlocked voltages, but be careful)
if it starts to reset your driver, try 20mhz less

0.6 * memclock should be your engine clock (e.g. 865mhz / 1400mhz)

if your system gets unstable, reduce your memclock and engineclock

[Marrs]

I've always heard that hitting good numbers mining LTC was a lot of playing around and now I know everybody was right. 

I think I may be able to do better but following everyone's advice I played around with mem clocks and settled on 1.6Ghz, engine running at 1Ghz and I was getting around 540khash/s.  I did the calculation on that and then set my engine down to 965 in CCC and my hashes instantly popped up to 634khash/s average.  No idea... I have no idea why the engine running slower would make it hash faster, I'm sure google would turn up something if I really wanted to know.

I'm going to continue playing around to see if I can tweak any more speed out of it.  Thanks guys!!

Ckolivas said that LTC is very timing sensitive, so a lower clock might bring a lot of the internal latencies into sync for a more efficient flow, where a high clock end up causing those minute latencies to stack up on top of each other, slowing everything down.

I've also noticed that sometimes a lower intensity results in a faster hash rate than a higher one, at least when you are running with multiple gpu threads.

Steiner11 - my hardware worked well at that ratio. Your specific hardware may be better at a slightly different ratio. Try raising and lower the core clock by a small amount. If your hash rate goes up, raise or lower it a bit more. Repeat until your rate doesn't go up anymore.

Also, if you get a driver crash while playing around, be sure to reboot. Even if the drivers recover, they don't always fully recover...

[Steiner11]

Marrs, that's exactly what I ended up doing, just playing around with the engine clocks to see where the sweet spot was at.

I had thought I found the right settings at about 975 engine and 1600 memory clocks, but after about 20 minutes or so the screen would go to a solid color, which I take it to mean the card had 'crashed'.  I figured the memory clock was just set too high and dropped it back down to 1375, of course my hash rate dropped quite a bit, so I started slowly raising it again, eventually coming up to 1425Mhz which is apparently where the best spot for my card is at.  I let that run all night and no more lockups and the hash rate is right at the 650khash/s range with nothing else going on with the system.  If I open a browser (in typing this) I saw the hash rate drop to 622khash/s, so I bumped the memory a little to 1440Mhz and the speed came back up into the 640khash/s range, which I'm happy with as long as I don't have any more lockups.

Thanks for the explanation crazyates on the internal latencies, that makes sense, and saves me from searching.

[Aggrophobia]

If I open a browser (in typing this) I saw the hash rate drop to 622khash/s, so I bumped the memory a little to 1440Mhz and the speed came back up into the 640khash/s range, which I'm happy with as long as I don't have any more lockups.

That is quite normal when your card "works". The best solution would be another GPU for your screen, so that your hashrate stays high.

[makbot]

getting between 750-1000 khs per card.  running 4 cards.

gets me between 3-4 ltc a day.

--scrypt --shaders 2048 --thread-concurrency 8192 --intensity 13 --worksize 256 -g 2 -v 1 --lookup-gap 2 --auto-fan --gpu-engine 1085 --gpu-memclock 1500


4 sapphire 7970 3gb cards.  runs between 75-82C for the most part.  creates a ton of heat so i had to create an exhaust to go straight out the window.  be good to heat the house for the winter.. haha

[Bitni]

Shapphire HD7970 3GB Dualx. 650Kh/s- 750Kh/s
Really don't know what are most of the options for, just copypasted a bit from every trhead and melted.

"intensity" : "13",
"vectors" : "1",
"worksize" : "256",
"kernel" : "scrypt",
"lookup-gap" : "2",
"thread-concurrency" : "8192",
"shaders" : "2048",
"api-port" : "4028",
"expiry" : "120",
"gpu-engine" : "1085",
"gpu-dyninterval" : "7",
"gpu-memdiff" : "0",
"gpu-memclock" : "1500",
"gpu-powertune" : "20",
"gpu-platform" : "0",
"gpu-threads" : "2",
"log" : "5",
"no-pool-disable" : true,
"scan-time" : "30",
"shares" : "0",
"scrypt" : true,
"kernel-path" : "/usr/local/bin",
"temp-target" : "75",
"temp-cutoff" : "85",
"temp-overheat" : "80",
"gpu-fan" : "80",
"temp-hysteresis" : "3"

[wreckish]

I've been running the gigabyte OC-version for +6 months getting 690-700Kh/s

I run it at core 1008, mem 1240

"intensity" : "13",
"worksize" : "256",
"thread-concurrency" : "8192"
"gpu-threads" : "2"

The old drivers 12.1 works best for me.

[robroest]

didn't see an awnser but i will post it anyway , i had alot of problems too with my card i have a 7970 model GV-R797OC-3GD to be exact.

i update my drivers to the latest (no betas)

running on windows 7 64 bit

these are my settings that work for me
this give me 720 kh/s

"intensity" : "13",
"vectors" : "1",
"worksize" : "256",
"kernel" : "scrypt",
"lookup-gap" : "0",
"thread-concurrency" : "8192",
"shaders" : "0",
"gpu-engine" : "1050-1050",
"gpu-fan" : "0-85",
"gpu-memclock" : "1500",
"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",
"expiry" : "120",
"gpu-dyninterval" : "7",
"gpu-platform" : "0",
"gpu-threads" : "2",
"hotplug" : "5",
"log" : "5",
"no-pool-disable" : true,
"queue" : "1",
"scan-time" : "30",
"scrypt" : true,
"temp-hysteresis" : "3",
"shares" : "0",
"kernel-path" : "/usr/local/bin"