Consolidated Litecoin Mining Guide for 5xxx, 6xxx, and 7xxx GPUs

[luffy]

thanks for the tips. i 've discovered that i have to increase RAM in my systems since i have 2 or 3 cards on board!
then i am forced to use low values in "thread-concurrency" and that gives HW faults and a lot of rejections!
But i wonder why in one system with 5850 the "thread-concurrency" key was not needed in cgminer command line!

[1714155917]

1714155917

[1714155917]

1714155917

[1714155917]

1714155917

[1714155917]

1714155917

[tacotime]

Cgminer tries to automatically pick values for your cards, it occasionally works.

[luffy]

success 
i have made different cgminer .bat files for every card and work great! i think cgminer don't even care about system RAM!

[flatronw]

Are they all in at least pcie 4x slots? I've mixed brands and cards before without problem

As stated use reaper for 79xx cards

Very interested to see the mention of 16x, 8x, 4x PCIe slots. Is this a verified issue? I have been trying to run 5 x 5850's on a MSI 990 FXA-GD80 M/board with 4GB of Ram. Obviously I am using PCIe extender cables, three 16x and two 1 x 16 extenders. Effectively two  full x16 one x8 and two x1. I am unsure how the M/B allocates the available resources but I doubt the two x16 slots are actually running at that.
All cards are underclocked and undervolted. This hardware setup (under volted and clocked) with CGMiner mines BTC flawlessly, trying to mine LTC at conservative settings gets, in my opinion, way too many rejects and HW errors. I have tried endless settings, nothing seems to clean it up. I had never considered the 1 x 16 extenders may be a problem?

By the way:- great topic, can we get it posted on Litecoin Forum as well? sort of parallel posts or something 

[tacotime]

The <4x PCI-e slot was a suspected issue for a long time but no one had actually verified it to my knowledge.  I've only ever mined on 4x or greater slots.  The transfer speed is only 500MB/s per lane with PCI-e 2.0, so it's possible that using only a single lane may cause a bottleneck.

You can verify this simply by moving around the cards.  If the same cards are fine in an x16 slot but awful in an x1 slot, then you probably need the extra bandwidth.

[squid]

Great thread, great tips

[pintycar]

The <4x PCI-e slot was a suspected issue for a long time but no one had actually verified it to my knowledge.  I've only ever mined on 4x or greater slots.  The transfer speed is only 500MB/s per lane with PCI-e 2.0, so it's possible that using only a single lane may cause a bottleneck.

You can verify this simply by moving around the cards.  If the same cards are fine in an x16 slot but awful in an x1 slot, then you probably need the extra bandwidth.

7970 in x1 slot on an asus p8-z68 gets exactly the same hashrate as 7970 in the x16 (x8 actual rate). YMMV

[meebs]

The <4x PCI-e slot was a suspected issue for a long time but no one had actually verified it to my knowledge.  I've only ever mined on 4x or greater slots.  The transfer speed is only 500MB/s per lane with PCI-e 2.0, so it's possible that using only a single lane may cause a bottleneck.

You can verify this simply by moving around the cards.  If the same cards are fine in an x16 slot but awful in an x1 slot, then you probably need the extra bandwidth.

7970 in x1 slot on an asus p8-z68 gets exactly the same hashrate as 7970 in the x16 (x8 actual rate). YMMV

this would make sense, as I can't imagine there being a whole lot of data needing to be transported between the card and the system.

however I DO wonder what the heck reaper is doing with the 2.9gb of system ram it hogs.. lol

[Jack1Rip1BurnIt]

I just tried to mine with my 3x 7950 rig for the first time. I used reaper and could only get it to give about 820 Kh total. I had to use 6144 thread concurrency because the whole system only has 2 gigs of memory. I have a single 7970 in a rig and I couldn't even get it to work, maybe because it only has 1 gig in it. Not too happy about the results with all of my 79xx series cards. I have 3x 5850s that I have been able to get more than 800 Kh. Oh well

[tacotime]

needs more ram.  it's cheap.

[meebs]

I just tried to mine with my 3x 7950 rig for the first time. I used reaper and could only get it to give about 820 Kh total. I had to use 6144 thread concurrency because the whole system only has 2 gigs of memory. I have a single 7970 in a rig and I couldn't even get it to work, maybe because it only has 1 gig in it. Not too happy about the results with all of my 79xx series cards. I have 3x 5850s that I have been able to get more than 800 Kh. Oh well

\

ya man... reaper ALONE uses 2.8 gigs for 2x 7950's.

go order a 2x4gb ram set and you'll be set


cgminer on two 7770's uses about 50mb... so ya.. its something about reaper that just requires a bit of ram.

once you get it working fine though... you'll get a nice hash rate.. i'm getting 1220khash from my 2x 7950 system.

[flatronw]

The <4x PCI-e slot was a suspected issue for a long time but no one had actually verified it to my knowledge.  I've only ever mined on 4x or greater slots.  The transfer speed is only 500MB/s per lane with PCI-e 2.0, so it's possible that using only a single lane may cause a bottleneck.

You can verify this simply by moving around the cards.  If the same cards are fine in an x16 slot but awful in an x1 slot, then you probably need the extra bandwidth.

I think that I am already actually doing that. Correct me if I am wrong. My system is all 5850 Reference cards, they are all on identical V,E & M settings on CGMiner and give near to identical performance in K/hs, HW errors and Rejects regardless of their using x16 or x1 extenders in their current slot positions. Therefore swapping them around would seem to be pointless. They are all underclocked to about 220 K/Hs. They dont seem to have a hardware problem.
It is the problem with stales that I am trying to resolve. 99% of all stales that I get occur just before and just after the LP. I have tried almost all the LTC Pools. I have found that if I use a pool with frequent LP's it just kills my profitability with "prev hash stale" stales. On a pool with fewer LP's I will still get stales at the LP but far fewer and sometimes none at the LP. It looks like the system cant cope with the transition/data from the LP. That's why I was interested when I saw the comment about the PCIe lanes maybe causing communication problems.

It is currently quite happily mining BTC at 2 Gh/s. I really dont want to start ripping the whole thing apart to "re-invent the wheel" if somebody has already solved this issue.

Edit:- After a fair bit of scratching around I found this command:-  lspci -vvv   Run that after sudo su and you will get mountains of info regarding the current PCIe status. Including what I think are the current lane allocations listed at LnkSta:
Run it without sudo su and you will not get all the available data.
There is more there than I care to try to understand.

[Jack1Rip1BurnIt]

ya man... reaper ALONE uses 2.8 gigs for 2x 7950's.

go order a 2x4gb ram set and you'll be set


cgminer on two 7770's uses about 50mb... so ya.. its something about reaper that just requires a bit of ram.

once you get it working fine though... you'll get a nice hash rate.. i'm getting 1220khash from my 2x 7950 system.

Yes, I am totally aware of what I would need to do in order to turn these particular rigs into efficient litecoin mining rigs. Today I just basically ran a test to see if I was ready to make the switch whenever my other doomsday devices arrive and render these puppies useless for bitcoin mining. I know exactly what I need to do and will adjust accordingly. Thank you for all of y'alls help. Much appreciated

[meebs]

I picked up a 7770 the other day ($90 after MIR, couldn't pass it up lol)..

Some notes: It appears that core speed is mostly irrelevant, for example, 1000mhz core is the same hash rate as 1200mhz, just crank up mem speed to the max and you should be set..

Hash speed in cgminer/bfgminer is... weird. 5 second average always tends to be on the low side, showing at 185 each... however when I look at the pool.. it shows about 220-230 each.

[markodude]

I dont see any settings for 7970s. I am using these for mines which get around 525-550KHs

worksize 256
aggression 13
threads_per_gpu 5
sharethreads 18
lookup_gap 2
gpu_thread_concurrency 10240

[tacotime]

I dont see any settings for 7970s. I am using these for mines which get around 525-550KHs

worksize 256
aggression 13
threads_per_gpu 5
sharethreads 18
lookup_gap 2
gpu_thread_concurrency 10240

Same as for 7950, you should get 700-800 kh/s

[jasinlee]

Here is what I have tried, the rig is only getting a little less than ~100khs per gpu though with this setup.

Code:

./cgminer --scrypt -o http://litecoinpool.org:9332 -u username -p password --thread-concurrency 5632 -I 18 -g 1 -w 64 --auto-fan 

I have 4 5970s on it on linuxcoin 0.2b, any suggestions?

[tacotime]

what are your core/memory clocks?

[jasinlee]

what are your core/memory clocks?

For bitcoin I was using 165/775.

[tacotime]

what are your core/memory clocks?

For bitcoin I was using 165/775.

Litecoin requires memory speeds to be at or above the core speeds