my test shown X13 is not that good because although in old days:

"In the case of algorithms besides Scrypt, we multiply by the average change in speed when a miner switches to that algorithm. Scrypt-N is 0.47 (a 1MHs miner on Scrypt does about 470kHs on Scrypt-N). X11 is 4.0 (a 1MHs Scrypt miner will normally do around 4MHs on X11). X13 is 3.5. This lets you look quickly at the "vsLTC" column to see how well each algorithm is performing on a normalized scale. "

The new situation is using best miner today, a 280X can do 4.2MHs on X11 but only 2.9-3.0 on X13. So I stick with X11 for now, and some new X13 coins became worthless after mining.
(except today, the X11 number looks bad)

could be
1) network issue,
2) first 18 hour he was mining high difficulty coins, now he is mining low difficulty fast new coins, that will cause the massive variance issue with d=0.16 , consider that.
3) check the scale of the graph

Hi there, I thought you are going with p2pool

Hint, set a lower difficulty in password like d=0.05, if you use 280x, so your hash rate will be more steady.
But don't set it too low.

My testing shown difficulty should be set according to the hash rate of per card, not the whole rig, for example, if you have 4 280x, setting a 0.16 difficulty will waste tons of your hash rate because waffle pool is fast at finding blocks, if it takes over 1 or 2 minute for each a card to submit a work, there is a big chance a new block has been found and your wasted lots of GPU time hashing the old block.  So set d=0.05 no matter how many 280x you have in your rig will work better.

I am talking about X11. The theory is each card was actually working individually with 2 threads (-g 2) on solving the assigned work, they don't work together, 4 cards in a rig means you have 8 threads working independently, but the difficulty set in stratum protocol was assigned to each card, not divided among the number of cards.

so, for example, rig 1 has 1 280x, I use d=0.05, rig 2 has 6 280x, I still use d=0.05, each card (thread) will submit a hash in same amount of time.

but if you use d=0.30 on that 6 280x rig, each card will take 2-3 minutes to submit a hash and there is a good chance it will be stale or being interrupted in the middle of working because a new block has been found.

correct me if I am wrong.  

BTW, I actually set those number with --gpu-engine 535 and --gpu-memclock 735

I think if you don't set, you were running at full speed, just lower -I 15

press g in your sgminer window to check

some advice guys about the following tip i got for my AMD R280X thats currently running at 4.3 MH/s for X11
(using SGminer v5 with -k darkcoin-mod) --> http://www.reddit.com/r/DRKCoin/comments/2bbqt3/holy_cow_batman_i_just_downloaded_the_new_sgminer/


*****************************************************************************************
press g on sgminer, then press x. after that type 500 and press enter. tell me if that boosts your hashrate.
I think you need more than 4-8gb of ram for it to run without hardware errors though. DDR3 ram to be exact.
I also have the new pci express i think its either 2.0 or 3.0 i have to check.
*****************************************************************************************

what does that x 500 (--xintensity 500) precisely do ? i dont feel like blowing up my graphic card just to find out...

btw.. this is my current batch file (i dont use conf. files) :

setx GPU_MAX_ALLOC_PERCENT 100
setx GPU_USE_SYNC_OBJECTS 1
c:\sgminerdarkcoin\sgminer.exe -k darkcoin-mod -o stratum+tcp://stratum-eu.suchpool.pw:3335 --failover-switch-delay 60 --no-submit-stale --gpu-dyninterval 7 --gpu-platform 0 --api-mcast-port 4028 --api-port 4028 --expiry 1 --log 5 --no-pool-disable --tcp-keepalive 30 --temp-hysteresis 3 --shares 0 --vectors 1 -u YrUsername.1 -p Yrpassword --expiry 1 --queue 0 --scan-time 1 --gpu-engine 1150 --gpu-memclock 1500 -g 2 -w 256 --lookup-gap 2 -I 18 --thread-concurrency 8192 --shaders 2048