[ANN] [ASIC-RESISTANT] UltraCoin (UTC) - Ultrafast 6 second transactions!!

[Protagonus]

Though I'm close, I'm using 4 different TC's for each card, my setup will be of no use to anyone for 290x, or perhaps one of the 4 might help 1 of 4 people... who knows.

Ye many miners may have to increase system RAM, hoping I can squeeze by with 8GB and 3 GPU;  gonna be close if so.

edit: think I'm at 0 HW's!! woot

care to share a pointer or two? i have 16GB RAM on my system with 3 7950s and keep figuring out new ways to BSOD

Sure, still playing but here's where I'm at with 0 HW's and getting 48.8 Kh/s per card.  WU showing 125 (3 cards in rig) but not long enough to settle.  Vitural RAM set @ default (same as system) 8192mb

For sapphire: 1060-1080 Clock and 1600Mhz RAM
For Powercolor: 1050 Clock and 1500 Mhz RAM

Code:

 ultracoinminer (pool info) --scrypt-jane --sj-nfmin 4 --sj-nfmax 30 --sj-time 1388361600 --failover-only --thread-concurrency 62720 --intensity 18 --lookup-gap 3 --gpu-powertune 20 --worksize 256 --gpu-dyninterval 5 --no-submit-stale -g 1 --scan-time 1 --expiry 10 --queue 0  

Proud to be running -g 1 still    

edit: not sure of catalyst / adl / sdk versions;  but they're probably 10 months old.

[Bluemustang]

Though I'm close, I'm using 4 different TC's for each card, my setup will be of no use to anyone for 290x, or perhaps one of the 4 might help 1 of 4 people... who knows.

Ye many miners may have to increase system RAM, hoping I can squeeze by with 8GB and 3 GPU;  gonna be close if so.

edit: think I'm at 0 HW's!! woot

Okay. I'm now at 56KH/s for each 280x accepting shares. Averaging 50 WU each,
But I got 1-2 HW errors on some and 0 on some others. Not perfect but close!
My 4 x 280x rigs only have 4GB RAM. I just had to raise the swap file to 32768.

Still tweaking....

How did you manage this?  Before the change i was getting 152 while everyone else seemed to be getting 135ish.  Now i cant get over 41. I've got 16gb ram using I11 and all TCs from 8k to 24k in 2048 increments.

[MasterCATZ]

As a temporary fix to get your rig running,
lower the intensity by 1 and reduce your total thread concurrency by a number
equal to the shaders count on your card.
Example: a 7950 card has 1792 shaders so the current TC should be 16384-1792= 14592

If your rig is still showing everything running normal, this is a false positive,
restart your rig and apply the new settings.

This is also a good time to check your virtual ram setting on windows systems
and set that at a size 3 or 4 times of what windows have set it at.

On a 4 gpu rig, lowering my TC actually increases my HW errors.  This has been true since the last N-factor.  

+1 decreasing has always given me HW increasing always made them go away
I found upping GPU cores helped last time

so far only dropped I 1 level
I also use small work sizes

using TC 65536 will tweak them later on
2560 shaders is what I think R9 290's had ..

got an excel sheet I worked on that’s got the new N factor setting on it somewhere ...

I think 3472MB was the MAX amount of RAM avail for GPU

[joawaywego]

Though I'm close, I'm using 4 different TC's for each card, my setup will be of no use to anyone for 290x, or perhaps one of the 4 might help 1 of 4 people... who knows.

Ye many miners may have to increase system RAM, hoping I can squeeze by with 8GB and 3 GPU;  gonna be close if so.

edit: think I'm at 0 HW's!! woot

care to share a pointer or two? i have 16GB RAM on my system with 3 7950s and keep figuring out new ways to BSOD

Sure, still playing but here's where I'm at with 0 HW's and getting 48.8 Kh/s per card.  WU showing 125 (3 cards in rig) but not long enough to settle.  Vitural RAM set @ default (same as system) 8192mb

For sapphire: 1060-1080 Clock and 1600Mhz RAM
For Powercolor: 1050 Clock and 1500 Mhz RAM

Code:

 ultracoinminer (pool info) --scrypt-jane --sj-nfmin 4 --sj-nfmax 30 --sj-time 1388361600 --failover-only --thread-concurrency 62720 --intensity 18 --lookup-gap 3 --gpu-powertune 20 --worksize 256 --gpu-dyninterval 5 --no-submit-stale -g 1 --scan-time 1 --expiry 10 --queue 0  

Proud to be running -g 1 still    

edit: not sure of catalyst / adl / sdk versions;  but they're probably 10 months old.

my cards seem happy after a few minutes of running. thanks again.

[cointradero]

Well, I couldn't get it to work right and I just don't have the patience to deal with it anymore since I'm not mining it currently as detailed in a previous post.  Though, I will point out a curiousity that happened last time I solo mined and it happened again.

I get a share well above the necessary amount to find a block.  I didn't see it happen.  All I know is it shows up in my best share area as 115k.  It doesn't show up in my wallet as a transaction at all.  Not an orphan, not anything.  No rejects or accepted in the miner.  It's like that share never happened.  The last 4 times I've had a share above the target while solo mining, I've seen no evidence that anything happened.  I can't figure it out.

I'm soloing another coin right now, all things are as expected.  Accepted shares show up as the current difficulty under A:.  Rejects do the same under R:   Accepted shares that are orphaned show up as transactions that never make it to my balance, or as orphans when in the debug screen. 

Something funny is going on with this coin or my wallet when solo mining. 

[Protagonus]

@joawaywego  

Glad to hear it worked out for you!!  I've finalized my settings and up'ed the TC a bit more.  





@ cointradero;  I'm pretty sure a "share above target" is not a valid find.  It's not really a reject or stale, though, so doesn't show up.  Someone else may clarify better

[projectdimpreza]

@joawaywego  

Glad to hear it worked out for you!!  I've finalized my settings and up'ed the TC a bit more.  


https://i.imgur.com/5VgLzK6.jpg


@ cointradero;  I'm pretty sure a "share above target" is not a valid find.  It's not really a reject or stale, though, so doesn't show up.  Someone else may clarify better

Is this for 280x card?

[Thirtybird]

For sapphire: 1060-1080 Clock and 1600Mhz RAM
For Powercolor: 1050 Clock and 1500 Mhz RAM

Code:

 ultracoinminer (pool info) --scrypt-jane --sj-nfmin 4 --sj-nfmax 30 --sj-time 1388361600 --failover-only --thread-concurrency 62720 --intensity 18 --lookup-gap 3 --gpu-powertune 20 --worksize 256 --gpu-dyninterval 5 --no-submit-stale -g 1 --scan-time 1 --expiry 10 --queue 0  

Proud to be running -g 1 still    

edit: not sure of catalyst / adl / sdk versions;  but they're probably 10 months old.

Those are good settings for that card - without 3GB onboard, you'd have to make even more compromises on lookup-gap.  You might squeak out 5-10kh/sec more if you utilized a miner with xintensity

-g any higher than 1 makes no difference at these N levels on cards with 1024 shaders or more... even on the low end cards, they're able to max out at g=1

*edit* saw your additional posts... increasing TC only increases the scrypt buffer size, which you do not need to do beyond where you were at.

[Protagonus]

Is this for 280x card?

7950's

Those are good settings for that card - without 3GB onboard, you'd have to make even more compromises on lookup-gap.  You might squeak out 5-10kh/sec more if you utilized a miner with xintensity

-g any higher than 1 makes no difference at these N levels on cards with 1024 shaders or more... even on the low end cards, they're able to max out at g=1

xD yea, I'm VERY tempted to try your yacminer now as I'm pretty sure I can get more as you say.  Your miner seems a lot better to fine tune with.  I'll have to check it out when I get more time to see.

Edit: I saw your edit.  Yea that last TC tweak only changed maybe 0.4 KH/s so not a big diff and I cannot go over that without other issues. 

[projectdimpreza]

Does anyone have the new settings for 280x?

[regnevacs]

close.. 46 khs no HW error.. 280x.. will try to tweak first before sharing it again.. i remember i shared the last N-factor settings and receive lots of donation

[projectdimpreza]

i dont mind giving donation 

[Thirtybird]

close.. 46 khs no HW error.. 280x.. will try to tweak first before sharing it again.. i remember i shared the last N-factor settings and receive lots of donation

That's going to be a tough one... it's so close to being able to use LG=3, but there just isn't going to be enough memory to get all 2048 shaders going full steam...

if you can use buffer-size 2458,
lg=3, w=128 R=7296
 would edge out
LG=4 W=256 X=32 (or I=18)
(aka --thread-concurrency 78656 --lookup-gap 4 -I 18)

if you can't go that high due to system ram, use LG=4 and TC=65536

[jakiman]

This is my setting for 280x to get about 56KH/s & 45 WU. Try it out and let me know.
It accumulates about 1 HW error every few minutes but so far, this is best I could do.

setx GPU_MAX_ALLOC_PERCENT 100

del scrypt-jane*.bin

C:\mining\cgminer-sj\cgminer.exe --scrypt-jane -o <pool> -u <id.worker> -p x -I 12 -g 2 -w 256 --thread-concurrency 28672 --lookup-gap 3 --gpu-engine 1025 --gpu-memclock 1500 --gpu-vddc 1.085 --gpu-powertune 10 --gpu-fan 80 --temp-cutoff 85 --sj-nfmin 4 --sj-nfmax 30 --sj-time 1388361600 --failover-only --no-submit-stale

If it helped, any donation is always welcome!  
UXLaLvo8L4eV9uC355WRzTX88GA7iFja2g

[eddaviking]

Getting 770 watts for a 4 280x gpu rig. Nice.  But just getting 34kh/s each with no HW errors.

[Thirtybird]

This is my setting for 280x to get about 56KH/s & 45 WU. Try it out and let me know.
It accumulates about 1 HW error every few minutes but so far, this is best I could do.

setx GPU_MAX_ALLOC_PERCENT 100

del scrypt-jane*.bin

C:\mining\cgminer-sj\cgminer.exe --scrypt-jane -o <pool> -u <id.worker> -p x -I 12 -g 2 -w 256 --thread-concurrency 29672 --lookup-gap 3 --gpu-engine 1025 --gpu-memclock 1500 --gpu-vddc 1.085 --gpu-powertune 10 --gpu-fan 80 --temp-cutoff 85 --sj-nfmin 4 --sj-nfmax 30 --sj-time 1388361600 --failover-only --no-submit-stale

If it helped, any donation is always welcome!  
UXLaLvo8L4eV9uC355WRzTX88GA7iFja2g

do you only have 2GB on the GPU?  You're leaving about 17% of the hashing power unused by using I=12.  Seriousy, try YACMiner or Kalroths with the patch so you can use Raw Intensity.

[ivanlabrie]

50khs no HW...280x at 1030/1500 

[noobyonekenobi]

Anyone have new settings for 270x?

[jakiman]

This is my setting for 280x to get about 56KH/s & 45 WU. Try it out and let me know.
It accumulates about 1 HW error every few minutes but so far, this is best I could do.

setx GPU_MAX_ALLOC_PERCENT 100

del scrypt-jane*.bin

C:\mining\cgminer-sj\cgminer.exe --scrypt-jane -o <pool> -u <id.worker> -p x -I 12 -g 2 -w 256 --thread-concurrency 29672 --lookup-gap 3 --gpu-engine 1025 --gpu-memclock 1500 --gpu-vddc 1.085 --gpu-powertune 10 --gpu-fan 80 --temp-cutoff 85 --sj-nfmin 4 --sj-nfmax 30 --sj-time 1388361600 --failover-only --no-submit-stale

If it helped, any donation is always welcome!  
UXLaLvo8L4eV9uC355WRzTX88GA7iFja2g

do you only have 2GB on the GPU?  You're leaving about 17% of the hashing power unused by using I=12.  Seriousy, try YACMiner or Kalroths with the patch so you can use Raw Intensity.

My GPUs all have 3GB. I'm using I=12 but that was what we used before with 280x so I just left it that's all.
I tried LG 4 but I end up not getting any shares even though I get decent hash rate and no HW errors. I will try YACMiner.

[Thirtybird]

50khs no HW...280x at 1030/1500 

PM me your settings, I'm curious how close I was with my guesstimation ...