CCminer(SP-MOD) Modded NVIDIA Maxwell / Pascal kernels. - page 125.

sp_

legendary

Activity: 2954

Merit: 1087

Team Black developer

Quote from: rednoW on November 23, 2017, 05:36:52 PM

I think the main thing that will hurt sp_ and his buyers of this mod is stale shares. sp_ isn't good in optimizing miner <=> pool communications.

I have improved the pool communication in sp-mod #1 optimized for the miningpoolhub pool.

sp-mod is currently testing with 1410.190 valid shares. 0 invalid...

sp_

legendary

Activity: 2954

Merit: 1087

Team Black developer

Quote from: CjMapope on November 23, 2017, 11:47:19 PM

with my 1080ti's doing 3300 easy, i wont worry about this SP_mod unless a 1070ti does 2500-2700, no point for me otherwise.

What are your clocks to reach 2300?. My mod can do 2350-2400 now, but I have the cheapest 1070ti card with a "reference" cooler so I guess 2500 is possible with the more expensive 1070ti cards.

sp_

legendary

Activity: 2954

Merit: 1087

Team Black developer

1070ti ~~2250~~ ~~2275~~ 2350 Can go higher with more oc...

zorvalth

full member

Activity: 223

Merit: 100

Quote from: wacko on November 24, 2017, 08:23:49 AM

Quote from: zorvalth on November 24, 2017, 02:24:29 AM

guys, did somebody compare the bitcore profit from calcs and pools? No matter the miner on both suprnova i yiimp i got massive difference like 40%. On all other algos i\m always between 0 and 10-15% difference tops...

For the past 24 hours I've got ~ 33% less btx than calcs promise. But the day before that it was ~ 10% less. Mining on yiimp.eu. I don't know, maybe the calcs are wrong on this one, maybe the diff was jumping a lot (althouth I don't see it going much above 10k for the last few days).

WTM uses last 24h average diff, so if yo compare it with your profit for the last day should be pretty correct.

sp_

legendary

Activity: 2954

Merit: 1087

Team Black developer

1070ti ~~2250~~ 2275 Can go higher with more oc...

stevascha

member

Activity: 312

Merit: 10

interesting kernel dev Grin

candyiman

full member

Activity: 308

Merit: 100

Quote from: wacko on November 24, 2017, 08:23:49 AM

Quote from: zorvalth on November 24, 2017, 02:24:29 AM

guys, did somebody compare the bitcore profit from calcs and pools? No matter the miner on both suprnova i yiimp i got massive difference like 40%. On all other algos i\m always between 0 and 10-15% difference tops...

For the past 24 hours I've got ~ 33% less btx than calcs promise. But the day before that it was ~ 10% less. Mining on yiimp.eu. I don't know, maybe the calcs are wrong on this one, maybe the diff was jumping a lot (althouth I don't see it going much above 10k for the last few days).

im mining on miningpoolhub but it has often invalid shares yet miner is solid

wacko

legendary

Activity: 1106

Merit: 1014

Quote from: zorvalth on November 24, 2017, 02:24:29 AM

guys, did somebody compare the bitcore profit from calcs and pools? No matter the miner on both suprnova i yiimp i got massive difference like 40%. On all other algos i\m always between 0 and 10-15% difference tops...

For the past 24 hours I've got ~ 33% less btx than calcs promise. But the day before that it was ~ 10% less. Mining on yiimp.eu. I don't know, maybe the calcs are wrong on this one, maybe the diff was jumping a lot (althouth I don't see it going much above 10k for the last few days).

pallas

legendary

Activity: 2716

Merit: 1094

Black Belt Developer

Quote from: bensam1231 on November 24, 2017, 01:44:09 AM

Quote from: rednoW on November 24, 2017, 01:37:08 AM

Quote from: bensam1231 on November 23, 2017, 07:52:44 PM

Low power usage is usually a sign that the algo isn't very well optimized.

say it to xmr miner developers )))

That is almost exclusively memory, same with Ethereum. Lyra2Z is not. :DDD

lyra2z uses a bigger table, hence why it uses gpu ram unlike lyra2rev[12]

Dr_Victor

sr. member

Activity: 954

Merit: 250

Quote from: zorvalth on November 24, 2017, 02:24:29 AM

guys, did somebody compare the bitcore profit from calcs and pools? No matter the miner on both suprnova i yiimp i got massive difference like 40%. On all other algos i\m always between 0 and 10-15% difference tops...

I have compared - everything is ok for me.

zorvalth

full member

Activity: 223

Merit: 100

guys, did somebody compare the bitcore profit from calcs and pools? No matter the miner on both suprnova i yiimp i got massive difference like 40%. On all other algos i\m always between 0 and 10-15% difference tops...

sp_

legendary

Activity: 2954

Merit: 1087

Team Black developer

Quote from: Schtroumpf on November 23, 2017, 07:39:20 PM

zcoin dev posted a public ccminer with fix for rates poolside problem Tongue

I have also fixed the poolside problem now.

bensam1231

legendary

Activity: 1764

Merit: 1024

Quote from: rednoW on November 24, 2017, 01:37:08 AM

Quote from: bensam1231 on November 23, 2017, 07:52:44 PM

Low power usage is usually a sign that the algo isn't very well optimized.

say it to xmr miner developers )))

That is almost exclusively memory, same with Ethereum. Lyra2Z is not. :DDD

rednoW

legendary

Activity: 1510

Merit: 1003

Quote from: bensam1231 on November 23, 2017, 07:52:44 PM

Low power usage is usually a sign that the algo isn't very well optimized.

say it to xmr miner developers )))

Dr_Victor

sr. member

Activity: 954

Merit: 250

Quote from: Schtroumpf on November 23, 2017, 07:39:20 PM

Quote from: sp_ on November 23, 2017, 05:52:50 PM

tpruvot's release seems to have lower rates poolside than in the miner window.

zcoin dev posted a public ccminer with fix for rates poolside problem Tongue

with this public ccminer (correct pool rates) :
1060 6gb (micron) (70% tdp): 1200 kH/s
1080ti 11gb (70% tdp): 3260 kH/s

will test on 70ti this week end ..

Link, please

crocozino

sr. member

Activity: 420

Merit: 250

Quote from: sp_ on November 23, 2017, 12:27:35 PM

Quote from: bensam1231 on November 21, 2017, 02:30:59 PM

Sp look at Lyra2Z for Zcoin. Do I need to slap you to look at multi-million dollar markets or do you have a truce with someone to not make a better miner for it?

Yes. I have decided so spend 50-100 hours on a mod. I have reached 2000 khash on the gtx 1070 so far.

right now gtx 1070 can push about 1730 kh/s, so your 2000 hash it is 10% gain, a good performance income, yes, but I think you could do better then this, SP
usually your miners brings about 20-25% gain

CjMapope

legendary

Activity: 1820

Merit: 1092

~Full-Time Minter since 2016~

Quote from: Schtroumpf on November 23, 2017, 07:39:20 PM

Quote from: sp_ on November 23, 2017, 05:52:50 PM

tpruvot's release seems to have lower rates poolside than in the miner window.

zcoin dev posted a public ccminer with fix for rates poolside problem Tongue

with this public ccminer (correct pool rates) :
1060 6gb (micron) (70% tdp): 1200 kH/s
1080ti 11gb (70% tdp): 3260 kH/s

will test on 70ti this week end ..

ya they posted that miner weeks ago ;p (its the one "kaylmist" has been trying to sell people lolz)
1070ti does 2300 kh/s, i own a few and have extensively tested

with my 1080ti's doing 3300 easy, i wont worry about this SP_mod unless a 1070ti does 2500-2700, no point for me otherwise.

tranquanghuy

member

Activity: 166

Merit: 10

Hshare (hcash) sp-mod #1 sendt to the donators.

#1 gtx 1070 (13.5MHASH)
#2 gtx 1060 (8.8MHASH)
#3 gtx 1060 3gb (8MHASH)

bensam1231

legendary

Activity: 1764

Merit: 1024

Quote from: Schtroumpf on November 23, 2017, 07:39:20 PM

Quote from: sp_ on November 23, 2017, 05:52:50 PM

tpruvot's release seems to have lower rates poolside than in the miner window.

zcoin dev posted a public ccminer with fix for rates poolside problem Tongue

with this public ccminer (correct pool rates) :
1060 6gb (micron) (70% tdp): 1200 kH/s
1080ti 11gb (70% tdp): 3260 kH/s

will test on 70ti this week end ..

If there is one it's not so 'publicly available'. My 1080tis aren't getting close to that, about 2900 with DJM34s cuda9 release. That also is showing up lower on pool by about 10%.

You don't need to adjust TDP, the algo itself uses almost 60% of full TDP.

Quote from: rednoW on November 23, 2017, 05:36:52 PM

Quote from: wacko on November 23, 2017, 02:24:12 PM

Quote from: rednoW on November 23, 2017, 12:44:02 PM

I have only 1780 with gtx1070. So 2k is good!

As always, there's a good chance that SP's "2k" is with the card overclocked at 100+ TDP, consuming almost like a gtx 1080.

And the actual gain compared to the free miner will be ~7-8%. Just like it was with keccak.

current djm's and tpruvot's ccminers are working with max oc (like 2050-2100mhz) consuming only 50-60% tdp. So it will not be so easy for sp_ to claim performance gain just with more oc with higher tdp without actual code optimization.
I think the main thing that will hurt sp_ and his buyers of this mod is stale shares. sp_ isn't good in optimizing miner <=> pool communications.

Low power usage is usually a sign that the algo isn't very well optimized.

Schtroumpf

newbie

Activity: 34

Merit: 0

Quote from: sp_ on November 23, 2017, 05:52:50 PM

tpruvot's release seems to have lower rates poolside than in the miner window.

zcoin dev posted a public ccminer with fix for rates poolside problem Tongue

with this public ccminer (correct pool rates) :
1060 6gb (micron) (70% tdp): 1200 kH/s
1080ti 11gb (70% tdp): 3260 kH/s

will test on 70ti this week end ..

Topic: CCminer(SP-MOD) Modded NVIDIA Maxwell / Pascal kernels. - page 125. (Read 2347601 times)