Pages:
Author

Topic: CCminer(SP-MOD) Modded NVIDIA Maxwell / Pascal kernels. - page 30. (Read 2347601 times)

jr. member
Activity: 213
Merit: 3
LoL, did you really just give results from a miner that just started up? Jesus dude.. And how is it the same settings? You have Alexis at 22 intensity, Trex at 20 intensity, and spmod at 24.4 intensity?

TOP INTENSITY LIMITS--

I haven't bothered to test c11 yet, but each miner has a top-end limit to the intensity setting.  It is likely not possible to raise the intensity setting on any of the three miners.  Further, the top intensity setting possible is often not the best performance setting.       --scryptr

I hear what you're saying, but to post screenshots of the first 3 seconds and use that as legit results? Come on..
jr. member
Activity: 186
Merit: 4
It would be a lot better if you would:
-post tests with your miner vs. other miners over a reasonable amount of time
-100%TDP means nothing, post Watts pls.
-you should know this.
legendary
Activity: 1797
Merit: 1028
LoL, did you really just give results from a miner that just started up? Jesus dude.. And how is it the same settings? You have Alexis at 22 intensity, Trex at 20 intensity, and spmod at 24.4 intensity?

TOP INTENSITY LIMITS--

I haven't bothered to test c11 yet, but each miner has a top-end limit to the intensity setting.  It is likely not possible to raise the intensity setting on any of the three miners.  Further, the top intensity setting possible is often not the best performance setting.       --scryptr
jr. member
Activity: 213
Merit: 3
LoL, did you really just give results from a miner that just started up? Jesus dude.. And how is it the same settings? You have Alexis at 22 intensity, Trex at 20 intensity, and spmod at 24.4 intensity?
sp_
legendary
Activity: 2926
Merit: 1087
Team Black developer
same card, same settings, 3 miners. ALGO: C11

alexis1.0, t-rex 0.6.1 (1% fee), spmod-git #9

testcard: asus 1080ti
coreclock: +100
memclock: +600
tdp: 100%



member
Activity: 312
Merit: 10
C11 spmod-git9  (cuda 9.2) vs t-rex 0.6.1  (cuda 9.1) vs enemy 1.17 (cuda 9.2) vs alexis 1.0 (cuda 7.5)

Latest nvidia driver.

Testcard: asus 1080ti

memclock: -500
coreclock: +100
tdp: 90%

sp-modgit      t-rex 0.6.1     enemy 1.17      alexis 1.0
 
32.2 MHASH  32.5MHASH    31.8MHASH       30.2MHASH

memclock: +500
coreclock: +100
tdp: 90%

sp-modgit9      t-rex 0.6.1     enemy 1.17   alexis 1.0

34MHASH   32MHASH    31.5MHASH   31.5 MHASH


-My free opensource miner is around +6% faster than the miners from the GPL violators. And free without any fee.
-If you oc the memory and run the old alexis miner it performs equal to the latest z-enemy miner. And free without any fee.



again, you're trying to deceive people, would you stop that?
you post screenshots from the console at the start. put a real comparison on the pool at two different addresses, one yours, the other z-enemy or t-rex on two identical rigs.
show that your miner gives more coins, and you can be trusted.
at the moment, in the console, your mainer draws a good speed, but at the output of coins loses to all miners

could you test spmod again? i am curious why sp post like that
big thanks if you want test it again  Smiley
sp_
legendary
Activity: 2926
Merit: 1087
Team Black developer
at the moment, in the console, your mainer draws a good speed, but at the output of coins loses to all miners

My code is based on ccminer 2.2.4 and opensource. The speed on the screen is correct, but if you mine a coin with a short blocktime you should reduce the intensity.
member
Activity: 392
Merit: 27
http://radio.r41.ru
C11 spmod-git9  (cuda 9.2) vs t-rex 0.6.1  (cuda 9.1) vs enemy 1.17 (cuda 9.2) vs alexis 1.0 (cuda 7.5)

Latest nvidia driver.

Testcard: asus 1080ti

memclock: -500
coreclock: +100
tdp: 90%

sp-modgit      t-rex 0.6.1     enemy 1.17      alexis 1.0
 
32.2 MHASH  32.5MHASH    31.8MHASH       30.2MHASH

memclock: +500
coreclock: +100
tdp: 90%

sp-modgit9      t-rex 0.6.1     enemy 1.17   alexis 1.0

34MHASH   32MHASH    31.5MHASH   31.5 MHASH


-My free opensource miner is around +6% faster than the miners from the GPL violators. And free without any fee.
-If you oc the memory and run the old alexis miner it performs equal to the latest z-enemy miner. And free without any fee.



again, you're trying to deceive people, would you stop that?
you post screenshots from the console at the start. put a real comparison on the pool at two different addresses, one yours, the other z-enemy or t-rex on two identical rigs.
show that your miner gives more coins, and you can be trusted.
at the moment, in the console, your mainer draws a good speed, but at the output of coins loses to all miners
sp_
legendary
Activity: 2926
Merit: 1087
Team Black developer
C11 spmod-git9  (cuda 9.2) vs t-rex 0.6.1  (cuda 9.1) vs enemy 1.17 (cuda 9.2) vs alexis 1.0 (cuda 7.5)

Latest nvidia driver.

Testcard: asus 1080ti

memclock: -500
coreclock: +100
tdp: 90%

sp-modgit      t-rex 0.6.1     enemy 1.17      alexis 1.0
 
32.2 MHASH  32.5MHASH    31.8MHASH       30.2MHASH

memclock: +500
coreclock: +100
tdp: 90%

sp-modgit9      t-rex 0.6.1     enemy 1.17   alexis 1.0

34MHASH   32MHASH    31.5MHASH   31.5 MHASH


-My free opensource miner is around +6% faster than the miners from the GPL violators. And free without any fee.
-If you oc the memory and run the old alexis miner it performs equal to the latest z-enemy miner. And free without any fee.


legendary
Activity: 2294
Merit: 1182
Now the money is free, and so the people will be
oh boy well do I have news for you hahahaha.  NVIDIA's Q1FY19 sheet says they sold 289$million of crypto-related GPU gear.  These are mostly high end, it comes out to 300k-350k 1080tis equivalents *most likely*.  Get this though, sold in the first quarter.  Who the hell bought so many gpus for  mining at that point is a big Huh for me and why.  When we look at total GPU sales, we see that OEM sales were *lower* in the Q4 last year, vs 1st quarter this year, and a little lower in the 1st quarter of last year.  This means that OEM crypto sales jumped for 1 quarter only.  Now guidance says 0 expected sales for crypto, its 20% of Q1 level now.


The big money boyz all bought their hashing power at the worst possible moment and for the worst possible price.
legendary
Activity: 1764
Merit: 1024
Both 1.44 and 1.03 are shit profit when you add in electricity costs. Mining is shit now..

It's been that way for a few months, January onward was the worst time to get into crypto mining, almost as bad as the end of 2014, only everything didn't go completely negative at $.1KWH. We're pretty fricking close, but not there yet.

If BTC hits about 5.2k~ I'm pretty sure we'll be negative. It's tough to tell as people with more expensive power have already shut off their rigs, so when BTC and altcoin prices increase, revenue rises, so will hashrate to compensate and we'll be stuck in limbo for awhile until things really take off. How much 'dark' hash is around now? I have no idea. But that's a pretty big sponge to soak up revenue from now till... whenever?

If you have expensive power you should really consider selling your rigs. Take that for what you will.
sp_
legendary
Activity: 2926
Merit: 1087
Team Black developer
X16r/X16s/C11 Spmod-git #9 has been released.


The C11 has been added. (around +33% faster thn sp-mod git#8).. +5.3% faster than T-rex 0.6.1  on the gtx 970.

This miner is free and has no fee.

Cuda 9.2 binary:

-Added optimalizations to C11 (+33% faster)
-faster X16R/X16S (faster bmw, shavite-cubehash)
-removed the other algos. (optimalizations not added yet)

https://github.com/sp-hash/suprminer/releases

source

https://github.com/sp-hash/suprminer/commits/master
jr. member
Activity: 112
Merit: 3
Linux Forever... Resistance is futile!!!
Is it possible to add support for ubuntu 18.04?
Giving error on compilation


Code:
./configure: line 6177: syntax error near unexpected token `,'
./configure: line 6177: `LIBCURL_CHECK_CONFIG(, 7.15.2, ,'
newbie
Activity: 93
Merit: 0
Is there a free phi2 release?
sp_
legendary
Activity: 2926
Merit: 1087
Team Black developer
Pushed a 30% faster C11 implementation to my github. free and opensource. Easy to add speed in the other algos as well.
You just need to copy and replace the old kernel calls, with the new sp-modded kernel calls from the x16Rmod/x16S.

https://github.com/sp-hash/suprminer/commits/master
member
Activity: 312
Merit: 10
start two identical rigs for 12 hours
z-enemy vs SP_
let's see which one will bring more coins (not hash)
even if you take into account the fee to developer 2% of the enemy, the enemy will bring more coins

do you already compare them?
yeah i dont care about 1 or 2% fee, all i need how many coins/day
test is off
SP_ lost on coins

thanks, so i dont need used spversion
member
Activity: 392
Merit: 27
http://radio.r41.ru
start two identical rigs for 12 hours
z-enemy vs SP_
let's see which one will bring more coins (not hash)
even if you take into account the fee to developer 2% of the enemy, the enemy will bring more coins

do you already compare them?
yeah i dont care about 1 or 2% fee, all i need how many coins/day
test is off
SP_ lost on coins
member
Activity: 312
Merit: 10
start two identical rigs for 12 hours
z-enemy vs SP_
let's see which one will bring more coins (not hash)
even if you take into account the fee to developer 2% of the enemy, the enemy will bring more coins

do you already compare them?
yeah i dont care about 1 or 2% fee, all i need how many coins/day
sp_
legendary
Activity: 2926
Merit: 1087
Team Black developer
spreadminer #16 released to the donators.

- added more speed. (around +5-10%)
- Fixed the intensity setting so you can set per gpu.
- improved the default launch configs on 1070ti/1070/p102/p104
legendary
Activity: 1797
Merit: 1028
HOLD THE PHONE--

There is a new nVidia architecture and graphics product line in town -- Turing architecture and RTX 20 Graphics seres.  All coding must be re-written for the new chips and be backwards compatible with Pascal, Maxwell, and maybe Kepler, too.

Hope it isn't too much trouble.       --scryptr
Pages:
Jump to: