CCminer(SP-MOD) Modded GPU kernels.

[pallas]

I don't think that he speaks chinese ;-)

Yeah!!!
//一個手前のスレッドからデータを貰う(同時に一個先のスレッドにデータを送る)

sp still didn't find the sources of this optimized lyra to include to his private? ))

"google translate" says it's japanese

[1714042374]

1714042374

[1714042374]

1714042374

[rednoW]

I don't think that he speaks chinese ;-)

Yeah!!!
//一個手前のスレッドからデータを貰う(同時に一個先のスレッドにデータを送る)

sp still didn't find the sources of this optimized lyra to include to his private? ))

"google translate" says it's japanese

You are right again ))

[sp_]

Spmod #7 lyra2v2 results

standard clocks

980ti (1280 core clock) 35600 (g1 gaming)
970   (1366 core clock) 22600 (g1 gaming)
960  (1350 core clock) 14500 (msi gaming)
750ti (1275 core clock) 6700 (asus strix)

so 970+960 would do around 37100 on 1350mhz core

Both tsiv and djm34 had kernals with the shuffle instruction a few days after the launch. This nicehash kernal has a simple crypt pass to avoid kernal extraction and reverse engineering. So my bet is that it is a private kernal has been used for a long time in binary form before it was bought by nicehash  and spread.

[rednoW]

Spmod #7 lyra2v2 results

standard clocks

980ti (1280 core clock) 35600 (g1 gaming)
970   (1366 core clock) 22600 (g1 gaming)
960  (1350) 14500 (msi gaming)

so 970+960 would do around 37100 on 1350mhz core

Both tsiv and djm34 had kernals with the shuffle instruction a few days after the launch. This nicehash kernal has a simple crypt pass to avoid kernal extraction and reverse engineering. So my bet is that it is a private kernal that has been bougth and spread.

Is that what we've seen in japaneese kernel a first use of nvidia shared memory technology to improve mining speeds?

[sp_]

Is that what we've seen in japaneese kernel a first use of nvidia shared memory technology to improve mining speeds?

Ofcourse not. Shared mem is already used in groestl, simd,echo, shavite (4 of the hashing algos in x11 in ccminer sp-mod)

[rednoW]

Spmod #7 lyra2v2 results

standard clocks

980ti (1280 core clock) 35600 (g1 gaming)
970   (1366 core clock) 22600 (g1 gaming)
960  (1350 core clock) 14500 (msi gaming)
750ti (1275 core clock) 6700 (asus strix)

so 970+960 would do around 37100 on 1350mhz core

Both tsiv and djm34 had kernals with the shuffle instruction a few days after the launch. This nicehash kernal has a simple crypt pass to avoid kernal extraction and reverse engineering. So my bet is that it is a private kernal that has been bougth and spread.

About what encryption are you talking about?
Isn't it an open source?
https://github.com/nicehash/ccminer-sp/commit/8fd0d35c314266c1b3186a3d16c79212df611f0a

[giagge]

Mhm good .

[sp_]

Doesn't make any sense..

d_hash2[thr_id] = (uint64_t*)(((uint64_t)d_hash2[thr_id] + 127)&~127);
lyra2v2_cpu_init(thr_id, throughput, d_hash2[thr_id]);
(....)
d_hash[thr_id] = (uint64_t*)(((uint64_t)d_hash[thr_id] + 127)&~127);


bether to remove it like this:

lyra2v2_cpu_init(thr_id, throughput, d_hash2[thr_id]);
(...)

[sp_]

About what encryption are you talking about?
Isn't it an open source?
https://github.com/nicehash/ccminer-sp/commit/8fd0d35c314266c1b3186a3d16c79212df611f0a

The binary has probobly been sold and spread earlier(closed source).  There is code that will corrupt the hash if you don't add the two lines of code outside the kernal code.

[sp_]

sp-mod #7 has been sendt to the donators. Check you inboxes.

[rednoW]

sp-mod #7 has been sendt to the donators. Check you inboxes.

Can you compare hashrates of nicehash lyra2v2 version and yours #7 private after incorporating this open source code in it?

[sp_]

Seems to work. Why don't you try yourself.

What hashrates are you getting?

[sp_]

i am playing with intensy and got this results :
970+960 core 1503 mem 3005

What do you get with sp-mod #7?

[fenomenhaa]

Spmod #7 lyra2v2 results peaked

970   (1500 boost core clock) 24500 (evga ssc gaming elpida)
750ti (1430 boost core clock) 7800 (zotac samsung)

[sp_]

Spmod #7 lyra2v2 results peakedi
970   (1500 boost core clock) 24500 (evga ssc gaming elpida)
750ti (1430 boost core clock) 7800 (zotac samsung)

Nice. Thanks for testing.

[thevictimofuktyranny]

Spmod #7 lyra2v2 results peakedi
970   (1500 boost core clock) 24500 (evga ssc gaming elpida)
750ti (1430 boost core clock) 7800 (zotac samsung)

Nice. Thanks for testing.

Peak hash difference 750TI extra 300Khs compared to Nicehash miner.

Peak hash difference GTX 950 extra 700Khs compared to Nicehash miner.

Peak hash difference GTX 960 extra 800Khs compared to Nicehash miner.

Across the Nvidia's cards SPMOD is best at LYRA2V2

Buy it Quick

[Amph]

Spmod #7 lyra2v2 results peaked

970   (1500 boost core clock) 24500 (evga ssc gaming elpida)
750ti (1430 boost core clock) 7800 (zotac samsung)

is that private? 24500 is better thna the one with ncicehash, but maybe it just oc

[rednoW]

Seems to work. Why don't you try yourself.

What hashrates are you getting?

I have no private. I have only one poor gtx750. All others are amd

[sp_]

Seems to work. Why don't you try yourself.

What hashrates are you getting?

I have no private. I have only one poor gtx750. All others are amd

My public private is working pretty good on the 28nm 750's in a grid.

[scryptr]

Spmod #7 lyra2v2 results peakedi
970   (1500 boost core clock) 24500 (evga ssc gaming elpida)
750ti (1430 boost core clock) 7800 (zotac samsung)

Nice. Thanks for testing.

Peak hash difference 750TI extra 300Khs compared to Nicehash miner.

Peak hash difference GTX 950 extra 700Khs compared to Nicehash miner.

Peak hash difference GTX 960 extra 800Khs compared to Nicehash miner.

Across the Nvidia's cards SPMOD is best at LYRA2V2

Buy it Quick

WHAT HE SAID!!! --

I am getting similar results on my cards.  Private #7 beats the NiceHash Lyra2v2 miner by a significant margin on Lyra2v2, and is easier to overclock.  It accepts higher intensity settings on the command line.       --scrryptr