Skip to content

Commit

Permalink
keccak second nonce, and higher intensity
Browse files Browse the repository at this point in the history
  • Loading branch information
@tpruvot
tpruvot committed Dec 4, 2017
1 parent cf886b5 commit 015d129
Show file tree
Hide file tree
Showing 3 changed files with 18 additions and 7 deletions.
15 changes: 13 additions & 2 deletions Algo256/keccak256.cu
View file
Original file line number Diff line number Diff line change
Expand Up @@ -72,7 +72,8 @@ extern "C" int scanhash_keccak256(int thr_id, struct work* work, uint32_t max_no

if(!use_compat_kernels[thr_id]) {
uint32_t intensity = 23;
if (strstr(device_name[dev_id], "GTX 1080")) intensity = 25;
if (strstr(device_name[dev_id], "GTX 1070")) intensity = 25;
if (strstr(device_name[dev_id], "GTX 1080")) intensity = 26;
throughput = cuda_default_throughput(thr_id, 1U << intensity);
keccak256_cpu_init(thr_id);
} else {
Expand Down Expand Up @@ -120,7 +121,17 @@ extern "C" int scanhash_keccak256(int thr_id, struct work* work, uint32_t max_no
if (vhash[7] <= ptarget[7] && fulltest(vhash, ptarget)) {
work->valid_nonces = 1;
work_set_target_ratio(work, vhash);
pdata[19] = work->nonces[0] + 1;
if (!use_compat_kernels[thr_id] && work->nonces[1] != UINT32_MAX) {
be32enc(&endiandata[19], work->nonces[1]);
keccak256_hash(vhash, endiandata);
if (vhash[7] <= ptarget[7] && fulltest(vhash, ptarget)) {
work->valid_nonces++;
bn_set_target_ratio(work, vhash, 1);
}
pdata[19] = max(work->nonces[0], work->nonces[1]) + 1;
} else {
pdata[19] = work->nonces[0] + 1;
}
return work->valid_nonces;
}
else if (vhash[7] > Htarg) {
Expand Down
2 changes: 1 addition & 1 deletion ccminer.vcxproj.filters
View file
Original file line number Diff line number Diff line change
Expand Up @@ -824,7 +824,7 @@
<Filter>Source Files\CUDA\Algo256</Filter>
</CudaCompile>
<CudaCompile Include="Algo256\keccak256.cu">
<Filter>Source Files\CUDA</Filter>
<Filter>Source Files\CUDA\Algo256</Filter>
</CudaCompile>
<CudaCompile Include="Algo256\cuda_blake256.cu">
<Filter>Source Files\CUDA\Algo256</Filter>
Expand Down
8 changes: 4 additions & 4 deletions res/ccminer.rc
View file
Original file line number Diff line number Diff line change
Expand Up @@ -60,8 +60,8 @@ IDI_ICON1 ICON "ccminer.ico"
//

VS_VERSION_INFO VERSIONINFO
FILEVERSION 2,2,2,0
PRODUCTVERSION 2,2,2,0
FILEVERSION 2,2,3,0
PRODUCTVERSION 2,2,3,0
FILEFLAGSMASK 0x3fL
#ifdef _DEBUG
FILEFLAGS 0x21L
Expand All @@ -76,10 +76,10 @@ BEGIN
BEGIN
BLOCK "040904e4"
BEGIN
VALUE "FileVersion", "2.2.2"
VALUE "FileVersion", "2.2.3"
VALUE "LegalCopyright", "Copyright (C) 2017"
VALUE "ProductName", "ccminer"
VALUE "ProductVersion", "2.2.2"
VALUE "ProductVersion", "2.2.3"
END
END
BLOCK "VarFileInfo"
Expand Down

9 comments on commit 015d129

@zeefporto
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Hi tpruvot,

i´m running ccminer-x64-2.2.3-cuda9 on an 1060 6gb to mining electroneum with no problems. :)

I have an portable that is in work basically doing nothing and trying there the same ccminer version.

He have an 820m nvidia gpu that i think is cuda 2.1. Ofcourse when i try to use this verion i got this error:

https://ibb.co/hGMUFG

Any tip for what version should i use to use this graphic or if its possible to use it?

Regards,

@zeefporto
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

no help ? :(

@eLvErDe
Copy link

@eLvErDe eLvErDe commented on 015d129 Dec 15, 2017 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

CUDA 9 is way too recent for you card.

I don't know anything regarding Windows stuff but CUDA 9 does not supports compute level under 3.0.

@zeefporto
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Hey,

Thanks for response,

So no hipotese of mining with my gpu maybe with other program? :)

Regards,

@kkkrackpot
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Hi,

Self-compiled 2.2.4 "cuda-9" branch on Linux with GTX 1060 3Gb.

I tried "keccak" (Maxcoin) on zpool.ca and noticed that the miner reports something like 400Mh/s, while on pool dashboard it's only about 40Mh/s (i.e. 10 times less).

Is it the pool's fault or miner's?

@eLvErDe
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Try adding --submit-stale

@tpruvot
Copy link
Owner Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

no, use -f 2 in this case... its pool factor seems 256 (but not the default set in yiimp sample)

@kkkrackpot
Copy link

@kkkrackpot kkkrackpot commented on 015d129 Dec 18, 2017 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@eLvErDe --submit-stale seems not working
@tpruvot the pool's default diff is 16 as far as I can understand. With -f 2 I see the same difficulty (16) in dashboard (and in the miner), but the miner starts working in "batch mode": a batch of low diff shares, then a high diff share, then "booo: diff too low", and then a batch of low diff shares again:

[2017-12-19 00:49:07] accepted: 367/373 (diff 0.077), 413.89 MH/s yes!
[2017-12-19 00:49:07] accepted: 368/374 (diff 0.077), 413.89 MH/s yes!
[2017-12-19 00:49:07] accepted: 369/375 (diff 0.077), 413.89 MH/s yes!
[2017-12-19 00:49:07] accepted: 370/376 (diff 0.077), 413.89 MH/s yes!
[2017-12-19 00:49:07] accepted: 371/377 (diff 0.077), 413.89 MH/s yes!
[2017-12-19 00:49:07] accepted: 372/378 (diff 0.077), 413.89 MH/s yes!
[2017-12-19 00:49:07] accepted: 373/379 (diff 0.077), 413.89 MH/s yes!
[2017-12-19 00:49:07] accepted: 374/380 (diff 0.077), 413.89 MH/s yes!
[2017-12-19 00:49:07] accepted: 375/381 (diff 0.077), 413.89 MH/s yes!
[2017-12-19 00:49:07] accepted: 376/382 (diff 0.077), 413.89 MH/s yes!
[2017-12-19 00:49:17] accepted: 377/383 (diff 0.169), 412.96 MH/s yes!
[2017-12-19 00:49:38] accepted: 377/384 (diff 0.102), 415.87 MH/s booooo
[2017-12-19 00:49:38] reject reason: Low difficulty share
[2017-12-19 00:49:48] accepted: 378/385 (diff 0.286), 414.90 MH/s yes!
[2017-12-19 00:49:58] accepted: 379/386 (diff 1.232), 414.68 MH/s yes!
[2017-12-19 00:49:58] accepted: 380/387 (diff 0.227), 414.68 MH/s yes!
[2017-12-19 00:49:58] accepted: 381/388 (diff 0.085), 414.68 MH/s yes!
[2017-12-19 00:49:58] accepted: 382/389 (diff 0.470), 414.68 MH/s yes!
[2017-12-19 00:49:58] accepted: 383/390 (diff 0.334), 414.68 MH/s yes!
[2017-12-19 00:49:58] accepted: 384/391 (diff 0.066), 414.68 MH/s yes!
[2017-12-19 00:49:58] accepted: 385/392 (diff 0.086), 414.68 MH/s yes!
[2017-12-19 00:49:58] accepted: 386/393 (diff 1.284), 414.68 MH/s yes!

In the miner I still see around 400Mh/s, on the dashboard it's now about 18Mh/s (and diff 16).
PS. Actually, I doubt 400Mh/s very much... If it were true, I'd get some crazy revenue, like 400x0.05=20 mBTC per day...
PS2. -f 2 gives a lot of rejected shares. Maybe it should be -m 2?

@auroracoin
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I use to get this with wildkeccak 2.2.2 version. And the pool owner worked on this and corrected it. But if I go into a different pool same algo the same thing will happen. Showing only 1/4 my hashrate. So is it the pool or the build or both?

Please sign in to comment.