tbearhere
Legendary
Offline
Activity: 3206
Merit: 1003
|
|
November 15, 2014, 01:00:15 PM |
|
still very new at compiling, where can i get curl to 7.38.0 windows 8.1 from a safe site please.
from the site of the author thanks djm34 going to fetch it
|
|
|
|
tbearhere
Legendary
Offline
Activity: 3206
Merit: 1003
|
|
November 15, 2014, 01:13:46 PM Last edit: November 15, 2014, 01:30:28 PM by tbearhere |
|
still very new at compiling, where can i get curl to 7.38.0 windows 8.1 from a safe site please.
from the site of the author thanks djm34 going to fetch it i cant find it
|
|
|
|
djm34
Legendary
Offline
Activity: 1400
Merit: 1050
|
|
November 15, 2014, 01:33:44 PM |
|
still very new at compiling, where can i get curl to 7.38.0 windows 8.1 from a safe site please.
from the site of the author thanks djm34 going to fetch it i cant find it google libcurl (no it isn't on microsoft page) it is open source
|
djm34 facebook pageBTC: 1NENYmxwZGHsKFmyjTc5WferTn5VTFb7Ze Pledge for neoscrypt ccminer to that address: 16UoC4DmTz2pvhFvcfTQrzkPTrXkWijzXw
|
|
|
Epsylon3
Legendary
Offline
Activity: 1484
Merit: 1082
ccminer/cpuminer developer
|
|
November 15, 2014, 02:39:38 PM |
|
I should probobly merge the latest changes in the main branch. (1.4.9) but I'm too lazy. My focus is on the kernals, and not the rest.
The work on the echo is not done. There is more to remove.
I almost didnt changed the .cu files recently, you can maybe refork my project
|
|
|
|
tbearhere
Legendary
Offline
Activity: 3206
Merit: 1003
|
|
November 15, 2014, 04:15:30 PM |
|
still very new at compiling, where can i get curl to 7.38.0 windows 8.1 from a safe site please.
from the site of the author thanks djm34 going to fetch it i cant find it google libcurl (no it isn't on microsoft page) it is open source i got it but it needs to be compiled im use to doing it bigjme way. so i cant do it. 6 hrs to try 1.4.9 no luck
|
|
|
|
djm34
Legendary
Offline
Activity: 1400
Merit: 1050
|
|
November 15, 2014, 04:28:04 PM |
|
still very new at compiling, where can i get curl to 7.38.0 windows 8.1 from a safe site please.
from the site of the author thanks djm34 going to fetch it i cant find it google libcurl (no it isn't on microsoft page) it is open source i got it but it needs to be compiled im use to doing it bigjme way. so i cant do it. 6 hrs to try 1.4.9 no luck hu ? still haven't compile it ? You know that you haven't anything to do (as epsilon told you... since libcurl has been put into the compat)
|
djm34 facebook pageBTC: 1NENYmxwZGHsKFmyjTc5WferTn5VTFb7Ze Pledge for neoscrypt ccminer to that address: 16UoC4DmTz2pvhFvcfTQrzkPTrXkWijzXw
|
|
|
sp_ (OP)
Legendary
Offline
Activity: 2954
Merit: 1087
Team Black developer
|
|
November 15, 2014, 04:37:26 PM |
|
You need to install the latest cuda 6.5 and visual studio 2013.
|
|
|
|
tbearhere
Legendary
Offline
Activity: 3206
Merit: 1003
|
|
November 15, 2014, 05:09:37 PM |
|
You need to install the latest cuda 6.5 and visual studio 2013.
i got it thank you djm34 and sp and all yes curl was build into it.. it was open zip to folder then sln..opens vs 2013 automatically then 64x release and done...3 minutes
|
|
|
|
sp_ (OP)
Legendary
Offline
Activity: 2954
Merit: 1087
Team Black developer
|
|
November 15, 2014, 05:15:43 PM |
|
On windows, ccminer runs faster when compiled for x86
|
|
|
|
tbearhere
Legendary
Offline
Activity: 3206
Merit: 1003
|
|
November 15, 2014, 05:20:01 PM Last edit: November 15, 2014, 05:45:32 PM by tbearhere |
|
On windows, ccminer runs faster when compiled for x86
i get less hash on x11 750ti and no improvement on quark. EDIT: I should say no improvement on x11
|
|
|
|
djm34
Legendary
Offline
Activity: 1400
Merit: 1050
|
|
November 15, 2014, 05:31:06 PM |
|
On windows, ccminer runs faster when compiled for x86
i get less hash on x11 750ti and no improvement on quark. compiled with x64 or x86 ?
|
djm34 facebook pageBTC: 1NENYmxwZGHsKFmyjTc5WferTn5VTFb7Ze Pledge for neoscrypt ccminer to that address: 16UoC4DmTz2pvhFvcfTQrzkPTrXkWijzXw
|
|
|
tbearhere
Legendary
Offline
Activity: 3206
Merit: 1003
|
|
November 15, 2014, 05:41:08 PM |
|
On windows, ccminer runs faster when compiled for x86
i get less hash on x11 750ti and no improvement on quark. compiled with x64 or x86 ? x86 EDIT: I should say no improvement on x11
|
|
|
|
sp_ (OP)
Legendary
Offline
Activity: 2954
Merit: 1087
Team Black developer
|
|
November 15, 2014, 05:57:29 PM |
|
the problem is that the throughput is set to be fast on the 980 in the source version. download the 1.4.9 source and replace the file: cuda_x11_echo.cu from my fork you should get a small boost in x11. On quark I have optimized bmw and blake a tinybit 1.4.9 with the intesity parameter is found here: https://github.com/tpruvot/ccminer
|
|
|
|
tbearhere
Legendary
Offline
Activity: 3206
Merit: 1003
|
|
November 15, 2014, 06:13:47 PM |
|
sp your older ccminer and im really pushing it. looking forward to your new one. the fastest hashing is 1.4.6
|
|
|
|
Epsylon3
Legendary
Offline
Activity: 1484
Merit: 1082
ccminer/cpuminer developer
|
|
November 16, 2014, 01:41:34 AM Last edit: November 16, 2014, 03:11:03 AM by Epsylon3 |
|
the problem is that the throughput is set to be fast on the 980 in the source version. download the 1.4.9 source and replace the file: cuda_x11_echo.cu from my fork you should get a small boost in x11. On quark I have optimized bmw and blake a tinybit 1.4.9 with the intesity parameter is found here: https://github.com/tpruvot/ccminerIndeed, +15kH on the 750 Ti (2ms improvement on your repo, its the biggest optimisation you have made on a single algo, was 0.5 before, i will pick it for the 1.5.0) on mine, i get +9 KH (2791 vs 2800KH in benchmark mode) but i didnt take the launch bounds change for the moment... 39.171ms before, 38.522ms before = 0.65ms on mine, enough for me (but not fully comparable) EDIT: but on windows :// seems to be lowered, investigating...
|
|
|
|
Schleicher
|
|
November 16, 2014, 06:36:00 AM |
|
Possible small optimization at the end of cuda_echo_round: for (int i = 0; i<15; i += 4) { W[i] ^= W[32 + i] ^ 512; W[i + 1] ^= W[32 + i + 1]; W[i + 2] ^= W[32 + i + 2]; W[i + 3] ^= W[32 + i + 3]; } W[15] ^= W[47] ^ 512;
(we don't need more than 16)
|
|
|
|
sp_ (OP)
Legendary
Offline
Activity: 2954
Merit: 1087
Team Black developer
|
|
November 16, 2014, 08:49:15 AM |
|
Indeed, +15kH on the 750 Ti (2ms improvement on your repo, its the biggest optimisation you have made on a single algo, was 0.5 before, i will pick it for the 1.5.0) on mine, i get +9 KH (2791 vs 2800KH in benchmark mode) but i didnt take the launch bounds change for the moment... 39.171ms before, 38.522ms before = 0.65ms on mine, enough for me (but not fully comparable) EDIT: but on windows :// seems to be lowered, investigating...
In addition to the launchbound change, did you remember to go from 256 to 320 threads when calling the kernal?. The launchbound will force the compiler to use 64 registers. We get more spills to memory, but it seems to run faster.
|
|
|
|
sp_ (OP)
Legendary
Offline
Activity: 2954
Merit: 1087
Team Black developer
|
|
November 16, 2014, 08:52:08 AM |
|
Possible small optimization at the end of cuda_echo_round: for (int i = 0; i<15; i += 4) { W[i] ^= W[32 + i] ^ 512; W[i + 1] ^= W[32 + i + 1]; W[i + 2] ^= W[32 + i + 2]; W[i + 3] ^= W[32 + i + 3]; } W[15] ^= W[47] ^ 512;
(we don't need more than 16) Thanks, it works. for (int i = 0; i<15; i += 4) { W ^= W[32 + i] ^ 512; W[i + 1] ^= W[32 + i + 1]; W[i + 2] ^= W[32 + i + 2]; W[i + 3] ^= W[32 + i + 3]; }
is enough.
|
|
|
|
sp_ (OP)
Legendary
Offline
Activity: 2954
Merit: 1087
Team Black developer
|
|
November 16, 2014, 11:47:07 AM Last edit: November 16, 2014, 01:15:15 PM by sp_ |
|
I have checked in some more performance improvements. I moved the precalc table in echo from constmem to the instruction cache. Improved registers/launchbounds on shavite. The 980 is now around 400KHASH faster than the release 6. on stock clocks.(x11). Here is the link: http://www.filedropper.com/release7The sourcecode is available here: https://github.com/sp-hash/ccminer
|
|
|
|
jpouza
Legendary
Offline
Activity: 2842
Merit: 1122
|
|
November 16, 2014, 12:14:46 PM |
|
I have checked in some more performance improvements. I moved the precalc table in echo from constmem to the instruction cache. Improved registers/launchbounds on shavite. The 980 is now around 400KHASH faster than the release 6. on stock clocks.(x11). Here is the link: http://www.filedropper.com/release7Nice, 9MH/s with 185+ on 980 GPUs. 10MH/s with extreme overclock GPU at 300+ and overvolted. 750Ti boost to 2.9MH/s with 135+ GPU 460+ MEM. Cheers
|
|
|
|
|