Search content
Sort by
Showing 3 of 3 results by Krypta
You may be right. May I suggest you run the kernel analyzer and see the disasm of both versions and see if the generated code is any different?
(This is because things like x=1; x=2; x=3; are easily converted by the compiler to just x=3; )
You are right, it's the same (This is because things like x=1; x=2; x=3; are easily converted by the compiler to just x=3; )
No improvement then...
K.
Thank you bitless, it seems you are right, they are never used, so I really don't know why these modifications halved my stales.
Here's another strange thing:
I think we can comment out
because they are changed on lines 265 and 266 before they are used, am I right?
Thanks.
Krypta
Here's another strange thing:
Code:
256 Vals[0]=H[0];
257 Vals[1]=H[1];
258 Vals[2]=H[2];
259 Vals[3]=H[3];
260 Vals[4]=H[4];
261 Vals[5]=H[5];
262 Vals[6]=H[6];
263 Vals[7]=H[7];
264
265 Vals[7] = 0xb0edbdd0 + K[0] + W[64] + 0x08909ae5U;
266 Vals[3] = 0xa54ff53a + 0xb0edbdd0 + K[0] + W[64];
257 Vals[1]=H[1];
258 Vals[2]=H[2];
259 Vals[3]=H[3];
260 Vals[4]=H[4];
261 Vals[5]=H[5];
262 Vals[6]=H[6];
263 Vals[7]=H[7];
264
265 Vals[7] = 0xb0edbdd0 + K[0] + W[64] + 0x08909ae5U;
266 Vals[3] = 0xa54ff53a + 0xb0edbdd0 + K[0] + W[64];
I think we can comment out
Code:
259 //Vals[3]=H[3];
263 //Vals[7]=H[7];
263 //Vals[7]=H[7];
because they are changed on lines 265 and 266 before they are used, am I right?
Thanks.
Krypta
Hy all, I'm new here and pretty new with this bitcoin thing.
I tried to modify phatk kernel as explained and I was able to speed up my mining as all of you.
Really don't know when and how this amd_bytealign function is called, anyway I think it should be better to write it as follows
#define Ma(x, y, z) amd_bytealign(((z) ^ (x)), (y), (x))
That said, I took a look at the rest of the phatk kernel and found what seems to be a sintax error. In my file it's on line 74:
#define r1(x) ((rot(x],15)^rot(x,13)^((x)>>10U)))
Mmmm... that poor and lonely square bracket... rewrite it as follows
#define r1(x) ((rot(x,15)^rot(x,13)^((x)>>10U)))
These modifications did not speed up my mining power but it seems that my rejecteds almost halved.
If this helps you as well, donations are obviously welcome
1CbRmFAYrGYhzQibmE5A1CWMp7srFT4upy
Krypta
I tried to modify phatk kernel as explained and I was able to speed up my mining as all of you.
Really don't know when and how this amd_bytealign function is called, anyway I think it should be better to write it as follows
#define Ma(x, y, z) amd_bytealign(((z) ^ (x)), (y), (x))
That said, I took a look at the rest of the phatk kernel and found what seems to be a sintax error. In my file it's on line 74:
#define r1(x) ((rot(x],15)^rot(x,13)^((x)>>10U)))
Mmmm... that poor and lonely square bracket... rewrite it as follows
#define r1(x) ((rot(x,15)^rot(x,13)^((x)>>10U)))
These modifications did not speed up my mining power but it seems that my rejecteds almost halved.
If this helps you as well, donations are obviously welcome
1CbRmFAYrGYhzQibmE5A1CWMp7srFT4upy
Krypta