3% faster mining with phoenix+phatk for everyone - page 6.

Soak

full member

Activity: 213

Merit: 100

WoW

310 to 380 Mhash (+18.42%) ! AWESOME! Thank you!

Code:

phoenix -u http://x:x@x:1234/;askrate=10 -k phatk VECTORS BFI_INT AGGRESSION=7 DEVICE=0
phoenix -u http://x:x@x:1234/;askrate=10 -k phatk VECTORS BFI_INT AGGRESSION=7 DEVICE=1

6970 @ Crossfire

Proof:

Before

After

theunforgiven

newbie

Activity: 13

Merit: 0

http://k.min.us/jdQi1Q.jpg

can't do it Sad

heroBitcoin

newbie

Activity: 12

Merit: 0

yeah, awesome information, thx guy

bitless

newbie

Activity: 28

Merit: 0

Quote from: PcChip on June 27, 2011, 12:12:40 PM

So m0mchill (or somebody?) already updated the code repository, and hasn't sent you a "thank you" PM yet ?

No, but it is me who should be thanking them for writing these awesome miners to begin with!

PcChip

sr. member

Activity: 418

Merit: 250

So m0mchill (or somebody?) already updated the code repository, and hasn't sent you a "thank you" PM yet ?

Hawkix

hero member

Activity: 531

Merit: 505

Great, went up promised 3%. So I sent you 3% of my daily income as a tip! Thanks!

Its great someone tries to improve on the OpenCL code. Maybe, if we could check the result of OpenCL compilation (VLIW assembler?), we could fine tune a bit more.

bitless

newbie

Activity: 28

Merit: 0

Quote from: theunforgiven on June 27, 2011, 11:46:56 AM

Gigabyte 5850 @ 1000/300 - from 390MHash/s to 400MHash/s with the mod applied.
Got a question, which boolean operator is used to sum all the outputs?
((y),(x|z),(z&x))

It is the CH function; given 3 arguments, a,b,c, CH returns B when A==0, and C when A==1. This is the same as (a&b)|(~a&c), which is the way BFI_INT is defined in ATI's docs.

BombaUcigasa

legendary

Activity: 1442

Merit: 1005

I have rewarded you with half of my increased bitcoin revenue margin, well done sir!

theunforgiven

newbie

Activity: 13

Merit: 0

Gigabyte 5850 @ 1000/300 - from 390MHash/s to 400MHash/s with the mod applied.

Got a question, which boolean operator is used to sum all the outputs?

((y),(x|z),(z&x))

"Y" , "X or Z", "Z and X"

xyz| (x|z) | (z&x) | y ? (x|z) ? (z&x) | Ma
000| 0 | 0 | ? | 0
001| 1 | 0 | ? | 0
010| 0 | 0 | ? | 0
011| 1 | 0 | ? | 1
------------------------------------------------
100| 1 | 0 | ¿ | 0
101| 1 | 1 | ¿ | 1
110| 1 | 0 | ¿ | 1
111| 1 | 1 | ¿ | 1

burp

member

Activity: 98

Merit: 10

10MH improvement, thanks! Shocked

Keninishna

hero member

Activity: 556

Merit: 500

woot works for me with guiminer. I edited the kernel.cl file in the phatk directory.

Kauwi

newbie

Activity: 2

Merit: 0

Quote from: bitless on June 27, 2011, 11:05:08 AM

Quote from: Kauwi on June 27, 2011, 10:37:01 AM

Quote from: bitless on June 27, 2011, 09:45:26 AM

I HIGHLY encourage one of us to go through the truth table in my original post and the C code that was used to generate it to make sure there's no errors. That's why I posted it to begin with. It would really suck if it didn't work.

You posted the truth-table (and modified the code) for the MA-function, but your C-code is for the CH-function.
But Your thruth-table for MA-function is correct.

Sorry about the confusion. CH = BFI_INT, since BFI does the same thing as the CH function in the hash spec. MA is built using CH/BFI, so there's a define for CH (copied from and docs) and the MA built on top of CH. Thank you for verifying this stuff.

Ah, i see. Didnt look close enough to the printf();
Sorry for that.

bitless

newbie

Activity: 28

Merit: 0

Quote from: Kauwi on June 27, 2011, 10:37:01 AM

Quote from: bitless on June 27, 2011, 09:45:26 AM

I HIGHLY encourage one of us to go through the truth table in my original post and the C code that was used to generate it to make sure there's no errors. That's why I posted it to begin with. It would really suck if it didn't work.

You posted the truth-table (and modified the code) for the MA-function, but your C-code is for the CH-function.
But Your thruth-table for MA-function is correct.

Sorry about the confusion. CH = BFI_INT, since BFI does the same thing as the CH function in the hash spec. MA is built using CH/BFI, so there's a define for CH (copied from and docs) and the MA built on top of CH. Thank you for verifying this stuff.

ananas5

newbie

Activity: 28

Merit: 0

Thanks! HD6850 270MHash/s -> 279Mhash/s

brunoshady

sr. member

Activity: 262

Merit: 250

Dubs Get

not too much increase here in 5850... about 4/5 mhash =(

Kauwi

newbie

Activity: 2

Merit: 0

Quote from: bitless on June 27, 2011, 09:45:26 AM

I HIGHLY encourage one of us to go through the truth table in my original post and the C code that was used to generate it to make sure there's no errors. That's why I posted it to begin with. It would really suck if it didn't work.

You posted the truth-table (and modified the code) for the MA-function, but your C-code is for the CH-function.
But Your thruth-table for MA-function is correct.

themike5000

member

Activity: 99

Merit: 10

Boosted me from 319MH/sec on my 6870 to 326MH/sec. Not bad.

I"m using GuiMiner v2011-06-09 because v2001-06-14 didn't have a BitcoinMiner.cl file. Anybody know where to find this file in the newest GUIminer release?

bitless

newbie

Activity: 28

Merit: 0

Heeey! This change is now in git repository! I'd still parenthesize (z^x) as ((z)^(x)), but regardless, thanks m0mchill and everybody!

https://github.com/m0mchil/poclbm/commit/3d43c4e4c2c476cf367822868202f02d168d6f3b

czz

newbie

Activity: 48

Merit: 0

Did not notice any increase in the reject ratio. Using 5870, came from 439 to 449 MH/s.
Many thanks.

bitless

newbie

Activity: 28

Merit: 0

Quote from: -ck on June 27, 2011, 05:18:18 AM

Thanks for your code. It's nice to see people get rewarded for their efforts (though it doesn't always happen). Well, I've tried it on my own miner (minerd) and have some very strange findings.

https://forum.bitcoin.org/index.php?topic=21275.0

First of all, minerd supports up to 4 vectors, and when I add this change to my kernel, it actually _slows down_ the 4 vector version. But when I override it to set 2 vectors, it speeds it up. However, once it's sped up, I then get runs of rejected shares. I tried it multiple times with and without and it does appear to be just this change that causes it, so I'm not sure what's going on.

Also, a very silly thing here -> I've posted #define Ma(x, y, z) amd_bytealign( (z^x), (y), (x) ), but I really should have put all expressions in parentheses. So, ((z)^(x)) and not (z^x) etc.... you know

Regardless. You got me worried about rejected shares. I HIGHLY encourage one of us to go through the truth table in my original post and the C code that was used to generate it to make sure there's no errors. That's why I posted it to begin with. It would really suck if it didn't work.

Topic: 3% faster mining with phoenix+phatk for everyone - page 6. (Read 36831 times)