3% faster mining with phoenix+phatk for everyone

[bitless]

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.

[Hawkix]

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.

[PcChip]

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

[bitless]

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!

[heroBitcoin]

yeah, awesome information, thx guy

[theunforgiven]

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

can't do it

[Soak]

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

[Fiyasko]

FYI, You can make the identical change to phoenix+poclbm with the same 2% increase in speed.

Do tell, If theres a simple file to replace that you can upload i'll definitly Tip you
Winx86 GUIminer-poclbm

[Rob P.]

Just tossing my confirmations out there:

Sapphire 5830s:

Pair 1:  264 Mhash/sec -> 271 Mhash/sec (each) at 875 MHz GPU, 900 MHz RAM (2.58% increase)
Pair 2:  273 Mhash/sec -> 281 Mhash/sec (each) at 900 MHz GPU, 600 MHz RAM (2.84% increase)

I'm also tracking Rejected shares.

Before Patch:
Pair 1:  ~2.6% rejects (after 800+ shares)
Pair 2:  ~3.2% rejects (after 800+ shares)

After Patch:
Pair 1:  ~3.0% rejects (after 650+ shares)
Pair 2:  ~4.0% (after 750+ shares)

So, the increase appears to also cause additional rejects that should be discounted from total increase gain, cutting it down a full 1% in my case.

[bitless]

FYI, You can make the identical change to phoenix+poclbm with the same 2% increase in speed.

Do tell, If theres a simple file to replace that you can upload i'll definitly Tip you
Winx86 GUIminer-poclbm

at the office now, no access to the miners, I'll take a look once I get home. In the meantime, search for .cl files and see if you have any Ma (or, even better, amd_bytealign) strings in them and take it from there.

[bitless]

Just tossing my confirmations out there:
...
So, the increase appears to also cause additional rejects that should be discounted from total increase gain, cutting it down a full 1% in my case.

This is really not good. I honestly do not know why this would happen; are you sure it is related to the patch and isn't a result in the general randomness when searching for solutions? Perhaps someone else, more familiar with bitcoin mining than me, can chime in?

[bitless]

can't do it

Yes, you can what you've done is the CH function, not MA. Now you can build MA on top of it. I don't think this belongs to this thread though, so you can pm me and we'll figure it out. Alternatively, you can ask someone around you for help - two heads are better than one.

[bitless]

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

Neat-o-rama! Somehow I still don't believe you (and my eyes)

[PcChip]

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

Neat-o-rama! Somehow I still don't believe you (and my eyes)

I noticed he's using Crossfire on those cards, perhaps the kernel patch somehow makes up for % loss due to using crossfire?

[Soak]

Neat-o-rama! Somehow I still don't believe you (and my eyes)

I do 370 Mhashes with poclbm and the new formula (365 without).
I do 380 Mhashes with phatk and the new formula (310 without).

So, really, +2.63% between poclbm and phatk with the new formula.

[TurdHurdur]

Just tossing my confirmations out there:
...
So, the increase appears to also cause additional rejects that should be discounted from total increase gain, cutting it down a full 1% in my case.

This is really not good. I honestly do not know why this would happen; are you sure it is related to the patch and isn't a result in the general randomness when searching for solutions? Perhaps someone else, more familiar with bitcoin mining than me, can chime in?

I'm not seeing this issue with my miners. They range from 0.12% to 2.07% stale. A variance that I'd say rules out your patch being the cause of any supposed increased stales.

[theunforgiven]

I want to minimize the expression applying Karnaugh. So I need the main expression that is used on the kernel. But i can't figure out with it.

[Hawkix]

Seriously, the Ma() function is soo deeply burried, so that if it was wrong, by the principle of good hashing function like SHA256 is, all the hashes would fail.

So, if it can find some good results, all are, with 99.999999999999% probability correct, too.

[bitless]

Seriously, the Ma() function is soo deeply burried, so that if it was wrong, by the principle of good hashing function like SHA256 is, all the hashes would fail.

So, if it can find some good results, all are, with 99.999999999999% probability correct, too.

More like 2^-128 if not less

[Clipse]

Nice, avg about 2% mhash boost at same clocks. I notice no additional stales, still hovering at 1%