What speed are your getting CPU mining TENEBRIX? - page 4.

tacotime

legendary

Activity: 1484

Merit: 1005

Quote from: ArtForz on October 06, 2011, 04:28:33 PM

Not sure what's going on there (aka "can't reproduce issue").
My only guess is it's possibly related to removing the sha256 algos, but... that was even before I started doing the compilers job for scrypt.
Not sure what to do other than general hints along the lines of "start with a clean tree, CFLAGS="-whatever" ./configure; make"
Hrrrm... I guess you could revert the sha256 removal or drop the new scrypt.c into Tenebrix-miner and see if that also causes the same issues.

There is no scrypt.c in your latest released source of 1.0.2... I'm looking at the tar.gz and it's missing it.

edit: It works just popping scrypt.c in, I'm getting 3.87kh/s now per core

tacotime

legendary

Activity: 1484

Merit: 1005

Quote from: grod on October 06, 2011, 09:15:36 PM

Up to 2.83 Khash/thread on an i7 [email protected] ghz. Trying different compilers and optimizations, no code changes. llvm is by far the worst, got a high of 2.3 Khash/thread with that compiler. Yuck.

That's pretty wild, faster than the 1055T if that's with hyperthreading on. The Intel people in this thread are going to want to know what your configure options and compiler was for that.

grod

full member

Activity: 154

Merit: 100

Up to 2.83 Khash/thread on an i7 [email protected] ghz. Trying different compilers and optimizations, no code changes. llvm is by far the worst, got a high of 2.3 Khash/thread with that compiler. Yuck.

TiagoTiago

hero member

Activity: 616

Merit: 500

Firstbits.com/1fg4i :)

1. I'm on a pool.

2. I guess, but i haven't looked into it yet

3.i don't think there is any GUI based miner, but the text based one that comes with it is relativelly friendly, just answer a few questions and it runs.

Beta-coiner1

hero member

Activity: 532

Merit: 500

I definitely have some questions concerning this considering I have rarely done CPU minin

1.Are you guys using mainly a pool or solo-mining ?

2.Are there any details towards optimizing further for Intel CPUs ?

3.What GUI based programs work with the Tenebrix fork ?

Getting 1.72 Kh/s per thread @ 3.4 Ghz i5

ArtForz

sr. member

Activity: 406

Merit: 257

Quote from: tacotime on October 06, 2011, 04:08:03 PM

Quote from: ArtForz on October 06, 2011, 03:45:09 PM

Just pushed another small tweak, gets another 3% or so on K10s.

Hey,

For some reason after compiling the new code, the program no longer takes command line arguments? Not sure what's happening. It just returns the -h line no matter what I input.

Also, I've been running the last version of your code at ~38kh/s and haven't gotten any blocks in about 6 hours. But maybe I'm just unlucky.

Not sure what's going on there (aka "can't reproduce issue").
My only guess is it's possibly related to removing the sha256 algos, but... that was even before I started doing the compilers job for scrypt.
Not sure what to do other than general hints along the lines of "start with a clean tree, CFLAGS="-whatever" ./configure; make"
Hrrrm... I guess you could revert the sha256 removal or drop the new scrypt.c into Tenebrix-miner and see if that also causes the same issues.

tacotime

legendary

Activity: 1484

Merit: 1005

Quote from: ArtForz on October 06, 2011, 03:45:09 PM

Just pushed another small tweak, gets another 3% or so on K10s.

Hey,

For some reason after compiling the new code, the program no longer takes command line arguments? Not sure what's happening. It just returns the -h line no matter what I input.

Also, I've been running the last version of your code at ~38kh/s and haven't gotten any blocks in about 6 hours. But maybe I'm just unlucky.

ArtForz

sr. member

Activity: 406

Merit: 257

Just pushed another small tweak, gets another 3% or so on K10s.

r3v3rs3

newbie

Activity: 22

Merit: 0

Atom 330 @ 2.16 GHz:
- 4 threads, 0.66 kH/s each
- 2 threads, 0.91 kH/s each

Phenom II @ 3.6 GHz:
- 4 threads, 3.62 kH/s each

Core 2 Duo (65 nm) @ 1.5 GHz:
- 2 threads, 1.35 kH/s each

Debian/sid x86_64, latest cpuminer, CFLAGS="-O3 -march=whatever_cpu"

tacotime

legendary

Activity: 1484

Merit: 1005

Using gcc4.6.1

Code:

CFLAGS="-march=amdfam10 -O3 -Wall -msse2" ./configure

Getting 3.8 kh/s per core for K10.5 @ 3.7GHz

Quote

Guys what about AVX and AES-NI instructions in Sandy 2600K !? How far can I push my 2600K !?

AVS won't make the cache size bigger, it's intended to facilitate floating point instructions more than anything... I think scrypt is mostly integer-based.

bulanula

hero member

Activity: 518

Merit: 500

Guys what about AVX and AES-NI instructions in Sandy 2600K !? How far can I push my 2600K !?

Moray

full member

Activity: 143

Merit: 100

1.38/1.4 to core (E2160@3000) Cheesy

TiagoTiago

hero member

Activity: 616

Merit: 500

Firstbits.com/1fg4i :)

i'm getting around between 0.9 and 1.2 per thread, x 8 cores (quadcore with hyperthreading)

Bobnova

full member

Activity: 210

Merit: 100

Aha, make clean, that's one thing I was missing.

Any flags I should pass on for intel?

Looks like 4 threads is faster than 6, I suspect the 256kb/core of L2 isn't enough to keep two threads worth hanging around and data gets exiled to L3 land.
PhII CPUs of course have 512k/core, and don't try to do the hyperthreading routine.
BD and it's 1mb/core should be nice.

EDIT:
So far looks like -O3 -Wall -msse2 is best.

ArtForz

sr. member

Activity: 406

Merit: 257

Quote from: Bobnova on October 05, 2011, 09:13:13 PM

I tried out the earlier one, it added 100hash/s to each 2600k thread on average, maybe 150.
I'll try out the new version now.
This makes me wish I had a CPU with real L2 cache like the PhIIs. I do love seeing a PhII CPU actually perform better at something though, it's refreshing.

As a sidenote: Ubuntu 11.04 doesn't seem to locate the libcurl.so file correctly, or at least doesn't set the @LIBCURL@ variable correctly. Editing the ./config file to remove the libcurl check and editing the makefile it generates to have the specific path works though.

EDIT:
How do you pass the -O3 and such flags to make? It doesn't like me.

CFLAGS="-whatever -somethingelse" ./configure
make clean; make

or just edit the CFLAGS= line in makefile after configuring

I like to always make clean so there's no objs from previous compilations with different flags hanging around Wink

Bobnova

full member

Activity: 210

Merit: 100

I tried out the earlier one, it added 100hash/s to each 2600k thread on average, maybe 150.
I'll try out the new version now.
This makes me wish I had a CPU with real L2 cache like the PhIIs. I do love seeing a PhII CPU actually perform better at something though, it's refreshing.

As a sidenote: Ubuntu 11.04 doesn't seem to locate the libcurl.so file correctly, or at least doesn't set the @LIBCURL@ variable correctly. Editing the ./config file to remove the libcurl check and editing the makefile it generates to have the specific path works though.

EDIT:
How do you pass the -O3 and such flags to make? It doesn't like me.

2.EDIT:
Think I found it. [email protected] now running 2xPhoenix+pandora+desktopstuff+6 threads averaging 2kh/s/thread.
Nice work Art Cheesy

tacotime

legendary

Activity: 1484

Merit: 1005

Quote from: ArtForz on October 05, 2011, 08:39:24 PM

Just pushed some more scrypt manglery, 3.62kH/s/core with -march=amdfam10 -O3

did you upload the source code already?

edit: okay, yeah I see you have... I will rebuild tomorrow and give it a go

ArtForz

sr. member

Activity: 406

Merit: 257

Just pushed some more scrypt manglery, 3.62kH/s/core with -march=amdfam10 -O3

worldinacoin

hero member

Activity: 756

Merit: 500

Can I test it on Windows? If possible can I have the link too?

Quote from: Lolcust on October 05, 2011, 03:08:15 PM

Quote from: bulanula on October 05, 2011, 03:04:30 PM

Quote from: tacotime on October 05, 2011, 02:53:57 PM

So, I remade with gcc-4.6.1 and there is no actual performance increase... not sure what is going on exactly.

edit: never mind, had to recompile with CFLAGS="-march=amdfam10 -O3" ./configure
and then with gcc-4.6.1

Now I'm at 3.40kh/s!! thanks!!

Man you guys are just murdering my stupid 2600K crappy Intel silicon. I want some of the love please !

Poster a few posts back : how did you get 1.66 khash/s per thread on the 2600K !? I only get around 1.15 khash/s per thread on my crappy chip.

Bulanula, are you on win or on Linux ? I'm about to pull an experimental Tenebrix Minerd optimization by ArtForz that is tailored to 64 bit linuxes (completely untested on win)
Can squeeze out up to 30% more.

tacotime

legendary

Activity: 1484

Merit: 1005

Quote from: Lolcust on October 05, 2011, 03:22:24 PM

Quote from: tacotime on October 05, 2011, 03:20:10 PM

I just got like 10 proofs of work (including one for 10 trillion TBX) trying to port this directly to another linux box... beeeee careful when you make the binaries lolcust.

Um, you mean you just copypasted the binaries built on one linux box to another lin and it misbehaved ? Or is something more sinister afoot ?

Also, AFAIK the max coinage a chain can deal with is around 130 billions....

I copypasted the binaries from a machine with gcc4.6.1 to a machine with gcc4.5.4... No idea what the wacky results were about, but they seem normal now that I recompiled it. But, yeah, the high performance binary seems to require the gcc4.6.1 libraries. you should get roughly 200% performance though if you compile them correctly.

Now I'm getting 34 kh/s on 10 AMD K10.5 cores.

I would post binaries but like I said, they don't translate well across variable gcc versions... You are best off compiling your own with gcc4.6.1.

Also, by the rate the hash rate seems to be increasing from block speeds I think we're into the hundreds of kh/s. I don't think there'll be a chance to 51% this if the adoption keep rising exponentially like it has been for the last week.

Topic: What speed are your getting CPU mining TENEBRIX? - page 4. (Read 13645 times)