[ANN] [QRK] Quark | Core 0.10 upgrade - page 399.

zvs

legendary

Activity: 1680

Merit: 1000

https://web.archive.org/web/*/nogleg.com

Quote from: eule on August 27, 2013, 05:28:07 PM

I could use some testers:
http://qrkpool.tk:8868/static/ (just registered so use the one below if this one does not work)
http://yacpool.tk:8868/static/
1% fee Tongue

I wonder if I should put Neisklar's node's IP to "BOOTSTRAP_ADDRS" in the networks.py? Great work Neisklar btw, I'll tip you tomorrow. Cheesy

edit to add: Interestingly the miner reports a much higher hashrate than the p2pool, not sure if that's a bug or if that happens on the other node too.

I tried to set this up but it kept getting errors about not locating quark_hasher or some such..

The p2pool website appeared, but it never loaded in any blocks? it connected to the rpc port fine

Code:

2013-08-27 21:03:32.398008 Initializing work...
2013-08-27 21:03:32.437453 RECV headers 0102000000da0a32ae29942b51b42888058e063ef820aaee5254d7b805285a9c6300000000d0d09b0e366e403a122db8adc2fe91e7b1554d757a79c5d57f08f0e59af59b0a535a1d52b8cc001d20069b0500
2013-08-27 21:03:32.470203 > Error handling message: (see RECV line)
2013-08-27 21:03:32.470249 > Traceback (most recent call last):
2013-08-27 21:03:32.470276 > File "/usr/lib/python2.7/dist-packages/twisted/internet/posixbase.py", line 586, in _doReadOrWrite
2013-08-27 21:03:32.470311 > why = selectable.doRead()
2013-08-27 21:03:32.470354 > File "/usr/lib/python2.7/dist-packages/twisted/internet/tcp.py", line 199, in doRead
2013-08-27 21:03:32.470388 > rval = self.protocol.dataReceived(data)
2013-08-27 21:03:32.470436 > File "/home/zevus/p2pool-quarkcoin/p2pool/util/p2protocol.py", line 27, in dataReceived
2013-08-27 21:03:32.470483 > self.dataReceived2(data)
2013-08-27 21:03:32.470527 > File "/home/zevus/p2pool-quarkcoin/p2pool/util/datachunker.py", line 40, in _DataChunker
2013-08-27 21:03:32.470558 > wants = receiver.send(buf.get(wants))
2013-08-27 21:03:32.470581 > --- ---
2013-08-27 21:03:32.470609 > File "/home/zevus/p2pool-quarkcoin/p2pool/util/p2protocol.py", line 57, in dataReceiver
2013-08-27 21:03:32.470632 > self.packetReceived(command, type_.unpack(payload, self.ignore_trailing_payload))
2013-08-27 21:03:32.470652 > File "/home/zevus/p2pool-quarkcoin/p2pool/util/p2protocol.py", line 71, in packetReceived
2013-08-27 21:03:32.470673 > handler(**payload2)
2013-08-27 21:03:32.470695 > File "/home/zevus/p2pool-quarkcoin/p2pool/bitcoin/p2p.py", line 130, in handle_headers
2013-08-27 21:03:32.470731 > header_hash = self.net.BLOCKHASH_FUNC(bitcoin_data.block_header_type.pack(header))
2013-08-27 21:03:32.470751 > File "/home/zevus/p2pool-quarkcoin/p2pool/bitcoin/networks.py", line 25, in
2013-08-27 21:03:32.470771 > BLOCKHASH_FUNC=lambda data: pack.IntType(256).unpack(__import__('quark_hash').getPoWHash(data)),
2013-08-27 21:03:32.470791 > exceptions.ImportError: No module named quark_hash

shazbits

member

Activity: 105

Merit: 10

Can the people who built it for Windows 64-bit, explain a bit how they've done it?
Have you used mingw? mingw-w64?
Did you have to update the default packages?
Did you install every needed package by hand (configure, make, make install)?
Any other details?

Thanks.

Edit: just found this https://bitcointalksearch.org/topic/m.2597100

BitJohn

hero member

Activity: 826

Merit: 1001

@Bit_John

Just wanted to thank our Quark traders with their own Cryptsy.com avatar. We will be giving coins away and merchandise randomly to folks who sport a cryptsy avatar on any coin forum that we spot.

trinsic

full member

Activity: 205

Merit: 100

Quote from: eule on August 27, 2013, 06:35:17 AM

Did you recieve many small p2pool transactions?

Yeah a weeks worth of p2pool transactions. I have since used settxfee and solo mining to send large amounts now.

eule

hero member

Activity: 756

Merit: 501

Thanks, a fresh git clone helped indeed! Cheesy

Does that mean people using an old version of the miner get lower payouts from the p2pool?
edit: guess not, the shares should determine payout.

K1773R

legendary

Activity: 1792

Merit: 1008

/dev/null

Quote from: eule on August 27, 2013, 05:28:07 PM

I could use some testers:
http://qrkpool.tk:8868/static/ (just registered so use the one below if this one does not work)
http://yacpool.tk:8868/static/

I wonder if I should put Neisklar's node's IP to "BOOTSTRAP_ADDRS" in the networks.py? Great work Neisklar btw, I'll tip you tomorrow. Cheesy

edit to add: Interestingly the miner reports a much higher hashrate than the p2pool, not sure if that's a bug or if that happens on the other node too.

pls patch your miners with this

Code:

commit b1af442712ee82fe9764df3812d134a99e11e3f2
Author: Neisklar
Date: Mon Aug 19 19:07:28 2013 +0200

* Move invariants out of loop (minor performance gain)

* Fix issue that miner could only work with diff 0.0039... or higher (not that that matters much now:)
* quick'n'dirty improve of hash pre-check

diff --git a/quark.c b/quark.c
index 0d2725d..01d8044 100644
--- a/quark.c
+++ b/quark.c
@@ -129,21 +164,88 @@ int scanhash_quark(int thr_id, uint32_t *pdata, const uint32_t *ptarget,
// applog(LOG_DEBUG, "Thr: %02d, firstN: %08x, maxN: %08x, ToDo: %d", thr_id, first_nonce, max_nonce, max_nonce-first_nonce);
// }

+ /* I'm to lazy to put the loop in an inline function... so dirty copy'n'paste.... */
+ /* i know that i could set a variable, but i don't know how the compiler will optimize it, not that then the cpu needs to load the value *everytime* in a register */
+ if (ptarget[7]==0) {
+ do {
+ pdata[19] = ++n;
+ be32enc(&endiandata[19], n);
+ quarkhash(hash64, &endiandata);
+ if (((hash64[7]&0xFFFFFFFF)==0) &&
+ fulltest(hash64, ptarget)) {
+ *hashes_done = n - first_nonce + 1;
+ return true;
+ }
+ } while (n < max_nonce && !work_restart[thr_id].restart);
+ }
+ else if (ptarget[7]<=0xF)
+ {
+ do {
+ pdata[19] = ++n;
+ be32enc(&endiandata[19], n);
+ quarkhash(hash64, &endiandata);
+ if (((hash64[7]&0xFFFFFFF0)==0) &&
+ fulltest(hash64, ptarget)) {
+ *hashes_done = n - first_nonce + 1;
+ return true;
+ }
+ } while (n < max_nonce && !work_restart[thr_id].restart);
+ }
+ else if (ptarget[7]<=0xFF)
+ {
+ do {
+ pdata[19] = ++n;
+ be32enc(&endiandata[19], n);
+ quarkhash(hash64, &endiandata);
+ if (((hash64[7]&0xFFFFFF00)==0) &&
+ fulltest(hash64, ptarget)) {
+ *hashes_done = n - first_nonce + 1;
+ return true;
+ }
+ } while (n < max_nonce && !work_restart[thr_id].restart);
+ }
+ else if (ptarget[7]<=0xFFF)
+ {
+ do {
+ pdata[19] = ++n;
+ be32enc(&endiandata[19], n);
+ quarkhash(hash64, &endiandata);
+ if (((hash64[7]&0xFFFFF000)==0) &&
+ fulltest(hash64, ptarget)) {
+ *hashes_done = n - first_nonce + 1;
+ return true;
+ }
+ } while (n < max_nonce && !work_restart[thr_id].restart);
+
+ }
+ else if (ptarget[7]<=0xFFFF)
+ {
+ do {
+ pdata[19] = ++n;
+ be32enc(&endiandata[19], n);
+ quarkhash(hash64, &endiandata);
+ if (((hash64[7]&0xFFFF0000)==0) &&
+ fulltest(hash64, ptarget)) {
+ *hashes_done = n - first_nonce + 1;
+ return true;
+ }
+ } while (n < max_nonce && !work_restart[thr_id].restart);
+
+ }
+ else
+ {
+ do {
+ pdata[19] = ++n;
+ be32enc(&endiandata[19], n);
+ quarkhash(hash64, &endiandata);
+ if (fulltest(hash64, ptarget)) {
+ *hashes_done = n - first_nonce + 1;
+ return true;
+ }
+ } while (n < max_nonce && !work_restart[thr_id].restart);
+ }

-
- do {
-
- pdata[19] = ++n;
- be32enc(&endiandata[19], n);
- quarkhash(hash64, &endiandata);
- if (((hash64[7]&0xFFFFFF00)==0) &&
- fulltest(hash64, ptarget)) {
- *hashes_done = n - first_nonce + 1;
- return true;
- }
- } while (n < max_nonce && !work_restart[thr_id].restart);
-
*hashes_done = n - first_nonce + 1;
pdata[19] = n;
return 0;

eule

hero member

Activity: 756

Merit: 501

I could use some testers:
http://qrkpool.tk:8868/static/ (just registered so use the one below if this one does not work)
http://yacpool.tk:8868/static/
1% fee Tongue

I wonder if I should put Neisklar's node's IP to "BOOTSTRAP_ADDRS" in the networks.py? Great work Neisklar btw, I'll tip you tomorrow. Cheesy

edit to add: Interestingly the miner reports a much higher hashrate than the p2pool, not sure if that's a bug or if that happens on the other node too.

ig0tik3d

legendary

Activity: 1246

Merit: 1000

Quote from: stonefoz on August 26, 2013, 03:31:42 PM

Quote from: ?? on ??

Quote from: stonefoz on August 26, 2013, 03:25:09 PM

it's the same, there is a binary posted about 3 pages back that someone was nice enough to make. If you feel adventures the source has been moved to github

I can upload a binary as well if there's interest.

That would be great, I would like to see it tested on something not AMD first. There is a "makeprofile.sh" script. It now has enough warnings that I worry it may have issues if not tested. Looks good right now, but the goal is fast, not necessarily good. Anyway, if the middle two lines match anything else the script prints out, at least two hashes where done correctly....

Oh, and there is a new option.
./configure CFLAGS="-msse2 -O3 -DCHEAT"
needs testing but the idea is that it's faster to just try a new nonce that run groestl twice. Anytime that section is found to be needed ran, it just give up. Half of the hashes end up being garbage. But, it now does more than twice the speed, for a net gain. EXPERIMENTAL, need to be ran to see if actually improves payouts as it may be losing shares. Still runs that section when required, but never twice.

who want can test this new option on win64
http://rghost.ru/48416928

Palmdetroit

legendary

Activity: 910

Merit: 1000

PHS 50% PoS - Stop mining start minting

Repost. Updated

AMD Bulldozer optimized client Win64bit and maybe some intels
---===---===---===---
built from StoneFoz version1 Which doubled speed and beyond

v2 was having many issues (+reject, +orphan for me) , bin below is version 1

8150 Stock 8 cores = .8~1.1 mhash
8350 Stock 8 cores = 1~?? mhash
4300 Stock 4 cores = .46+ mhash=+

Bulldozer 1 (more compatable) and Bulldozer gen 2 (more specialized)
https://sourceforge.net/projects/philosopherstone/files/QRK/

Tips appreciated: QRK: QQVzchEu7aqAfc71NZa1TYi3toih6abCtk

Re-posting because this is not AMD bulldozer only, and may in fact work on the newer intels as well.

this is not the uber optimized client, will post when over 2mhash per chip

mrmanager

newbie

Activity: 50

Merit: 0

Quote

is compiled for windows 64 ( from source code http://stonefoz.myfastmail.com/cpuminer-quark.zip)
other link https://docs.google.com/file/d/0B9cvOfoOekSdVzZZcThLZHg4bjA/edit?usp=sharing

Wow, this is way faster on my Windows boxes.

topminingcontracts

hero member

Activity: 532

Merit: 500

TaaS is a closed-end fund designated to blockchain

What miner you recomend for Ubuntu - Centos?

Any binary?

TMC

glongsword

full member

Activity: 314

Merit: 100

Quote from: ig0tik3d on August 27, 2013, 04:02:49 AM

Quote from: glongsword on August 26, 2013, 06:27:18 PM

cloned the git repo, and got an error message:

user@computer:~/quarkcoin-cpuminer# ./autogen.sh
configure.ac:117: `PTHREAD_FLAGS' includes configure substitution `@PTHREAD_FLAGS@'
configure.ac:117: and is referred to from `minerd_SOURCES';
configure.ac:117: configure substitutions are not allowed in _SOURCES variables

Thoughts?

Thanks!

in Makefile.am line 20
change

Code:

quark.c \

to

Code:

quark.c

enjoy))

Worked like a charm! Thanks!

zvs

legendary

Activity: 1680

Merit: 1000

https://web.archive.org/web/*/nogleg.com

Quote from: Spoetnik on August 27, 2013, 07:18:53 AM

Quote from: stonefoz on August 26, 2013, 03:25:09 PM

it's the same, there is a binary posted about 3 pages back that someone was nice enough to make. If you feel adventures the source has been moved to github.
Compared to the original miner though, it is about 85% faster, uses SSE2 along with ASM for two of the rounds.

uhhhhhhhhhhh 85% faster on AMD Bulldozer CPU's only ?
am i wasting my time ripping my hair out trying to compile this newer version (on windows / for windows) for no speed gain on my Intel CPU ?

don't know how fast/slow original miner is. the one in the program itself? quarkcoind got about 100khash, quarkcoinqt got about 200khash, and this minerd64_sse2 gets over 1mhash on an i7 4770K

ig0tik3d

legendary

Activity: 1246

Merit: 1000

Quote from: Spoetnik on August 27, 2013, 10:10:47 AM

Quote from: ig0tik3d on August 27, 2013, 08:55:03 AM

Quote from: Spoetnik on August 27, 2013, 07:18:53 AM

Quote from: stonefoz on August 26, 2013, 03:25:09 PM

it's the same, there is a binary posted about 3 pages back that someone was nice enough to make. If you feel adventures the source has been moved to github.
Compared to the original miner though, it is about 85% faster, uses SSE2 along with ASM for two of the rounds.

uhhhhhhhhhhh 85% faster on AMD Bulldozer CPU's only ?
am i wasting my time ripping my hair out trying to compile this newer version (on windows / for windows) for no speed gain on my Intel CPU ?

https://bitcointalksearch.org/topic/m.2993417

dead link = 503 Service Unavailable

also you have made two comments and still not said what it is..

is it optimized for Intel and was the old one ?

can someone shed some light on this please ?
and a windows build would be great too

is compiled for windows 64 ( from source code http://stonefoz.myfastmail.com/cpuminer-quark.zip)
other link https://docs.google.com/file/d/0B9cvOfoOekSdVzZZcThLZHg4bjA/edit?usp=sharing

Spoetnik

legendary

Activity: 1540

Merit: 1011

FUD Philanthropist™

Quote from: ig0tik3d on August 27, 2013, 08:55:03 AM

Quote from: Spoetnik on August 27, 2013, 07:18:53 AM

Quote from: stonefoz on August 26, 2013, 03:25:09 PM

it's the same, there is a binary posted about 3 pages back that someone was nice enough to make. If you feel adventures the source has been moved to github.
Compared to the original miner though, it is about 85% faster, uses SSE2 along with ASM for two of the rounds.

uhhhhhhhhhhh 85% faster on AMD Bulldozer CPU's only ?
am i wasting my time ripping my hair out trying to compile this newer version (on windows / for windows) for no speed gain on my Intel CPU ?

https://bitcointalksearch.org/topic/m.2993417

dead link = 503 Service Unavailable

also you have made two comments and still not said what it is..

is it optimized for Intel and was the old one ?

can someone shed some light on this please ?
and a windows build would be great too

Lauda

legendary

Activity: 2674

Merit: 3000

Terminated.

Time to buy cheap AMD CPUs Cheesy

ig0tik3d

legendary

Activity: 1246

Merit: 1000

Quote from: Spoetnik on August 27, 2013, 07:18:53 AM

Quote from: stonefoz on August 26, 2013, 03:25:09 PM

it's the same, there is a binary posted about 3 pages back that someone was nice enough to make. If you feel adventures the source has been moved to github.
Compared to the original miner though, it is about 85% faster, uses SSE2 along with ASM for two of the rounds.

uhhhhhhhhhhh 85% faster on AMD Bulldozer CPU's only ?
am i wasting my time ripping my hair out trying to compile this newer version (on windows / for windows) for no speed gain on my Intel CPU ?

https://bitcointalksearch.org/topic/m.2993417

Spoetnik

legendary

Activity: 1540

Merit: 1011

FUD Philanthropist™

Quote from: stonefoz on August 26, 2013, 03:25:09 PM

it's the same, there is a binary posted about 3 pages back that someone was nice enough to make. If you feel adventures the source has been moved to github.
Compared to the original miner though, it is about 85% faster, uses SSE2 along with ASM for two of the rounds.

uhhhhhhhhhhh 85% faster on AMD Bulldozer CPU's only ?
am i wasting my time ripping my hair out trying to compile this newer version (on windows / for windows) for no speed gain on my Intel CPU ?

eule

hero member

Activity: 756

Merit: 501

Quote from: trinsic on August 26, 2013, 09:46:14 PM

Any idea what's going on with this?

Code:

root@quarky:~/quark/src# ./quarkcoind sendtoaddress QWghDKW59cFJTitJg4zBp7BnjUWM3H4wcF 10000
error: {"code":-4,"message":"Transaction too large"}
root@quarky:~/quark/src# ./quarkcoind sendtoaddress QWghDKW59cFJTitJg4zBp7BnjUWM3H4wcF 5000
error: {"code":-4,"message":"Transaction too large"}
root@quarky:~/quark/src# ./quarkcoind sendtoaddress QWghDKW59cFJTitJg4zBp7BnjUWM3H4wcF 2000
error: {"code":-4,"message":"Transaction too large"}
root@quarky:~/quark/src# ./quarkcoind sendtoaddress QWghDKW59cFJTitJg4zBp7BnjUWM3H4wcF 1000
20bfb1a939064973f7e8d08dee352d34f757b625b3e8bac14da98e457675f675
root@quarky:~/quark/src# ./quarkcoind sendtoaddress QWghDKW59cFJTitJg4zBp7BnjUWM3H4wcF 1000
54b405e8d532842a7ef1386d3cdeea972ec7ffd4fac965ca89308e6865f395dd
root@quarky:~/quark/src# ./quarkcoind sendtoaddress QWghDKW59cFJTitJg4zBp7BnjUWM3H4wcF 1000
error: {"code":-4,"message":"Transaction too large"}

Did you recieve many small p2pool transactions?

cryptohunter

legendary

Activity: 2100

Merit: 1167

MY RED TRUST LEFT BY SCUMBAGS - READ MY SIG

I think this coin has been one of the most stable in price and hash since i have been watching alt coins. I wouldn't sell any of mine for fiat until the reward has halved a few more times and these are starting to get quite scarce. I have traded some for primes although actually i wish i had not now since prime has sunk since i bought them and this has continued to rise.

Topic: [ANN] [QRK] Quark | Core 0.10 upgrade - page 399. (Read 1031025 times)