Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - Ykw

Pages: 1 2 [3] 4 5 6
31
BitShares PTS / Re: fast AMD OpenCL PTS/NRS miner released
« on: April 29, 2014, 10:13:49 pm »
NaN... great work.. I found this version.. of catalyst 14 with more room for improvement than with catalyst 13... I think they got the power usage better... so.. more room for overclocking... under same col/m performance.

I am still not sure if there something "way badly behaving" with the CPU+GPU communication under PTS/NRS... with the Bitcoin miners.. I get much higher temps.. and less stress over CPU.. much higher TEMPs.. but better (lower) with Catalyst 14 again... so.. I might guess there is more efficient code.. running over the driver.. still.. there is room for improvement from AMD...

Looking at the comm.. of the driver.. there are too much repetition... from AMD compared with NVIDIA... this is... what I found to be the main "lack of HP force" from AMD compared with NVIDIA.. but apart from that.. the drivers.. history.. are way better! I think.. that.. AMD is trying to play the APU card with this issue in time... to overcome their excessive communication.. (in my opinion). NVIDIA.. is somehow playing this type of game since versions 500.. already...

well.. I like what I see.. =) heeeee =)

32
BitShares PTS / Re: fast AMD OpenCL PTS/NRS miner released
« on: April 29, 2014, 09:57:46 pm »
Quote from: JoeyD on April 29, 2014, 09:32:42 pm
Normally I'd rather not mess with a running system, but is the upgrade to Catalyst14 really that big of an improvement?

You must have a "test" system.. then.. you move to the miners..

And indeed it's better..

After changing to latest... Catalyst available.. I got...
290X (2 thread) - 4679 col/m (from 4579 for 30 min)...
270X (1 thread) - 2465 col/m (from 2279 for 30 min)...

still stabilizing...


EDIT: I have a X58 Board overclocked... and I found it.. limiting in terms of some CPU+GPU memory intensive communication... so.. I get probably lower results than others...

33
BitShares PTS / Re: fast AMD OpenCL PTS/NRS miner released
« on: April 29, 2014, 07:15:45 pm »
Version 1.2 (NSR) - Catalyst 13 (Windows 7 64bit)

290X (2 thread) - 4550 col/m
270X (1 thread) - 2250 col/m

34
General Discussion / Re: BitShares X Status Update
« on: April 13, 2014, 03:53:41 pm »
Quote from: bytemaster on April 10, 2014, 04:38:08 pm
Quote from: sumantso on April 10, 2014, 04:04:14 pm
Quote from: bytemaster on April 10, 2014, 03:41:56 pm
No lost confidence.  Just had to solve pos first. 

Has this been solved now wrt coding?

Yes, I released a white paper on DPOS last week and have implemented a large part of it and will be testing it in the days ahead.   Resolving the security model in a way that one can thoroughly understand and prove is secure was a major challenge.  You can follow discussions on this forum where about a half dozen variations / iterations where vetted and found wanting...  Let me list the evolution:

1) Mining with Momentum... tossed due to centralization and non profitability
2) Transactions as Proof of Stake white paper (Nov 28th, 2013) adjusted mining difficulty by stake.  Suffered from weakness on deciding who should produce the next block.
3) Ripple Consensus + TaPOS was the concept in January at the Miami event through Feb, but after attempted implementation we realized it had several deal breakers:
      a) the unique node list was an invite-only group and a point of trust
      b) the UNL was not compensated in any way
      c) the bandwidth requirements / scalability were an issue
4) Fall back on a simple mining + TaPOS and various variations on equations for option 2.  All of these variations suffered from many potential attacks and would require many confirmations.  It became too complex to analyze.  (Feb through March)
5) Single trustee which could be fired... the process was too manual but otherwise sound.
6) Delegated POS this solution has been reviewed, found to be decentralized, fast, and secure without ambiguity.  We have finally resolved the issue.

In that time we also identified problems and found solutions to several market manipulation attacks.  These need to be implemented, but then we should be good to go.

As you can see R&D is not predictable when you are at the bleeding edge of innovation and solving problems as they are discovered.


Very nice paper (http://107.170.30.182/security/delegated-proof-of-stake.php)...

Although I must admit that still needs some visual explanation for broader public understanding. It also would help consolidate by commenting it (visual drawings).

Things like this necessarily have to change/adapt with time essence... we face edge developing ideas that get reactions... when implemented. So documenting them is a MUST. Consolidating their stable audience.. also. Otherwise developing something that only we understand about.. and where the focus is for everyone to adopt.. will be worth nothing.. I understand that time races against your availability, but felt right to me in pointing out this.

This is just my way of extrapolating the effect of knowledge specialization. That we do require as a source for developing new core ideas.. but that with time if canalized to the right direction can potentially create a positive diversion.. for mass adoption.

35
BitShares PTS / Re: Why not change the PTS mining method from POW convert POS?
« on: April 13, 2014, 03:04:55 pm »
For me also makes sense for PTS to be TaPOS... as coins like BC and now FAC are taking market share.. also people only move more money into crypto once BTC regains popularity and stability... so China market will need to regain trust first ... My guess is that not only their anti-others capitalism fight will not work in long term.. but also because multi-altcoin will flood their micro-transaction markets... their influence will somehow be diluded into their play-role in rejecting/non-supporting cryptocurrencies..

So, my prediction is for capitalistic cryptos to be better china "influential" crypto than "minable" only ones. Therefore.. more tempting for them to adopt.

36
BitShares PTS / Re: BitShares-PTS 1.0 Release (2.18.2014)
« on: April 10, 2014, 10:15:15 am »
Quote from: testz on April 10, 2014, 09:57:33 am
Quote from: Ykw on April 10, 2014, 09:32:26 am
Any updates regarding new release of openSSL 1.0.1g ?

Yes, at next release Open SSL will be updated, but it's not critical because Open SSL hole not really affected wallet software:
http://www.cryptocoinsnews.com/news/openssl-heartbleed-security-bug/2014/04/08

So only mining to wallets could be affected... anything else (like transfering PTS) is not being affected by this bug?

37
BitShares PTS / Re: BitShares-PTS 1.0 Release (2.18.2014)
« on: April 10, 2014, 09:32:26 am »
Any updates regarding new release of openSSL 1.0.1g ?

38
BitShares PTS / Re: fast AMD OpenCL PTS miner released
« on: April 03, 2014, 07:09:53 pm »
Quote from: ozzy on April 03, 2014, 07:03:06 pm
Quote from: Ykw on April 03, 2014, 06:42:44 pm
Quote from: ozzy on April 03, 2014, 04:34:24 pm
Quote from: wsnnwa on April 02, 2014, 12:48:37 pm
Just launch mine.sh, like this --> ./mine.sh
oz@oz:~/Documents/miner2$ ./mine.sh
bash: ./mine.sh: Permission denied

chmod +x ./mine.sh
sudo ./mine.sh

.oz@oz:~/Documents/miner2$ sudo ./mine.sh
[sudo] password for oz:
./clpts: /lib/x86_64-linux-gnu/libc.so.6: version `GLIBC_2.17' not found (required by ./clpts)
oz@oz:~/Documents/miner2$

i think i am missing some file

You don't have a compatible distribution... or you have older glibc...

Do:
cat /etc/issue

39
BitShares PTS / Re: fast AMD OpenCL PTS miner released
« on: April 03, 2014, 06:42:44 pm »
Quote from: ozzy on April 03, 2014, 04:34:24 pm
Quote from: wsnnwa on April 02, 2014, 12:48:37 pm
Just launch mine.sh, like this --> ./mine.sh
oz@oz:~/Documents/miner2$ ./mine.sh
bash: ./mine.sh: Permission denied

chmod +x ./mine.sh
sudo ./mine.sh

40
Keyhotee / Re: Keyhotee Status Update
« on: April 02, 2014, 10:09:19 pm »
Successfully recreated profile and identity..

New Pub key: 7W73zQHPmKuApHzzNzRVzq8FScdDd65UpMyVbxhoeYXGCQ4mdC

Points show up.. =)

41
BitShares PTS / Re: fast AMD OpenCL PTS miner released
« on: March 31, 2014, 05:53:16 pm »
Quote from: Alec246 on March 31, 2014, 05:36:17 pm
I don't want to mess too much with my Linux because it's working flawless for CgMiner! If I Start doing stuff to it, I will probably crash it, and since I have no Linux experience, the only way will be a complete Format :/

Here's the tutorial I followed

https://www.weminecryptos.com/forum/topic/2187-guide-housing-xubuntu-setup-amd-drivers-kdev-sd-cgminer-monitoring-and-vbios-flashing/

The output is simply No GPU Device with the id.

What you have in this "aticonfig --lsa"...


EDIT: Share the actual output of the miner when you try to run it

42
BitShares PTS / Re: fast AMD OpenCL PTS miner released
« on: March 31, 2014, 05:39:03 pm »
Hi folks....

You have to understand that if you want help you have to give back something.. just saying that it does not work.. doesn't help anyone nor it will be enough for anyone from the community to help.

Try to be logic... when you have a problem.. share the errors, configs, when problem started to happen, when it did not.. etc etc...

There are no magic within devs...  they are all hard working geeks =)

43
BitShares PTS / Re: fast AMD OpenCL PTS miner released
« on: March 31, 2014, 05:06:44 pm »
Quote from: Alec246 on March 31, 2014, 04:38:40 pm
Tried with Sudo.

Pretty sure everything is OK otherwise I wouldn't be able to mine with cgminer.

Can you post the output? Are you using/booting into single user mode (init 3) ? Is "startx" command working?

44
BitShares PTS / Re: fast AMD OpenCL PTS miner released
« on: March 31, 2014, 05:01:24 pm »
Quote from: ozzy on March 31, 2014, 08:48:35 am
Quote from: chaosz88 on March 30, 2014, 12:56:28 am
Quote from: ozzy on March 28, 2014, 07:13:02 pm
Quote from: NaN on March 28, 2014, 06:50:07 pm
I used C++11 for the new miner. Is your glibc up to date? Perhaps I should revert to boost and statically link it. Which distro do you use?
EDIT: Oh, the correct syntax is
./clpts -o 0,6 Pezz4JUPfpAFTSSwWgp1TMpCcS1YvEqVsn -u ozzy.PTS_1 -p x -t 0,0,1,1,2,2 -a 2
yea i also tried the ./clpts -o 0,6 Pezz4JUPfpAFTSSwWgp1TMpCcS1YvEqVsn -u ozzy.PTS_1 -p x -t 0,0,1,1,2,2 -a 2

but it didnt work .And i dont know if i have the latest c++11 i dont know much about linux (i am using ubuntu 12.04 )
edit i manage to install but :
installed
                      Conflicts: libstdc++6-4.5-dbg but 4.5.3-12ubuntu2 is to be installed
                      Conflicts: libstdc++6-4.6-dbg but 4.6.4-1ubuntu1~12.04 is to be installed
 libstdc++6-4.7-doc : Conflicts: libstdc++6-4.4-doc but 4.4.7-1ubuntu2 is to be installed
                      Conflicts: libstdc++6-4.5-doc but 4.5.3-12ubuntu2 is to be installed
                      Conflicts: libstdc++6-4.6-doc but 4.6.4-1ubuntu1~12.04 is to be installed
 libstdc++6-4.8-dbg : Conflicts: libstdc++6-4.4-dbg but 4.4.7-1ubuntu2 is to be installed
                      Conflicts: libstdc++6-4.5-dbg but 4.5.3-12ubuntu2 is to be installed
                      Conflicts: libstdc++6-4.6-dbg but 4.6.4-1ubuntu1~12.04 is to be installed
                      Conflicts: libstdc++6-4.7-dbg but 4.7.3-2ubuntu1~12.04 is to be installed
E: Unable to correct problems, you have held broken packages.
I got the same error with Ubuntu 12.04, I solved it by

sudo apt-get install libstdc++6
it also doesnt work
i am geting :
libstdc++6 is already the newest version.
0 upgraded, 0 newly installed, 0 to remove and 57 not upgraded.

btw i am using amd-catalyst-14.1-betav1.3-linux-x86.x86_64 drivers


You have libstdc++6-4.4 installed and probably have multiple sources.. with different names of the same package... try to remove all the libstdc++ and dependent packages and then install them again.. but using only one source. This is not a miner problem.. it's an OS config know how...

If you are uncomfortable with ubunto.. try other OSs you are ok with... you have fedora, centos, windows.. etc etc...

Otherwise.. post more details. (the more you can get)

45
BitShares PTS / Re: fast AMD OpenCL PTS miner released
« on: March 30, 2014, 06:12:55 pm »
Quote from: Mfqrs3 on March 30, 2014, 05:57:56 pm
Quote from: NaN on March 30, 2014, 05:49:55 pm
Quote from: Mfqrs3 on March 30, 2014, 04:00:52 pm
Hi NaN,
Thanks for the latest miner!

It's working fine in Win7 64bit with latest 14.3 beta drivers, but it doesn't work with win 8.1 64bit with 13.12 drivers, R9 270, R9 280x or R9 290 cards.

Error: Compilation from LLVMIR binary to IL text failed!
As you may have noticed there are different files for Catalyst 13.12 and Catalyst 14.x drivers. You have to use the *Catalyst13 files for Catalyst 13.12

You mean Gpu_hash files or driver files ? If driver files which and how shoul I replace them ? Just copy over the old ones or ?
Tnx

NaN wanted you to understand that in his download links... there are 2 set of files (miner+gpu_hash files), ones are for 13.12 drivers.. the other for the beta 14.x version. So you should download the correct version to run over the correct driver you are using.

The gpu_hash switch trick was before he posted that fix.

Pages: 1 2 [3] 4 5 6

Page created in 0.196 seconds with 23 queries.