The folks on ypool just pointed out to me that gmp 6.0.0a has been released. (Thanks!)
fastrie is only 5% faster with 6.0.0a than with 5.1.3 (I have SandyBridge)
11% faster with a custom ARM based miner.
--
bsunau7
Topic: [ANN][RIC] Riecoin: constellations POW *CPU* HARD FORK successful, world record - page 180. (Read 685381 times)
March 29, 2014, 07:36:54 AM
fastrie is only 5% faster with 6.0.0a than with 5.1.3 (I have SandyBridge)
11% faster with a custom ARM based miner.
--
bsunau7
March 29, 2014, 02:15:00 AM
That near 40k buy wall though 
March 28, 2014, 11:02:20 PM
If Chinese looks like Greek to you,
BTC38 launched an English version of their exchange
RIC/CNY market linked.
Looks like btc38 had about twenty times the volume of Mintpal the last 24 hours...
Edit: looks like they only support Facebook for English log in.
BTC38 launched an English version of their exchange
RIC/CNY market linked.
Looks like btc38 had about twenty times the volume of Mintpal the last 24 hours...
Edit: looks like they only support Facebook for English log in.
March 28, 2014, 10:48:01 PM
Finally, something to convince me that I need to upgrade to xeons with AVX2 !
March 28, 2014, 09:58:09 PM
Need more buy orders on poloniex
March 28, 2014, 04:29:14 PM
Coming Now Soon for RiecoinFoundation.org is now uploaded!
March 28, 2014, 02:35:33 PM
The folks on ypool just pointed out to me that gmp 6.0.0a has been released. (Thanks!)
fastrie is only 5% faster with 6.0.0a than with 5.1.3 (I have SandyBridge)
Correct. But the gmp speedup mixes nicely with another improvement I'm working on (it's currently buggy still). I'm seeing 10-15% faster on AVX2 overall.
March 28, 2014, 02:09:49 PM
The folks on ypool just pointed out to me that gmp 6.0.0a has been released. (Thanks!)
fastrie is only 5% faster with 6.0.0a than with 5.1.3 (I have SandyBridge)
March 28, 2014, 10:51:33 AM
I know, didn't someone say Cryptsy was going to add Riecoin to the voting list, but they never did...
You should try in the community for people to ask the following
Try to know more
Good luck to you
March 28, 2014, 10:41:50 AM
Thanks Gatra! . . . but now how am I going to get any video editing work done?!
March 28, 2014, 10:41:26 AM
The folks on ypool just pointed out to me that gmp 6.0.0a has been released. (Thanks!)
I'll get a binary build out this weekend with two sets of improvements:
- gmp 6.0 integration (just affects the binary)
- better support for sieving past 500m (binary and source)
The latter steals a trick from a00k's release that I liked - for the biggest numbers, scan once and discard them if the offset is too large.
In the meantime, if you build your own version linked against gmp 6 and run the tuneup function, you should see decent speed gains on avx and avx2-compatible processors.
I'll get a binary build out this weekend with two sets of improvements:
- gmp 6.0 integration (just affects the binary)
- better support for sieving past 500m (binary and source)
The latter steals a trick from a00k's release that I liked - for the biggest numbers, scan once and discard them if the offset is too large.
In the meantime, if you build your own version linked against gmp 6 and run the tuneup function, you should see decent speed gains on avx and avx2-compatible processors.
March 28, 2014, 10:38:02 AM
Difficulty has backed off for a little bit so get in while you can get some shares and keep pumping that extra CPU power to the network. Otherwise what are you doing with that CPU power?
March 28, 2014, 09:15:09 AM
New status update: https://bitcointalksearch.org/topic/m.5943221
New ideas to explore:
- Chebyshev's bias: is it large enough to use it for miner optimization? at first sight, it doesn't look that way, but it's effect may be amplified by the fact that we use prime 6tuples, not just primes... needs more research. Optimized miners use a hardcoded value for the remainder (or "offset") of their primorial, but is it the best value?
- If we'll be hardforking to include constant difficulty adjustments, maybe we could use the opportunity to include searching for 7 or 8tuples along with 6 tuples.
Best regards to all!
gatra
New ideas to explore:
- Chebyshev's bias: is it large enough to use it for miner optimization? at first sight, it doesn't look that way, but it's effect may be amplified by the fact that we use prime 6tuples, not just primes... needs more research. Optimized miners use a hardcoded value for the remainder (or "offset") of their primorial, but is it the best value?
- If we'll be hardforking to include constant difficulty adjustments, maybe we could use the opportunity to include searching for 7 or 8tuples along with 6 tuples.
Best regards to all!
gatra
Does the stratum miner have a way for the worker thread to determine if a new block has arrived without exiting its mining loop? If so, I can throw mine into the stratum miner pretty easily at this point.
"work_restart[thr_id].restart" is global and will change from false to true (for all elements of the work_restart array - ie for all threads) when a new block has arrived. Then each thread will set it to false again when it has updated it's block data struct.
Without stratum, that restart was only getting set if a miner exited its processing loop and called getwork again.
I see that the stratum thread can set it also. Cool. That was a huge pain with getwork. March 28, 2014, 08:08:04 AM
This coin design is very good-looking, will be very popular
March 28, 2014, 08:00:33 AM
New status update: https://bitcointalksearch.org/topic/m.5943221
New ideas to explore:
- Chebyshev's bias: is it large enough to use it for miner optimization? at first sight, it doesn't look that way, but it's effect may be amplified by the fact that we use prime 6tuples, not just primes... needs more research. Optimized miners use a hardcoded value for the remainder (or "offset") of their primorial, but is it the best value?
- If we'll be hardforking to include constant difficulty adjustments, maybe we could use the opportunity to include searching for 7 or 8tuples along with 6 tuples.
Best regards to all!
gatra
New ideas to explore:
- Chebyshev's bias: is it large enough to use it for miner optimization? at first sight, it doesn't look that way, but it's effect may be amplified by the fact that we use prime 6tuples, not just primes... needs more research. Optimized miners use a hardcoded value for the remainder (or "offset") of their primorial, but is it the best value?
- If we'll be hardforking to include constant difficulty adjustments, maybe we could use the opportunity to include searching for 7 or 8tuples along with 6 tuples.
Best regards to all!
gatra
Does the stratum miner have a way for the worker thread to determine if a new block has arrived without exiting its mining loop? If so, I can throw mine into the stratum miner pretty easily at this point.
"work_restart[thr_id].restart" is global and will change from false to true (for all elements of the work_restart array - ie for all threads) when a new block has arrived. Then each thread will set it to false again when it has updated it's block data struct.
March 28, 2014, 06:51:59 AM
New status update: https://bitcointalksearch.org/topic/m.5943221
New ideas to explore:
- Chebyshev's bias: is it large enough to use it for miner optimization? at first sight, it doesn't look that way, but it's effect may be amplified by the fact that we use prime 6tuples, not just primes... needs more research. Optimized miners use a hardcoded value for the remainder (or "offset") of their primorial, but is it the best value?
- If we'll be hardforking to include constant difficulty adjustments, maybe we could use the opportunity to include searching for 7 or 8tuples along with 6 tuples.
Best regards to all!
gatra
New ideas to explore:
- Chebyshev's bias: is it large enough to use it for miner optimization? at first sight, it doesn't look that way, but it's effect may be amplified by the fact that we use prime 6tuples, not just primes... needs more research. Optimized miners use a hardcoded value for the remainder (or "offset") of their primorial, but is it the best value?
- If we'll be hardforking to include constant difficulty adjustments, maybe we could use the opportunity to include searching for 7 or 8tuples along with 6 tuples.
Best regards to all!
gatra
Does the stratum miner have a way for the worker thread to determine if a new block has arrived without exiting its mining loop? If so, I can throw mine into the stratum miner pretty easily at this point.
March 28, 2014, 04:02:17 AM
Hey my first ever mined block on ypool, shame it wasn't solo 
March 28, 2014, 12:38:30 AM
Was just wondering, how powerful a computer/cpu is needed to at least be able to mine a block? I'm currently on ypool with Dave's optimized miner and my workers never find any at all.
How many shares do you have? I've found 2 blocks and 26.000 shares since the optimized xpt miners came out. Before that I found 1 block and 2000 shares.Is the number of shares a metric to mining riecoin?
March 28, 2014, 12:30:17 AM
Was just wondering, how powerful a computer/cpu is needed to at least be able to mine a block? I'm currently on ypool with Dave's optimized miner and my workers never find any at all.
How many shares do you have? I've found 2 blocks and 26.000 shares since the optimized xpt miners came out. Before that I found 1 block and 2000 shares. March 28, 2014, 12:25:54 AM
Was just wondering, how powerful a computer/cpu is needed to at least be able to mine a block? I'm currently on ypool with Dave's optimized miner and my workers never find any at all.
Jump to: