[XPM] Primecoin Built-in Miner Sieve Performance Issue

[ivanlabrie]

I'm hoping I can run bfgminer on my vps systems...I tweaked mine and am pulling 400 per instance and 1000 in others.
Total is 6400pps atm, and looking to increase it to 9900pps soon.

Where are the pools!? :p

Soon

yay

I hope reward halves soon, so we can see price rise...cpu only pool would be fun. xD

I don't think we will get to diff 10 any time soon.

Who knows when?
It's a nice experiment, and deviation from the other copy/paste coins.
I'm enjoying it, and I hope it succeeds to fulfill it's potential.

[1714911154]

1714911154

[1714911154]

1714911154

[Lauda]

I'm hoping I can run bfgminer on my vps systems...I tweaked mine and am pulling 400 per instance and 1000 in others.
Total is 6400pps atm, and looking to increase it to 9900pps soon.

Where are the pools!? :p

Soon

Not soon enough.

[tacotime]

Alright, so just updated my version (currently on github) such that each thread an independent evolving weave timing parameter.  To compare to mine with Sunny's most recent update, I used the testnet where my version found 30 confirmed blocks in 10 minutes while the original code found 16 confirmed blocks.  I feel that this is a legitimate comparison because there were no other nodes on the test net currently mining (I know this because my client found every continuous block in both cases).  This comparison was performed with a t61p IBM laptop with a T9300 Core 2 Duo processor.  The current difficulty on the testnet is 5.4426.  

Going to test this with the 8 threads on my Core i7 next.

Mind linking to your github? The speed this thread is updating is a bit overwhelming. Thanks in advance.

Updated my profile website to link directly to it, just fyi so you dont have to keep coming back here...

https://github.com/Chemisist/primecoin

Good news, got 4 blocks today on the main net with your fork!

[dudeguy]

Alright, so just updated my version (currently on github) such that each thread an independent evolving weave timing parameter.  To compare to mine with Sunny's most recent update, I used the testnet where my version found 30 confirmed blocks in 10 minutes while the original code found 16 confirmed blocks.  I feel that this is a legitimate comparison because there were no other nodes on the test net currently mining (I know this because my client found every continuous block in both cases).  This comparison was performed with a t61p IBM laptop with a T9300 Core 2 Duo processor.  The current difficulty on the testnet is 5.4426.  

Going to test this with the 8 threads on my Core i7 next.

Mind linking to your github? The speed this thread is updating is a bit overwhelming. Thanks in advance.

Updated my profile website to link directly to it, just fyi so you dont have to keep coming back here...

https://github.com/Chemisist/primecoin

Good news, got 4 blocks today on the main net with your fork!

Linux? Windows 7? What processor? How many PPS/Blocks per 10min? Using Cheisist's build right now on combined 1000PPS. Nothing.


BTW to anyone who misunderstood my post earlier about how "WE NEED" I wasn't trying to demand. I was just trying to vocalize a common request in the community. You should check out Reddit. There are plenty of people there very excited about this coin and even doubly excited to see where devs. like yourselves can take it: http://www.reddit.com/r/primecoin/    So thanks for all your hard work. I've already donated to one of the devs.

[tacotime]

Linux? Windows 7? What processor? How many PPS/Blocks per 10min? Using Cheisist's build right now on combined 1000PPS. Nothing.

Linux (Ubuntu 64-bit)
Core i7 2700K @ 4.4 GHz
~2200 PPS on average
I'm getting a block every 4 or so hours

[fabrizziop]

Linux? Windows 7? What processor? How many PPS/Blocks per 10min? Using Cheisist's build right now on combined 1000PPS. Nothing.

Linux (Ubuntu 64-bit)
Core i7 2700K @ 4.4 GHz
~2200 PPS on average
I'm getting a block every 4 or so hours

Are you having the client stops syncing blockchain problem?.

[taturii]

Block generation time is changing rapidly. this morning 150 blocks were generated in  less than 30 min; right now 150 blocks need more than 40 min.

[Carra23]

Is the 0.1.1 client still the best? I haven't got anything in 2 days and wondering whether there are better versions out there.

[wibtc]

Linux? Windows 7? What processor? How many PPS/Blocks per 10min? Using Cheisist's build right now on combined 1000PPS. Nothing.

Linux (Ubuntu 64-bit)
Core i7 2700K @ 4.4 GHz
~2200 PPS on average
I'm getting a block every 4 or so hours

lucky you

[tacotime]

Linux? Windows 7? What processor? How many PPS/Blocks per 10min? Using Cheisist's build right now on combined 1000PPS. Nothing.

Linux (Ubuntu 64-bit)
Core i7 2700K @ 4.4 GHz
~2200 PPS on average
I'm getting a block every 4 or so hours

Are you having the client stops syncing blockchain problem?.

No

Here is my binary

[BitcoinFX]

Updated windows build using the new Chemisist source. Tuned for Intel Sandy and Ivy Bridge but compatible with other architecture.

https://www.dropbox.com/s/4k0xmuajxf5i4ly/primecoin0712v3-avx.zip

I'm seeing lower PPS than my v2 builds but I think that weaving will be better overall.

I'm running the v3 build on an i7 3930k (6-core 12-thread - Sandy Bridge-E - Windows 7) and have found 2 blocks so far running with 11 threads.

Found a block with primecoin0712 (v1 build for avx from Sunny's 0.1.1 updated code - from another thread) and also found a block with the v2 build before switching to v3. pps is around 2200 to 2500.

Maybe a bit of luck, but just confirming that I've found blocks with all three builds. I've not experienced any network or block 'lag' either (yet), as reported elsewhere with those builds. Seems to be all OK to me, until I find time to review and build.

[AgentME]

How high should the difficulty be at 4M pps?

PPS is not a good metric for actual mining performance, so it's not a good metric for difficulty either.

[hasle2]

I did some block generation testing on testnet of whether using more threads than cores helps the mining speed. On a core 2 duo (2 cores) running on Chemisist's code, standard build (O3):

4 threads: 47 blocks found in 32.67 minutes = 1.44 blocks/minute

2 threads: 116 blocks in 76.43 minutes = 1.52 blocks/minute

On the stock 0.1.1 from Sunny:

2 threads: 69 blocks in 64.6 minutes = 1.07 blocks/minute

Using 2x threads basically made no difference and is possibly worse. Running for a few hours would give better data but I've got what I wanted.

This somewhat moot now with mikaelh's new code. His looks to be over 3x faster than Chemisist's.

https://bitcointalk.org/index.php?topic=255782.0

[markm]

Speaking of core 2 Duo, i just bought a machine that has such a CPU, but it got only 21 PPS using latest high performance code compiled with -march=core2. My other core 2 machines, none of which mention anything about "duo" do much much better than that.

Is a core 2 duo maybe not really a core2 at all as far as gcc's optimise by architecture is concerned or something weird like that?

It is now showing 1753 PPS using same high performance version code compiled without telling gcc to optimise for anything in particular.

(In both cases though, I have -O3 instead of -O2 in the makefile. I don't do the compiling on the target machines, I just bring them home put same old fedora 17 on them my compile box runs, and have the compile box compile with -march that seems it maybe is talking about the same kind of CPU that the new box has in it.)

-MarkM-

[Sunny King]

I pushed a new commit with sieve optimizations. It appears to generate larger performance improvement on systems not heavily crippled by v0.1.0 miner.

Give it a try. Please note this version has a slower warmup period.

https://github.com/primecoin/primecoin/commit/ab9fe91d711c92527fcde2fa34563b0aefd16742

[mika1977]

I pushed a new commit with sieve optimizations. It appears to generate larger performance improvement on systems not heavily crippled by v0.1.0 miner.

Give it a try. Please note this version has a slower warmup period.

https://github.com/primecoin/primecoin/commit/ab9fe91d711c92527fcde2fa34563b0aefd16742

windows build  ?

[daserpent1]

Guys, HOW do I mine this coin?

Where can I download the client from? And how do i start mining then?

[Lauda]

Guys, HOW do I mine this coin?

Where can I download the client from? And how do i start mining then?

Read first post.

[daserpent1]

Those links show a code of some kind.

How do i use that text code to mine?

[paultramarine]

I pushed a new commit with sieve optimizations. It appears to generate larger performance improvement on systems not heavily crippled by v0.1.0 miner.

Give it a try. Please note this version has a slower warmup period.

https://github.com/primecoin/primecoin/commit/ab9fe91d711c92527fcde2fa34563b0aefd16742

I had a noticeable decrease in PPS from this update ( ~1/10th of PPS from original release).  I'm running a Core2 Quad.  I realize PPS isn't a great measure of efficacy but the decrease seems pretty consistent and dramatic (15min thus far). 

Is the "slower warmup period" longer than this, Sunny King??