[XPM] [ANN] Primecoin Release - First Scientific Computing Cryptocurrency - page 119.

Zalfrin

sr. member

Activity: 401

Merit: 250

Quote from: willphase on July 11, 2013, 09:45:22 PM

just a warning, when I tried compiling on VS2012 with optimizations, I found that my serialization code was off kilt, and run into all sorts of issues...

if you're playing with the reference client, it might be worth just trying to optimize prime.c (via CFLAGS override in Makefile) since that's where all the logic for calculations are - and you won't risk messing with any of the other important network code.

Will

Probably a good idea, testing with -O3 only on prime.

craslovell

legendary

Activity: 1470

Merit: 1021

Quote from: CaptChadd on July 11, 2013, 07:51:54 PM

How do we get on the right chain?

Looks like some of the custom builds were causing these issues.

craslovell

legendary

Activity: 1470

Merit: 1021

Quote from: de_xt on July 11, 2013, 06:20:01 PM

Nobody noticed Vircurex just added XPM?

They went down shortly after that, but now they are fine and dandy again Wink

Looks like coins-e was abandoned because of this Tongue

craslovell

legendary

Activity: 1470

Merit: 1021

Quote from: jpmi1 on July 11, 2013, 05:23:47 PM

Quote from: jpmi1 on July 11, 2013, 03:03:22 PM

How long does it usually take for a transaction to appear in your wallet?

I sent 2 separate transactions over 2 hours ago and still nothing Sad

any idea?

Did you add .01 tx fee in your wallet settings?

Wolf0

member

Activity: 81

Merit: 1002

It was only the wind.

Quote from: Lauda on July 11, 2013, 04:00:55 PM

Quote from: Ethera on July 11, 2013, 03:47:03 PM

Quote from: Wolf0 on July 11, 2013, 03:40:00 PM

Quote from: Lauda on July 11, 2013, 03:32:31 PM

Quote from: Wolf0 on July 11, 2013, 03:31:15 PM

Quote from: Lauda on July 11, 2013, 03:24:38 PM

Quote from: Wolf0 on July 11, 2013, 03:23:37 PM

Quote from: Lauda on July 11, 2013, 03:20:25 PM

Has there been an update to the client so far?

Yes.

Where, link!

Github.

If i have to compile myself, it's rubbish.

Who knows, it could be a 1000x speedup. I would be careful before saying it's rubbish. Compile and see.

he means "i lack the knowledge to do anything like that, so liek you dont give me free mining so liek lol.. (idiocracy was a great movie, still watching it time to time to remind me what kind of society we are going for)...

Really? That makes no sense. I just don't wanna bother compiling, my VS 2012 hasn't been used for a while.

Then you "don't wanna bother" getting the latest updates, which could include speedups.

96redformula

full member

Activity: 224

Merit: 100

That was incredibly strange, 17754 was taking a long long time it seemed. All of a sudden the block solve rate just slowed tremendously from a glance at the block count. It went from one every 10 seconds to 20-30 seconds plus.

willphase

hero member

Activity: 767

Merit: 500

just a warning, when I tried compiling on VS2012 with optimizations, I found that my serialization code was off kilt, and run into all sorts of issues...

if you're playing with the reference client, it might be worth just trying to optimize prime.c (via CFLAGS override in Makefile) since that's where all the logic for calculations are - and you won't risk messing with any of the other important network code.

Will

96redformula

full member

Activity: 224

Merit: 100

I am starting to get orphan city using the a compiled build from github....Something must have happened, maybe optimized miners out now or GPUs being used. It seems a bit fishy.

Wolf0

member

Activity: 81

Merit: 1002

It was only the wind.

Quote from: Ethera on July 11, 2013, 03:47:03 PM

Quote from: Wolf0 on July 11, 2013, 03:40:00 PM

Quote from: Lauda on July 11, 2013, 03:32:31 PM

Quote from: Wolf0 on July 11, 2013, 03:31:15 PM

Quote from: Lauda on July 11, 2013, 03:24:38 PM

Quote from: Wolf0 on July 11, 2013, 03:23:37 PM

Quote from: Lauda on July 11, 2013, 03:20:25 PM

Has there been an update to the client so far?

Yes.

Where, link!

Github.

If i have to compile myself, it's rubbish.

Who knows, it could be a 1000x speedup. I would be careful before saying it's rubbish. Compile and see.

he means "i lack the knowledge to do anything like that, so liek you dont give me free mining so liek lol.. (idiocracy was a great movie, still watching it time to time to remind me what kind of society we are going for)...

Well, if he lacks the knowledge, he better go get it, because no one's going to cater to him.

mrmanager

newbie

Activity: 50

Merit: 0

Quote from: Zalfrin on July 11, 2013, 09:05:41 PM

Quote from: mrmanager on July 11, 2013, 09:00:40 PM

Quote from: Zalfrin on July 11, 2013, 08:56:10 PM

Just the processor architecture that the compiler optimized the code for. 11 might be faster but won't run on as many processors. 12 should be more widely supported, it's a more generic instruction set.

Your initial release is solid, and I've mined a few blocks with it. I haven't noticed any blockchain issues on any of my machines.

Even so, I'd recommend switching over to one of these new versions. Primespersec look to be about the same, and there's definitely something to be said for being on the official codebase versus the code I hacked together based on some optimization suggestions. That was actually done before Sunny released his sieve code update, which made the most difference by far anyways.

I should have been more specific. I'm using a build with the updates from Sunny. That update translated into 4x the speed.

Zalfrin

sr. member

Activity: 401

Merit: 250

Quote from: mrmanager on July 11, 2013, 09:00:40 PM

Quote from: Zalfrin on July 11, 2013, 08:56:10 PM

Just the processor architecture that the compiler optimized the code for. 11 might be faster but won't run on as many processors. 12 should be more widely supported, it's a more generic instruction set.

Your initial release is solid, and I've mined a few blocks with it. I haven't noticed any blockchain issues on any of my machines.

Even so, I'd recommend switching over to one of these new versions. Primespersec look to be about the same, and there's definitely something to be said for being on the official codebase versus the code I hacked together based on some optimization suggestions. That was actually done before Sunny released his sieve code update, which made the most difference by far anyways.

mrmanager

newbie

Activity: 50

Merit: 0

Quote from: Zalfrin on July 11, 2013, 08:56:10 PM

Just the processor architecture that the compiler optimized the code for. 11 might be faster but won't run on as many processors. 12 should be more widely supported, it's a more generic instruction set.

Your initial release is solid, and I've mined a few blocks with it. I haven't noticed any blockchain issues on any of my machines. I have a Sandy Bridge CPU working phenomenally with your i5 optimized build. The rest are using your first build. I don't know if I'm unlucky, but I've rebooted all of the machines that haven't found a block in a few hours.

Zalfrin

sr. member

Activity: 401

Merit: 250

Quote from: RustyShackleford1950 on July 11, 2013, 08:53:39 PM

Quote from: Zalfrin on July 11, 2013, 08:41:23 PM

Please download updated clients, these are using the official code base versus the "optimized" code on my branch which may be causing problems.

These replace my releases 5 and 8 respectively. Let me know if there are any other releases you were using.

https://www.dropbox.com/s/76eij0zcjx091am/release11_sandybridge_i5.zip
https://www.dropbox.com/s/lxm220qqztpgps5/release12_x86-64.zip

Thanks, what's the main differences between the two above?

Just the processor architecture that the compiler optimized the code for. 11 might be faster but won't run on as many processors. 12 should be more widely supported, it's a more generic instruction set.

Wolf0

member

Activity: 81

Merit: 1002

It was only the wind.

Quote from: Lauda on July 11, 2013, 03:32:31 PM

Quote from: Wolf0 on July 11, 2013, 03:31:15 PM

Quote from: Lauda on July 11, 2013, 03:24:38 PM

Quote from: Wolf0 on July 11, 2013, 03:23:37 PM

Quote from: Lauda on July 11, 2013, 03:20:25 PM

Has there been an update to the client so far?

Yes.

Where, link!

Github.

If i have to compile myself, it's rubbish.

Who knows, it could be a 1000x speedup. I would be careful before saying it's rubbish. Compile and see.

Zalfrin

sr. member

Activity: 401

Merit: 250

Quote from: hiddenvalleyranch on July 11, 2013, 08:50:06 PM

Anyone have an i5 haswell optimized windows binary? I'll shoot you some XPM.

I would probably try the x86-64 build (12), unless you see something on this list that is closer: http://gcc.gnu.org/onlinedocs/gcc/i386-and-x86_002d64-Options.html

RustyShackleford1950

sr. member

Activity: 266

Merit: 251

Quote from: Zalfrin on July 11, 2013, 08:41:23 PM

Please download updated clients, these are using the official code base versus the "optimized" code on my branch which may be causing problems.

These replace my releases 5 and 8 respectively. Let me know if there are any other releases you were using.

https://www.dropbox.com/s/76eij0zcjx091am/release11_sandybridge_i5.zip
https://www.dropbox.com/s/lxm220qqztpgps5/release12_x86-64.zip

Thanks, what's the main differences between the two above?

hiddenvalleyranch

member

Activity: 103

Merit: 10

Anyone have an i5 haswell optimized windows binary? I'll shoot you some XPM.

Zalfrin

sr. member

Activity: 401

Merit: 250

Please download updated clients, these are using the official code base versus the "optimized" code on my branch which may be causing problems.

These replace my releases 5 and 8 respectively. Let me know if there are any other releases you were using.

https://www.dropbox.com/s/76eij0zcjx091am/release11_sandybridge_i5.zip
https://www.dropbox.com/s/lxm220qqztpgps5/release12_x86-64.zip

tadakaluri

hero member

Activity: 616

Merit: 500