correction it wont allow the active x to install

pikajedi is running on windows 7 just fine - are you sure you gave it permission to run?

Also, did you have the CPU miner running about 90 minutes ago? I noticed a small amount of activity then but haven't been able to find who it was running it, and your question just now implies it may be you - if so, leave it running overnight and close it when you need to use the computer again.

81037/80284 khps

it says 8 shares submitted after about 10 minutes work. whatever that means

Here's me with my little old Q6600 doing  290k/262k



...Of course I also have a 5870 in there, but still, not bad for a 5 year old system.

Edit the shortcut, add a space and then -D after minerd.exe:
minerd.exe -D

Then start it again and give me the error message

That's easy, that's about 1-2 days earnings

That's more interesting - budget thoughts?

It was a "top of the head" thought.

It'd have to have oversight, and it'd have to be very carefully meted out.

Obviously we're not going to be generating enough to fly transatlantic (alas), but within the UK it could be used to pay up a % of railfare for the less well off or those further from the meeting point.

I have a top-end graphics card. At Idle (sort of) I draw 130w, When mining it's about 220-230w. Not quite doubling, but certainly an upswing.

Newer cards are, naturally, more efficient. GPU crunching is more efficient and quicker than CPU.

As Gareth mentioned, you can fiddle this - ATi cards can be overclocked, but they can also be underclocked - I would assume nVidia has the same.

a translation of that would be nice

There are multiple rounds, at the end of the round the number of shares contributed by any one miner (and the AFF pool is seen upstream as just one miner) is divided by the total number of shares contributed by all other miners to calculate the payout (each round is 50BTC).
At current rates, the AFF pool is generating about 2BTC per day.

Depending on what else you're doing with your system and depending on the difficulty on the bitcoin network you will see variations like this, it's nothing unusual at all.

I'm not sure if SLI is required for using OpenCL in that setup, but generally it should autodetect and use all available GPUs. You can always try it and see.

NOOOOO!

ONLY activate SLI if your main card is the EXACT SAME CHIP AND SPEED. 8800GTS 640mb cannot be SLI'd with 8800GTS 320 or 512, or an 8800GT or GTX. 8600s cannot be SLI'd with 8800s and so on.

This is an area wherein ATi has an advantage - any two cards from the same generation can be Crossfired - 5770 can be slaved to a 5870, for example.

They're actually different chips - the 8800GTS and GTX had a G80, the GT an 86, and then the second generation of the 8000 series had various iterations of the G90 and G92.

There are rare exceptions for "mixed SLI" configurations on some cards that only have a matching core codename - for example, earlier I said that the 8800GTS 640 and 320 could not be SLI'd - that's incorrect, since they're both G80s, but the GTS 512 is a G92 so it cannot be SLI'd. In this example, the 320 will become the master card and the 640 will slave itself at the lower speed.

Source (nVidia, slizone.com)

I've seen that one - it's what inspired my version, which by the way i'm happy to give source code out for to anyone who wants it (beyond what i'm legally obligated to do too - it's only partly GPL).

Does this mean...?

*BUMP*

Honestly, pikajedi is better to ask than me - but it should in theory be using all available GPUs. The easiest way to test is to enable it, check hashrate, and then disable it and check again while keeping all other conditions the same - if you get a higher hashrate with the nforce on, then obviously it's doing some good.

If it's not using it i'd consider that a bug in the client and will look into fixing it. I didn't write the java OpenCL binding myself though, and i'm not even sure if the java API will let me fix it if there is indeed a bug (personally I prefer PyCUDA). But i'll investigate either way.

It depends on multiple factors, including the current network difficulty - the best way to evaluate performance is the raw hashrate rather than shares submitted.

By the way, i'm pondering doing the first payouts manually on a trust system and then automating it afterwards - anyone who wants to claim their share i'll need a screenshot of you running the client showing the hashrate plus shares submitted.

You'll need to stick a bitcoin address into your profile (click user CP and edit profile). I'll then calculate a % of the amount generated so far with 40% going to people who have contributed as per my previous posting.

You could trivially cheat this, but I don't imagine any regulars here will. Even if someone does try to be cheeky about it, I will be able to detect more obvious cases of cheating - but again, I trust that nobody here is dishonest enough to pull that one. Once the automated system is in place it'll be cheat-proof and payouts will happen automatically.

I went to disable it for CUDA, and at the same time I displayed the browser window with the client running, and at the same moment I disabled it I saw the client report a result for that GPU!  "Block 287 found on nForce 980a/780a SLI (#2)".  That's the first time I've seen it report a result for it!  Was it BECAUSE I had just disabled CUDA?  I think it might have coincidentally happened JUST before I applied the change.  I immediately re-enabled it.  The client isn't bypassing CUDA somehow such that enabling CUDA would create a problem, is it?  So it's been used all along and I couldn't tell simply because I wasn't sitting there watching the client display at the right moment?  See, this is the problem: the client just isn't providing enough feedback.

The client doesn't keep track of total shares from one restart to the next, so how can we calculate it?  It starts over from 0 every time I restart it.  I assumed that information was being collected and saved on the server side.  I'm not really gunning to get "my share", though.  I thought the idea was a donation.

It uses CUDA, just with a different API - OpenCL is a cross-platform standard, but on nvidia chips it uses the same approach in the hardware as CUDA does. Since the client configures the GPUs at startup I would expect that if you didn't restart the client then it was likely a coincidence.

It will be collected server side eventually, don't worry about calculating the total yourself though - i'll go primarily off the hashrate when calculating the payout.
The setup we have is a bit complex as the AFF pool is essentially a metapool - it submits shares itself further upstream to one of about 4 other pools, only 1 of which actually tracks total shares. This allows switching to whichever pool is the most efficient at present - for example, the CPU miner is configured to use whatever upstream pool is paying out highest for lower hashrates at present.

When I automate this it'll go off your AFF forum account, so basically you make sure you're logged in on AFF and open the miner applet (i'll add something to the UI to display your username to avoid people accidentally doing it while logged out) and the server will track shares per user. There's no need to track shares on the client side, but if people want this it's trivial to implement.

Only per person - the bitcoin address is how you'll receive your payout, you can then choose to keep those coins or you can sell them for USD or spend them on a site like bitmunchies.com (I recommend their beef jerky). To get a bitcoin address just download the client from bitcoin.org and run it - it'll generate an address for you which you can stick in your profile here.

Indeed I didn't restart the client... yet.  That would have been my next step.  I didn't see a noticeable change in the hashrate just from disabling CUDA on #2 without restarting the client, but then the hashrate is kindaall over the place any way!  It ranges anywhere from 10000 to 25000.  It gyrates even while I'm just glancing at it.

I have no idea how much money is involved in translation, but I assume it's "bus fare" money.  Since my free Aspie Affection Gold account expired, maybe something could be done about that instead?     I was never even vaguely close to hooking up, but what the heck.

Since I installed and briefly ran the standard CPU client, I guess I already have an assigned address.

Take a screenshot when your client is running at peak and i'll go by that.
As for the address - you're correct, the client will generate one for you - just make sure you take care of wallet.dat (it's somewhere in c:\users\yourname\Application data\bitcoin on windows) - if you lose that you lose the coins.

Current exchange rate is about $7 per bitcoin. Check yourself at mtgox.com

Wow!  And how many shares per bitcoin?

I haven't been running the standalone CPU client except once for a few hours, and with my Phenom it didn't even generate 0.01!  So there aren't any coins from that.  Still, it sounds like 6 months of this could be serious money if the exchange rate doesn't tank (which it might just like my housing market ).

---

On a vaguely related GPU note, I finally decided to try using my soldering-rework station to try to permanently fix the GEForce Go 7900 GS video module from my laptop.  You know... one of the same ones that spurred the class action lawsuit against NVidia and HP?  Well, my laptop is a Toshiba, and the b4st4rds somehow managed to avoid being a defendant in the lawsuit.

I got the thing apart, was cleaning off the paste etc. before deciding how to proceed with the heat gun, when I noticed that the BOARD IS WARPED, by 2 or maybe even 3 millimeters!  This is a small board, too.  I've had it disassembled to this point before, and I'm embarrassed I missed it.  Long story short, I've postponed any heat gun fun so that I can first call Toshiba on Monday and then a lawyer if need be.  A warped board with BGA chips on it is a defect they'd better not ignore.

So, the second one is averaged over the entire runtime it looks like

If they are both G92s it SHOULD be working. Easy way to tell; Note down how many kh/s you're doing, deactivate the onboard, see what the difference is.

SC2 isn't as intensive as SC:FA was. Granted, the system still takes quite a beating, but still...

The 8800 runs hot. Very hot. So hot that my old 8800GTS would actually cause the rest of the system to overheat (and this in a midi-case).

Out of interest; What AV are you running?

where in the profile do you want our bitcoin address?

I have SC:FA, too (and tons of goodies), and that's been played plenty on this system as well (just not lately, SC2 is is some ways more approachable).  No troubles with that, either.  Well, not after I applied the /3GB switch when I was still using Windows XP.

I know.  I think I mentioned my special RivaTuner cooling profile?  I've re-pasted it with Arctic Silver as well.  I've always intended to shift back to watercooling at some point (have an old Koolance), but the wallet isn't fat enough.  I could mod it with a better fan (there's case room), but that might hurt my chances of getting a few bucks for it later on Craigslist.

I'm using MSE.  Yeah, yeah, I know, it's not hip, but it's unobtrusive.  Actually I've had the realtime monitoring disabled for months, so that can't be it, either.

...No, not like that, it simulates the movement a physics of, well, hair and fur.

Well I might try that, but not today since I'm busy trying to destroy the GPU and heatsink from my laptop.  

Why?

It's weird. I looked for the driver for my Nvidia quadro FX 2800m and all I saw was the driver for the 2700m.

VulcanTourist Wrote:

Well I might try that, but not today since I'm busy trying to destroy the GPU and heatsink from my laptop.  

Why?

I was being literal (imagine that).  The destruction part wasn't intentional.  You do NOT want to try to mess with a heat pipe SOLDERED to a heat sink and fins.

It had apparently overheated enough at one point that the heat pipe had partially separated from the heat sink.  It needed to be reattached, and I foolishly figured the heat gun could handle it.  I was wrong.  I made such a *$&#@^%@ mess of things trying to get it all reattached I'm not sure it will ever work again.  Didn't really even get to trying to reflow the chips... probably a GOOD thing considering how wrong the rest of this went.  I've still got at least another hour removing the oxidation on the heatsink, from having to stick the whole contraption in the oven at 475 as a last resort.  The fins on one end and heat sink on the other was an impossible job to do with a heat gun.

I should have stuck to what I said earlier and called Toshiba first on Monday.  Sure, they most likely would/will blow me off and I'd have just repeated these same mistakes a couple days later, but it doesn't make me any less disgusted with myself.  I can't afford those mistakes.  I also had to stop in the middle of all that to try to repair a dead Dremel tool; just when I thought I'd figured out and solved the problem and was trying to reassemble it, I broke off a critical internal pin and wasted the whole effort.  I'll have to buy a new one I can't afford.

I'm not feeling very proud right now.  

Acetone can be used to clean heatsinks of oxidization, and quickly.

Hmmm, why was I under the impression they were in difficulty..?

Ah, I misread - you'd be surprised though, I've spoken to people who've tried to do stuff to sinks whilst they're still on the board.

Those other people aren't ME!  I make mistakes, but the ignorance that causes them is usually minor rather than profound.

Except in social matters... there my ignorance is pretty profound.

It is?!

Acetone is great stuff - I'm making Fibreglass armour at the moment, and it's about the only thing I've found that lets me just rub resin off of skin or clothes. I've also used it to clean heatsinks in the past - both the contact surface and, because I have an unoxidised copper sink, cleaning the fins (though I suspect my dremel might be faster at this, but that might damage it).

Just....for gods sake, don't oxidise it. It's more unstable than Nitroglycerin.

If there's something to flush, then yes.

*snerk*

DFG, we have no way of telling because we haven't implemented Per User-tracking. If it says it's doing x number of kh/s, it's working.

I didn't see this post. I may have been typing as you were posting this. I go afk a lot. 0oops. The console screen say k/hashe/sec. So it is working.

Do you mean there's a current problem with it or you're finding it difficult to use?

getting the message that it is not hooking up with the bitcoin pools, on several different computers

Specific symptoms?
Low hashrate or what?

The new server is a VPS, so it may not be taking the load of database requests every 10 seconds as well as the old one. If the pool gets more profitable I could put a new server up, for now though the income generated can just about cover annual domain fees.

hash rate goes to zero and it just sits there, until refreshed. Then it repeats the same.

External upstream pool dieing it looks like - deepbit has been having issues lately and most of the traffic heads there

^ ^     **bump**

It quit on sunday nite

thanks.

Java Runtime Enviroment + google

edit. 4 pages already? Guess I'm late...

Yeah, having more problems than I thought.
The idea was to move this from the pool setup we've been using that's very database intensive to a proxy relay - of course nginx isn't liking bitcoin for whatever reason.

I'll get this fixed somehow - just have to do so in a way that doesn't use too many server resources as the hosting provider dislikes that.

When I  first set this up I used my personal wallet which is backed up on a regular basis, I then manually tracked what bits belong to AFF and what belongs to me personally. Now i'm moving it to a dedicated AFF wallet on the server which will be backed up with everything else.

may I ask what is your method of backup for your personal wallet?

Paper printouts of the crypto keys

LOL, and I thought it refreshed because I was sitting there staring at the screen and going refresh, refresh. ( too lazy to walk over to the computer )

Hy guys,

nice programm. I would really like to so see the source, can you upload it again plz?

Greetings, Alex

nginx index is a bit broken but i'll happily email it to anyone who asks.
gareth@garethnelson.com

Last person on earth who'd violate the GPL is me.