Bitcoin Forum
May 24, 2024, 12:38:47 PM *
News: Latest Bitcoin Core release: 27.0 [Torrent]
 
  Home Help Search Login Register More  
  Show Posts
Pages: « 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 [16] 17 18 19 20 21 »
301  Other / CPU/GPU Bitcoin mining hardware / Re: OC'd video cards at ~70% fans for 9+ months sustainable? on: August 26, 2011, 03:08:57 PM
If the ref cards fan doesn't die within the first month or so (= manufacturing defect), it'll probably last near forever.
The sapphire extreme 5830/5850... not so much, those have really cheap sleeve bearing fans. I'd be surprised if it lasted much over 6 months at 70%.
302  Other / CPU/GPU Bitcoin mining hardware / Re: I killed my 5970 on: August 26, 2011, 12:07:36 AM
Right under Products -> Cooling -> VGA -> Accelero XTREME 5970 -> Support

"Advice for installation on ATI HD 5970"
http://www.arctic.ac/en/support/support/detail/question/95-advice-for-installation-on-ati-hd-5970.htm

"Manual : Accelero XTREME 5970"
http://www.arctic.ac/index.php?eID=tx_mm_bccmsbase_zip&id=14140256894e56258d0a992

And yeah, the backplate isn't held on by anything once you remove the screws.
303  Other / CPU/GPU Bitcoin mining hardware / Re: Will my 850W PSU power 5970 + 5870 + 5870? on: August 25, 2011, 11:57:12 PM
If they aren't overvolted... I'd say "yes, but".
The AX850 is a pretty good PSU, just look at JGs review Wink
OCed 5870 at stock V is 200-220W, like 180 with mem @ 300
OCed 5970 at stock V is ~300W, about 270 with mem @ 300
Stock wattage is ~ the same as OCed with underclocked mem.
So that's < 700W load on a PSU rated for 840W continous on 12V.
now, theres the "but" part... simple, I wouldn't want to run a PSU at > 80% rated 24/7 for months.
304  Other / CPU/GPU Bitcoin mining hardware / Re: I killed my 5970 on: August 25, 2011, 11:46:00 PM
Wait, you tried installing a accelero 5970 with NO INSTRUCTIONS?

Sounds like you killed your card improperly installing it then.
The installation instructions that come with it (and are also available on ACs website) state VERY clearly to leave the backplate on and to *NOT* fully tighten the 4 x-plate screws around each GPU or the screws mounting the main cooler to the board (it's ok to fully tighten the 2 screws holding the VRM heatsink).

Overtighten the x-plates and you kill your DVI outputs.
Overtighten the standoffs (especially with no backplate) and you bend the board so badly it's usually a paperweight.
305  Other / CPU/GPU Bitcoin mining hardware / Re: 1500w silverstone psu safe for 4 x 5970? on: August 25, 2011, 09:47:57 PM
5970s at stock V are < 300W.
at 850/300 and stock V they're also < 300W.
930/1000 at 1150mV ~400W each
So that PSU should be fine as long as you don't overvolt em.
306  Other / CPU/GPU Bitcoin mining hardware / Re: Problems joining two PSU's (old parts) is a 300w PSU enough for a 5770? on: August 25, 2011, 09:44:31 PM
Yep, does it power on with just green/black shorted?
Also, it's generally a VERY bad idea to only load old/cheap 300/350W PSUs on +12.
Those are pretty much all group-regulated designs taking feedback from +5V or +3.3V -> load +12 significantly with +5/+3.3 unloaded and +12 ends up *way* low, I've personally seen < 9.5V with 60% rated load on +12 and nothing on +5/+3.3.
307  Alternate cryptocurrencies / Altcoin Discussion / Re: New Ixcoin fork -> I0coin on: August 25, 2011, 07:05:21 AM
Something like this should do the trick (disable the broken one-week crap, switch to solidcoin algo after block 14639).

Code:
diff -Nur I0coinClientv1.0-linux-stable-orig/src/main.cpp I0coinClientv1.0-linux-stable-dev/src/main.cpp
--- I0coinClientv1.0-linux-stable-orig/src/main.cpp 2011-08-18 23:15:50.000000000 +0200
+++ I0coinClientv1.0-linux-stable-dev/src/main.cpp 2011-08-25 08:23:52.143725519 +0200
@@ -644,7 +644,7 @@
     return nSubsidy + nFees;
 }
 
-unsigned int static GetNextWorkRequired(const CBlockIndex* pindexLast)
+unsigned int static GetNextWorkRequired_OLD(const CBlockIndex* pindexLast)
 {
     const int64 nTargetTimespan = 7 * 24 * 60 * 60; // two weeks
     const int64 nTargetSpacing = 5 * 60;
@@ -659,6 +659,7 @@
     // Only change once per interval
  if ( nRemaining != 0)
  {
+/* HORRIBLY BROKEN, *NEVER* use time() in here
  const CBlockIndex* pindexFirst = pindexLast;
  for (int i = 0; pindexFirst && i < nRemaining-1; i++)
  pindexFirst = pindexFirst->pprev;
@@ -666,6 +667,7 @@
 
  int64 rema = GetAdjustedTime() - pindexFirst->GetBlockTime();
  if(rema < nTargetTimespan)
+*/
  return pindexLast->nBits;
  }
 
@@ -691,6 +693,71 @@
 
     if (bnNew > bnProofOfWorkLimit)
         bnNew = bnProofOfWorkLimit;
+
+    /// debug print
+    printf("GetNextWorkRequired RETARGET\n");
+    printf("nTargetTimespan = %"PRI64d"    nActualTimespan = %"PRI64d"\n", nTargetTimespan, nActualTimespan);
+    printf("Before: %08x  %s\n", pindexLast->nBits, CBigNum().SetCompact(pindexLast->nBits).getuint256().ToString().c_str());
+    printf("After:  %08x  %s\n", bnNew.GetCompact(), bnNew.getuint256().ToString().c_str());
+
+    return bnNew.GetCompact();
+}
+
+//blatantly stolen from SolidCoin
+unsigned int static GetNextWorkRequired(const CBlockIndex* pindexLast)
+{
+    const int64 nTargetTimespan = 12 * 60 * 60; // 12 hours
+    const int64 nTargetSpacing = 3 * 60;    //3 minute blocks
+    const int64 nInterval = nTargetTimespan / nTargetSpacing;
+
+    // Genesis block
+    if (pindexLast == NULL)
+        return bnProofOfWorkLimit.GetCompact();
+
+//okay, maybe not this line
+    if ((pindexLast->nHeight+1) < 14640)
+        return GetNextWorkRequired_OLD(pindexLast);
+
+    // Only change once per interval
+    if ((pindexLast->nHeight+1) % nInterval != 0)
+        return pindexLast->nBits;
+
+    // Go back by what we want to be 14 days worth of blocks
+    const CBlockIndex* pindexFirst = pindexLast;
+    for (int i = 0; pindexFirst && i < nInterval-1; i++)
+        pindexFirst = pindexFirst->pprev;
+    assert(pindexFirst);
+
+    // Limit adjustment step
+    int64 nActualTimespan = pindexLast->GetBlockTime() - pindexFirst->GetBlockTime();
+    int64 nTwoPercent = nTargetTimespan/50;
+    //printf("  nActualTimespan = %"PRI64d"  before bounds\n", nActualTimespan);
+
+    if (nActualTimespan < nTargetTimespan)  //is time taken for a block less than 3minutes?
+    {
+         //limit increase to a much lower amount than dictates to get past the pump-n-dump mining phase
+        //due to retargets being done more often it also needs to be lowered significantly from the 4x increase
+        if(nActualTimespan<(nTwoPercent*16)) //less than a minute?
+            nActualTimespan=(nTwoPercent*45); //pretend it was only 10% faster than desired
+        else if(nActualTimespan<(nTwoPercent*32)) //less than 2 minutes?
+            nActualTimespan=(nTwoPercent*47); //pretend it was only 6% faster than desired
+        else
+            nActualTimespan=(nTwoPercent*49); //pretend it was only 2% faster than desired
+
+        //int64 nTime=nTargetTimespan-nActualTimespan;
+        //nActualTimespan = nTargetTimespan/2;
+    }
+    else if (nActualTimespan > nTargetTimespan*4)   nActualTimespan = nTargetTimespan*4;
+
+    // Retarget
+    CBigNum bnNew;
+    bnNew.SetCompact(pindexLast->nBits);
+    bnNew *= nActualTimespan;
+    bnNew /= nTargetTimespan;
+
+
+    if (bnNew > bnProofOfWorkLimit)
+        bnNew = bnProofOfWorkLimit;
 
     /// debug print
     printf("GetNextWorkRequired RETARGET\n");
308  Alternate cryptocurrencies / Altcoin Discussion / Re: New Ixcoin fork -> I0coin on: August 25, 2011, 06:08:54 AM
The one week part of "one week or 2016 blocks" retargeting is b0rked.
Downloading the chain with a fresh datadir makes it try to retarget at block 1.
Then it tries to go back 2016 blocks. from block 1. Boom, NULL pointer.
With a client that already has the blockchain up to some point in the past, it'll start disagreeing on what the target should be (again, mistrigger of the after-7-days stuff)... around now.
Easy Fix: pull a Thomas and blatantly rip off solidcoins retargeting algo.
309  Other / CPU/GPU Bitcoin mining hardware / Re: Sapphire 5850 Fan Dead on: August 25, 2011, 02:11:20 AM
If they're anything like the fans on these: http://www.newegg.com/Product/Product.aspx?Item=N82E16814102873
Once they start "growling" the single sleeve bearing (!) is totally shot. Takes about 3-5 months at 100%.
And there's no way to lubricate them.
Whoever decided to put a cheap-ass sleeve bearing fan right on top of the hottest part of the heatsink was a true genius.
310  Economy / Computer hardware / Re: Custom FPGA Board for Sale! on: August 25, 2011, 01:19:42 AM
That was a loooong time ago, nowadays it's < 1%.
311  Economy / Computer hardware / Re: Custom FPGA Board for Sale! on: August 24, 2011, 11:42:17 PM
50 != bulk.
Hell, that's not even one tray.
312  Bitcoin / Bitcoin Discussion / Re: Can the same bitcoin be spent twice (or more) in one block? on: August 24, 2011, 12:42:21 AM
Clients don't spend 0/unconfirmed coins they get from somewhere else.
They will spend their own 0/unconfirmed change if left no other option.
Original reason for that, "viral bitdust":
A transaction that split 0.01 btc into 100s of tiny outputs and had no fee, so it staid at 0/unconf "forever".
Clients that received one of the outputs sooner or later used them as inputs in their own transactions... and as a tx can only go into a block if all its inputs are in the same or older blocks...
Yeah.
313  Alternate cryptocurrencies / Altcoin Discussion / Re: Seriously guys, why? on: August 23, 2011, 04:31:30 AM
Simple, anyone mining back then and thinking like that most likely sold at $1.
314  Alternate cryptocurrencies / Altcoin Discussion / Re: [ANNOUNCE] SolidCoin - new and improved block chain. Secure from pools on: August 23, 2011, 01:12:48 AM
Is this evaluation ok?
Does it have no correlation with about 32SC ~ 3min and 50BTC ~ 10min ?
Target time/block doesn't enter this calculation (it only affects what difficulty a network will end up at at a given hashrate).
Just block reward (thats where the 32 and 50 come from), difficulty and exchange rate.
another way to put per-block reward, difficulty and exchange rate together:

1 share on BTC gets you about 50/1806098 BTC = 0.000027684 btc/share
1 share on SC at a exchage rate of 0.0017btc/sc gets you ... 32/2703*0.0017 BTC = 0.000020126 btc/share
0.000020126 / 0.000027684 = 0.726990319 -> you're making about 27.3% less mining SC and selling em than mining BTC.
now... what exchange rate do we need to come out even?

50/1806098 = 32/2703*X
or...
X = (50 / 1806098) / (32 / 2703)
X = 0.002338432
you can also swap things around to suit you
X = (2703 / 32) / (1806098 / 50)
X = 2703 / 1806098 * 50 / 32
or replace the * 50 / 32 with * 1.5625
X = 1.5625 * 2703 / 1806098
... and it's still
X = 0.002338432
Wonders of arithmetic!
315  Alternate cryptocurrencies / Altcoin Discussion / Re: [ANNOUNCE] SolidCoin - new and improved block chain. Secure from pools on: August 23, 2011, 12:21:11 AM
Oh i see.  Excellent explanation. Thanks again Smiley  So it seems with my 2 GH/s i make slightly more mining solidcoins if the price is 0.0017 BTC/SC.  I guess I'll wait 12 hours like you suggested and check it out then.
Err... No.
At the time of your post, SC difficulty was about 2700

(2.7k/32)/(1807k/50) = 0.00233467

that's what btc/sc would have to be to end up +-0.

at 0.0017btc/sc, you're making roughly 30% less mining SC.
316  Other / CPU/GPU Bitcoin mining hardware / Re: What's the purpose of providing 12v to the pcie extender ? is 5 GPU too much ? on: August 22, 2011, 11:55:59 PM
24pin ATX has 2 +12V pins and 8 GND pins, rated for 6A each.
So thats a max current of 12A on the 12V side, 48A on the return.
That's the short version of why splitting out GND isn't needed.
317  Other / CPU/GPU Bitcoin mining hardware / Re: Regarding the theoretical maximum Performance of GPUs on: August 20, 2011, 06:59:55 PM
You realize a bitcoinhash is *2* sha256 blocks operations, right?
Well, not exactly 2 thanks to some optimizations possible
you can drop the last 3 rounds completely (they don't change H), and lose part of the previous round (you only need the E output of the 4th-to-last)
Initial rounds can be optimized as well, as the last DWORD of hMerkleRoot and nTime/nBits don't change between loops, so you can drop the equivalent of ~3 rounds there as well.
Same thing goes for optimizing/precalculating parts of the W mangling, as we're feeding in quite a bit of constants.
Register access is basically free on GPUs (they mask reg r/w by pipelining 4 "threads" on the shader pipeline).
Ch() can be done in 1 cycle, and Maj() in 2.
Also, what LUT accesses? just hardcode the K constants in the instruction stream.

So while you came up with a somewhat reasonable result, you did so by pure chance using invalid assumptions and numbers.
318  Bitcoin / Mining / Re: Computer freezes every few days, help! on: August 20, 2011, 06:47:06 PM
I'd pile onto the "OC too high" heap, but...
Look at the gpu-z log, that card is running stock 725/1000 clocks at stock voltage.

Edit:
core and vddc temps look ok, maybe a tad high for my taste.
But 30% fanspeed? what do the VDDC temps on the slave GPU look like?
319  Bitcoin / Bitcoin Discussion / Re: What version of the Bitcoin client should we be using right now? on: August 19, 2011, 06:22:22 PM
Well, or just make sure to always leave at least a 0.01 balance or empty your wallet completely.
320  Bitcoin / Bitcoin Discussion / Re: selling testnet BTC. on: August 19, 2011, 03:43:37 PM
There you go, now the faucet is giving away 500 again Wink
Pages: « 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 [16] 17 18 19 20 21 »
Powered by MySQL Powered by PHP Powered by SMF 1.1.19 | SMF © 2006-2009, Simple Machines Valid XHTML 1.0! Valid CSS!