16 chips board from BitFury store only run at 25GHz though, while 16 chips board from burnin run at 40+ GHz at least.
With some experience and solder skill you can tune up "reference design" from shop up to 2.7Gh/s per chip. Or more using extra cooling and mod DC converter. with just a pencil those 2 reference boards (with 2 dead chips and one half dead) that where hashing at under 20 GH each ... went to around 58 GH ... and remeber the price for one h-board was at 350 € for the early october delivery ^^ speed:1652 noncerate[GH/s]:58.526 (1.829/chip) hashrate[GH/s]:58.746 +1 these boards are actually quite interesting/strange to work with. I had a poor performance chip recently start working almost like new (went tfrom effective 0.4GH to 1.7GH). These boards can be modified for 40% gains with good airflow and some small heatsinks (~35Ghash) and annecdotal reports indicate that the chips can handle some heat and the converter presumably has even more headroom. i would not be surprised if october h-boards were manufactured to run at 30GHash by default
|
|
|
Im using bitminter with difficulty of 16 for a starter and it works great. The card has a hashrate of 36.5GH with the noncerate (and pool reported) being 33.25GHash. help requested:
1) chip 6 is producing a few miso-errors (about 4%) plus the same number of hardware/stale errors. its pretty low, and only 2 other chips report miso-errors (1% and 2% rates). should i be concerned/fixing this? 2) a chip (varies time to time) has stopped a few times recently, usually in the first 5 minutes from reboot (if they start at all). Its running okay now after using stop miner/start miner via web gui, but is there a known reason for the issue or a solution? the chip is manually tuned at 54? 3) when tuning, what error rate should we go for to get the best result? Generally, ive tuned anything with <10% to run at 55, and over 10% i tuned to 54 or even 53. In most cases, tuning down to 54 seems to barely increase the effective noncerate from the drop in both hashrate and error rate. Is 10% errors a good threshold for tuning, or should i use lower (ie 5% or maybe 7%)?
The card is cooled somewhat aggressively, with tiny heatsinks on each chip (same size as chip) plus 5 slightly larger heatsinks (about 4x the surface area) centered on the back of the board, opposite the capacitor banks (that way, each sink is centered for the 4 nearest chips). a 120mm fan is aimed normal to the board about 5" away. nothing gets hot, just reasonably warm
|
|
|
Kit #1: 3 cards: 1: 30.609GH/s 9: 31.339GH/s 13: 29.335GH/s
Kit#2: 2 cards: Bank 1 1: 30.723GH/s 9: 30.093GH/s
i penciled the resistor to about 1230ohms.
Should I have done that or go lower?
I was able to do 1.18k Ohm... I've been steady at around 33Gh/s. I tuned down to 1.165k, getting around 35Ghash, and 33Ghash/noncerate at the pool. Amazingly, chip #13 which was running at ~300mhash with tons of errors now works just fine.
|
|
|
watch bitfury. they have some good headroom to overclock and are power efficient enough to battle many gen2 designs
|
|
|
Huh, I now read it as Blue PCB and Red LED. I'm sure it said Blue LED earlier...
it did, haha! Id love to get a few of these, but the price needs to be slightly lower i think, more in line with ~0.65 BTC for orders of 10+
|
|
|
if the 30A converter is the limiting factor on chip power (presumably up to about 0.84V), does setting a poor-performance chip (<0.6Ghash) "free up" ~6% more power headroom for the other chips to use? (for example, pushing them to 0.88V with added cooling)
|
|
|
Same problem here. My miner is running at 25 ghash, but Slush's stats show just 1.5 ghash. hey what pool(s) are people using for bitfury?
Slush's is my [previous] preference, and has a 30day luck >100% Bitminter has fewer fees + namecoin (this is about a 4% bonus, correct?), but the luck recently has been around 90%
my slush pool with bitfury is reporting only about 1000Mh per worker (3Ghash) why?
i noticed that in the last 2 days at some point they changed the accounts page and removed the ability to set a difficulty. perhaps that is related?
|
|
|
hey what pool(s) are people using for bitfury?
Slush's is my [previous] preference, and has a 30day luck >100% Bitminter has fewer fees + namecoin (this is about a 4% bonus, correct?), but the luck recently has been around 90%
my slush pool with bitfury is reporting only about 1000Mh per worker (3Ghash) why?
|
|
|
did the accounts page just get modified? I no longer see any namecoin info, and it does not seem to show the difficulty/suggested difficulty anymore...?
|
|
|
^That looks good what sort of heatsinks, tiny individual ones or larger plate(s)?
|
|
|
1 AIfDSo 55 1.532 1.807 107 5 0 0 171 $ 2 AIfDSo 55 1.890 1.871 132 8 0 0 177 $ 3 AIfDSo 55 1.747 1.807 122 7 0 0 171 $ 4 AIfDSo 55 1.460 1.691 102 14 0 0 160 $ 5 AIfDSo 55 1.417 1.829 99 14 0 0 173 $ 6 AIfDSo 55 1.518 1.797 106 3 0 0 170 $ 7 AIfDSo 55 1.661 1.871 116 14 0 0 177 $ 8 AIfDSo 55 1.546 1.733 108 16 0 0 164 $ 9 AIfDSo 55 1.947 1.850 136 8 0 0 175 $ 10 AIfDSo 55 1.904 1.871 133 6 0 0 177 $ 11 AIfDSo 55 1.718 1.776 120 13 0 0 168 $ 12 AIfDSo 55 1.589 1.744 111 12 0 0 165 $ 13 AiFDso 55 0.172 0.486 12 11 0 0 46 $ 14 AIfDSo 55 1.603 1.797 112 5 0 0 170 $ 15 AIfDSo 55 1.632 1.765 114 8 0 0 167 $ 16 AIfDSo 55 1.475 1.755 103 10 0 0 166 $ speed:825 noncerate[GH/s]:24.811 (1.551/chip) hashrate[GH/s]:27.449 good:1733 e$ 0: 825 24.811 27.449 1733 154 0 0 15 1 $
suggested tweaks? Obviously #13 is a poorly performing chip, should i disable it?
|
|
|
up and running now 15Ghash and climbing very nice! what did you have to do to get it going? re-download the image and redo the SD card? yeah, downloaded the full image, redid the SD card, edited the ip setting using VMware ubuntu system and the following: sudo sed -i 's/10.10.11.249/192.168.1.249/' etc/network/interfaces sudo sed -i 's/255.0.0.0/255.255.255.0/' etc/network/interfaces sudo sed -i 's/10.1.1.1/192.168.1.1/' etc/network/interfaces bitfury page reports 25ghash +/-2 and im waiting to see how slushs pool handles it
|
|
|
on further inspection, it seems dropbox connection got lost mid-download, resulting in what appeared to be a full .img file but only 1.6GB. Ive redownloaded and now have a 3.63GB file that is formatting now. crossing my fingers.
perhaps a secondary upload location (particularly with the 'correct' ip settings) would be a good idea?
got the full image now, and am working at 15+ Ghash and climbing. ill report back once it stabilizes at slushes pool
|
|
|
up and running now 15Ghash and climbing
|
|
|
on further inspection, it seems dropbox connection got lost mid-download, resulting in what appeared to be a full .img file but only 1.6GB. Ive redownloaded and now have a 3.63GB file that is formatting now. crossing my fingers.
perhaps a secondary upload location (particularly with the 'correct' ip settings) would be a good idea?
|
|
|
Since you were messing with it as an external device to modify the networking, I am wondering if that caused the problem? Since you have a keyboard/monitor now maybe try to re-burn the image fresh to the sd card and see if it works that way? Then you can change the networking again. I am not too familiar with linux so I am not sure. I am downloading the v2 and going to throw it on a spare RPi I have to see if it happens to me too. It will be about 30 minutes or so though because my game of League of Legends is now starting well i finally got a liveusb ubuntu going, sudo edited the ip fields (confirmed with explorer) and no luck. wiped the card and reinstalled the image with the bad ip - still no luck. i am re-downloading the image as a final resort, but this is really starting to bug me... edit: looks like i cant get a stable download. my first image file was only 1.6GB (expanded to about 3.69GB on the SD) and my re-download 'completed' at only 548mb. whats up with that? can someone post a different link maybe?
|
|
|
bringing my tech problems back to this thread (i got some fast replies last time):
i got a monitor and keyboard now to manually get the rpi running, and cannot even login.
It appears to load the entire code, and gets to
"(none) login:"
I enter pi as the username and get this response:
"unable to determine your tty name. Debian GNU/Linux 7 (none) ttyl
(none) login: "
whats going on?
Something wrong there and I think you better to redo the SD card again then try to login. I redid it twice now. my card appears to be supported for use (PNY 8GB microSDHC 10), and i am using win32diskimager
|
|
|
its all about the resale for these, personal use will never turn a profit sadly. I bought 8 of them for about 1.3 BTC (after tax, shipping, etc), and just resold 3 of them for a total 0.8 BTC, putting me at a relative cost of 0.1BTC per unit i still have. These have each mined almost 0.01BTC and are still up for sale
|
|
|
bringing my tech problems back to this thread (i got some fast replies last time):
i got a monitor and keyboard now to manually get the rpi running, and cannot even login.
It appears to load the entire code, and gets to
"(none) login:"
I enter pi as the username and get this response:
"unable to determine your tty name. Debian GNU/Linux 7 (none) ttyl
(none) login: "
whats going on?
ps: i apologize for harassing multiple threads. the bitfury US guide thread is the best place to reply, i just want to get my error seen so that someone can help me fix it. Its been over 24hrs that ive been unable to mine with my bitfury system because of this
|
|
|
bringing my tech problems back to this thread (i got some fast replies last time):
i got a monitor and keyboard now to manually get the rpi running, and cannot even login.
It appears to load the entire code, and gets to
"(none) login:"
I enter pi as the username and get this response:
"unable to determine your tty name. Debian GNU/Linux 7 (none) ttyl
(none) login: "
whats going on?
|
|
|
|