X6500 Custom FPGA Miner

[cypherdoc]

You need mini-USB to USB cable (http://www.amazon.com/AmazonBasics--Male-Mini-B-Cable-Meters/dp/B001TH7GUK/ref=sr_1_2?ie=UTF8&qid=1327971683&sr=8-2).

i have a bunch of these from my Blackberry's of the past.  will they work just as well?

[1714254327]

1714254327

[1714254327]

1714254327

[1714254327]

1714254327

[1714254327]

1714254327

[1714254327]

1714254327

[teflone]

Yes, a lot of compound..

Rule of thumb is to never put it on in such a way as a bubble can form on the chip..

Thats why a circular blob in the middle is the correct method to put on and it squeezed out evenly

[TheHarbinger]

Yes, a lot of compound..

Rule of thumb is to never put it on in such a way as a bubble can form on the chip..

Thats why a circular blob in the middle is the correct method to put on and it squeezed out evenly

Not even a blob, more like a drop.  The compound is only meant to fill in surface imperfections between the chip and the heat-sink.  Personally, I use a vinyl applicator to apply a very, very thin film across the top of the chip and back of the heat-sink, then mate them together.  I mean so thin it becomes semi transparent.

[nbtcminer]

Criticism noted:

I wasn't going to get another chance to apply thermal grease on the outer-rim of the unit and from reading / testing, it looked like a plausible solution. However I will agree that maybe too much epoxy / adhesive went on. Performance is tapering off at around 390mh/s so I don't imagine it affected my performance much. Going to take it down for the night to cool and after 12 hours see how it performs again.


Cheers,

nbtcminer

Yes, a lot of compound..

Rule of thumb is to never put it on in such a way as a bubble can form on the chip..

Thats why a circular blob in the middle is the correct method to put on and it squeezed out evenly

[TheSeven]

A new mining software that supports the X6500 board has entered public beta: https://bitcointalk.org/index.php?topic=62823.0
This might finally get us rid of those nasty d2xx drivers/libraries. Feel free to try it out and report possible issues in the thread linked above.

[nbtcminer]

@Everyone running the 200 mhz bitstream:

Are you seeing speeds above 400mh/s over a 24 hour period? my mh/s seems to be dipping to 380 mh/s and going as high as 450mh/s during certain points in the day (seemingly random). The miner shows I'm hitting 400mh/s on average but my mining results are showing another thing. Maybe I'm just experiencing a lucky period?

Also seeing a small amount of duplicates on btcguild and deepbit atm. They seem to randomly pop up (the miner seems to pause for a second, then reports an invalid share)

[TheSeven]

@Everyone running the 200 mhz bitstream:

Are you seeing speeds above 400mh/s over a 24 hour period? my mh/s seems to be dipping to 380 mh/s and going as high as 450mh/s during certain points in the day (seemingly random). The miner shows I'm hitting 400mh/s on average but my mining results are showing another thing. Maybe I'm just experiencing a lucky period?

Also seeing a small amount of duplicates on btcguild and deepbit atm. They seem to randomly pop up (the miner seems to pause for a second, then reports an invalid share)

x6500-miner calculates the hash rate by counting shares, so it's affected by some major variance. What you're observing is most likely just luck, the FPGA will always be running at a constant exact 200MH/s with that bitstream, possibly with some calculation errors though if it isn't cooled properly.

Duplicates could be a result of your actual getwork interval (including things like network latencies and possibly some lag here and there) exceeding 20 seconds. Try setting that interval shorter (to like 12-15 seconds) and check if the duplicates go away. Or just use MPBM which handles all of this automatically

[nbtcminer]

@theseven:

Cool beans, I'll give your miner a try tonight when I get back! So far I think getworks intervals / network latency are my main issues now.


Cheers!


Edit:
This is how things are looking 12 hours later:

Code:

Run Summary:
-------------
Device: 0
Number of FPGAs: 2
Running time: 12h14m
Getwork interval: 15 secs
FPGA 0:
  Accepted: 2118
  Rejected: 1 (0.05%)
  Invalid: 0 (0.00%)
  Hashrate (all nonces): 206.96 MH/s
  Hashrate (valid nonces): 206.96 MH/s
  Hashrate (accepted shares): 206.28 MH/s
FPGA 1:
  Accepted: 2035
  Rejected: 8 (0.39%)
  Invalid: 8 (0.39%)
  Hashrate (all nonces): 200.14 MH/s
  Hashrate (valid nonces): 199.36 MH/s
  Hashrate (accepted shares): 198.19 MH/s
Total hashrate for device: 407.11 MH/s / 406.33 MH/s / 404.48 MH/s

@Everyone running the 200 mhz bitstream:

Are you seeing speeds above 400mh/s over a 24 hour period? my mh/s seems to be dipping to 380 mh/s and going as high as 450mh/s during certain points in the day (seemingly random). The miner shows I'm hitting 400mh/s on average but my mining results are showing another thing. Maybe I'm just experiencing a lucky period?

Also seeing a small amount of duplicates on btcguild and deepbit atm. They seem to randomly pop up (the miner seems to pause for a second, then reports an invalid share)

x6500-miner calculates the hash rate by counting shares, so it's affected by some major variance. What you're observing is most likely just luck, the FPGA will always be running at a constant exact 200MH/s with that bitstream, possibly with some calculation errors though if it isn't cooled properly.

Duplicates could be a result of your actual getwork interval (including things like network latencies and possibly some lag here and there) exceeding 20 seconds. Try setting that interval shorter (to like 12-15 seconds) and check if the duplicates go away. Or just use MPBM which handles all of this automatically

[O_Shovah]

Just for quick notice.

While running the 200mhz bitstream on my x6500 i hear a high pitched squiiqing noise from the voltage regulators.
I havent determinded yet wich one it is. I will also check vie my thermal camera to be shure.

Has anybody recognized similar behavior ?



I also wanted to point out the

Modular Python Bitcoin Miner
from TheSeven

It supports the x6500 by default and gives a nice interface
and a lot of additional features.

Please have a look.
https://bitcointalk.org/index.php?topic=62823.0

[freshzive]

I've noticed that noise as well, though it seems to only happen some of the time or maybe with only a few of my boards. Going to test out the modular miner. Cheers!

[thirdlight]

I have the modular miner working on X6500's, on windows. I had to install libusb-win32, which stopped the "official" software working. MPBM seems to be working well, even acceptably with p2pool. TheSeven suggests you look at the options detailed in each script, I would say it's a requirement! You'll probably want to set at least one option from /worker/fpgamining/x6500hotplug.py

It programs the boards, & then mines. It's fault tolerance seems quite good - it'll keep trying until it succeeds. 

[TheSeven]

I have the modular miner working on X6500's, on windows.

Which one of the curses libraries are you using?

It programs the boards, & then mines. It's fault tolerance seems quite good - it'll keep trying until it succeeds. 

That's only true if you use the hotplug module at the moment, which should work well for most users.
I'll need to think of some tricks to make autorecovery work properly with individual board modules as well.

[shad]

had MPBM running over night without problems on openWRT

[thirdlight]

Which one of the curses libraries are you using?

Your second suggestion, the one at http://www.lfd.uci.edu/~gohlke/pythonlibs/#curses

[nbtcminer]

had MPBM running over night without problems on openWRT

@Shad:
 Very exciting news!!! That would be awesome (mining via an X6500 through a router is a huge WIN in terms of energy savings!)

@OShovah:
I don't get the high pitched noise, but then again I'm only a few days in with my card. How are you cooling your card? I've got a 120mm fan directly blowing air down onto the card (the fan is right against the heatsinks). Most CPU's are cooled in this fashioned and I figured it would work for the x6500.


Cheers,

nbtcminer

[fizzisist]

I think we can all agree that MPBM is awesome! I think MPBM will become our recommended miner, and we will support and develop it into the future. We'll most likely be phasing out development on x6500-miner over time.

The documentation guides will be updated (and simplified, hopefully) for the switch over to MPBM. This will also give me the chance to make a more clear and detailed tutorial (read: screenshots ).

Also, we are now sold out of boards from this batch. We're about to begin production on X6500 rev. 3, and expect to be shipping in 3-4 weeks. This new revision is going to have some nice additions that I think will be a big step forward! I'll give a list of what's changing and what's staying the same in another post soon.

[nbtcminer]

@ Fizzisist:

Will a 12v / 5v molex adapter (2.5a) work for powering the X6500 and a fan through the molex connector? Here is what I'm thinking of using:

http://www.ebay.ca/itm/160715008901?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1497.l2649

I'm planning on putting together an enclosure for my X6500 and its current setup with isn't ideal. I'm looking to put a molex Y splitter so that I can power both a 120mm casefan and the X6500 off the adapter I've listed above. This way I can have the Y adapter mounted closer together to the USB port so that I can build some sidings around the case fan / bottom mounting plate. Thanks for your help so far!


Cheers,

ntbcminer

[TheSeven]

Will a 12v / 5v molex adapter (2.5a) work for powering the X6500 and a fan through the molex connector? Here is what I'm thinking of using:

http://www.ebay.ca/itm/160715008901?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1497.l2649

Might work, but I've seen some of those which are designed in a way that makes them behave badly when only one of the voltages is used (due to them using a single transformer for both voltage outputs). Also the one I have isn't particularly efficient (it gets rather hot, probably wastes more energy than it actually outputs), so that might not be the ideal solution for an energy-efficient FPGA mining board

Depending on how the x6500 is wired, it might be possible to power it using a standard barrel connector wall wart, and take power for the fan from the molex connector on the board. You'll probably have to use a 5V fan adapter though, because the barrel connector is wired to the 5V rail of the molex IIRC.

[nbtcminer]

@Theseven: That's a very interesting idea! I wasn't sure if the on-board molex was capable of supplying power to the fan, but it sounds worthy of a try. My only concern would be that the on-board molex adapter isn't wired properly for any reverse supply of power from the barrel connector (which might damage my x6500). BTW cudos to you for making MBPM! I'm still struggling with getting it to work on my computer at home, but hopefully will have it working soon P.S. I'm currently using the barrel connector (I guess wall wort lol) I got from cablesauraus's website and it tends to get pretty warm as it is. Probably about 30-40 degress C, very likely not much different from yours.

@Fizzisist:
Sorry to bombard you with so many questions; is what theseven proposing possible and will it cause any damage to the unit to draw power from the molex connector while powering the unit from the barrel connector? I'm using the 12v 2a barrel connector on cablesauraus's website.

Will a 12v / 5v molex adapter (2.5a) work for powering the X6500 and a fan through the molex connector? Here is what I'm thinking of using:

http://www.ebay.ca/itm/160715008901?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1497.l2649

Might work, but I've seen some of those which are designed in a way that makes them behave badly when only one of the voltages is used (due to them using a single transformer for both voltage outputs). Also the one I have isn't particularly efficient (it gets rather hot, probably wastes more energy than it actually outputs), so that might not be the ideal solution for an energy-efficient FPGA mining board

Depending on how the x6500 is wired, it might be possible to power it using a standard barrel connector wall wart, and take power for the fan from the molex connector on the board. You'll probably have to use a 5V fan adapter though, because the barrel connector is wired to the 5V rail of the molex IIRC.

[shad]

had MPBM running over night without problems on openWRT

@Shad:
 Very exciting news!!! That would be awesome (mining via an X6500 through a router is a huge WIN in terms of energy savings!)

there is some work to do to make it perfect, daemon(ize) it, webpage or sendmail option to keep an eye on it,
i reduced powerusage of the hole minig system about 11Watt with switching from celeron900 to arm-router