X6500 Custom FPGA Miner

[TheSeven]

Sorry for not reading the thread through but how well do these boards work on P2Pool compared to a normal PPS pool.  Thanks.

x6500s work the same assuming adequate available resources on the host machine.

Would a RaspberryPi be enough computing resource's for an X6500 on P2Pool?

My experience with a pandaboard shows that p2pool doesn't really work well on these platforms.
One suspicion is that SD card access might be playing a role, but I'm not sure.

[1714328628]

1714328628

[1714328628]

1714328628

[1714328628]

1714328628

[1714328628]

1714328628

[O_Shovah]

My Pandaboard ES runs p2pool with mpbm without major problems.

I just invested into some 50 € SD card a that does the job well.
I consider the SD I/O crucial for the systems performance.

[matthewh3]

Sorry for not reading the thread through but how well do these boards work on P2Pool compared to a normal PPS pool.  Thanks.

x6500s work the same assuming adequate available resources on the host machine.

Would a RaspberryPi be enough computing resource's for an X6500 on P2Pool?

My experience with a pandaboard shows that p2pool doesn't really work well on these platforms.
One suspicion is that SD card access might be playing a role, but I'm not sure.

OK thanks do you think a cheap second hand PC would be OK for running a farm of X6500's on P2Pool?

[TheSeven]

My Pandaboard ES runs p2pool with mpbm without major problems.

I just invested into some 50 € SD card a that does the job well.
I consider the SD I/O crucial for the systems performance.

What's your DOA rate? It was 4-7% for me before I moved p2pool to a Core2Duo box. Now it's like 1-2%.

[O_Shovah]

My Pandaboard ES runs p2pool with mpbm without major problems.

I just invested into some 50 € SD card a that does the job well.
I consider the SD I/O crucial for the systems performance.

What's your DOA rate? It was 4-7% for me before I moved p2pool to a Core2Duo box. Now it's like 1-2%.

I was testing p2pool for ~ 12 days.
I got 3.1 % DOA. Seemed ok to me.

[coretechs]

I also get about 1-2% DOA with the X6500 mining with p2pool.  This is using a semi-dedicated PC with an i3/ssd to run p2pool/bitcoind and a separate PC controlling the FPGAs.

[nbtcminer]

Finally posting pics / notes of the X6500 Rev 3.0 after what seems like FOREVER after recieving it (like 4 weeks!):

First off I wanted to say: Fizzisist, you are the man! He was super helpful with all my troubles and worked hard to get me my Rev 3.0!


Having said that; here are some pics!


The Naked Rev 3.0 (*fap fap fap*)


With one of the Zalman "Flower" North bridge coolers on it (apologies for the brightness):


The final product with both Zalmans attached and hashing away (Still trying to figure out the best configuration for a final enclosed setup):



Notes on the Rev 3.0:

1.) I've had this thing running for about 4 weeks now with a variety of weird / good / bad fan setups (80mm / 120mm / 2 x 80 mm / 2 x 120mm) and through out all my configurations I've yet to encounter a single invalid share with the 200 mhz ztexmerge bitstream! The push-pin mounts were definitely a great addition to the Rev 3.0!

2.) The "Flower" heatsinks are good but have some pretty big cons to them: A.) They extend over some of the mounting holes and the 3-pin fan power ports. B.) Two fans are needed to maintain a consistent tempurature on both FPGA chip (according to temp readings from MPBM).

3.) This thing ran the 200mhz ztexmerge bitstreams out of the box without any trouble!! I ended up using MPBM v0.1.0 to get both my Rev 2.0 and Rev 3.0 going at the same time and man am I happy so far!


Big props to the fpgamining crew and to my man Fizzist!

[coretechs]

1.) I've had this thing running for about 4 weeks now with a variety of weird / good / bad fan setups (80mm / 120mm / 2 x 80 mm / 2 x 120mm) and through out all my configurations I've yet to encounter a single invalid share with the 200 mhz ztexmerge bitstream! The push-pin mounts were definitely a great addition to the Rev 3.0!

What is the longest you have had it mining without interruption using the 200mhz bitstream?

I'm using the same bitstream and it usually runs with no invalids for 2-3 days but eventually they show up.  I've had one board running for 5 days with no invalids and then on the 6th day each core got 3 invalids within a few hours.  The temps were measuring in the same range as before (25-30C) so I assume it's just near the limit at 200mhz.

I'd like to the boards to run reliably for a long time and I'm not sure if a handful of invalids every few days is worth worrying about.  I may just run the latest bitstream at 190-195 mhz for a tiny margin of safety as I don't consider an extra 10-20mhash to be worth running on the edge.  Are my concerns warranted?  Any opinions on the risk of electromigration from running them at this speed over a period of months/years?

[fizzisist]

Finally posting pics / notes of the X6500 Rev 3.0 after what seems like FOREVER after recieving it (like 4 weeks!):

First off I wanted to say: Fizzisist, you are the man! He was super helpful with all my troubles and worked hard to get me my Rev 3.0!


Having said that; here are some pics!


The Naked Rev 3.0 (*fap fap fap*)


With one of the Zalman "Flower" North bridge coolers on it (apologies for the brightness):


The final product with both Zalmans attached and hashing away (Still trying to figure out the best configuration for a final enclosed setup):



Notes on the Rev 3.0:

1.) I've had this thing running for about 4 weeks now with a variety of weird / good / bad fan setups (80mm / 120mm / 2 x 80 mm / 2 x 120mm) and through out all my configurations I've yet to encounter a single invalid share with the 200 mhz ztexmerge bitstream! The push-pin mounts were definitely a great addition to the Rev 3.0!

2.) The "Flower" heatsinks are good but have some pretty big cons to them: A.) They extend over some of the mounting holes and the 3-pin fan power ports. B.) Two fans are needed to maintain a consistent tempurature on both FPGA chip (according to temp readings from MPBM).

3.) This thing ran the 200mhz ztexmerge bitstreams out of the box without any trouble!! I ended up using MPBM v0.1.0 to get both my Rev 2.0 and Rev 3.0 going at the same time and man am I happy so far!


Big props to the fpgamining crew and to my man Fizzist!

Awesome, nbtcminer! Thanks for the kind words! And thanks for sharing the pics and your experiences with it so far.

I have one more configuration that might be worth trying: a single 92 mm fan blowing down on the board. The two mounting holes that aren't covered by the Zalman heatsinks are spaced perfectly for this. I actually tried this out myself recently:



Note that the two screws alone are strong enough to support the fan by themselves, so the other two standoffs aren't needed at all. I'd be interested to hear how this works out for you.

[nbtcminer]

What is the longest you have had it mining without interruption using the 200mhz bitstream?

I'm using the same bitstream and it usually runs with no invalids for 2-3 days but eventually they show up.  I've had one board running for 5 days with no invalids and then on the 6th day each core got 3 invalids within a few hours.  The temps were measuring in the same range as before (25-30C) so I assume it's just near the limit at 200mhz.

I'd like to the boards to run reliably for a long time and I'm not sure if a handful of invalids every few days is worth worrying about.  I may just run the latest bitstream at 190-195 mhz for a tiny margin of safety as I don't consider an extra 10-20mhash to be worth running on the edge.  Are my concerns warranted?  Any opinions on the risk of electromigration from running them at this speed over a period of months/years?

@Coretechs:

Longest I've had the Rev 3 running is about 12 days straight  (24/7) and no invalids so far (*fingers crossed*). Having said that, are you using the Rev 2 or Rev 3 X6500? I found that my Rev 2 does produce invalids after 1 - 2 days but I'm only seeing 0.5% invalids after having it run for that amount of time. I wouldn't be too worried about the invalids as they are pretty low in number but I would recommend looking at a couple other things:

1.) Power - The Wall wort from cablesaurus is good but I think a decent PSU would be a good replacement (either use the molex for power or splice a barrel connector). The other thing you may want to consider if you haven't already done this is get a UPS to condition the power for your units / Host PC.

2.) Cooling: My X6500 Rev 3 runs at about 30-31C, so you should be good in terms of cooling. The only thing you may want to play with is how close the fan(s) are to your heatsinks. I know the Delta fans has some pretty strong magnets in them that could possibly affect things so that's something to consider.

Overall your setup looks good and I don't think you need to be too worried about the invalids but I'm going to leave the electromigration question up to someone from the fpgaminging team. (BTW I plan on running my units for a full 24/7 /365 days a year if I can)

@Fizzisit:
What brand fan is that?

[fizzisist]

Overall your setup looks good and I don't think you need to be too worried about the invalids but I'm going to leave the electromigration question up to someone from the fpgaminging team. (BTW I plan on running my units for a full 24/7 /365 days a year if I can)

I really don't know about this, and I don't think anyone else really has data on this since mining with the Spartan 6 hasn't been going on very long. I do recommend running the "overclocker" bitstream even if it means losing 5-10% of hashrate because it will be safer. MPBM will automatically reduce the clock if the FPGA starts to heat up (e.g., from a stopped fan) or invalids get over 1%. This peace of mind is worth losing a few BTC for me, in my opinion.

@Fizzisit:
What brand fan is that?

It happens to be this one: http://www.newegg.com/Product/Product.aspx?Item=N82E16835226039

There's also a little bit slower, quieter, and cheaper one that might be worth trying: http://www.newegg.com/Product/Product.aspx?Item=N82E16835226038

Or, even cheaper, but faster and louder: http://www.newegg.com/Product/Product.aspx?Item=N82E16835150103

Not enough air? How about this? http://www.newegg.com/Product/Product.aspx?Item=N82E16811999613

[Garr255]

Not enough air? How about this? http://www.newegg.com/Product/Product.aspx?Item=N82E16811999613

[Qoheleth]

Think I just snapped up the last X6500 in this batch. It wouldn't let me buy two, and then once I checked out it wouldn't let me buy another single.

[coblee]

Think I just snapped up the last X6500 in this batch. It wouldn't let me buy two, and then once I checked out it wouldn't let me buy another single.

I have 2 used ones for sale if anyone is interested: https://bitcointalk.org/index.php?topic=79602.0

[bonks]

Here's my 10GH/s rig I just set up today using 25 x6500 FPGA boards:

[Garr255]

Cool! Now the Cognitive setup has a big brother

[tgmarks]

That is a ridiculously awesome system. Although, it seems like an awe fully expensive way to get to that much power.

[Turbor]

Here's my 10GH/s rig I just set up today using 25 x6500 FPGA boards:

A nice setup you have there. very clean !

[Garr255]

That is a ridiculously awesome system. Although, it seems like an awe fully expensive way to get to that much power.

In the long run, its extremely cheap!

[bulanula]

That is a ridiculously awesome system. Although, it seems like an awe fully expensive way to get to that much power.

In the long run, its extremely cheap!

Nobody wants to think about it but ... what happens when / if BTC fails "in that long run"

Loving to see you try sell those FPGAs