Topic: Algorithmically placed FPGA miner: 255MH/s/chip, supports all known boards (Read 119439 times)

ok, thanks for the clarification
I will watch as much as possible

The signcryption network was down today from 31-Jul-2013 11:49 UTC until 22:24:35 UTC.  Service has been restored.

There are two "backend" servers in secure locations which hold copies of the signcryption keys.  The frontend machines load balance signcryption requests between them automatically.

Last night one of the two backend machines had a hardware failure.  Unfortunately unbeknownst to me the other server had not been participating for the last several days due to a software problem (runit got stuck in a peculiar state), but I had not noticed because the first one had been handling the whole load without problems.

I have added to my monitoring software to let me know immediately when either one of the backend servers stops servicing requests rather than waiting until requests stop being serviced completely.  This is a bit tricky since one of the machines is slower (worse network connection) than the other and the clients will always favor whichever one is responding most quickly, so even under normal circumstances the load is very heavily skewed.

I apologize for the inconvenience.

what could it be? all servers are unavailable?
an hour ago it worked without problems


unable to connect: java.net.SocketTimeoutException: connect timed out
             exception in work queueing thread: java.net.SocketTimeoutException: connect timed out
             opening signcryption connection to limp.tricone-mining.com/96.8.120.116:7778
               unable to connect: java.net.NoRouteToHostException: No route to host
             exception in work queueing thread: java.net.NoRouteToHostException: No route to host
             opening signcryption connection to limp.tricone-mining.com/109.75.176.36:7778
               unable to connect: java.net.SocketException: Connection reset
             exception in work queueing thread: java.net.SocketException: Connection reset
             opening signcryption connection to limp.tricone-mining.com/23.23.249.133:7778
               unable to connect: java.net.SocketTimeoutException: Read timed out
             exception in work queueing thread: java.net.SocketTimeoutException: Read timed out
             opening signcryption connection to limp.tricone-mining.com/54.248.103.36:7779
               unable to connect: java.net.SocketTimeoutException: Read timed out
             exception in work queueing thread: java.net.SocketTimeoutException: Read timed out
             opening signcryption connection to limp.tricone-mining.com/23.23.249.133:7779
               unable to connect: java.net.SocketTimeoutException: Read timed out
             exception in work queueing thread: java.net.SocketTimeoutException: Read timed out
             opening signcryption connection to limp.tricone-mining.com/54.248.103.36:7779
               unable to connect: java.net.SocketTimeoutException: Read timed out


thanks

I'm surprised it worked at all.  The code you sent me wasn't able to talk to the board.

Hi Chris, I sent you this email back in October:


… and never heard back from you.  It's basically impossible for me to debug the low-level code that talks to the board unless I have physical access to one (no, I don't want a board), especially if it is a board with jumpers like the Enterpoint board.  So I was waiting for you to get it at least to the "hello world" point where it could at least read the magic number.

Like I said, I never heard back from you.

Due to cooling/voltage/board design probably.

There you are! lol ET said you never got back in contact with him? Anyways ET also says there is a DCM watchdog already implemented in TML.

Worked at a speed slower than opensource bitstreams.

https://bitcointalksearch.org/topic/m.1047343

Didn't work?

Could be!  I released all the code I had to him, never saw anything in return or info back.

Im preety sure some poeple managed to get it working - I think I posted the link in this thread or search enterpoints thread.

I never got it to work reliably enough to be worth releasing.  the DCM had issues and he seemed to have little interest in getting it working, so it wasn't worth the time it wasn't hashing away reliably to further develop it.

Its not that, but it could be the signcryption part of the TML.

It's only a problem with a BFL FPGA - since it only replies with results after it completes the full nonce range, in ~5s (vs the p2pool 10s LP)

Any of these devices: CPU, GPU, Icarus, ModMinerQuad, Ztex, Lancelot, Cairnsmore1 - it's no issue ... unless you are doing something wrong.
(and most likely most other devices are OK also)

p2p is the one with really fast blocks right, I can see how that could be problematic due to various reasons.

Yes, no problem with non-p2pool connections (e.g. eclipsemc)

Does this miner work on a stratum pool or is a proxy needed? 

Hey, how much revenue is this generating now?

At what difficulty would you release the bitstream for free?

It's not so much that I'm cheap (all spartan owners could pool a substantial amount of BTC if you would...), as I don't like to run code that does work it doesn't need to, like DRM and stuff.

Also the bitstream doesn't improve the x1.15 much I guess, don't wan't to risk my setup to try before difficulty is maybe double that of today / I receive my ASIC...

Does it work well with non p2p?

Did anybody have any luck with TML and p2pool?
I just gave it a try and received 100% rejected shares :-/