Pages:
Author

Topic: Algorithmically placed FPGA miner: 255MH/s/chip, supports all known boards - page 20. (Read 119444 times)

donator
Activity: 980
Merit: 1004
felonious vagrancy, personified
So, good news and bad news.

Bad news first: I woke up this morning to find the mine giving back 100% invalids on more than half the boards.

Good news: having a huge sample set helped me isolate a really subtle clocking bug that is present -- to varying degrees -- in all bitstreams released so far.  It has to do with the clock domain crossing between the ring clock and the host-interface clock.  How serious the effects are varies widely from board to board, which probably explains why kakobreklaa and punin were reporting that some bitstreams were "lucky" for them.

I've fixed the bug, and my mine is now performing above its pre-switchover hashrate (which was based on a six-week-old pre-0.9 bitstream) but the non-nexus6-only bitstreams won't finish building for another 10-12 hours.  I'm stuck in meetings again all day tomorrow, but I will try to post a release candidate tomorrow evening or Friday mid-day at the latest.
donator
Activity: 980
Merit: 1004
felonious vagrancy, personified
hmm... it's now hard for me to help you futher...

But, you can open a cmd and you can start jtag in any directory? and if yes you can type "tdo" after connecting to the cable and get a "0" or "1" back?

yup. Maybe ET will know better, I'm not too good at debugging java.

Sorry about that; it was probably this bug; please re-download tml-0.999w.jar.
donator
Activity: 980
Merit: 1004
felonious vagrancy, personified
@eldentyrell

Please correct line 77 in UrJtag.java:
Code:
       if (i != '\r') i = br.read();

but it should be:
Code:
       if (i == '\r') i = br.read();

*smacks forehead*

That's what I get for being in a hurry.

I have fixed it and re-uploaded tml-0.999w.jar.  Here is the md5 for the new jar file:


MD5 (tml-0.999w.jar) = 7fcbd56ceb48de27be634c9f091492fb
hero member
Activity: 556
Merit: 500
hmm... it's now hard for me to help you futher...

But, you can open a cmd and you can start jtag in any directory? and if yes you can type "tdo" after connecting to the cable and get a "0" or "1" back?

yup. Maybe ET will know better, I'm not too good at debugging java.
hero member
Activity: 1596
Merit: 502
member
Activity: 89
Merit: 10
sr. member
Activity: 397
Merit: 500
hmm... it's now hard for me to help you futher...

But, you can open a cmd and you can start jtag in any directory? and if yes you can type "tdo" after connecting to the cable and get a "0" or "1" back?
hero member
Activity: 556
Merit: 500
ok, same what i use and it's working here until i get the errors.

And my unit is only working with the shipping bitstream, with twin_test it is not working, so my unit is a faulty one.

Your unit is working ok with twin_test etc.?

working fine with twin test
sr. member
Activity: 397
Merit: 500
ok, same what i use and it's working here until i get the errors.

And my unit is only working with the shipping bitstream, with twin_test it is not working, so my unit is a faulty one.

Your unit is working ok with twin_test etc.?
hero member
Activity: 556
Merit: 500
strange...

what command line you are using?

java -jar tml-0.999mod.jar urjtag:FT2232 http://blah:[email protected]:8337
sr. member
Activity: 397
Merit: 500
strange...

what command line you are using?
hero member
Activity: 556
Merit: 500
Code:
[urjtag:0:0] programming FPGA
Exception in thread "main" java.io.EOFException
        at com.triconemining.jtag.UrJtag.getTDO(UrJtag.java:67)
        at com.triconemining.jtag.JtagChainFromRawWires$JtagChainFromRawWiresTAP
.shiftDR(JtagChainFromRawWires.java:192)
        at com.triconemining.jtag.XilinxIdCode.getUserCode(XilinxIdCode.java:37)

        at com.triconemining.board.JtagBoard$JtagChip.loadBitstream(JtagBoard.ja
va:68)
        at com.triconemining.bitcoin.miner.Miner.(Miner.java:31)
        at com.triconemining.bitcoin.miner.Main$1.(Main.java:414)
        at com.triconemining.bitcoin.miner.Main.main(Main.java:414)

I'm going to try no_reprogramming

I programmed only fpga3 (0,1 and 2 is twin_test setting). I have controller version 1.2.

Dip settings:
SW1/SW6: all on
SW2: all on
SW3: 12 off, 34 on
SW4: 12 off, 34 on
SW5: all on <- davis bitstream

hmmm in your log your at least getting a usercode. I'm erroring out before that.

also I am using those same switch settings and I updated the controller to 1.2
sr. member
Activity: 397
Merit: 500
Code:
[urjtag:0:0] programming FPGA
Exception in thread "main" java.io.EOFException
        at com.triconemining.jtag.UrJtag.getTDO(UrJtag.java:67)
        at com.triconemining.jtag.JtagChainFromRawWires$JtagChainFromRawWiresTAP
.shiftDR(JtagChainFromRawWires.java:192)
        at com.triconemining.jtag.XilinxIdCode.getUserCode(XilinxIdCode.java:37)

        at com.triconemining.board.JtagBoard$JtagChip.loadBitstream(JtagBoard.ja
va:68)
        at com.triconemining.bitcoin.miner.Miner.(Miner.java:31)
        at com.triconemining.bitcoin.miner.Main$1.(Main.java:414)
        at com.triconemining.bitcoin.miner.Main.main(Main.java:414)

I'm going to try no_reprogramming

I programmed only fpga3 (0,1 and 2 is twin_test setting). I have controller version 1.2.

Dip settings:
SW1/SW6: all on
SW2: all on
SW3: 12 off, 34 on
SW4: 12 off, 34 on
SW5: all on <- davis bitstream
sr. member
Activity: 397
Merit: 500
@eldentyrell

Please correct line 77 in UrJtag.java:
Code:
       if (i != '\r') i = br.read();

but it should be:
Code:
       if (i == '\r') i = br.read();
hero member
Activity: 556
Merit: 500
You need to use the windows version of tml. there is a tml-0.999w.jar posted from eldentyrell yesterday

I am running tml-0.999w.jar



Ok sorry, I compiled it myself and didn't tested the "fixed" .jar from eldentyrell. I will upload my self compiled jar in a sec. Pls wait...

I downloaded bluej and fixed the one line of code hah now it runs but I get

Code:
[urjtag:0:0] programming FPGA
Exception in thread "main" java.io.EOFException
        at com.triconemining.jtag.UrJtag.getTDO(UrJtag.java:67)
        at com.triconemining.jtag.JtagChainFromRawWires$JtagChainFromRawWiresTAP
.shiftDR(JtagChainFromRawWires.java:192)
        at com.triconemining.jtag.XilinxIdCode.getUserCode(XilinxIdCode.java:37)

        at com.triconemining.board.JtagBoard$JtagChip.loadBitstream(JtagBoard.ja
va:68)
        at com.triconemining.bitcoin.miner.Miner.(Miner.java:31)
        at com.triconemining.bitcoin.miner.Main$1.(Main.java:414)
        at com.triconemining.bitcoin.miner.Main.main(Main.java:414)

I'm going to try no_reprogramming
sr. member
Activity: 397
Merit: 500
You need to use the windows version of tml. there is a tml-0.999w.jar posted from eldentyrell yesterday

I am running tml-0.999w.jar



Ok sorry, I compiled it myself and didn't tested the "fixed" .jar from eldentyrell. I will upload my self compiled jar in a sec.

Temporary download: tml-0.999mod for windows
hero member
Activity: 556
Merit: 500
You need to use the windows version of tml. there is a tml-0.999w.jar posted from eldentyrell yesterday

I am running tml-0.999w.jar

sr. member
Activity: 397
Merit: 500
You need to use the windows version of tml. there is a tml-0.999w.jar posted from eldentyrell yesterday
hero member
Activity: 556
Merit: 500
Here is my patched urjtag -> patched urjtag incl. dll's

You have to install first the official 0.10 urjtag from the webside, then unpack my archive and copy over all files. Please make sure the PATH Variable contains the installation directory of urjtag.

To test it, open a cmd and type jtag
Then type "cable FT2232" in the jtag prompt, now follow the discription on the tml webside to test idcode etc.

If this is working type "q" to quit the jtag programm, then you can start tml-0.999.jar and play with it.

EDIT:
This binary has been compiled to work with cairnsmore1, for other applications I recommend the original binary!

hmmm I flashed the fpgas with davis bitstream using enterpoint tools ok but when I run java -jar tml-0.999w.jar urjtag:FT2232 http://user:[email protected]:8337/ I still get

Exception in thread "main" java.io.IOException: garbled output from urjtag; did
'ou rememeber to apply the patch?  got character '
        at com.triconemining.jtag.UrJtag.getTDO(UrJtag.java:79)
        at com.triconemining.jtag.JtagChainFromRawWires.(JtagChainFromRawW
ires.java:42)
        at com.triconemining.board.UrJtagHost.getBoard(UrJtagHost.java:27)
        at com.triconemining.bitcoin.miner.Main.main(Main.java:413)

(and yes urjtag is working with the tdo command)
sr. member
Activity: 397
Merit: 500
Yeah, that sounds right.  If you get past the "magic number check" but get read-acknowledgement errors later on, it means that there is some flaky hardware somewhere on the communication path.  Try slowing down the JTAG clock to something really slow like 500khz.

Ok, how do I slowing down the JTAG clock? In the source code or is there a hidden option?
Pages:
Jump to: