Bitcoin Forum
March 19, 2024, 07:43:16 AM *
News: Latest Bitcoin Core release: 26.0 [Torrent]
 
   Home   Help Search Login Register More  
Pages: « 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 [24] 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 »
  Print  
Author Topic: Algorithmically placed FPGA miner: 255MH/s/chip, supports all known boards  (Read 119415 times)
eldentyrell (OP)
Donator
Legendary
*
Offline Offline

Activity: 980
Merit: 1004


felonious vagrancy, personified


View Profile WWW
June 16, 2012, 07:24:55 AM
 #461

Xilinx guarantees their chips can tolerate operating junction temperatures up to 125°C (see page 2 of DS162) without damage, for all temperature grades (they're manufactured identically, then sorted by testing).
That's the Absolute Maximum Rating. If you look at the footnote it warns: "Exposure to Absolute Maximum Ratings conditions for extended periods of time might affect device reliability." So they guarantee it won't die immediately, but not that it won't eventually fail in a few weeks or months as a result of running it out of spec which is what ztex is worried about.

I think you missed the part of my post where I showed that it takes 26 Amps to get the junction up to 125C with a heatsink+fan.

The printing press heralded the end of the Dark Ages and made the Enlightenment possible, but it took another three centuries before any country managed to put freedom of the press beyond the reach of legislators.  So it may take a while before cryptocurrencies are free of the AML-NSA-KYC surveillance plague.
1710834196
Hero Member
*
Offline Offline

Posts: 1710834196

View Profile Personal Message (Offline)

Ignore
1710834196
Reply with quote  #2

1710834196
Report to moderator
1710834196
Hero Member
*
Offline Offline

Posts: 1710834196

View Profile Personal Message (Offline)

Ignore
1710834196
Reply with quote  #2

1710834196
Report to moderator
Whoever mines the block which ends up containing your transaction will get its fee.
Advertised sites are not endorsed by the Bitcoin Forum. They may be unsafe, untrustworthy, or illegal in your jurisdiction.
eldentyrell (OP)
Donator
Legendary
*
Offline Offline

Activity: 980
Merit: 1004


felonious vagrancy, personified


View Profile WWW
June 16, 2012, 07:28:55 AM
Last edit: June 16, 2012, 07:56:34 AM by eldentyrell
 #462

you can offer to replace/refund FPGA's that fail during usage of your software.

Intriguing.  But how am I supposed to tell this apart from a chip damaged by failure to bother installing a heatsink and fan?  Unless I can do that, I'll simply end up taking over your warranty liabilities...


The printing press heralded the end of the Dark Ages and made the Enlightenment possible, but it took another three centuries before any country managed to put freedom of the press beyond the reach of legislators.  So it may take a while before cryptocurrencies are free of the AML-NSA-KYC surveillance plague.
eldentyrell (OP)
Donator
Legendary
*
Offline Offline

Activity: 980
Merit: 1004


felonious vagrancy, personified


View Profile WWW
June 16, 2012, 07:31:37 AM
 #463

Well, I am sitting here staring at PAR grinding slowly along.  I don't know if I'll be able to stay awake until it finishes.

Assuming nothing goes wrong (big if), preview bitstreams in the morning.

The printing press heralded the end of the Dark Ages and made the Enlightenment possible, but it took another three centuries before any country managed to put freedom of the press beyond the reach of legislators.  So it may take a while before cryptocurrencies are free of the AML-NSA-KYC surveillance plague.
eldentyrell (OP)
Donator
Legendary
*
Offline Offline

Activity: 980
Merit: 1004


felonious vagrancy, personified


View Profile WWW
June 16, 2012, 07:37:16 AM
Last edit: June 16, 2012, 09:02:32 AM by eldentyrell
 #464

That quote's from a thread about a Virtex-5 device. As I recall those have metal heatspreaders. So a different chip, different packaging

No.

The whole reason for using junction temperature is that it's package-independent.

The package determines the relationship between the air/board/case temperature and the junction temperature.  Junction temperature is all that matters, but you can't measure it directly, so you compute it using thermal constants from the package.

, and built on a different process (65nm rather than 45nm).

Good point, but if the 45nm process really was more easily damaged by temperature you'd see that reflected in lower maximum junction temperatures in the datasheet.  The fact that Xilinx didn't change them means it's unlikely there has been a major change in temperature tolerance.  And we're only talking about one generation difference in process here -- it's not like 180nm vs 22nm or anything like that.

The printing press heralded the end of the Dark Ages and made the Enlightenment possible, but it took another three centuries before any country managed to put freedom of the press beyond the reach of legislators.  So it may take a while before cryptocurrencies are free of the AML-NSA-KYC surveillance plague.
eldentyrell (OP)
Donator
Legendary
*
Offline Offline

Activity: 980
Merit: 1004


felonious vagrancy, personified


View Profile WWW
June 17, 2012, 12:22:16 AM
 #465

Hrm.

So, I have a bitstream that will run error-free on the ztex board at 170mhz as long as I only use one of the three rings.  I can also run any one ring at 170mhz and the other two really slow (like 50mhz slow).  But if I use all three rings at full speed, I get errors all the way down to some pretty embarrassingly-poor hash rates.  I experienced a similar phenomenon on my own boards, but it wasn't nearly this severe and the optimal clock frequencies were still giving me 245+MH/s on my SG-2 boards (ztex uses faster SG-3 chips).

I'll be doing some more experiments on the clock-rate/error relationship this evening, but the important questions require a new build in order to answer, and that's going to take 24-48 hours (sorry, folks).  Still lots of tricks up my sleeve, but they take (build) time.

The printing press heralded the end of the Dark Ages and made the Enlightenment possible, but it took another three centuries before any country managed to put freedom of the press beyond the reach of legislators.  So it may take a while before cryptocurrencies are free of the AML-NSA-KYC surveillance plague.
rjk
Sr. Member
****
Offline Offline

Activity: 448
Merit: 250


1ngldh


View Profile
June 17, 2012, 12:49:08 AM
 #466

What's the specs on the box you are building on? Always good to know for comparison.

Mining Rig Extraordinaire - the Trenton BPX6806 18-slot PCIe backplane [PICS] Dead project is dead, all hail the coming of the mighty ASIC!
kano
Legendary
*
Offline Offline

Activity: 4438
Merit: 1794


Linux since 1997 RedHat 4


View Profile
June 17, 2012, 01:07:04 AM
 #467

As of today, such a bitstream change would have to be manually handled.

No.  There is an option to quit the miner if it is unable to contact the signcryption server.  So you launch it from a three-line shell script:


#!/bin/bash
run-tml-miner
run-old-miner


problem solved.

When you submit free patches to all the major mining software packages to support automatic failover to backup bitstreams I will agree with you.

I hereby open-source the above three-line shell script.
quit and never go back - simplicity at it's best Smiley
I like it.

Pool: https://kano.is - low 0.5% fee PPLNS 3 Days - Most reliable Solo with ONLY 0.5% fee   Bitcointalk thread: Forum
Discord support invite at https://kano.is/ Majority developer of the ckpool code - k for kano
The ONLY active original developer of cgminer. Original master git: https://github.com/kanoi/cgminer
Inspector 2211
Sr. Member
****
Offline Offline

Activity: 448
Merit: 250



View Profile
June 17, 2012, 02:13:52 AM
 #468

Hrm.

So, I have a bitstream that will run error-free on the ztex board at 170mhz as long as I only use one of the three rings.  I can also run any one ring at 170mhz and the other two really slow (like 50mhz slow).  But if I use all three rings at full speed, I get errors all the way down to some pretty embarrassingly-poor hash rates.  I experienced a similar phenomenon on my own boards, but it wasn't nearly this severe and the optimal clock frequencies were still giving me 245+MH/s on my SG-2 boards (ztex uses faster SG-3 chips).

I'll be doing some more experiments on the clock-rate/error relationship this evening, but the important questions require a new build in order to answer, and that's going to take 24-48 hours (sorry, folks).  Still lots of tricks up my sleeve, but they take (build) time.

Bitfury experienced a similar thing.
It's probably ground bounce INTERNALLY to the FPGA.
Or something like that.
Xilinx never designed their FPGAs in such a way that 95% of all flip-flops could switch at the same time.
They just didn't.
But that's what a miner does.

               ▄█▄
            ▄█ ▀█▀
     ▄ ▄███▄▄████▄▀ ▄▄▀▄
    ▀█▄████
██████▀▄█████▀▄▀
   ▄█▀▄
███████████████████▄
 ▄██▀█▀
▀▀▀███▀▀▀█████▄▄▄▀█▀▄
 ▄█▀▀   ▀█
███▀▄████████ █▀█▄▄
██▀  ▀ ▀ ▀
██████████▄   ▄▀▀█▄
     ▀ ▀
  ███▀▀▀▀▀████▌ ▄  ▀
          ████████████▌   █
        █████████████▀
        ▀▀▀██▀▀██▀▀
           ▀▀  ▀▀
BTC-GREEN       ▄▄████████▄▄
    ▄██████████████▄
  ▄██████
██████████████▄
 ▄███
███████████████████▄
▄█████████████████████████▄
██████████████████████████
███████████████████████████
███████████████████████████
▀█████████████████████████▀
 ▀███████████████████████▀
  ▀█████████████████████▀
    ▀█████████████████
       ▀▀█████████▀▀
Ecological Community in the Green Planet
❱❱❱❱❱❱     WHITEPAGE   |   ANN THREAD     ❰❰❰❰❰❰
           ▄███▄▄
       ▄▄█████████▄
      ▄████████████▌
   ▄█████████████▄▄
 ▄████████████████████
███████████████▄
▄████████████████████▀
███████████████████████▀
 ▀▀██████▀██▌██████▀
   ▀██▀▀▀  ██  ▀▀▀▀▀▀
           ██
           ██▌
          ▐███▄
.
pusle
Member
**
Offline Offline

Activity: 89
Merit: 10


View Profile
June 17, 2012, 09:59:34 AM
 #469


You have probably thought of this stuff already but here goes:

Hookup an oscilloscope to the vccint, close as possible to the fpga.
Make it look smooth on the scope at all times by:

-Make sure new midstate load etc doesn't results in spikes.

-Stagger the rings start time/midstate load/nonce wrap

-Use phase offset to interleave clock transitions for the different rings

-Ramp the clocks up gradually from idle


It could also be the PLL suffering from too much noise. Try changing the loopfilter/bandwidth of the PLL.
Might be hard but try an external high-speed clock source (connection/termination to the board is critical)


ShadesOfMarble
Donator
Hero Member
*
Offline Offline

Activity: 543
Merit: 500



View Profile
June 17, 2012, 07:43:33 PM
 #470

I don't know if I missed a statement for this question completely or it has not yet been answered...

What about eldentyrell's ("tricone mining") bitstream? Are you going to support it? Will it be the first bitstream working* with CM? Or are you focusing on your own bitstream?

(* working means more than 700 MH/s)

You really should be asking eldentyrell this question.  Given his plans for a commission structure, it makes no sense for anyone other than himself to work on implementations.
So, eldentyrell, what do you say? Wink (CM = Cairnsmore board by Enterpoint)

Review of the Spondoolies-Tech SP10 „Dawson“ Bitcoin miner (1.4 TH/s)

[22:35] <Vinnie_win> Did anyone get paid yet? | [22:36] <Isokivi> pirate did!
makomk
Hero Member
*****
Offline Offline

Activity: 686
Merit: 564


View Profile
June 17, 2012, 08:34:42 PM
 #471

On ZTEX boards, the FPGA's JTAG signals are not even connected to the Cypress FX2 microcontroller.
That's unfortunate. I'm guessing he's not broken out the appropriate pins to allow the two to be connected either.

Quad XC6SLX150 Board: 860 MHash/s or so.
SIGS ABOUT BUTTERFLY LABS ARE PAID ADS
eldentyrell (OP)
Donator
Legendary
*
Offline Offline

Activity: 980
Merit: 1004


felonious vagrancy, personified


View Profile WWW
June 18, 2012, 05:40:55 AM
 #472

Hookup an oscilloscope to the vccint, close as possible to the fpga.

Y'know, I was never any good with an oscilliscope.  One of these days….

-Make sure new midstate load etc doesn't results in spikes.

Check.  I deliberately don't stop the rings when loading nonces for this very reason; I just let garbage fly out the back end due to half-loaded work.  The noise caused by that huge change in power consumption is not worth it.

-Stagger the rings start time/midstate load/nonce wrap

-Use phase offset to interleave clock transitions for the different rings

Well, they're on different clocks.  However, I will build one where they all use the same clock so I can try this -- good ideas.


-Ramp the clocks up gradually from idle

Yes, already doing this.

It could also be the PLL suffering from too much noise. Try changing the loopfilter/bandwidth of the PLL.

Since it's only a jitter filter I have it on the lowest bandwidth setting.

I'm also going to try dropping it altogether after finding a comment by Austin saying that Xilinx's PLLs are very sensitive to activity in nearby logic

Might be hard but try an external high-speed clock source (connection/termination to the board is critical)

Unfortunately I don't have boards that can do that (SMA connectors, right?)

The printing press heralded the end of the Dark Ages and made the Enlightenment possible, but it took another three centuries before any country managed to put freedom of the press beyond the reach of legislators.  So it may take a while before cryptocurrencies are free of the AML-NSA-KYC surveillance plague.
eldentyrell (OP)
Donator
Legendary
*
Offline Offline

Activity: 980
Merit: 1004


felonious vagrancy, personified


View Profile WWW
June 18, 2012, 05:42:32 AM
 #473

Ok, so, it appears that I can get the top and bottom rings running at the rated speed (I'm still using 150mhz builds because they finish fast).  But the middle ring only runs at 60% of expected speed unless the top+bottom rings are switched off (or running super slow).

If it runs stable overnight I will launch a high-frequency build and post those bitstreams when they finish.  It won't be the predicted hashrate, but it should still be an improvement over what people have right now.  And no commissions until I figure out wtf is really going on here.

The printing press heralded the end of the Dark Ages and made the Enlightenment possible, but it took another three centuries before any country managed to put freedom of the press beyond the reach of legislators.  So it may take a while before cryptocurrencies are free of the AML-NSA-KYC surveillance plague.
eldentyrell (OP)
Donator
Legendary
*
Offline Offline

Activity: 980
Merit: 1004


felonious vagrancy, personified


View Profile WWW
June 18, 2012, 05:45:21 AM
 #474

Bitfury experienced a similar thing.

Yeah, I know… once I have fewer things on my to-do list I think me and him and anybody else interested ought to heckle forums.xilinx.com until they own up to this issue.  I have been seeing the very same "center of the fabric drops out first" phenomenon, but until I read about his experiences I had it chalked up to my crappy homemade boards.  Now that I'm seeing it on ztex's boards too I am kinda disappointed with X.

Xilinx never designed their FPGAs in such a way that 95% of all flip-flops could switch at the same time.
They just didn't.

Maybe, but they steadfastly refuse to post maximum current ratings for their devices, and say over and over "run our power analysis tools, and if the tool says it's ok, it's ok".

Well, all my designs pass their power analyses.  Yet the voltage near the center of the chip is clearly sagging.

Basically, the power analysis tools are effectively "part of the datasheet" and Xilinx has a serious datasheet error here.

The printing press heralded the end of the Dark Ages and made the Enlightenment possible, but it took another three centuries before any country managed to put freedom of the press beyond the reach of legislators.  So it may take a while before cryptocurrencies are free of the AML-NSA-KYC surveillance plague.
eldentyrell (OP)
Donator
Legendary
*
Offline Offline

Activity: 980
Merit: 1004


felonious vagrancy, personified


View Profile WWW
June 18, 2012, 05:49:26 AM
 #475

quit and never go back - simplicity at it's best Smiley
I like it.

Wink

Or, at least, manual intervention required to go back.

I suppose a better idea would be a 3-line script that emails the operator to let him/her know that it has "downshifted".

The printing press heralded the end of the Dark Ages and made the Enlightenment possible, but it took another three centuries before any country managed to put freedom of the press beyond the reach of legislators.  So it may take a while before cryptocurrencies are free of the AML-NSA-KYC surveillance plague.
DiabloD3
Legendary
*
Offline Offline

Activity: 1162
Merit: 1000


DiabloMiner author


View Profile WWW
June 18, 2012, 06:06:23 AM
 #476

Ok, so, it appears that I can get the top and bottom rings running at the rated speed (I'm still using 150mhz builds because they finish fast).  But the middle ring only runs at 60% of expected speed unless the top+bottom rings are switched off (or running super slow).

If it runs stable overnight I will launch a high-frequency build and post those bitstreams when they finish.  It won't be the predicted hashrate, but it should still be an improvement over what people have right now.  And no commissions until I figure out wtf is really going on here.

Sounds like you need prime numbers.

pieppiep
Hero Member
*****
Offline Offline

Activity: 1596
Merit: 502


View Profile
June 18, 2012, 06:47:35 AM
 #477

What if you shift the clock of the middle ring?
Maybe the voltage internally in the chip in the middle drops to much each clock edge.
ztex
Donator
Sr. Member
*
Offline Offline

Activity: 367
Merit: 250

ZTEX FPGA Boards


View Profile WWW
June 18, 2012, 01:54:14 PM
 #478

Bitfury experienced a similar thing.

Yeah, I know… once I have fewer things on my to-do list I think me and him and anybody else interested ought to heckle forums.xilinx.com until they own up to this issue.  I have been seeing the very same "center of the fabric drops out first" phenomenon, but until I read about his experiences I had it chalked up to my crappy homemade boards.  Now that I'm seeing it on ztex's boards too I am kinda disappointed with X.

Inspector 2211 already mentioned it and it is also hidden in the datasheet ("Simultaneous switching" issue): The internal GND traces of the S6 seem to be a little bit weak.



pusle
Member
**
Offline Offline

Activity: 89
Merit: 10


View Profile
June 18, 2012, 02:57:03 PM
 #479


Err, isn't "Simultaneous switching" issue about I/O pins?  not internal core logic.
ngzhang
Hero Member
*****
Offline Offline

Activity: 592
Merit: 501


We will stand and fight.


View Profile
June 18, 2012, 03:48:54 PM
 #480

What if you shift the clock of the middle ring?
Maybe the voltage internally in the chip in the middle drops to much each clock edge.

a trade-off is: the added GCLKs will consume more power.
but i think that's worth trying. Smiley
Pages: « 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 [24] 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 »
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.19 | SMF © 2006-2009, Simple Machines Valid XHTML 1.0! Valid CSS!