Welcome, Guest. Please login or register.

Author Topic: FPGA Replay Board  (Read 825035 times)

Description:

0 Members and 10 Guests are viewing this topic.

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: FPGA Replay Board
« Reply #314 from previous page: May 15, 2013, 01:19:06 PM »
@mikej, Is there coding space in the ARM and time slices left over to make use of its built in A/D ..?
(I guess it might be hard to save to SD-card if it's too busy sampling, otoh there's RAM)

When do you estimate boards will be available for buy?
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: FPGA Replay Board
« Reply #315 on: May 15, 2013, 09:26:47 PM »
The core of MikeJ offering is the FPGA Arcade and perhaps in the feature 060-addon. To do special one off nonstandard offerings is likely to cost more.
(as I understand it)
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: FPGA Replay Board
« Reply #316 on: May 15, 2013, 11:42:08 PM »
FPGA Arcade with 128 MB RAM and 120 MHz CPU?

Not what I recall.
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: FPGA Replay Board
« Reply #317 on: May 16, 2013, 04:33:51 AM »
"We can't be threesome in this marriage, you will have to end your relationship with this miss FP.GA!" ;)
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: FPGA Replay Board
« Reply #318 on: May 16, 2013, 02:13:28 PM »
BitcoinFPGA uses LX150-3FGG484C  (overview / datasheet) at 100 MHz which can use DCM internally to get 375 MHz (ds-page 59). It performs at 400 MH/s at an unknown frequency. And has an equalient number of logic cells of 147 443.

FPGA Arcade uses XC3S1600E-FG320-4 at 27 MHz which can be up clocked internally to at least 180 MHz via DCM or 311 MHz if the chip is a stepping 1 (page 140).  And has equalient number of logic cells of 33 192 (page 2).
(schematic page8 CDCE906 HC49-27.000)

So given the larger matrix and higher clock speed the LX150-3 would be capable to (147443/33192)*(375/180) = 9.3 times the number crunching capacity. However the LX150-3 based firmware (src) is limited to 186 MHz thus max (147443/33192)*(186/180) = 4.6 times the performance.

The board in question using LX150-3 does however employ two FPGA chips but this won't change the per chip performance.

So expect no more than 43 MH/s on FPGA Arcade. Perhaps someone can translate that into USD/s ..?

@MikeJ, Question, is the FPGA-chip used a stepping 0 or stepping 1 version ..?
« Last Edit: May 16, 2013, 02:25:19 PM by freqmax »
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: FPGA Replay Board
« Reply #319 on: May 16, 2013, 03:06:51 PM »
I did some calculus using the defaults on Difficulty and Bitcoins per block and it would take 4.9 years to break even economicly for the FPGA Arcade ;)

Perhaps it's better to make other use of the FPGA Arcade.. :D
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: FPGA Replay Board
« Reply #320 on: May 16, 2013, 03:17:41 PM »
With a FPGA Arcade cost of 283 USD and a profit of  0.22 USD/day it would take 3.5 years to break even. It would be more profitable to put the money in state bonds ;)
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: FPGA Replay Board
« Reply #321 on: May 17, 2013, 12:08:28 PM »
English site version plz? ;)
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: FPGA Replay Board
« Reply #322 on: May 17, 2013, 01:37:18 PM »
Quote from: Mrs Beanbag;735203
or just get a job :/


Better to buy money ;)
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: FPGA Replay Board
« Reply #323 on: May 20, 2013, 11:46:12 PM »
The designer is recharging the mental batteries in order to be a more productive citizen at other times ;)
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: FPGA Replay Board
« Reply #324 on: May 22, 2013, 12:01:59 AM »
Bewerage meditation #80056454 :D
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: FPGA Replay Board
« Reply #325 on: May 24, 2013, 05:23:38 PM »
What is the cycle time on the wires to the DRAM chip? Took a look at the clock generator. And it's only 27 MHz, which should mean that EMI issues are few. Compared to any standard DVI-VGA-LVDS TFT controller board which seems to be full of series and parallell impedance corrections using passive components (DVI seems transistor boosted however).
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: FPGA Replay Board
« Reply #326 on: May 25, 2013, 12:01:55 AM »
I meant the physical kind of EMI such that the signals reach their destination in "good shape" at all. The EMC regulation is more like a letter of indulgence payed to the emperor.

The reason for asking is that I had a look at the more powerful A/D converters available. The catch is that they demand a solid signal integrity on the PCB and anything more than dual layer co$t..
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: FPGA Replay Board
« Reply #327 on: May 25, 2013, 04:45:22 PM »
Is it still a 64 MByte DDRx16 memory ? (so the drawing says)
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: FPGA Replay Board
« Reply #328 on: May 26, 2013, 01:36:44 AM »
How few layers would you think could make a functional FPGA Arcade? I wonder how far one could push signals (MHz-wise) on a 2-layer board. Otoh, it's the dV/dt that is the catch not the cycle time per se.
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: FPGA Replay Board
« Reply #329 on: May 28, 2013, 01:31:55 AM »
Asfair, the reason for the delay is that the boards need to be checked and shipped. The motivation is not profit per se. I think the mailed requests shows quite clearly how large the interest is.