Bitcoin Forum
March 19, 2024, 07:45:05 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 »
  Print  
Author Topic: BTCMiner - Open Source Bitcoin Miner for ZTEX FPGA Boards, 215 MH/s on LX150  (Read 161481 times)
ztex (OP)
Donator
Sr. Member
*
Offline Offline

Activity: 367
Merit: 250

ZTEX FPGA Boards


View Profile WWW
August 29, 2011, 02:20:52 PM
Last edit: July 12, 2012, 10:12:30 AM by ztex
 #1

BTCMiner -- Bitcoin Miner for ZTEX FPGA Boards,  215 MH/s on LX150

BTCMiner runs on Linux and Windows and supports USB-FPGA Modules 1.15b and 1.15d, USB-FPGA Modules 1.15x and USB-FPGA Modules 1.15y. These FPGA Boards contain an USB controller which is used for data transfers and device programming. Thus, no additional hardware like JTAG adapters is required and low cost mining clusters can be build using standard USB hubs.

BTCMiner supports dynamic overclocking based on error measurement, i.e. the software scales the frequency such that a maximum rate of valid hashes is generated.
A cluster mode with hot-plugging support allows to run large mining rigs from one software instance.

For complete list of Features please visit the BTCMiner Homepage.

FPGA BoardFPGA TypeMinimum hash rateTypical effective hash ratePower requirement
USB-FPGA-Module_1.15bXilinx XC6SLX75-384 MH/s90 MH/s5.0 W @ 184 MHz
USB-FPGA-Module_1.15dXilinx XC6SLX150-3190 MH/s215 MH/s9.7 W @ 216 Mhz
USB-FPGA-Module_1.15xXilinx XC6SLX150-3/N3190 MH/s215 MH/s9.7 W @ 216 MHz
USB-FPGA-Module_1.15y4 x Xilinx XC6SLX150-N3760 MH/s860 MH/s39 W @ 216 MHz

For bitcoin mining I recommend the USB-FPGA-Module 1.15x (1 x XC6SLX150) or USB-FPGA-Module 1.15y (4 x XC6SLX150) which are optimized for such purposes.

Prices (including volume prices) can be found in the hardware thread: https://bitcointalk.org/index.php?topic=49180.0


USB-FPGA Module 1.15x with cooler.

UPDATE Oct 20 2011:

UPDATE Nov 14 2011:
  • New software release (number 111114): Support of backup pools and improved behavior at bad server connectivity

UPDATE Dec 14 2011:
  • New software release (number 111214) with improved performance

UPDATE Jan 26 2012:

UPDATE Feb 08 2012:

UPDATE Apr 18 2012:

UPDATE Jul 05 2012: Typical (=average) hash rate updated because recent FPGA's are faster, see https://bitcointalk.org/index.php?topic=49180.0 for measurements

UPDATE Jul 12 2012: New software release (number 120703), see https://bitcointalk.org/index.php?topic=40047.msg1025153#msg1025153

1710834305
Hero Member
*
Offline Offline

Posts: 1710834305

View Profile Personal Message (Offline)

Ignore
1710834305
Reply with quote  #2

1710834305
Report to moderator
Make sure you back up your wallet regularly! Unlike a bank account, nobody can help you if you lose access to your BTC.
Advertised sites are not endorsed by the Bitcoin Forum. They may be unsafe, untrustworthy, or illegal in your jurisdiction.
1710834305
Hero Member
*
Offline Offline

Posts: 1710834305

View Profile Personal Message (Offline)

Ignore
1710834305
Reply with quote  #2

1710834305
Report to moderator
1710834305
Hero Member
*
Offline Offline

Posts: 1710834305

View Profile Personal Message (Offline)

Ignore
1710834305
Reply with quote  #2

1710834305
Report to moderator
1710834305
Hero Member
*
Offline Offline

Posts: 1710834305

View Profile Personal Message (Offline)

Ignore
1710834305
Reply with quote  #2

1710834305
Report to moderator
rannici
Newbie
*
Offline Offline

Activity: 38
Merit: 0


View Profile
August 29, 2011, 02:29:26 PM
 #2

Quote
For amounts of 100 or larger a license production program can be offered which allow to build large clusters at competitive prices in comparison to GPUs.

whoa.
ElectricMucus
Legendary
*
Offline Offline

Activity: 1666
Merit: 1057


Marketing manager - GO MP


View Profile WWW
August 29, 2011, 02:34:46 PM
 #3

Have you proof of these numbers? The power requirements can't be.

You don't just get 200%+ more efficient with another board design or different hdl code.
ztex (OP)
Donator
Sr. Member
*
Offline Offline

Activity: 367
Merit: 250

ZTEX FPGA Boards


View Profile WWW
August 29, 2011, 02:59:21 PM
 #4

Quote
The power requirements can't be. You don't just get 200%+ more efficient with another board design or different hdl code.

I measured it on Friday: 310mA @ 11.7V. 

In order to go sure, I'll measure it with a different device.




huayra.agera
Full Member
***
Offline Offline

Activity: 154
Merit: 100



View Profile
August 29, 2011, 03:03:14 PM
 #5

Can you ship internationally? This is awesome for us with high electricity bills.

BTC: 1JMPScxohom4MXy9X1Vgj8AGwcHjT8XTuy
newMeat1
Full Member
***
Offline Offline

Activity: 210
Merit: 100



View Profile
August 29, 2011, 03:17:54 PM
Last edit: August 29, 2011, 03:44:32 PM by newMeat1
 #6

Those power and Mhash/s claims are a little outrageous. Be wary about this one until we have some better data

Maybe the power consumption #'s could be caused if ztex only measured the core power usage. The 6.8W number from our FPGA board was the total power usage of the board- including fan and regulator inefficiencies.

EDIT- and just after I post this, he changes the power #'s!! Fishy. From 2.8 to 5.9W on the LX150

ztex (OP)
Donator
Sr. Member
*
Offline Offline

Activity: 367
Merit: 250

ZTEX FPGA Boards


View Profile WWW
August 29, 2011, 03:43:40 PM
Last edit: August 29, 2011, 04:04:52 PM by ztex
 #7

I measured the power again, first with the old ampere meter (now I got 2.8W on the XC6SLX150 board), then with a 2 Ohm resister in series (voltage across the resistor). This seems to be more accurate: 3.6W on the LX75 Board and 5.9W on the LX150 board.



ztex (OP)
Donator
Sr. Member
*
Offline Offline

Activity: 367
Merit: 250

ZTEX FPGA Boards


View Profile WWW
August 29, 2011, 03:48:36 PM
 #8

Can you ship internationally? This is awesome for us with high electricity bills.

yes, of course. The boards are shipped from Germany. Private customers from EU have to pay 19% German VAT.

newMeat1
Full Member
***
Offline Offline

Activity: 210
Merit: 100



View Profile
August 29, 2011, 04:00:42 PM
 #9

Well, a 2-Ohm resistor would suck about half a watt of power. So that should make your # go up by 0.5W. But I can't think of any other reason why it would change your reading. We should be dealing with almost pure DC here.

Syke
Legendary
*
Offline Offline

Activity: 3878
Merit: 1190


View Profile
August 29, 2011, 04:13:14 PM
 #10

Hardware:The FPGA Boards presented above is not optimal since it contains features which are not required for bitcoin mining. A cost optimized single board version with on-bord voltage regulators, but without DDR2 SDRAM, configuration booster CPLD, microSD and GPIO pins is scheduled for second half of September. The expected single unit price is 330 EUR (471 USD), expected 25 unit price is 260 EUR (370 USD).
Let's assume a best case scenario. $370 USD for 140 MH/s with $9 BTC/USD means it will take 1.5 years for this to pay for itself, assuming free electricity. If you have to pay for electricity, the time is even longer. Does this device come with a 2-year warranty?

Buy & Hold
ztex (OP)
Donator
Sr. Member
*
Offline Offline

Activity: 367
Merit: 250

ZTEX FPGA Boards


View Profile WWW
August 29, 2011, 04:17:05 PM
 #11

Well, a 2-Ohm resistor would suck about half a watt of power. So that should make your # go up by 0.5W. But I can't think of any other reason why it would change your reading. We should be dealing with almost pure DC here.

I measured the voltage across the 2 Ohm resistor: 1.03 V and the voltage across the the Experimental Board (i.e. behind the resistor): 11.5V. This results in 11.5V * 1.03V / 2 Ohm = 5.92 W.

The ampere-meter, I used on Friday, seems to be absolutely inaccurate or damaged. Sorry for that wrong values.

ztex (OP)
Donator
Sr. Member
*
Offline Offline

Activity: 367
Merit: 250

ZTEX FPGA Boards


View Profile WWW
August 29, 2011, 04:54:57 PM
 #12

Let's assume a best case scenario. $370 USD for 140 MH/s with $9 BTC/USD means it will take 1.5 years for this to pay for itself, assuming free electricity. If you have to pay for electricity, the time is even longer. Does this device come with a 2-year warranty?

To be honest, this scenario is still too optimistic  Sad You have to take into account that the earnings / MHash will be reduced over the time and you also have to consider the risk of a crash.

I wrote:
Quote
For amounts of 100 or larger a license production program can be offered which allow to build large clusters at competitive prices in comparison to GPUs.
In this case the costs would be reduced to 50% - 70% or even less, depending on the amounts you want to produce.



randomguy7
Hero Member
*****
Offline Offline

Activity: 527
Merit: 500


View Profile
August 29, 2011, 05:19:28 PM
 #13

I've a question about the USB-FPGA Module 1.15d. Version 1.15d isn't listed as supported by the SDK and the description of the v1.15d module says that the Xilinx ISE License is required. The free edition doesn't work with this fpga and the non-free license is freaking expensive Sad. Is there a ready2go bitstream available?
ztex (OP)
Donator
Sr. Member
*
Offline Offline

Activity: 367
Merit: 250

ZTEX FPGA Boards


View Profile WWW
August 29, 2011, 05:43:22 PM
 #14

I've a question about the USB-FPGA Module 1.15d. Version 1.15d isn't listed as supported by the SDK

The SDK description has to be updated. If you download the SDK, you can find a lot of examples for the 1.15d Board.

Quote
and the description of the v1.15d module says that the Xilinx ISE License is required. The free edition doesn't work with this fpga and the non-free license is freaking expensive Sad. Is there a ready2go bitstream available?
2 x yes.

XC6LX150 FPGA's require a license or a 30 day evaluation license.

Bitstreams are included in the jar archive. If yo unpack it (using zip or so) you will find it in the FPGA subdirectory.

ElectricMucus
Legendary
*
Offline Offline

Activity: 1666
Merit: 1057


Marketing manager - GO MP


View Profile WWW
August 29, 2011, 06:52:11 PM
 #15

While measuring power consumption is not trivial (Current compensated or voltage compensated) it should be clear to anyone being able to design a board. You learn this very early in electrical engineering.

This makes me wonder.
Protip: Practical power consumption should be measured at 115V/230V level before the power adapter since that is what you actually pay for.
It's not this big of a deal since the cos(phi) should be >0.99 anyway so eventual losses could be calculated but if you use a calibrated power meter it should still be the most accurate measurement.
ztex (OP)
Donator
Sr. Member
*
Offline Offline

Activity: 367
Merit: 250

ZTEX FPGA Boards


View Profile WWW
August 29, 2011, 07:50:35 PM
 #16

Those power and Mhash/s claims are a little outrageous. Be wary about this one until we have some better data

Maybe the power consumption #'s could be caused if ztex only measured the core power usage. The 6.8W number from our FPGA board was the total power usage of the board- including fan and regulator inefficiencies.

EDIT- and just after I post this, he changes the power #'s!! Fishy. From 2.8 to 5.9W on the LX150

O.k. the initial power values where wrong because I used a faulty ampere meter as described above. It changed them from 3.6W (not 2.8W) to 5.9W on LX150 board. Sorry for that.

For those, who don't (want) to believe my results, here are some logs generated with the current version (http://www.ztex.de/btcminer/). These logs also show the
the overclocking feature in action.

This log was made with an USB-FPGA Module 1.15d started in single mode.

2011-08-29T18:13:34: ztex_ufm1_15d1-04A32DDEAF: New device: bitfile=ztex_ufm1_15d1   f_default=128,00MHz  f_max=176,00MHz
2011-08-29T18:13:35: ztex_ufm1_15d1-04A32DDEAF: FPGA configuration time: 194 ms
2011-08-29T18:13:35: ztex_ufm1_15d1-04A32DDEAF: Set frequency to 128,00MHz
2011-08-29T18:13:50: ztex_ufm1_15d1-04A32DDEAF: f=128,00MHz,  errorRate=0,00%,  submitted 0 new nonces,  measured hash rate 0,0MH/s
2011-08-29T18:14:07: ztex_ufm1_15d1-04A32DDEAF: Error: connect timed out: Disabeling miner for 60s
2011-08-29T18:15:07: ztex_ufm1_15d1-04A32DDEAF: f=128,00MHz,  errorRate=0,00%,  submitted 0 new nonces,  measured hash rate 0,0MH/s
2011-08-29T18:15:22: ztex_ufm1_15d1-04A32DDEAF: f=128,00MHz,  errorRate=0,00%,  submitted 0 new nonces,  measured hash rate 0,0MH/s
2011-08-29T18:15:37: ztex_ufm1_15d1-04A32DDEAF: f=128,00MHz,  errorRate=0,00%,  maxErrorRate=0,00%,  submitted 3 new nonces,  measured hash rate 105,4MH/s
2011-08-29T18:15:52: ztex_ufm1_15d1-04A32DDEAF: f=128,00MHz,  errorRate=0,00%,  maxErrorRate=0,00%,  submitted 1 new nonces,  measured hash rate 125,1MH/s
2011-08-29T18:16:06: ztex_ufm1_15d1-04A32DDEAF: Set frequency to 136,00MHz
2011-08-29T18:16:07: ztex_ufm1_15d1-04A32DDEAF: f=136,00MHz,  submitted 1 new nonces,  measured hash rate 140,8MH/s
2011-08-29T18:16:23: ztex_ufm1_15d1-04A32DDEAF: f=136,00MHz,  errorRate=0,00%,  submitted 0 new nonces,  measured hash rate 128,2MH/s
2011-08-29T18:16:38: ztex_ufm1_15d1-04A32DDEAF: f=136,00MHz,  errorRate=0,00%,  submitted 4 new nonces,  measured hash rate 211,4MH/s
2011-08-29T18:16:53: ztex_ufm1_15d1-04A32DDEAF: f=136,00MHz,  errorRate=0,00%,  maxErrorRate=0,00%,  submitted 1 new nonces,  measured hash rate 217,0MH/s
2011-08-29T18:17:08: ztex_ufm1_15d1-04A32DDEAF: f=136,00MHz,  errorRate=0,00%,  maxErrorRate=0,00%,  submitted 0 new nonces,  measured hash rate 201,7MH/s
2011-08-29T18:17:23: ztex_ufm1_15d1-04A32DDEAF: f=136,00MHz,  errorRate=0,00%,  maxErrorRate=0,00%,  submitted 0 new nonces,  measured hash rate 188,4MH/s
2011-08-29T18:17:23: ztex_ufm1_15d1-04A32DDEAF: Set frequency to 144,00MHz
2011-08-29T18:17:38: ztex_ufm1_15d1-04A32DDEAF: f=144,00MHz,  errorRate=3,93%,  maxErrorRate=1,97%,  submitted 0 new nonces,  measured hash rate 176,7MH/s
2011-08-29T18:17:46: ztex_ufm1_15d1-04A32DDEAF: Set frequency to 136,00MHz
2011-08-29T18:17:53: ztex_ufm1_15d1-04A32DDEAF: f=136,00MHz,  errorRate=0,00%,  maxErrorRate=0,00%,  nextMaxErrorRate=6,11%,  submitted 0 new nonces,  measured hash rate 166,4MH/s
2011-08-29T18:18:20: ztex_ufm1_15d1-04A32DDEAF: Error: connect timed out: Disabeling miner for 60s
2011-08-29T18:19:20: ztex_ufm1_15d1-04A32DDEAF: f=136,00MHz,  errorRate=0,00%,  maxErrorRate=0,00%,  nextMaxErrorRate=6,11%,  submitted 0 new nonces,  measured hash rate 124,4MH/s
2011-08-29T18:19:35: ztex_ufm1_15d1-04A32DDEAF: f=136,00MHz,  errorRate=0,00%,  maxErrorRate=0,00%,  nextMaxErrorRate=6,11%,  submitted 1 new nonces,  measured hash rate 131,1MH/s
2011-08-29T18:19:50: ztex_ufm1_15d1-04A32DDEAF: f=136,00MHz,  errorRate=0,00%,  maxErrorRate=0,00%,  nextMaxErrorRate=6,11%,  submitted 1 new nonces,  measured hash rate 137,3MH/s


Here is the same for USB-FPGA Module 1.15b

2011-08-29T19:05:36: ztex_ufm1_15b-04A32DCB9A: New device: bitfile=ztex_ufm1_15b   f_default=128,00MHz  f_max=176,00MHz
2011-08-29T19:05:37: ztex_ufm1_15b-04A32DCB9A: FPGA configuration time: 112 ms
2011-08-29T19:05:37: ztex_ufm1_15b-04A32DCB9A: Set frequency to 128,00MHz
2011-08-29T19:05:52: ztex_ufm1_15b-04A32DCB9A: f=128,00MHz,  errorRate=0,00%,  submitted 0 new nonces,  measured hash rate 0,0MH/s
2011-08-29T19:06:07: ztex_ufm1_15b-04A32DCB9A: f=128,00MHz,  errorRate=0,00%,  submitted 1 new nonces,  measured hash rate 142,4MH/s
2011-08-29T19:06:22: ztex_ufm1_15b-04A32DCB9A: f=128,00MHz,  errorRate=0,00%,  maxErrorRate=0,00%,  submitted 0 new nonces,  measured hash rate 94,9MH/s
2011-08-29T19:06:37: ztex_ufm1_15b-04A32DCB9A: f=128,00MHz,  errorRate=0,00%,  maxErrorRate=0,00%,  submitted 0 new nonces,  measured hash rate 71,2MH/s
2011-08-29T19:06:51: ztex_ufm1_15b-04A32DCB9A: Set frequency to 136,00MHz
2011-08-29T19:06:52: ztex_ufm1_15b-04A32DCB9A: f=136,00MHz,  submitted 1 new nonces,  measured hash rate 113,8MH/s
2011-08-29T19:07:07: ztex_ufm1_15b-04A32DCB9A: f=136,00MHz,  errorRate=0,00%,  submitted 2 new nonces,  measured hash rate 189,4MH/s
2011-08-29T19:07:22: ztex_ufm1_15b-04A32DCB9A: f=136,00MHz,  errorRate=0,00%,  submitted 0 new nonces,  measured hash rate 162,6MH/s
2011-08-29T19:07:38: ztex_ufm1_15b-04A32DCB9A: f=136,00MHz,  errorRate=0,00%,  maxErrorRate=0,00%,  submitted 0 new nonces,  measured hash rate 141,9MH/s
2011-08-29T19:07:53: ztex_ufm1_15b-04A32DCB9A: f=136,00MHz,  errorRate=0,00%,  maxErrorRate=0,00%,  submitted 0 new nonces,  measured hash rate 126,2MH/s
2011-08-29T19:08:07: ztex_ufm1_15b-04A32DCB9A: Set frequency to 144,00MHz
2011-08-29T19:08:08: ztex_ufm1_15b-04A32DCB9A: f=144,00MHz,  submitted 0 new nonces,  measured hash rate 113,6MH/s
2011-08-29T19:08:23: ztex_ufm1_15b-04A32DCB9A: f=144,00MHz,  errorRate=0,00%,  submitted 0 new nonces,  measured hash rate 103,3MH/s
2011-08-29T19:08:38: ztex_ufm1_15b-04A32DCB9A: f=144,00MHz,  errorRate=0,00%,  submitted 0 new nonces,  measured hash rate 94,7MH/s
2011-08-29T19:08:53: ztex_ufm1_15b-04A32DCB9A: f=144,00MHz,  errorRate=0,00%,  maxErrorRate=0,00%,  submitted 0 new nonces,  measured hash rate 87,4MH/s
2011-08-29T19:09:08: ztex_ufm1_15b-04A32DCB9A: f=144,00MHz,  errorRate=0,00%,  maxErrorRate=0,00%,  submitted 0 new nonces,  measured hash rate 81,2MH/s
2011-08-29T19:09:22: ztex_ufm1_15b-04A32DCB9A: Set frequency to 152,00MHz
2011-08-29T19:09:23: ztex_ufm1_15b-04A32DCB9A: f=152,00MHz,  submitted 0 new nonces,  measured hash rate 75,8MH/s
2011-08-29T19:09:38: ztex_ufm1_15b-04A32DCB9A: f=152,00MHz,  errorRate=0,00%,  submitted 0 new nonces,  measured hash rate 71,1MH/s
2011-08-29T19:09:53: ztex_ufm1_15b-04A32DCB9A: f=152,00MHz,  errorRate=0,00%,  submitted 0 new nonces,  measured hash rate 66,9MH/s
2011-08-29T19:10:09: ztex_ufm1_15b-04A32DCB9A: f=152,00MHz,  errorRate=0,00%,  maxErrorRate=0,00%,  submitted 0 new nonces,  measured hash rate 63,2MH/s
2011-08-29T19:10:24: ztex_ufm1_15b-04A32DCB9A: f=152,00MHz,  errorRate=0,00%,  maxErrorRate=0,00%,  submitted 0 new nonces,  measured hash rate 59,9MH/s
2011-08-29T19:10:36: ztex_ufm1_15b-04A32DCB9A: Set frequency to 160,00MHz
2011-08-29T19:10:39: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  submitted 1 new nonces,  measured hash rate 71,1MH/s
2011-08-29T19:10:54: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=0,00%,  submitted 0 new nonces,  measured hash rate 67,8MH/s
2011-08-29T19:11:09: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=0,00%,  submitted 1 new nonces,  measured hash rate 77,6MH/s
2011-08-29T19:11:24: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=0,74%,  maxErrorRate=0,96%,  submitted 0 new nonces,  measured hash rate 74,3MH/s
2011-08-29T19:11:39: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=0,47%,  maxErrorRate=0,96%,  submitted 0 new nonces,  measured hash rate 71,2MH/s
2011-08-29T19:11:50: ztex_ufm1_15b-04A32DCB9A: Set frequency to 168,00MHz
2011-08-29T19:11:54: ztex_ufm1_15b-04A32DCB9A: f=168,00MHz,  submitted 0 new nonces,  measured hash rate 68,4MH/s
2011-08-29T19:11:57: ztex_ufm1_15b-04A32DCB9A: Set frequency to 160,00MHz
2011-08-29T19:12:09: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=0,27%,  maxErrorRate=0,96%,  nextMaxErrorRate=5,86%,  submitted 1 new nonces,  measured hash rate 76,7MH/s
2011-08-29T19:12:24: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=0,69%,  maxErrorRate=0,96%,  nextMaxErrorRate=5,86%,  submitted 2 new nonces,  measured hash rate 95,0MH/s
2011-08-29T19:12:39: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=0,49%,  maxErrorRate=0,96%,  nextMaxErrorRate=5,86%,  submitted 1 new nonces,  measured hash rate 101,8MH/s
2011-08-29T19:12:54: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=0,36%,  maxErrorRate=0,96%,  nextMaxErrorRate=5,86%,  submitted 0 new nonces,  measured hash rate 98,3MH/s
2011-08-29T19:13:09: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=0,26%,  maxErrorRate=0,96%,  nextMaxErrorRate=5,86%,  submitted 0 new nonces,  measured hash rate 95,0MH/s
2011-08-29T19:13:24: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=0,19%,  maxErrorRate=0,96%,  nextMaxErrorRate=5,86%,  submitted 0 new nonces,  measured hash rate 92,0MH/s
2011-08-29T19:13:39: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=0,14%,  maxErrorRate=0,96%,  nextMaxErrorRate=5,86%,  submitted 0 new nonces,  measured hash rate 89,1MH/s
2011-08-29T19:13:54: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=0,51%,  maxErrorRate=0,96%,  nextMaxErrorRate=5,86%,  submitted 0 new nonces,  measured hash rate 86,4MH/s
2011-08-29T19:14:09: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=0,38%,  maxErrorRate=0,96%,  nextMaxErrorRate=5,86%,  submitted 0 new nonces,  measured hash rate 83,9MH/s
2011-08-29T19:14:24: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=0,28%,  maxErrorRate=0,96%,  nextMaxErrorRate=5,86%,  submitted 1 new nonces,  measured hash rate 89,6MH/s
2011-08-29T19:14:39: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=0,21%,  maxErrorRate=0,96%,  nextMaxErrorRate=5,86%,  submitted 0 new nonces,  measured hash rate 87,2MH/s
2011-08-29T19:14:54: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=0,60%,  maxErrorRate=0,96%,  nextMaxErrorRate=5,86%,  submitted 2 new nonces,  measured hash rate 100,2MH/s
2011-08-29T19:15:09: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=0,93%,  maxErrorRate=0,97%,  nextMaxErrorRate=5,86%,  submitted 0 new nonces,  measured hash rate 97,6MH/s
2011-08-29T19:15:22: ztex_ufm1_15b-04A32DCB9A: Set frequency to 168,00MHz
2011-08-29T19:15:24: ztex_ufm1_15b-04A32DCB9A: f=168,00MHz,  maxErrorRate=5,86%,  submitted 1 new nonces,  measured hash rate 102,4MH/s
2011-08-29T19:15:26: ztex_ufm1_15b-04A32DCB9A: Set frequency to 160,00MHz
2011-08-29T19:15:39: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=1,47%,  maxErrorRate=1,54%,  nextMaxErrorRate=6,63%,  submitted 1 new nonces,  measured hash rate 107,0MH/s
2011-08-29T19:15:54: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=1,57%,  maxErrorRate=1,70%,  nextMaxErrorRate=6,63%,  submitted 1 new nonces,  measured hash rate 111,4MH/s
2011-08-29T19:16:09: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=1,16%,  maxErrorRate=1,70%,  nextMaxErrorRate=6,63%,  submitted 0 new nonces,  measured hash rate 108,7MH/s
2011-08-29T19:16:24: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=1,32%,  maxErrorRate=1,70%,  nextMaxErrorRate=6,63%,  submitted 0 new nonces,  measured hash rate 106,2MH/s
2011-08-29T19:16:34: ztex_ufm1_15b-04A32DCB9A: Set frequency to 168,00MHz
2011-08-29T19:16:38: ztex_ufm1_15b-04A32DCB9A: Set frequency to 160,00MHz
2011-08-29T19:16:39: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=2,06%,  maxErrorRate=2,06%,  nextMaxErrorRate=7,43%,  submitted 0 new nonces,  measured hash rate 103,8MH/s
2011-08-29T19:16:54: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=1,98%,  maxErrorRate=2,41%,  nextMaxErrorRate=7,43%,  submitted 1 new nonces,  measured hash rate 107,8MH/s
2011-08-29T19:17:09: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=1,48%,  maxErrorRate=2,41%,  nextMaxErrorRate=7,43%,  submitted 0 new nonces,  measured hash rate 105,5MH/s
2011-08-29T19:17:24: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=1,11%,  maxErrorRate=2,41%,  nextMaxErrorRate=7,43%,  submitted 0 new nonces,  measured hash rate 103,2MH/s
2011-08-29T19:17:39: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=0,83%,  maxErrorRate=2,41%,  nextMaxErrorRate=7,43%,  submitted 0 new nonces,  measured hash rate 101,1MH/s
2011-08-29T19:17:54: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=0,62%,  maxErrorRate=2,41%,  nextMaxErrorRate=7,43%,  submitted 0 new nonces,  measured hash rate 99,0MH/s
2011-08-29T19:18:09: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=0,46%,  maxErrorRate=2,41%,  nextMaxErrorRate=7,43%,  submitted 1 new nonces,  measured hash rate 102,8MH/s
2011-08-29T19:18:24: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=0,35%,  maxErrorRate=2,41%,  nextMaxErrorRate=7,43%,  submitted 0 new nonces,  measured hash rate 100,8MH/s
2011-08-29T19:18:39: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=0,26%,  maxErrorRate=2,41%,  nextMaxErrorRate=7,43%,  submitted 0 new nonces,  measured hash rate 98,8MH/s
2011-08-29T19:18:54: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=0,65%,  maxErrorRate=2,41%,  nextMaxErrorRate=7,43%,  submitted 0 new nonces,  measured hash rate 97,0MH/s
2011-08-29T19:19:09: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=0,89%,  maxErrorRate=2,41%,  nextMaxErrorRate=7,43%,  submitted 1 new nonces,  measured hash rate 100,5MH/s
2011-08-29T19:19:24: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=0,67%,  maxErrorRate=2,41%,  nextMaxErrorRate=7,43%,  submitted 0 new nonces,  measured hash rate 98,6MH/s
2011-08-29T19:19:39: ztex_ufm1_15b-04A32DCB9A: f=160,00MHz,  errorRate=0,88%,  maxErrorRate=2,41%,  nextMaxErrorRate=7,43%,  submitted 1 new nonces,  measured hash rate 102,0MH/s

"measured hash rate" denotes the hash rate derived from the amount of submitted nonces.  nextMaxErrorRate denotes the maximum error rate at the next frequnecy step (+8MHz). (errorRate is sometimes larger than maxErrorRate due to a different weight function)

One can see that the 1.15d board achieves 136 MHz. I tested several of these boards. Some achieve 136 MHz, some 144 MHz. At 144MHz there is usually a small error rate of a few percent. Only very few do the 144MHz without errors. Altogether the rate of valid hashes of all tested 1.15d board is between 135 MH/s and 144 MH/s.

I only tested one 1.15b board. But this one achieves 160MHz at an average error rate of about 1 percent, makes about 79 valid MH/s.  Smiley

It is very unlikely that the difference of the "overclockability" comes from the difference of the design. (routability?) Nevertheless , I will try it out (two LX75 pipelines on one LX150). Probably the LX150 FPGA's are just slower.

Here are the par results from the from current implementation:
http://www.ztex.de/btcminer/ztex_ufm1_15d1.par
http://www.ztex.de/btcminer/ztex_ufm1_15b.par
The design goal was 160MHz in both cases. Of course, this is not achievable. The "Best case achievable" value from the par files is the clock period "guaranteed" by Xilinx. i.e. it is the minimum stable frequency.

newMeat1,  you denote my hash rates as outrageous. Maybe you want to show us some results for your board (at least a par file).

randomguy7
Hero Member
*****
Offline Offline

Activity: 527
Merit: 500


View Profile
August 29, 2011, 08:24:05 PM
 #17

ztex, just to be sure, the prices you mentioned in your first post are excl. VAT? And do you accept bitcoins for payment? Grin
ztex (OP)
Donator
Sr. Member
*
Offline Offline

Activity: 367
Merit: 250

ZTEX FPGA Boards


View Profile WWW
August 29, 2011, 08:39:02 PM
Last edit: August 29, 2011, 08:51:29 PM by ztex
 #18

ztex, just to be sure, the prices you mentioned in your first post are excl. VAT? And do you accept bitcoins for payment? Grin

Prices are without VAT. VAT has to by payed only by private customers from the EU.

EDIT - BTW, if customers from the EU order from outside the EU, they also have to pay import VAT.  Sad

FPGA boards (+ Base boards for the power supply) can be ordered at http://shop.ztex.de.

If you want to purchase more than 4 units or if you want a quotation in bitcoins (currently not supported in the shop due to the volatility) please use the contact Contact form.



gusti
Legendary
*
Offline Offline

Activity: 1099
Merit: 1000


View Profile
August 29, 2011, 09:57:36 PM
 #19

Things are getting interesting for FPGA mining.
Are you open to sell me a couple of boards with a nice discount, and get a deep review on it ?
I'm a 25 ghps experienced miner.

If you don't own the private keys, you don't own the coins.
makomk
Hero Member
*****
Offline Offline

Activity: 686
Merit: 564


View Profile
August 29, 2011, 11:12:17 PM
 #20

BTCMiner supports dynamic overclocking based on error measurement, i.e. the software scales the frequency such that a maximum rate of valid hashes is generated.
Aha - I remember you mentioning this on IRC. Cunning! Didn't realise that you were the person behind the ZTEX FPGA boards though...

Quad XC6SLX150 Board: 860 MHash/s or so.
SIGS ABOUT BUTTERFLY LABS ARE PAID ADS
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 »
  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!