Pages:
Author

Topic: FPGA development board "Icarus" - DisContinued/ important announcement - page 36. (Read 207285 times)

sr. member
Activity: 410
Merit: 252
Watercooling the world of mining
I wantted to share some empiric data regarding heat.
I took thermografic pictures from the Icarus board at full speed.

Circumstances:
No additional airflow(still room air)
Ambient temperature 21 °C

Nice pictures Smiley

Can you make a picture of the flipside of the board while mining ?

I messured the highest temps at the backside of the spartan6 chips.
Ok will try to take one next week when i get the camera back.
legendary
Activity: 892
Merit: 1002
1 BTC =1 BTC
I wantted to share some empiric data regarding heat.
I took thermografic pictures from the Icarus board at full speed.

Circumstances:
No additional airflow(still room air)
Ambient temperature 21 °C

Nice pictures Smiley

Can you make a picture of the flipside of the board while mining ?

I messured the highest temps at the backside of the spartan6 chips.
hero member
Activity: 527
Merit: 500
Got my boards today Smiley (thx ngzhang). The boards seem to work well but I'm constantly getting the message "reader get queue", not sure if this should be the case. I use the queue version of the miner by starting the python script directly (without the shell script).
Btw, does this miner support LP? Has someone already tried to mine on p2pool?
sr. member
Activity: 410
Merit: 252
Watercooling the world of mining
Hello everybody.

I wantted to share some empiric data regarding heat.
I took thermografic pictures from the Icarus board at full speed.

Circumstances:
No additional airflow(still room air)
Ambient temperature 21 °C

All parts are nicely cool and within safe margins:)




legendary
Activity: 3080
Merit: 1080
Excellent. Thanks for the info. So far it seems the board is not running hot at all, but then again I've had it mining for only 2 hours. I may still replace it with the ball bearing version of the same fan (for longer lifespan and peace of mind).

hero member
Activity: 592
Merit: 501
We will stand and fight.

1, the single 4CM fan model get a higher temperature than the 2* 2.5CM fan model, but also acceptable, and not find any error rate increase during the compare test. the ambient temperature is 25-26 centigrade.
2, the single 4CM fan model uses a 4200RPM slave bearing fan, made by SUNON. MODEL NO. is KD1204PFS3.  here is the datasheet:
http://datasheet.octopart.com/KD1204PFS3-H-Sunon-Fans-datasheet-146198.pdf
3, 4CM fan model's noise is much lower than the 2 fans model.
4, the 4CM fan is fasten by 2 screw, you can change it by yourself easily.


Hmm I see. Well to me noise is no issue so I think I will replace the single 4 cm fan with 2 x 2.5 cm ones if I can find them somewhere.

By the way if anyone is wondering the USB to Serial chipset on the board is a Prolific PL2303 which can be found here:

http://www.prolific.com.tw/eng/downloads.asp?id=31

This is necessary on Windows XP (and perhaps older) machines. I'm not sure about windows 7, but apparently the windows action center actually located the proper manufacturer and it even gives you a link to follow to download the drivers. Needless to say I did not actually install it on a windows 7 machine yet.

update: did some checking and apparently the link to the proper usb driver is located in the README file for the Tcl miner but not for the python beta miner.


Now the next challenge is getting the miner.py program to actually work. Do I need a specific version of python or any will do? JSON-RPC addon perhaps?

I know for linux this will be a lot easier but it would be cool if there was a guide for windows out there. 

For now I think I shall see if I can get the Tcl version working at least.



notice that the 4CM fan i select is a low speed model. change to a faster model will sure better than even use 2 X 2.5CM fans.
the FPGA is -I temperature grade, and can run under 100 C core temp.

all the need files i'm updating them onto github.
legendary
Activity: 3080
Merit: 1080
Just wanted to thank ngzhang for the excellent support. With a bit of assistance from him I was able to get the board mining.
legendary
Activity: 3080
Merit: 1080

1, the single 4CM fan model get a higher temperature than the 2* 2.5CM fan model, but also acceptable, and not find any error rate increase during the compare test. the ambient temperature is 25-26 centigrade.
2, the single 4CM fan model uses a 4200RPM slave bearing fan, made by SUNON. MODEL NO. is KD1204PFS3.  here is the datasheet:
http://datasheet.octopart.com/KD1204PFS3-H-Sunon-Fans-datasheet-146198.pdf
3, 4CM fan model's noise is much lower than the 2 fans model.
4, the 4CM fan is fasten by 2 screw, you can change it by yourself easily.


Hmm I see. Well to me noise is no issue so I think I will replace the single 4 cm fan with 2 x 2.5 cm ones if I can find them somewhere.

By the way if anyone is wondering the USB to Serial chipset on the board is a Prolific PL2303 which can be found here:

http://www.prolific.com.tw/eng/downloads.asp?id=31

This is necessary on Windows XP (and perhaps older) machines. I'm not sure about windows 7, but apparently the windows action center actually located the proper manufacturer and it even gives you a link to follow to download the drivers. Needless to say I did not actually install it on a windows 7 machine yet.

update: did some checking and apparently the link to the proper usb driver is located in the README file for the Tcl miner but not for the python beta miner.


Now the next challenge is getting the miner.py program to actually work. Do I need a specific version of python or any will do? JSON-RPC addon perhaps?

I know for linux this will be a lot easier but it would be cool if there was a guide for windows out there. 

For now I think I shall see if I can get the Tcl version working at least.

hero member
Activity: 592
Merit: 501
We will stand and fight.
Does the single fan on the 2'nd batch Icarus board provide adequate cooling to the fpgas? (compared to the dual fan solution)

edit: Also is the current fan sleeve or ball bearing?

1, the single 4CM fan model get a higher temperature than the 2* 2.5CM fan model, but also acceptable, and not find any error rate increase during the compare test. the ambient temperature is 25-26 centigrade.
2, the single 4CM fan model uses a 4200RPM slave bearing fan, made by SUNON. MODEL NO. is KD1204PFS3.  here is the datasheet:
http://datasheet.octopart.com/KD1204PFS3-H-Sunon-Fans-datasheet-146198.pdf
3, 4CM fan model's noise is much lower than the 2 fans model.
4, the 4CM fan is fasten by 2 screw, you can change it by yourself easily.

would you consider to have a storefront to accept order in the near future?

Also is the 2nd batch sold out yet? I'm looking for 2 boards.

i have no plan for a personal website for selling, here in china i can use "taobao" for a Intermediary Transactions. but for international selling i have no idea.

2nd batch is all sold out. now accept 3rd batch pre-order, planed shipping date is mid - Feb.
full member
Activity: 168
Merit: 100
would you consider to have a storefront to accept order in the near future?

Also is the 2nd batch sold out yet? I'm looking for 2 boards.
legendary
Activity: 3080
Merit: 1080
Does the single fan on the 2'nd batch Icarus board provide adequate cooling to the fpgas? (compared to the dual fan solution)

edit: Also is the current fan sleeve or ball bearing?
legendary
Activity: 1134
Merit: 1005
please PM me once you start accepting new orders.
hero member
Activity: 592
Merit: 501
We will stand and fight.
Hi ngzhang

I got this error while doing the first step( synthesize the stuff under ./miner_core). nothing relate to 2nd step.  Wink I do get this file sha256_top.ngc. but it give me 1 error. should I take care of this error or just ignore. jump to 2nd step?

thanks ngzhang
xiangfu

no map need at the first step. just sync and get a NGC file.

thanks ngzhang.

another question. how can I create the smartguide file: ./src/miner_top.ncd? it is a binary file. I want know which sources generate this smartguide file?  by read the help of xilinx. it's should be one of the files when running synthesis.

this is generate by smartXplorer flow.
actually, it's used for a higher P&R speed. you certainly can run with out this guild file, and turn to smartXplorer, but it cost a lot of time...
hero member
Activity: 592
Merit: 501
We will stand and fight.


Damn! 451 peak!  Shocked

I wonder if there is a way to increase the average speed (ie shorten the gap between the peak and the minimum). 400 Mh/s average would be really sweet to have.

a close beta ver bitsteam is already reached that speed (400M average), but looks like not 100% stable on all -2 device.
so ... Cheesy

So you are saying that the beta ver bitstream would be 100% stable on all -3 devices or are you saying that only SOME -2 devices will work 100% stable with this  beta version bitstream?

Personally I would take my chances loading the bitstream on the Icarus board if it meant the reward would be consistent 400 MH/s performance.


no, please do not deduce. i do not have a -3 device, so there is no test under them.

and, 380 to 400 only a 5% difference, it there really necessary drive the FPGAs to such a limiting case?

after the shipment over, i will clear up all the code and put them on github, if any one want to try a higher speed version, just do it. anyway Icarus present is a  warranty invalid product.  Grin
full member
Activity: 120
Merit: 100
Hi ngzhang

I got this error while doing the first step( synthesize the stuff under ./miner_core). nothing relate to 2nd step.  Wink I do get this file sha256_top.ngc. but it give me 1 error. should I take care of this error or just ignore. jump to 2nd step?

thanks ngzhang
xiangfu

no map need at the first step. just sync and get a NGC file.

thanks ngzhang.

another question. how can I create the smartguide file: ./src/miner_top.ncd? it is a binary file. I want know which sources generate this smartguide file?  by read the help of xilinx. it's should be one of the files when running synthesis.
hero member
Activity: 592
Merit: 501
We will stand and fight.
Hi

anyone try to synthesis the bitsteam from scratch? I follow the README:https://github.com/ngzhang/Icarus/blob/master/FPGA_project/README.txt
 1. first synthesize the stuff under ./miner_core, then you got a NGC file, named sha256_top.ngc
     I do get this file name 'sha256_top.ngc'. but it give some 1 error[1]. is there any problem about this error? should I take care of it?

do you copy the "sha256_top.ngc" NGC file(first step you got) to the 2nd step "right_with_bigID_and_ZERONonceMod2" folder?

Hi ngzhang

I got this error while doing the first step( synthesize the stuff under ./miner_core). nothing relate to 2nd step.  Wink I do get this file sha256_top.ngc. but it give me 1 error. should I take care of this error or just ignore. jump to 2nd step?

thanks ngzhang
xiangfu

no map need at the first step. just sync and get a NGC file.
full member
Activity: 120
Merit: 100
Hi

anyone try to synthesis the bitsteam from scratch? I follow the README:https://github.com/ngzhang/Icarus/blob/master/FPGA_project/README.txt
 1. first synthesize the stuff under ./miner_core, then you got a NGC file, named sha256_top.ngc
     I do get this file name 'sha256_top.ngc'. but it give some 1 error[1]. is there any problem about this error? should I take care of it?

do you copy the "sha256_top.ngc" NGC file(first step you got) to the 2nd step "right_with_bigID_and_ZERONonceMod2" folder?

Hi ngzhang

I got this error while doing the first step( synthesize the stuff under ./miner_core). nothing relate to 2nd step.  Wink I do get this file sha256_top.ngc. but it give me 1 error. should I take care of this error or just ignore. jump to 2nd step?

thanks ngzhang
xiangfu
legendary
Activity: 3080
Merit: 1080


Damn! 451 peak!  Shocked

I wonder if there is a way to increase the average speed (ie shorten the gap between the peak and the minimum). 400 Mh/s average would be really sweet to have.

a close beta ver bitsteam is already reached that speed (400M average), but looks like not 100% stable on all -2 device.
so ... Cheesy

So you are saying that the beta ver bitstream would be 100% stable on all -3 devices or are you saying that only SOME -2 devices will work 100% stable with this  beta version bitstream?

Personally I would take my chances loading the bitstream on the Icarus board if it meant the reward would be consistent 400 MH/s performance.
hero member
Activity: 784
Merit: 1009
firstbits:1MinerQ
Hi

anyone try to synthesis the bitsteam from scratch? I follow the README:https://github.com/ngzhang/Icarus/blob/master/FPGA_project/README.txt
 1. first synthesize the stuff under ./miner_core, then you got a NGC file, named sha256_top.ngc
     I do get this file name 'sha256_top.ngc'. but it give some 1 error[1]. is there any problem about this error? should I take care of it?

 2. now I am doing the step [put this file to ./miner , than run the flow by using Synplify E-2011.03-SP2 as synthesizer and ./src/miner_top.ncd as smartguide file]


I've tried synthesizing the Ztex core but had no success. That message you get seems to imply that no pins are connected. Maybe you need to add a UCF file to define I/O pins so they don't all get removed?

When I try to synthesize I get thousands of warnings like,

WARNING:Xst:1895 - Due to other FF/Latch trimming, FF/Latch (without init value) has a constant value of 0 in block . This FF/Latch will be trimmed during the optimization process.

and
WARNING:Xst:2677 - Node of sequential type is unconnected in block .
and others. 53018 warnings total.

I have no idea if that is supposed to happen or if I've just done something wrong. In the end I get a message about 118% of LUT resources being used and mapping fails.

Is there a reason that Synplify is used instead of Xilinx tools? Will the Xilinx tools not work here?
hero member
Activity: 592
Merit: 501
We will stand and fight.


Damn! 451 peak!  Shocked

I wonder if there is a way to increase the average speed (ie shorten the gap between the peak and the minimum). 400 Mh/s average would be really sweet to have.

a close beta ver bitsteam is already reached that speed (400M average), but looks like not 100% stable on all -2 device.
so ... Cheesy
Pages:
Jump to: