I tried 180 for a short while, got invalids & rejects - back to 166.
Be very careful with the 180 !
Topic: X6500 Custom FPGA Miner - page 36. (Read 220050 times)
Hi,
Was running the 180MHz file for about 30 seconds, then the miner aborted with IO errors after some succesful hashes.
I did a complete reset but the board does not come back
I had two 40mm fans above each heatsink, the did not get hot, therefore I don't believe they are burnt.
Maybee something about the power supply went wrong?
Can you gibe me some instruction what to do? Can I check the Voltage regulators?
Was running the 180MHz file for about 30 seconds, then the miner aborted with IO errors after some succesful hashes.
I did a complete reset but the board does not come back
I had two 40mm fans above each heatsink, the did not get hot, therefore I don't believe they are burnt.
Maybee something about the power supply went wrong?
Can you gibe me some instruction what to do? Can I check the Voltage regulators?
Code:
root@esprimo:~/x6500-miner-org# python listDevices.py
0 AH00WOVU *
* means this device is currently available
root@esprimo:~/x6500-miner-org# ./program.py -v ztexmerge_166mhz.bit
2012-01-19 09:39:03 | Opening bitstream file: ztexmerge_166mhz.bit
2012-01-19 09:39:03 | Bitstream file opened:
2012-01-19 09:39:03 | Design Name: fpgaminer_top.ncd;HW_TIMEOUT=FALSE;UserID=0xFFFFFFFF
2012-01-19 09:39:03 | Part Name: 6slx150fgg484
2012-01-19 09:39:03 | Date: 2012/01/03
2012-01-19 09:39:03 | Time: 23:50:25
2012-01-19 09:39:03 | Bitstream Length: 3704054
2012-01-19 09:39:03 | Try open...
2012-01-19 09:39:03 | Device 0 opened (AH00WOVU)
2012-01-19 09:39:03 | Discovering JTAG chain 0 ...
Traceback (most recent call last):
File "./program.py", line 167, in
jtag[chain].detect()
File "/home/snpr/x6500-miner-org/jtag.py", line 81, in detect
raise NoDevicesDetected
jtag.NoDevicesDetected
root@esprimo:~/x6500-miner-org#
A lot of invalids with the 180MHz bitstreem. I guess I will stick with the 166.
Code:
Running time: 2h2m
Getwork interval: 15 secs
FPGA 0:
Accepted: 255
Rejected: 2 (0.78%)
Invalid: 27 (9.47%)
Hashrate (all nonces): 166.84 MH/s
Hashrate (valid nonces): 151.03 MH/s
Hashrate (accepted shares): 149.28 MH/s
FPGA 1:
Accepted: 230
Rejected: 2 (0.86%)
Invalid: 35 (13.11%)
Hashrate (all nonces): 156.30 MH/s
Hashrate (valid nonces): 135.81 MH/s
Hashrate (accepted shares): 134.64 MH/s
Total hashrate for device: 323.15 MH/s / 286.85 MH/s / 283.93 MH/s
Yeah, 166MHz was quite stable for me using a passive cooling system I like to call "putting it in an unheated garage in winter". But 180MHz tipped the scales and produced bad data. Amazing what a <9% boost can do. Time to get one of those little cooling fans.
Board is way too hot. I think i'll stick with the 166 bitstream for now. I was hoping to run this board close to silent, hence, with no sideways case fan contraption. very stable so far with latest github release.
My opinions: I think the installed sinks are inadequate (they do look pretty though
) and the pro long thermal abuse on the chips are going to cause damage and shorten their life. Might actually be a disservice to customers by shipping with that sink, especially now that the released bitsteams are getting up there in speed. The use of thermal tape to attach sink is robbing some of the boards efficiency. There really needs to be mount holes on the pcb for a proper cooling solution. An on board fan plug would also be nice. Other than the things I mentioned above, I like the speeds I am getting with this board. Hopeful for the next gen fpga boards if the current ones are any indication on where fpga mining is heading, great work.
My opinions: I think the installed sinks are inadequate (they do look pretty though
) and the pro long thermal abuse on the chips are going to cause damage and shorten their life. Might actually be a disservice to customers by shipping with that sink, especially now that the released bitsteams are getting up there in speed. The use of thermal tape to attach sink is robbing some of the boards efficiency. There really needs to be mount holes on the pcb for a proper cooling solution. An on board fan plug would also be nice. Other than the things I mentioned above, I like the speeds I am getting with this board. Hopeful for the next gen fpga boards if the current ones are any indication on where fpga mining is heading, great work.
188mhz is getting really hot on the down-side
i would love to see someone selling a costum watercooling block, but i guess i have to build a air solution :-(
where do i get the board "feets" with double lenght so that i can screw it from both sides?
i would love to see someone selling a costum watercooling block, but i guess i have to build a air solution :-(
where do i get the board "feets" with double lenght so that i can screw it from both sides?
I have 2 of these pointing at 2 x6500's: http://www.amazon.com/Thermaltake-Mobile-External-Cooling-AF0007/dp/B002OJN250/ref=sr_1_4?ie=UTF8&qid=1326912959&sr=8-4
That should make them cool enough, right?
That should make them cool enough, right?
its the development tree
but thirdlight wrote its stable on his board(s) and no one other has a problem, i think it would be worth a release, maybe 0.3
but thirdlight wrote its stable on his board(s) and no one other has a problem, i think it would be worth a release, maybe 0.3
You're absolutely right, I should have tagged that as v0.3 as soon as thirdlight confirmed that it fixed his problem. Compared to v0.2, we fixed to two bugs and increased the speed of communication with the FPGA which for most people resulted in an increase in effective hashrate. This upgrade is definitely recommended for everyone, and you can download it now from http://fpgamining.com/software .
And while we're at it, why not release a new bitstream, too? I just posted the 180MHz for download at http://fpgamining.com/bitstreams . This comes along with an even more stern warning about cooling. If you're using a dinky fan, running it at a reduced voltage, and just kinda pointing it at the board, it's probably not enough for this clock rate. After loading the new bitstream, keep a close eye on it for the first few hours. Touch the PCB with your finger to see if it feels hot. Watch the invalid rate and watch the overall hashrate. Either of those can be indicators of an overheating FPGA.
Well, instead of just stern warnings how about some general recommended guidelines that we can follow when looking for a cooling solution? What do you consider adequate, and what is a "dinky fan" to you? How about making a version of the board with mounting holes so we can attach northbridge coolers (heatsink + fan) or maybe something bigger. Maybe come up with an optional cooling solution for your fpga that customers can purchase - this way they can rest assured that the fpga board will be cooled properly at 180 Mhz.
And while we're at it, why not release a new bitstream, too? I just posted the 180MHz for download at http://fpgamining.com/bitstreams .
If I had a daughter, you'd be her husband!
And read the warning people! Hot = bad. keep em cool!
cheers
its the development tree
but thirdlight wrote its stable on his board(s) and no one other has a problem, i think it would be worth a release, maybe 0.3
but thirdlight wrote its stable on his board(s) and no one other has a problem, i think it would be worth a release, maybe 0.3
You're absolutely right, I should have tagged that as v0.3 as soon as thirdlight confirmed that it fixed his problem. Compared to v0.2, we fixed to two bugs and increased the speed of communication with the FPGA which for most people resulted in an increase in effective hashrate. This upgrade is definitely recommended for everyone, and you can download it now from http://fpgamining.com/software .
And while we're at it, why not release a new bitstream, too? I just posted the 180MHz for download at http://fpgamining.com/bitstreams . This comes along with an even more stern warning about cooling. If you're using a dinky fan, running it at a reduced voltage, and just kinda pointing it at the board, it's probably not enough for this clock rate. After loading the new bitstream, keep a close eye on it for the first few hours. Touch the PCB with your finger to see if it feels hot. Watch the invalid rate and watch the overall hashrate. Either of those can be indicators of an overheating FPGA.
its the development tree
but thirdlight wrote its stable on his board(s) and no one other has a problem, i think it would be worth a release, maybe 0.3
but thirdlight wrote its stable on his board(s) and no one other has a problem, i think it would be worth a release, maybe 0.3
found it. thanks. maybe a more prominent placement for the github link on fpgaminers.com might help for blind bastards like myself
Thank you for the suggestion.
Latest version of what? bitstream? program script?
i looked that division/0 up and i guess you are not on the latest version because it is fixed
because it is fixedLatest version of what? bitstream? program script?
Maybe time for fizzisist to make a release?
i looked that division/0 up and i guess you are not on the latest version because it is fixed
because it is fixed
A few of us have seen that "ZeroDivision" error when pre-processing an unprocessed bitstream. Just run program.py again & it should work fine.
The other errors may be due to running mine.py on an unprogrammed board. Fix the first & you may well fix the second!
The other errors may be due to running mine.py on an unprogrammed board. Fix the first & you may well fix the second!
Any clue why I get these errors? It's on vista.
For the stock heatsinks, we are using thermal tape. Mounting holes would have been nice, though. If you really want them, you might be able to build something that could use the existing mount holes. I'm not sure exactly how that would work out.
If you're really wanting the one with installed heatsinks, Cablesaurus will have them in stock again in a few days. Sorry, but we got a little backlogged!
If you're really wanting the one with installed heatsinks, Cablesaurus will have them in stock again in a few days. Sorry, but we got a little backlogged!
It's ok for now, I'm just pondering how I'd attach something like http://www.xilence.net/en/products/cpu-cooler/intel-cpu-cooler/product/224802.html
I guess it's too big
running so perfect since the last update
308.98 MH/s | 0: 2785/24/0 0.9%/0.0% | 1: 2794/27/0 1.0%/0.0% | 20h37m | 322.79MH/s <=overall hashrate
big thx to all the hardware & software developer
308.98 MH/s | 0: 2785/24/0 0.9%/0.0% | 1: 2794/27/0 1.0%/0.0% | 20h37m | 322.79MH/s <=overall hashrate
big thx to all the hardware & software developer
Jump to: