Author

Topic: ANTMINER S5: 1155GH(+OverClock Potential), In Stock $0.319/GH & 0.51W/GH - page 104. (Read 451266 times)

legendary
Activity: 872
Merit: 1010
Coins, Games & Miners
The controller looks like a plain old SPI + I2C adapter and GPIO button card than anything else to me. The SPI port is used to communicate to all the chips in a chained SPI bus, and the I2C bus is used to communicate to the temp chips.
From what it would seem there is more going on than just that, the bitmain_spi.ko kernel module appears to be built from the following source files. It looks like there is a FPGA of sorts as well that it interfaces with which would explain the fpga source files being a part of the bitmain_spi.ko kernel module.

bitmain-asic-drv.c
bitmain-asic.h
fpga.c
fpga.h
sha2.c
sha2.h
spi.c
spi.h
Makefile

Yes, there's an FPGA (An Altera Cyclone 4 specifically, i took some photos and uploaded them here, but they got blatantly ignored). However, Altera Cyclones get programmed with VHDL or Verilog source, not ".c" files, so what you see there is just some verification code from the CGMiner side of things. If/when Bitmain releases their S5 cgminer sources, we can get some more hints about how the communication is done, and would probably get more info about its functionality.

As it stands, i'm pretty much convinced that the FPGA contoller board does at least these things:
  • Enumerate the SPI chain (i've connected 4 hashing boards to one controller successfully).
  • Communicate via Chained SPI with all the BM1384 chips, at most 120 in total.
  • Provide A0-A1-A2 for each hashing board. This corresponds to each BM1384 address pins configuration and would also work for the temp sensor.
  • Use I2C to read temps from the onboard sensor.
  • Interrupt the beagle bone as nonces come from the SPI chain, as there's no interrupt channel set for this kind of bus, and it would probably needs to busy poll the BM1384s.
newbie
Activity: 24
Merit: 0
somewhere buried deep in this forum is the answer to my question. unfortunately, i cannot find the answer so i a reposting it. sorry

what are the hardware error thresholds for this miner? i am at 657 hardware errors for 17 hours of work, that is about .0038%. i think anything around .02 is a concern but i cannot remember. dose anyone know?

thanks

We'd just be making up arbitrary numbers to give you a threshold, but say 0.5%. Higher than average but still a negligible amount. If you can't find an obvious cause then you can and should still ignore it.

great, thanks dogie!

somewhere buried deep in this forum is the answer to my question. unfortunately, i cannot find the answer so i a reposting it. sorry

what are the hardware error thresholds for this miner? i am at 657 hardware errors for 17 hours of work, that is about .0038%. i think anything around .02 is a concern but i cannot remember. dose anyone know?

thanks

The amount of errors you have is so low, its costing you around .2 cents per day.

thanks digitalmocking
legendary
Activity: 1666
Merit: 1185
dogiecoin.com
somewhere buried deep in this forum is the answer to my question. unfortunately, i cannot find the answer so i a reposting it. sorry

what are the hardware error thresholds for this miner? i am at 657 hardware errors for 17 hours of work, that is about .0038%. i think anything around .02 is a concern but i cannot remember. dose anyone know?

thanks

We'd just be making up arbitrary numbers to give you a threshold, but say 0.5%. Higher than average but still a negligible amount. If you can't find an obvious cause then you can and should still ignore it.
full member
Activity: 224
Merit: 100
somewhere buried deep in this forum is the answer to my question. unfortunately, i cannot find the answer so i a reposting it. sorry

what are the hardware error thresholds for this miner? i am at 657 hardware errors for 17 hours of work, that is about .0038%. i think anything around .02 is a concern but i cannot remember. dose anyone know?

thanks

The amount of errors you have is so low, its costing you around .2 cents per day.
newbie
Activity: 24
Merit: 0
somewhere buried deep in this forum is the answer to my question. unfortunately, i cannot find the answer so i a reposting it. sorry

what are the hardware error thresholds for this miner? i am at 657 hardware errors for 17 hours of work, that is about .0038%. i think anything around .02 is a concern but i cannot remember. dose anyone know?

thanks
member
Activity: 99
Merit: 10
i have really bad luck with my s5. have replaced a total of 2 hashboards and just returned the beaglebone for RMA. whats with the quality....

my s3 is running fine. so i doubt it is the psu causing all these problem
full member
Activity: 181
Merit: 100
Looks like a new firmware version may be incoming soon (hopefully) for the Antminer S5. Hopefully this will have a newer version of CGMiner and may solve some of the pool / lockup problems people have been having.

https://bitmaintech.com/support.htm?pid=00720150101032800325X8cMcY2y068F
newbie
Activity: 9
Merit: 0
From what it would seem there is more going on than just that, the bitmain_spi.ko kernel module appears to be built from the following source files. It looks like there is a FPGA of sorts as well that it interfaces with which would explain the fpga source files being a part of the bitmain_spi.ko kernel module.

bitmain-asic-drv.c
bitmain-asic.h
fpga.c
fpga.h
sha2.c
sha2.h
spi.c
spi.h
Makefile
One thing I found was that sha2.c and sha2.h are the exact same GPL licensed implementations that are found in cgminer(I confirmed the fuction names all match using objdump).
sr. member
Activity: 285
Merit: 250
Turning money into heat since 2011.
You might want to double check and see if the 9V claim is still present on their website. I think they may well have realized that's it's a bogus claim and removed it. Too bad it still claims the S5 is "quiet"......

The ability to run higher efficiency was a huge part of why I bought from them.  Removing it from the web site later doesn't make it OK...

Best I can tell, they got greedy and make unrealistic claims to sell more units.  9V?  Quiet?  BS!
newbie
Activity: 9
Merit: 0
The controller looks like a plain old SPI + I2C adapter and GPIO button card than anything else to me. The SPI port is used to communicate to all the chips in a chained SPI bus, and the I2C bus is used to communicate to the temp chips.
From what it would seem there is more going on than just that, the bitmain_spi.ko kernel module appears to be built from the following source files. It looks like there is a FPGA of sorts as well that it interfaces with which would explain the fpga source files being a part of the bitmain_spi.ko kernel module.

bitmain-asic-drv.c
bitmain-asic.h
fpga.c
fpga.h
sha2.c
sha2.h
spi.c
spi.h
Makefile
alh
legendary
Activity: 1846
Merit: 1052
You might want to double check and see if the 9V claim is still present on their website. I think they may well have realized that's it's a bogus claim and removed it. Too bad it still claims the S5 is "quiet"......
sr. member
Activity: 285
Merit: 250
Turning money into heat since 2011.
If spondoolies was ever going to do something now would be the time.

Makes you wonder whether they were spooked by the same competitor that bitmain were worried about ... but were they, like bitmain, brave enough to place firm orders to have their chips fabricated?

More likely they didn't want to invest further in manufacturing current-gen technology whose bottom-end efficiency is pretty much on par with current or upcoming competition's stock efficiency, so they could put money into R&D for the next-gen stuff to actually compete directly with whoever Bitmain's worried about in another month or two.

Additionally, when factoring longevity into S5 calculations recall that BM1384 chips are rated to run around 0.25W/GH with sufficient undervolt/underclock. That takes the S5 from a 600W 1100GH machine to a 125W 500GH machine.
Did you ever get further into your project to make a DC/DC supply for the S5? I haven't updated the firmware on mine since I got it to see if the issues with running at reduced volts was a firmware limit, but I'd be interested to know if anyone's actually gotten them down to 9V stable.

 
Hi, would be interested in getting these lower for Summer and longer.  So a DC/DC supply would be great if the price works out. 

Have a question, but let me start by saying that I know nothing about pretty much everything so forgive me if the question is stupid.

I have read that the undervolt/underclock is not stable.  I am wondering is the issue possibly the controller board for these since they receive power from the Boards at a fixed voltage.  Could the Controller boards not be getting enough power and that is causing the stability issues?  Or could it be the fan drawing more power as the boards heat up? 

I tested an S5 powered through a pair of adjustable buck converters a couple months ago.  The claim that they will run @9V is bogus.  I don't have the numbers in front of me, but 10.5V was about the minimum I could get it to run @100MHz... at ~40GH... and the efficiency was crap.  At 11-11.5V it would run @100MHz, but no increase in efficiency.

I'd like to hear Bitmain clarify how/why they claimed 9V was possible.  Can it be done with a firmware or config change?  Some engineer miscalculated?  Was it just marketing BS?
hero member
Activity: 556
Merit: 500
Quick question. How can I split the hash power of the S5 ?? I will like to mine half somewhere and half somewhere else.

Thanks

unfortunately the S5 doesn't have the balanced option that the s3 has..

any particular reason you want to mine 2 places at once?


What do you mean? It worked fine for me. In the password field, if your password is '123' enter '123 --load-balance' for the first password entry. That will split it 3 ways. I guess if the 3rd pool entry was invalid you might get a 50/50 split.
donator
Activity: 792
Merit: 510
that seems to be a low hasher guy.

Please print your Bitmaintech.org Order Confirmation Page (where it shows the original shipping tracking number and the delivery address)

Print a brief note what is wrong with your S5.

Then return to

Bitmain Warranty
3700 Quebec Street
Unit 100-239
Denver, Colorado 80207
USA

Tel: 844-248-6246

just in case, please check the thermo paste behind the hashing PCB.  May be dried up, if so you may be able to solve the problem by applying a bit more thermo paste.  If not, unit needs to be worked on by the tech.

also, please try a different pool, such as antpool.com, kano.is btcguild.com and such and see if you see a different result.

some P2pool may show higher error and lower hash rate.  Also the pool that uses extranonce subscription may cause similar error, too


BitMainWarranty:

i ordered 3 S5, two are working at spec, one of them not so much, i can`t seem to pass the 1TH no matter what i try,
i tried, OC, under OC, better ventallation, different pools, different PSU etc etc. while my other 2 units are working well at 1.14 TH this one is running at ~1.02 th with 10x the HW my other 2 units have, tought 0.1% is not much, it`s still 10X.

i live in Canada, so how should i proceed for an RMA to Colorado ( If Applicable ) btw this is a B5 that i just received last week. ordered from Bitmaintech.
legendary
Activity: 1652
Merit: 1067
Christian Antkow
Food for thought: At the current BTC price of ~$220 USD/BTC, you are likely far better off buying Bitcoin on Coinbase than one of these miners.

FWIW, I've been buying on the way down below $250 over the last couple weeks instead of investing in new miners.
sr. member
Activity: 805
Merit: 250
BitMainWarranty:

i ordered 3 S5, two are working at spec, one of them not so much, i can`t seem to pass the 1TH no matter what i try,
i tried, OC, under OC, better ventallation, different pools, different PSU etc etc. while my other 2 units are working well at 1.14 TH this one is running at ~1.02 th with 10x the HW my other 2 units have, tought 0.1% is not much, it`s still 10X.

i live in Canada, so how should i proceed for an RMA to Colorado ( If Applicable ) btw this is a B5 that i just received last week. ordered from Bitmaintech.
legendary
Activity: 872
Merit: 1010
Coins, Games & Miners
Yeah I know they can keep some stuff. But we at least need the source code to be able to work on the chip.
We are currently working to use the BM1384 for a new board. If we cant get more documentation on the controller and the chip, Im not sure we will be able to continue the project, unfortunatly.

The controller looks like a plain old SPI + I2C adapter and GPIO button card than anything else to me. The SPI port is used to communicate to all the chips in a chained SPI bus, and the I2C bus is used to communicate to the temp chips.
legendary
Activity: 1484
Merit: 1004
Yeah I know they can keep some stuff. But we at least need the source code to be able to work on the chip.
We are currently working to use the BM1384 for a new board. If we cant get more documentation on the controller and the chip, Im not sure we will be able to continue the project, unfortunatly.
newbie
Activity: 9
Merit: 0
We signed a NDA with bitmain and we still dont have the documentation.
Look like they want to keep that for them.  Embarrassed
That is certainly a bit strange, technically they aren't required to release the documentation, it is just the source code that they are required to release.
legendary
Activity: 1484
Merit: 1004
I just came across this post from a while back, looks like they ignored a request for the cgminer source code for a while but eventually got the message. Bitmain needs to understand that the firmware development environment source code must be released for the same reason they had to release cgminer source code. In order to be compliant with the GPL license the development environment must produce the firmware files that Bitmain provides for download from their source code.

Dear Community,

Our truly apologies for the mis-understanding reported in this Support Thread. We are very willing to share the cgminer code to the Community.
Just reviewed our internal resource again, there is a little mis-communicaton between Customer Service Team and R&D Department. We thought this task was done 3 weeks ago.. 
However, R&D Department is repacking the cgminer code and will upload it to GitHub.com in 3 hours.

Any advice from the Community and Global Customers is appreciated, feel free to contact us via PM or [email protected].

Thank you!


We signed a NDA with bitmain and we still dont have the documentation.
Look like they want to keep that for them.  Embarrassed
Jump to: