Ultra Under-overclock image for A2 Innosilicon by Emdje - V5.0

[musetheater]

Hey guys, just had a few quick questions about the A2 in general, and wanted to get a few opinions. I know this isn't an "A2 general support" thread, but everywhere else I've looked is super sketchy, so I thought that I should try.

So I ordered an A2 110MH/s unit from Zoomhash and I got it in the mail this past Monday. Tested it out with the stock image provided on the SD card to get a general idea of an average hashrate. Had it running for about 3 hours on the "default" 1100MHz frequency, and lo and behold the PSU died while mining. I've contacted Zoomhash about getting a replacement, but I'm starting to get nervous because I haven't gotten a reply in almost 3 days. I'll still hold out hope until the end of the week, but it's looking like I'm going to be replacing the PSU on my own dime, which I'm definitely not thrilled about an extra $220 on top of the $900 already spent...

A few questions I have in regards to using the Version 3.1 or 4.0 firmware with my A2:

- During my bench testing, I received an extremely erratic "average" hashrate while mining Litecoin (slow and steady coin, wasn't on coin-changing multipool). Over the 4 hour period, my 30-minute average ranged from 65MH/s to 90MH/s, spiking back and forth between the two every 10 minutes or so. I'm not at all familiar with how Raspberry Pi's work, but could this problem have been caused by a bad (outdated/defective) image? Or possibly by the near-death power supply? The PSU it was sold with is "supposedly" a 1100W Gold unit, but I wouldn't think that running it on 1100MHz would push the PSU to die like that. I have not opened the case because doing so would "void the warranty" just in case they get back to me with a replacement.

- If I were to replace the power supply myself with a reliable unit, what range of power/efficiency rating should I go with? Based on the answer to that question, what would be the optimum frequency setting I should use with the overclock image? I'd like to get as close as I can to the *advertised* 110MH. Also, assume that the chips and boards that I have are completely untouched and not modded in any way, and I'd prefer to keep it that way since I'm not the best (or luckiest) working with electricity and circuits.

Any help would be greatly appreciated. I might as well learn about the A2 and anticipate what I need to do while I wait to get it fixed. Thanks guys!

[ZeroGee]

We'll do what we can.

The power supply that came with mine was unable to support more than 4 cards at any freq above 1000MHz. From what I can tell you're having the same problem. It's probably most prominent widespread problem that these machines have. We can talk about specific lower cost server power supply replacements if you'd like, such as the 39Y7408/9 that would power both your machines off of 240V. Since you're not wanting to do too much hardware work, I'd recommend a breakout board from J4bberwock. Packages like this would be the easiest recommendation. I'd PM him and ask for 12x PCIe cables. That, or you could use individual 1.2kW-1.5kW supplies.

If you're talking about V3.1 or V4.0 versions posted in this thread, they do not work on A2 110s as far as I can tell. (Please correct me if I'm wrong.) For the time being until emdje releases a A2 110 compatible version, I would recommend using MarkAz's Anx image. Though, the index.php does not have clock speed options above 1200MHz by default, but it can be edited to allow the options.

I never really ran my machine at 1100MHz for any extended period of time, so I don't know what to expect out of it. Though, while running at 1200MHz I was getting about 90-95MH/s poolside if I remember correctly. I've found that without the modification I posted earlier, it cannot run above 1200MHz and yield a higher accepted hashrate. Though, in your situation you would have to rectify your power supply issue by either replacement or giving it a bit of help by feeding some of the cards with other power supplies before you can run above 1000MHz. The case of my unit was slightly damaged in shipping and this was an early test. I've since replaced those three side supplies with a 1kW server supply.

As far as getting 110MH/s poolside, from my findings, you'd have to overvolt and overclock. Locally mine hashes at 119-121MH/s. My machine uses about 1200W at 12V, and draws about 1450W from the wall, ~83% efficient. Producing about 113MH/s poolside yields 12.8W/MH efficiency. I think the stock power supply is rated 80 Plus Bronze. I know the one I'm using to help it is, and the numbers support this. I did not take any pre-overvoltage measurements running at 1100MHz or 1200MHz. Long story short you should buy the highest 80 Plus rated supply for the money that you want to spend.

[Zoomhash_michael]

Hey guys, just had a few quick questions about the A2 in general, and wanted to get a few opinions. I know this isn't an "A2 general support" thread, but everywhere else I've looked is super sketchy, so I thought that I should try.

So I ordered an A2 110MH/s unit from Zoomhash and I got it in the mail this past Monday. Tested it out with the stock image provided on the SD card to get a general idea of an average hashrate. Had it running for about 3 hours on the "default" 1100MHz frequency, and lo and behold the PSU died while mining. I've contacted Zoomhash about getting a replacement, but I'm starting to get nervous because I haven't gotten a reply in almost 3 days. I'll still hold out hope until the end of the week, but it's looking like I'm going to be replacing the PSU on my own dime, which I'm definitely not thrilled about an extra $220 on top of the $900 already spent...

A few questions I have in regards to using the Version 3.1 or 4.0 firmware with my A2:

- During my bench testing, I received an extremely erratic "average" hashrate while mining Litecoin (slow and steady coin, wasn't on coin-changing multipool). Over the 4 hour period, my 30-minute average ranged from 65MH/s to 90MH/s, spiking back and forth between the two every 10 minutes or so. I'm not at all familiar with how Raspberry Pi's work, but could this problem have been caused by a bad (outdated/defective) image? Or possibly by the near-death power supply? The PSU it was sold with is "supposedly" a 1100W Gold unit, but I wouldn't think that running it on 1100MHz would push the PSU to die like that. I have not opened the case because doing so would "void the warranty" just in case they get back to me with a replacement.

- If I were to replace the power supply myself with a reliable unit, what range of power/efficiency rating should I go with? Based on the answer to that question, what would be the optimum frequency setting I should use with the overclock image? I'd like to get as close as I can to the *advertised* 110MH. Also, assume that the chips and boards that I have are completely untouched and not modded in any way, and I'd prefer to keep it that way since I'm not the best (or luckiest) working with electricity and circuits.

Any help would be greatly appreciated. I might as well learn about the A2 and anticipate what I need to do while I wait to get it fixed. Thanks guys!

May I please have the order number?

[mjgraham]

I think I'll start a new thread so we can talk hardware a bit more, I do plan on releasing the code I have for the FPGA, I am trying to pretty up my setup today, got a cheap cpld board from china that is a lot smaller and mounting everything down at least.

[emdje]

The driver should have something to do with it at least. emdje's version supports 8 chips per board as a max in the current releases, according to him. If it's anything like the A1 source code, it should matter.

Well if he is using the 'standard' version of my software is should work, as the number of maximum chips is coded in the software. However when he uses the 'mini' version of my software that should work.

[emdje]

I am working on a new version that will run from 400 MHz to 1500MHz. Plus it should be set up to run on the newer A2's as well with a higher chip count. (this would require testing from someone who has one)

Mine is currently running on 1440MHz without problems and doing 17.6 Mh/s on average with an average temperature of 34 degrees Celsius and at ~ 0.92 Volts (https://www.nicehash.com/?p=miners&a=0&addr=1EsEWk5XtcsmRdpbjciTFHL91n439NTa6d&l=0 <-- yes I only have one blade left ). This would result in an overclocked 6 blade, 10 chip terminator to run at 132 Mh/s

[naron]

Sound great. When are you planning to release version for the newer A2's 110 Mh/s ?

[emdje]

Sound great. When are you planning to release version for the newer A2's 110 Mh/s ?

Most probably today. It is however untested for newer A2's (I don't have one so I can't test that).

[emdje]

New version uploaded: V5.0  

This version should work on the newer A2's as well. It takes up to 12 chips, so even the more exotic A2 versions should work on it.
Clock options are: 400, 500, 600, 700, 800, 900, 1000, 1100, 1200, 1280, 1300, 1320, 1340, 1360, 1380, 1400, 1420, 1440, 1460, 1480, 1500.
Default clock value has been set to 1200. Note that overclocking your A2 requires Voltmodding.

https://mega.nz/#!rIdCwaYI!eh55V6G3eAWILmRJbXZA__gzSQcsBePByxpIqEr4cyU

Note that the default frontend is set-up for 6 chip blades, meaning that the MH/s values and so on. I will upload other versions of the frontend later that you can manually replace.


Also be carefull when overclocking and using higher clock settings. When using 1440 MHz for example the blade draws 200Watts:



And the connectors can get hot locally:

[emdje]

Could anyone who tried this version on 10 of 12 chip bladed report back if it works or not?

[nomoco]

Hi emdje,

I would like to try your new software, but for download I need a password?

[emdje]

Hi emdje,

I would like to try your new software, but for download I need a password?

I see it sh*t sorry, will fix that.

[emdje]

Hi emdje,

I would like to try your new software, but for download I need a password?

I misunderstood the info when I copied the link, I now added the right download link. Thanks for pointing it out!

[mjgraham]

I'll check it out tomorrow and see, I actually have been running mine at 1.0V here lately, definitely drops the HW errors, as for connectors I pulled them off and just put some screw headers on there and got rid of the molex plugs. I had a bad experience with those PCI-E cables a long time ago in the GPU days, sure the outside of the cable says 18 AWG and is the same diameter but inside the wire was like 24 AWG, nice Chinese cables.

EDIT:
I tired the new version it did not work with my boards, what had you changed to try and fix this?

[nomoco]

Now the download works, thank you emdje for your work.
Everything is working fine for me (I have 8 blades of the old 8-chip design). 

[brobot_5000]

Hello i was wondering if i O/Cd my A2 (older black version) to 1280 or 1300mhz if i could do this WITHOUT volt modding my machine?
sorry if this is a silly question.

thank you and i appreciate any responses

[emdje]

Hello i was wondering if i O/Cd my A2 (older black version) to 1280 or 1300mhz if i could do this WITHOUT volt modding my machine?
sorry if this is a silly question.

thank you and i appreciate any responses

There are no silly questions

Technically you can run your chip at any frequency without voltmodding, only it will not run very well. It won't break at 1400 mhz, it will just result in a lot of hardware errors. Reducing those hardware errors requires voltmodding (increasing voltage to the chip). On the other side, if you want to run it at a much lower frequency your chip does not need that much voltage and you can reduce the voltage to the chip.

[isbjorninn]

Hello, I am trying to use the new 5.0 for a2 mega (i think, bought second hand on ebay) and it starts, and then slowly dies.... i am trying to run it at super low level, 700mhz to see if the hw errors go away.. but to no avail.

starts fine after power-on boot,

Code:

[2016-02-03 05:18:02] Started cgminer 3.9.0
 [2016-02-03 05:18:02] Run Reset=1
 [2016-02-03 05:18:02] ST MCU hardware reset start
 [2016-02-03 05:18:06] SPI Speed 4000 kHz
 [2016-02-03 05:18:06] ST MCU - Enable (Pre-header)
 [2016-02-03 05:18:06] A2 = 700,24
 [2016-02-03 05:18:06] A2 PLL Clock = 700MHz
 [2016-02-03 05:18:06] A2 = 700,24
 [2016-02-03 05:18:06] A2 PLL Clock = 700MHz
 [2016-02-03 05:18:06] A2 = 700,24
 [2016-02-03 05:18:06] A2 PLL Clock = 700MHz
 [2016-02-03 05:18:06] A2 = 700,24
 [2016-02-03 05:18:06] A2 PLL Clock = 700MHz
 [2016-02-03 05:18:06] A2 = 700,24
 [2016-02-03 05:18:06] A2 PLL Clock = 700MHz
 [2016-02-03 05:18:06] A2 = 700,24
 [2016-02-03 05:18:06] A2 PLL Clock = 700MHz
 [2016-02-03 05:18:06] AUTO GPIO CS
 [2016-02-03 05:18:07] spidev0.0(cs0): Found 12 A2 chips
 [2016-02-03 05:18:07] Found chip 1 with 54 active cores
 [2016-02-03 05:18:07] Found chip 2 with 54 active cores
 [2016-02-03 05:18:07] Found chip 3 with 54 active cores
 [2016-02-03 05:18:07] Found chip 4 with 54 active cores
 [2016-02-03 05:18:07] Found chip 5 with 54 active cores
 [2016-02-03 05:18:07] Found chip 6 with 0 active cores
 [2016-02-03 05:18:07] Found chip 7 with 54 active cores
 [2016-02-03 05:18:07] Found chip 8 with 54 active cores
 [2016-02-03 05:18:07] Found chip 9 with 54 active cores
 [2016-02-03 05:18:07] Found chip 10 with 54 active cores
 [2016-02-03 05:18:07] Found chip 11 with 54 active cores
 [2016-02-03 05:18:07] Found chip 12 with 0 active cores
 [2016-02-03 05:18:07] Found 12 chips with total 540 active cores
 [2016-02-03 05:18:08] spidev0.0(cs1): Found 12 A2 chips
 [2016-02-03 05:18:08] Found chip 1 with 54 active cores
 [2016-02-03 05:18:08] Found chip 2 with 54 active cores
 [2016-02-03 05:18:08] Found chip 3 with 54 active cores
 [2016-02-03 05:18:08] Found chip 4 with 54 active cores
 [2016-02-03 05:18:08] Found chip 5 with 54 active cores
 [2016-02-03 05:18:08] Found chip 6 with 0 active cores
 [2016-02-03 05:18:08] Found chip 7 with 54 active cores
 [2016-02-03 05:18:08] Found chip 8 with 54 active cores
 [2016-02-03 05:18:08] Found chip 9 with 54 active cores
 [2016-02-03 05:18:08] Found chip 10 with 54 active cores
 [2016-02-03 05:18:08] Found chip 11 with 54 active cores
 [2016-02-03 05:18:08] Found chip 12 with 0 active cores
 [2016-02-03 05:18:08] Found 12 chips with total 540 active cores
 [2016-02-03 05:18:08] spidev0.0(cs2): Found 12 A2 chips
 [2016-02-03 05:18:08] Found chip 1 with 54 active cores
 [2016-02-03 05:18:08] Found chip 2 with 54 active cores
 [2016-02-03 05:18:08] Found chip 3 with 54 active cores
 [2016-02-03 05:18:08] Found chip 4 with 54 active cores
 [2016-02-03 05:18:08] Found chip 5 with 54 active cores
 [2016-02-03 05:18:08] Found chip 6 with 0 active cores
 [2016-02-03 05:18:08] Found chip 7 with 54 active cores
 [2016-02-03 05:18:08] Found chip 8 with 54 active cores
 [2016-02-03 05:18:08] Found chip 9 with 54 active cores
......etc

then I get a lot of invalid nonce errors...


and then it slowly dies when trying to reset the boards

Code:

[2016-02-03 05:19:32] Failed to reset bc from chip(cs4) 0
 [2016-02-03 05:19:32] Reinit board (cs4)
 [2016-02-03 05:19:32] Accepted 01d1b814 Diff 36K/64 BA1 4 pool 0
 [2016-02-03 05:19:32] Accepted 036af226 Diff 74/64 BA1 1 pool 0
 [2016-02-03 05:19:32] Accepted 037eb80e Diff 73/64 BA1 3 pool 0
 [2016-02-03 05:19:32] Accepted 018460f3 Diff 168/64 BA1 3 pool 0
(5s):64.33M (avg):64.75Mh/s (pool):59.76Mh/s | A:73536  R:0  HW:114  WU:54709.1/m^M [2016-02-03 05:19:32] Accepted 03372d18 Diff 20.4K/15360 BA1 2 pool 0
 [2016-02-03 05:19:33] Failure(cs4)(240): missing ACK for cmd 0x04
 [2016-02-03 05:19:33] ACK(cs4) timeout:cmd_RESET_BCAST-1.14901s
 [2016-02-03 05:19:33] Reinit board failure (cs4)

and when it has tried to reinit the boards, it just sits there..

if I try to restart cgminer manually this is the error:

Code:

[2016-02-03 05:42:45] Started cgminer 3.9.0
 [2016-02-03 05:42:45] Run Reset=1
 [2016-02-03 05:42:45] ST MCU hardware reset start
 [2016-02-03 05:42:49] SPI Speed 4000 kHz
 [2016-02-03 05:42:49] ST MCU - Enable (Pre-header)
 [2016-02-03 05:42:49] A2 = 700,24
 [2016-02-03 05:42:49] A2 PLL Clock = 700MHz
 [2016-02-03 05:42:49] A2 = 700,24
 [2016-02-03 05:42:49] A2 PLL Clock = 700MHz
 [2016-02-03 05:42:49] A2 = 700,24
 [2016-02-03 05:42:49] A2 PLL Clock = 700MHz
 [2016-02-03 05:42:49] A2 = 700,24
 [2016-02-03 05:42:49] A2 PLL Clock = 700MHz
 [2016-02-03 05:42:49] A2 = 700,24
 [2016-02-03 05:42:49] A2 PLL Clock = 700MHz
 [2016-02-03 05:42:49] A2 = 700,24
 [2016-02-03 05:42:49] A2 PLL Clock = 700MHz
 [2016-02-03 05:42:49] AUTO GPIO CS
 [2016-02-03 05:42:50] Failure(cs0)(20): missing ACK for cmd 0x04
 [2016-02-03 05:42:50] ACK(cs0) timeout:cmd_RESET_BCAST-1.2750s
 [2016-02-03 05:42:51] Failure(cs1)(20): missing ACK for cmd 0x04
 [2016-02-03 05:42:51] ACK(cs1) timeout:cmd_RESET_BCAST-1.3020s
 [2016-02-03 05:42:52] Failure(cs2)(20): missing ACK for cmd 0x04
 [2016-02-03 05:42:52] ACK(cs2) timeout:cmd_RESET_BCAST-1.2607s
 [2016-02-03 05:42:53] Failure(cs3)(20): missing ACK for cmd 0x04
 [2016-02-03 05:42:53] ACK(cs3) timeout:cmd_RESET_BCAST-1.4347s
 [2016-02-03 05:42:54] Failure(cs4)(20): missing ACK for cmd 0x04
 [2016-02-03 05:42:54] ACK(cs4) timeout:cmd_RESET_BCAST-1.2680s
 [2016-02-03 05:42:55] Failure(cs5)(20): missing ACK for cmd 0x04
 [2016-02-03 05:42:55] ACK(cs5) timeout:cmd_RESET_BCAST-1.2720s
 [2016-02-03 05:42:55] No any A2 board

No any A2 board

any ideas?

[mjgraham]

I have a few but I could be wrong, I would like to know a bit more I see that it is running with the --hwreset , do you notice if the green LEDs on all the boards go off then one cycle of red flashing then green (not when you first turn it on, every time you run cgminer command)?, I know on my A2 boards the reset pin was not connected to the MCU on board so using --hwreset to reset them does nothing, instead they need a software reset like the official innosilicon binary does, while with a small amount of soldering it is fixable if your comfortable with that. Anyway I could be wrong . I actually have a real A2 on the way not just a bunch of hashing boards I bread together into a miner, I may be way off and just adapted to what worked. Also are they green boards or blue?

[prossport]

Why can not I run the image? I do not see it on the network. My raspberry Pi work good.