S17+ S9k/SE T17 S17/Pro DR5 B7 Others All Overclock per chip and Voltage Control

[efudd]

See, this is what a run of Z11 properly modded for voltage and frequencies should look like after a few hours.... You should note that there are no errors, no spamming of a failed scanning thread running every 2 minutes, blahblah. You'll note the frequencies here in this example are explicitly set by the user based on their environmental conditions and frequencies can be set and will lock PLL at anything from 200-850Mhz.

From a statistics perspective, things would look something like this:

Elapsed       R/T           Avg
5h17m4s   152.03   143.29

... This particular system has been running in an unfortunately hot area with chip temps at 80/81C, which is throttling it down some. Overnight it'll speed up again as it cools back down into the 60C's.

Here's a picture of the server side of my firmware over a few hours (with some downtime in the middle for some upgrades)


And here's a picture of a firmware running auto-tuning over that same time period


What you'll note here is that the hash rates running my firmware server side are very stable and almost never drop below the 140kSol line... yet if you run an image that has auto-tuning or scanning enabled, the results of the second pictures are true... fluctuations and drops, general hash instability as the control loop for scanning/auto-tune is poor. Even when the dip from my downtime roll off, the rates are stable.

Anyway...

Code:

Apr 30 19:09:34 (none) local0.err cgminer: Miner compile time: Tue Aug 21 16:00:32 CST 2018 (v2.2) type: Antminer Z11
Apr 30 19:09:34 (none) local0.warn cgminer: Started cgminer 4.9.0
Apr 30 19:09:34 (none) local0.warn cgminer[29510]: bitmain_ZCASH_init
Apr 30 19:09:34 (none) local0.notice cgminer[29510]: No 2 Fan find, check again
Apr 30 19:09:35 (none) local0.notice cgminer[29510]: fan-num 2 fan-map 3
Apr 30 19:09:35 (none) local0.notice cgminer[29510]: check_chain
Apr 30 19:09:35 (none) local0.notice cgminer[29510]: Chain 0 existed!
Apr 30 19:09:35 (none) local0.notice cgminer[29510]: Chain 1 existed!
Apr 30 19:09:35 (none) local0.notice cgminer[29510]: Chain 2 existed!
Apr 30 19:09:35 (none) local0.notice cgminer[29510]: chain0 every_chain_reset_PIC16F1704_pic_new
Apr 30 19:09:37 (none) local0.notice cgminer[29510]: reset_PIC16F1704_pic_new ok
Apr 30 19:09:37 (none) local0.notice cgminer[29510]: chain1 every_chain_reset_PIC16F1704_pic_new
Apr 30 19:09:38 (none) local0.notice cgminer[29510]: reset_PIC16F1704_pic_new ok
Apr 30 19:09:39 (none) local0.notice cgminer[29510]: chain2 every_chain_reset_PIC16F1704_pic_new
Apr 30 19:09:40 (none) local0.notice cgminer[29510]: chain0 every_chain_jump_from_loader_to_app_PIC16F1704_new
Apr 30 19:09:41 (none) local0.notice cgminer[29510]: jump_from_loader_to_app_PIC16F1704_new ok
Apr 30 19:09:41 (none) local0.notice cgminer[29510]: chain1 every_chain_jump_from_loader_to_app_PIC16F1704_new
Apr 30 19:09:41 (none) local0.notice cgminer[29510]: jump_from_loader_to_app_PIC16F1704_new ok
Apr 30 19:09:41 (none) local0.notice cgminer[29510]: chain2 every_chain_jump_from_loader_to_app_PIC16F1704_new
Apr 30 19:09:42 (none) local0.notice cgminer[29510]: jump_from_loader_to_app_PIC16F1704_new ok
Apr 30 19:09:43 (none) local0.notice cgminer[29510]: reset_all_hash_board_low 4204
Apr 30 19:09:43 (none) local0.notice cgminer[29510]: pic_heart_beat_func_new
Apr 30 19:09:43 (none) local0.notice cgminer[29510]: every_chain_disable_PIC16F1704_dc_dc_new
Apr 30 19:09:44 (none) local0.notice cgminer[29510]: disable_PIC16F1704_dc_dc_new ok
Apr 30 19:09:44 (none) local0.notice cgminer[29510]: disable_PIC16F1704_dc_dc_new ok
Apr 30 19:09:44 (none) local0.notice cgminer[29510]: disable_PIC16F1704_dc_dc_new ok
Apr 30 19:09:49 (none) local0.notice cgminer[29510]: every_chain_enable_PIC16F1704_dc_dc_new
Apr 30 19:09:49 (none) local0.notice cgminer[29510]: enable_PIC16F1704_dc_dc_new ok
Apr 30 19:09:49 (none) local0.notice cgminer[29510]: enable_PIC16F1704_dc_dc_new ok
Apr 30 19:09:49 (none) local0.notice cgminer[29510]: enable_PIC16F1704_dc_dc_new ok
Apr 30 19:09:50 (none) local0.notice cgminer[29510]: reset_all_hash_board_high 4210
Apr 30 19:09:50 (none) local0.notice cgminer[29510]: tty_init
Apr 30 19:09:50 (none) local0.notice cgminer[29510]: tty_init_chain chainid = 0
Apr 30 19:09:50 (none) local0.err cgminer[29510]: create rx read thread for chain 0 ok
Apr 30 19:09:50 (none) local0.notice cgminer[29510]: Start A New Asic Response.Chain Id:[0]
Apr 30 19:09:50 (none) local0.err cgminer[29510]: create tx read thread for chain 0 ok
Apr 30 19:09:50 (none) local0.notice cgminer[29510]: open device over
Apr 30 19:09:50 (none) local0.notice cgminer[29510]: tty_init_chain chainid = 1
Apr 30 19:09:50 (none) local0.err cgminer[29510]: create rx read thread for chain 1 ok
Apr 30 19:09:50 (none) local0.notice cgminer[29510]: Start A New Asic Response.Chain Id:[1]
Apr 30 19:09:50 (none) local0.err cgminer[29510]: create tx read thread for chain 1 ok
Apr 30 19:09:50 (none) local0.notice cgminer[29510]: open device over
Apr 30 19:09:50 (none) local0.notice cgminer[29510]: tty_init_chain chainid = 2
Apr 30 19:09:50 (none) local0.err cgminer[29510]: create rx read thread for chain 2 ok
Apr 30 19:09:50 (none) local0.notice cgminer[29510]: Start A New Asic Response.Chain Id:[2]
Apr 30 19:09:51 (none) local0.err cgminer[29510]: create tx read thread for chain 2 ok
Apr 30 19:09:51 (none) local0.notice cgminer[29510]: open device over
Apr 30 19:09:55 (none) local0.notice cgminer[29510]: calculate_address_interval:addrInterval = 64
Apr 30 19:09:57 (none) local0.notice cgminer[29510]: check_every_chain_asic_number
Apr 30 19:09:58 (none) local0.notice cgminer[29510]: check_asic_reg: chain J0 has 3 ASIC
Apr 30 19:10:00 (none) local0.notice cgminer[29510]: check_asic_reg: chain J1 has 3 ASIC
Apr 30 19:10:02 (none) local0.notice cgminer[29510]: check_asic_reg: chain J2 has 3 ASIC
Apr 30 19:10:02 (none) local0.notice cgminer[29510]: chainid0 set_frequency_chain freq=750
Apr 30 19:10:02 (none) local0.notice cgminer[29510]: chainid1 set_frequency_chain freq=750
Apr 30 19:10:02 (none) local0.notice cgminer[29510]: chainid2 set_frequency_chain freq=750
Apr 30 19:10:02 (none) local0.notice cgminer[29510]: set_ticket_mask_chain chainID0 ticket_mask = 0x0000000f
Apr 30 19:10:02 (none) local0.notice cgminer[29510]: set_ticket_mask_chain chainID1 ticket_mask = 0x0000000f
Apr 30 19:10:02 (none) local0.notice cgminer[29510]: set_ticket_mask_chain chainID2 ticket_mask = 0x0000000f
Apr 30 19:10:02 (none) local0.notice cgminer[29510]: software_set_address_chain chainID0 asics 3 addrInterval 64
Apr 30 19:10:02 (none) local0.notice cgminer[29510]: software_set_address_chain chainID1 asics 3 addrInterval 64
Apr 30 19:10:03 (none) local0.notice cgminer[29510]: software_set_address_chain chainID2 asics 3 addrInterval 64
Apr 30 19:10:03 (none) local0.notice cgminer[29510]: chainid0 set_core_timeout_chain: core timeout = 00014820
Apr 30 19:10:03 (none) local0.notice cgminer[29510]: chainid1 set_core_timeout_chain: core timeout = 00014820
Apr 30 19:10:03 (none) local0.notice cgminer[29510]: chainid2 set_core_timeout_chain: core timeout = 00014820
Apr 30 19:10:03 (none) local0.notice cgminer[29510]: enable_read_temperature_from_asic_chain: reg_value=0x07006061 chainid 0
Apr 30 19:10:03 (none) local0.notice cgminer[29510]: enable_read_temperature_from_asic_chain: reg_value=0x07006061 chainid 1
Apr 30 19:10:03 (none) local0.notice cgminer[29510]: enable_read_temperature_from_asic_chain: reg_value=0x07006061 chainid 2
Apr 30 19:10:03 (none) local0.notice cgminer[29510]: dev.timeout = 10000000 us
Apr 30 19:10:04 (none) local0.notice cgminer[29510]: bitmain_ZCASH_init: fan_eft : 1  fan_pwm : 100
Apr 30 19:10:06 (none) local0.notice cgminer[29510]:  424116one
Apr 30 19:10:06 (none) local0.err cgminer[29510]: cgminer time error total_secs = 1556651406.912049 last_total_secs = 1.000000
Apr 30 19:10:07 (none) local0.info cgminer[29510]: Pool 0 extranonce set to a800e7ee60
Apr 30 19:10:07 (none) local0.info cgminer[29510]: Pool 1 extranonce set to ae0e45a3be
Apr 30 19:10:07 (none) local0.info cgminer[29510]: Pool 2 extranonce set to a1013ef095
Apr 30 19:10:12 (none) local0.warn cgminer[29510]: API running in IP access mode on port 4028 (7)

[1715178905]

1715178905

[1715178905]

1715178905

[1715178905]

1715178905

[chipless]

Well your wrong or you did something wrong with the second image you post because this is what my graph looks like. It also looks like I am getting a couple more k/sol then you are. Temps here are 80-90 outside the last few days. Miners are running in 70-80's

The graph is 4x mini's, 2x Z9 Full, and 2x Z11's

flypool
145.9 KH/s   5472 (100%)   0 (0%)   a minute ago

machine
Elapsed   KSol/S(RT)   KSol/S(avg)
15h15m58s   149.73   145.4



edit.....
must be Cooling off outside
flypool
Hashrate
464.3 KH/s 460.1 KH/s

miner
Elapsed   KSol/S(RT)   KSol/S(avg)
16h4m47s   143.34   145.46   

[efudd]

If we want to play the poolside game: 148.0 KH/s   5549 (100%)   0 (0%)   a minute ago

-j

[GarySeven]

I love it.

The more you overclock and overvolt the Z11's the greater the difficulty becomes and the less profit you make.

Brilliant!

[chipless]

If we want to play the poolside game: 148.0 KH/s   5549 (100%)   0 (0%)   a minute ago

-j

Sorry I dont play games, I have proven you wrong now a few times today just accept it for what it is and work on your own rather then sit here trying to say mine is bogus shit.

You wanted me to use my own shit and leave yours alone so I did, now thats not enough for you. If you want to pay me off I will stop public release of free files, if not everything I do is gonna be put out for free. Your starting to act like Bitmain and want to corner the market.

Fact is machine side I am getting 2k more then you in speed, the server side varies I know that, so we will deal with only the machine side from here out. You may not agree with how I changed the code and thats fine it is a beta 1 first release and I know there is tweaking to do. I have 1 test version that held closer to 150k for awhile.

You need to do like I did and invest close to 25k in miners and test away without depending on others. I also dont know how many times I have to tell you dont under estimate what I know how to and what I dont know how to do. I am retired and have many hours to play with these things and the firmware.

[chipless]

I love it.

The more you overclock and overvolt the Z11's the greater the difficulty becomes and the less profit you make.

Brilliant!

The difficulty is going up no matter what so you either get the profit now or you lose profit anyways. Since last August I have received half my costs back already. The faster I recover that the faster I start to profit.

[efudd]

*shrug*.

-j

[chipless]

*shrug*.

-j

So what are you referring to let see some times and averages etc shit i can reboot mine and it boosts upto over 160k that wold be aboout 53.3k each but it last for a few minutes till it settles in at 145....I am still waiting for the pm and link for the safe cgminer file to test. You never sent a link so it must be another thing you say then dont do. Wasnt I suppose to look at it and see how wrong I was.

[efudd]

*shrug*.

-j

So what are you referring to let see some times and averages etc shit i can reboot mine and it boosts upto over 160k that wold be aboout 53.3k each but it last for a few minutes till it settles in at 145....I am still waiting for the pm and link for the safe cgminer file to test. You never sent a link so it must be another thing you say then dont do. Wasnt I suppose to look at it and see how wrong I was.

22m14s   167.11   150.32
34m59s   157.93   149.85

*shrug*.

-j

[chipless]

*shrug*.

-j

So what are you referring to let see some times and averages etc shit i can reboot mine and it boosts upto over 160k that wold be aboout 53.3k each but it last for a few minutes till it settles in at 145....I am still waiting for the pm and link for the safe cgminer file to test. You never sent a link so it must be another thing you say then dont do. Wasnt I suppose to look at it and see how wrong I was.

22m14s   167.11   150.32
34m59s   157.93   149.85

*shrug*.

-j

Not very long, hope it stays any improvement is good improvement. I know I still have some tweaking I can do and hopefully can get a little more out of them.

[efudd]

*shrug*.

-j

So what are you referring to let see some times and averages etc shit i can reboot mine and it boosts upto over 160k that wold be aboout 53.3k each but it last for a few minutes till it settles in at 145....I am still waiting for the pm and link for the safe cgminer file to test. You never sent a link so it must be another thing you say then dont do. Wasnt I suppose to look at it and see how wrong I was.

22m14s   167.11   150.32
34m59s   157.93   149.85

*shrug*.

-j

Not very long, hope it stays any improvement is good improvement. I know I still have some tweaking I can do and hopefully can get a little more out of them.

1h0m44s   150.07   149.57
...
9h8m9s   152.03   147.09

Shrug.

-j

[chipless]

Z11-Chipless-v1-Beta 2 Released

Fix restart bug when setting pools and updated to 775m speeds are now around 143-148k/sol

[efudd]

Z11-Chipless-v1-Beta 2 Released

Fix restart bug when setting pools and updated to 775m speeds are now around 143-148k/sol

From your first post: "A bit of what to look forward to...…… I have set manually to 775 right now. The holdback so far is voltage if it cannot be overcome
these machines wont hold much more then 143-148k/sol as far as I can tell right now"

No, the limit is not voltage.
Also, you should find that 775 as a fixed frequency will not work for many hash boards.

-j

[chipless]

Z11-Chipless-v1-Beta 2 Released

Fix restart bug when setting pools and updated to 775m speeds are now around 143-148k/sol

From your first post: "A bit of what to look forward to...…… I have set manually to 775 right now. The holdback so far is voltage if it cannot be overcome
these machines wont hold much more then 143-148k/sol as far as I can tell right now"

No, the limit is not voltage.
Also, you should find that 775 as a fixed frequency will not work for many hash boards.

-j

Well thank you for that input. I know there is more to do to improve my image. Its better to get it out there so many can use it if they choose and as with your fw as time goes on it will get better. The best part is it's free. Even if no one uses it the fw has increased my own speed and that is what really matters to me.

I find a couple things funny about this....First you kept reporting you couldn't get past the 142k and hold it and second how you tried telling me you cant change a 7 to an 8 and get more power. Then all of a sudden you start saying your getting 150k+ after I released an image. So that either makes you a liar, a bullshitter or both. Your image was suppose to be released 2 days ago per your words. I am beginning to think you enjoy making statements then not following thru with them. That is like 4+ times now you haven't stuck with what you say.

The only thing holding back the 142k was clock and power. I know power was the main holdback because as soon as I set the 7 to an 8 I was getting more speed. Hmm... You must either still be having trouble with the small problems or your spending a lot of time trying to protect the code.

[efudd]

Z11-Chipless-v1-Beta 2 Released

Fix restart bug when setting pools and updated to 775m speeds are now around 143-148k/sol

From your first post: "A bit of what to look forward to...…… I have set manually to 775 right now. The holdback so far is voltage if it cannot be overcome
these machines wont hold much more then 143-148k/sol as far as I can tell right now"

No, the limit is not voltage.
Also, you should find that 775 as a fixed frequency will not work for many hash boards.

-j

Well thank you for that input. I know there is more to do to improve my image. Its better to get it out there so many can use it if they choose and as with your fw as time goes on it will get better. The best part is it's free. Even if no one uses it the fw has increased my own speed and that is what really matters to me.

I find a couple things funny about this....First you kept reporting you couldn't get past the 142k and hold it and second how you tried telling me you cant change a 7 to an 8 and get more power. Then all of a sudden you start saying your getting 150k+ after I released an image. So that either makes you a liar, a bullshitter or both. Your image was suppose to be released 2 days ago per your words. I am beginning to think you enjoy making statements then not following thru with them. That is like 4+ times now you haven't stuck with what you say.

The only thing holding back the 142k was clock and power. I know power was the main holdback because as soon as I set the 7 to an 8 I was getting more speed. Hmm... You must either still be having trouble with the small problems or your spending a lot of time trying to protect the code.

Lol.

As I said, I won’t release “shit”.

There are things you clearly have not realized in how those specific changes operate on a broader range of machines.

Really, power is not the limit.
And really, you will find machines do not respond consistently or predictably to those specific changes you have made. Either across machines or even within the same unit from restart to restart....

Now, back to my trollin’

-j

[mindtrip]

so efudd when do you plan on releasing your Z11 firmware? If you need anyone to do any validation testing I have a few units I could test them on with you if your interested. Your work on the Z9 Mini firmwares was great 

[efudd]

so efudd when do you plan on releasing your Z11 firmware? If you need anyone to do any validation testing I have a few units I could test them on with you if your interested. Your work on the Z9 Mini firmwares was great  

@mindtrip - the release is waiting on some "administration" stuff at the moment. It's been technically ready for a while now, but it is being released in partnership with one of top equihash pools to provide a more seamless integration between the hardware, software, and the mining pools. Additionally there will be no developer fee in conjuction with that pool and lower pool side fees as well. Simultaneous releases for Z9, Mini and Z11 are ready to go just as soon as we can get the joint release stuff completed.

Another way to look at it is if you are currently paying X percent for mining fees + dev fees (let's call it 6% if you are using antpool + my dev release), you'll be 1% or less combined with the new firmware and pool which is a good trade off for limited overclock and a win if there is good overclock.

From an expectations perspective, the Z11 firmware is not going to initially be a "big overclock" release like it was for the Z9. It'll be similar to the mini in terms of small improvements out the door.

The feature set between Z9/Mini/Z11 will be in sync...so if you have used and tuned on one, you can use and tune on another.

Soon(tm) although it doesn't help that I'm sick at the moment either though which is delaying things a bit.

thank you,

Jason

[chipless]

** Latest Speed Update  5/3/2019 12:59 CST ***
Elapsed      KSol/S(RT)           KSol/S(avg)
36m57s      153                     149.08

[MrMiner13]

https://ibb.co/JqwrcS6


Chipless your firmware ran for 2hr5m without any major issues. The average hash rate did increase 5-6k over stock. A big increase on HW errors, hundreds more on each board. The increase in voltage made ASIC chains 0 & 2 5c hotter and 3c hotter on chain 1 from normal stock firmware from 4/23. After 2hr5m all 3 ASIC chains went offline, hash rate dropped to zero, after several restarts the miner would NOT remain stable for more than 20mins and I had to revert back to non-OC firmware. Good job so far, getting close.

[chipless]

Chipless your firmware ran for 2hr5m without any major issues. The average hash rate did increase 5-6k over stock. A big increase on HW errors, hundreds more on each board. The increase in voltage made ASIC chains 0 & 2 5c hotter and 3c hotter on chain 1 from normal stock firmware from 4/23. After 2hr5m all 3 ASIC chains went offline, hash rate dropped to zero, after several restarts the miner would NOT remain stable for more than 20mins and I had to revert back to non-OC firmware. Good job so far, getting close.

Thank you for the input. I am working on an update that should fix many issues that may occur in certain situations. The speed should be increased alo to about 150k....There will be more heat no matter what firmware you use when you start to overclock them. I suggest manually setting the fan higher or cooling it better for now.