|
Sorry folk, I've been trying to track down a bug and have been distracted the last couple of days.
Unfortunately the bug is with flypool -- rather, I hope it is a bug. There are conditions where flypool will acknowledge receipt of a valid share and never associate it with your wallet. I have no idea what wallet it goes to, but it isn't the one configured and accepted.
That is bad, very very bad.
I posted this on a zcash/flypool thread:
--snip--
Why does flypool not follow the stratum protocol when it comes to mining.authorize? Per the protocol, a single connection should allow multiple mining.authorize calls. Flypool does not follow this, and worse, with some finagling, it will provide a valid authorization for more than one wallet, but never pay out, all the while returning “true” to the submitted shares. Which leads to my next question…
Why does flypool accept shares for wallets and then never associate the shares with the wallet? The simplest case is one where you mining.authorize walletA but perform all of the mining.submits with walletB.
In that scenario, flypool acknowledges the share acceptance for walletB, but never pays walletB. Where do those shares go? Who is keeping them? Why does flypool respond with an acceptance of the shares?
I’ve been using flypool a while and recommending it strongly, but until I can understand this behavior, I can no longer make that recommendation.
This is how flypool responds to a mining.submit with an address it did not accept as authorized:
{“id”:2001,“result”:true,“error”:null}
This is how a proper pool should respond to a mining.submit with an address not authorized:
{“id”:6,“version”:“2.0”,“result”:false,“error”:[24,“Unauthorized User”,null],“Notification”:{“method”:"",“params”:null}}
The difference is in the first case, flypool seems to take the share. Are there other conditions where flypool silently takes shares?
While trying to investigate this, I was banned from the flypool API.
--snip--
I've tried to reach out to them, but for the time being I am personally going to remove work from flypool until I can get an explanation.
Thank you,
Jason
|
|
|
|
|
Update:
Much to my dismay, I am not going to hit my self imposed deadline of tonight for the next update (paid users included). I am continuing to work on it and will release as soon as I am able to get all of the pieces together. This next update is reasonably large in terms of infrastructure support and will unify the paid and dev-fee supported images, as well as the Z9 and Mini images into one (not 100% locked on the mini yet).
The mini image has been running for about 24 hours stable, although I am not yet happy with the results as it requires a lot of effort to manually tune the units at the moment.
From an "internals" point of view, I am making very good progress on gaining control of other components of the system.
I apologize for the (continued) delay.
Thank you,
Jason
|
|
|
|
So are you saying the only version that will be released is the dev-fee version unless you have a certain amount of machines then you will do a non-dev fee version offered privately?
I'm willing to discuss options... I just simply haven't decided, however. When I was selling, people said they preferred a dev fee, when I've released a dev fee, people have said they wanted to buy. I may end up with a solution for all of these use cases.... coding on it all as quick as I my fingers and brain will work right now. Jason |
|
|
|
Already see the end of the week....  How is progress with the new firmware versions and how much will the version for the Z9 mini cost? likely this will be dev-fee only going forward. options exist for volume customers, pm to discuss. Jason |
|
|
|
If I already paid for your software can I get this version without the fee?
Post #247 3) Z9 pre-paid users: I'm hoping to complete the work this weekend and get that release out.
Once it has been done, Z9 prepaid, dev-fee, and mini users will all be on equivalent functionality. Future updates/features will be a lot quicker across the releases once I get things figured out on one platform.
The weekend ends around 3am eastern Monday AM from my perspective. ;-) -j |
|
|
|
Does anyone know the difference between a rev 1.0.11 and rev 2x acic hashboard between the batch 1 and batch 2 z9 minis. Some batch 2 and 3 have a revision 2.x while the batch 1 has a rev 1.0.11
something else is the minis have a 1 to 3 ratio of accepted vs discarded while the z9 has a 2 to 1 or so ratio less discarded on the z9 compared to the mini
mini: 2716 accepted, 840 discarded Z9: 2567 accepted, 4380 discarded random sampling. -j my z9 has 34k accepted and 23k discarded where 1 mini has 1390 accepted and 3470 discarded other mini is similar. I am running my minis at 756mhz have rand the one as high as 800 but I board falls out the z9 is at 656mhz. My biggest curiosity is what did they change between the batch 1 and 2 boards revisions. Are you sure you have your versions correct? Here are a couple of examples I have data for.... Batch1: Board BEZ10601_Hashboard-4 V1.1 Model# Z9 mini_V1.1010 - Production date 2018-06-11 / JF0018BB4 BIN1 AV32 TMP45 XXXX Batch2: Board BEZ10601_Hashboard-4 V1.1 Model# Z9 mini_V1.1011 - Production date 2018-07-19 / SF0018BM1 BIN1 A03 TMP451 XXXX High resolution pictures of the front and back of the boards would be very useful. Thank you, Jason |
|
|
|
Does anyone know the difference between a rev 1.0.11 and rev 2x acic hashboard between the batch 1 and batch 2 z9 minis. Some batch 2 and 3 have a revision 2.x while the batch 1 has a rev 1.0.11
something else is the minis have a 1 to 3 ratio of accepted vs discarded while the z9 has a 2 to 1 or so ratio less discarded on the z9 compared to the mini
mini: 2716 accepted, 840 discarded Z9: 2567 accepted, 4380 discarded random sampling. -j |
|
|
|
Zcash - scam...
Brand new account, first post -- definitely plausible commentary. Jason |
|
|
|
eagerly awaiting the Z9 mini software change! get these baby to ROI and make some income before it's a doorstop next year!
I'm doing some testing with a batch 2 mini now. I need to return more than 3% improvement at a minimum, and if I can do that on the "batch 2" units, things should be good on the other batches as they are less finicky. Z9 paid release is still on target for this weekend. Thank you, Jason |
|
|
|
|
Folk,
Quick update;
1) Fan controls -- I am making headway on this, but not quite there yet. These units do not handle fan controls like past antminers.
2) Z9 Mini -- loaner Z9 arrived today. Initial modifications to the firmware have been made and are being tested - they will run at least overnight to observe and identify any potential issues that need to be resolved.
3) Z9 pre-paid users: I'm hoping to complete the work this weekend and get that release out.
Once it has been done, Z9 prepaid, dev-fee, and mini users will all be on equivalent functionality. Future updates/features will be a lot quicker across the releases once I get things figured out on one platform.
Thank you,
Jason
|
|
|
|
I give you props for the pretty UI... you beat me to that. However there is no way you have per hashboard frequency control.
Good luck!
-j
i just want to say something very negative right now. Yeah, my response was kinda douche. My bad.... I just know how much work I have put into additional functionality in these binary only releases and have a very hard time believing someone else Is at that point. -j |
|
|
|
|
Update on the mini; a loaner mini is arriving tomorrow... I should be able to get the release out the same day.
Paid variant is ongoing...
FAN control is making progress... bitmain did something much different than past miners with the fan control. I’m still working on reverse engineering that functionality.
Jason
|
|
|
|
i have been running this FW on a Z9 batch 2 @650m for 17hrs to test it out, so far no errors here everything working as normal including temps.
Awesome -- hope it works out well for you! If not, you know where to look. :-) Jason 24hr using this firmaware with around 56k avg @650m and made the same as using the original firmware with turbo mode 46k avg @550m. something not right i guess, gonna try your firmware tonight at 12 and let it run for a day and see the difference. That's weird, regardless of the firmware -- are you on antpool, by chance? Jason prohashing, it made $15 in 24hr . is that even good ? I don’t look at it in USD... what was your coin earning and which coin? 0.00242502 BTC - the settings are to were it switch to the highest coin at the moment . so random coins. im using your firmware since 12am to test it out. do you recommend mining a specific coin instead of the auto switching ? Yeah, I suggest not auto switching... switching takes time, costs hash, and is often more gimmick than success with asic miners.... my personal experience at least. I personally mine just ZEC and sell it when it makes sense to. -j |
|
|
|
|
I give you props for the pretty UI... you beat me to that. However there is no way you have per hashboard frequency control.
Good luck!
-j
|
|
|
|
i have been running this FW on a Z9 batch 2 @650m for 17hrs to test it out, so far no errors here everything working as normal including temps.
Awesome -- hope it works out well for you! If not, you know where to look. :-) Jason 24hr using this firmaware with around 56k avg @650m and made the same as using the original firmware with turbo mode 46k avg @550m. something not right i guess, gonna try your firmware tonight at 12 and let it run for a day and see the difference. That's weird, regardless of the firmware -- are you on antpool, by chance? Jason prohashing, it made $15 in 24hr . is that even good ? I don’t look at it in USD... what was your coin earning and which coin? |
|
|
|
Really late 2019 by the time that comes the difficulty will be outrageous and asic miners will fork the chain this is just zooko blowing smoke up GPU miners asses hoping to regain the support they lost, here's another coin that's centralized and controlled by one man that's gotten rich of the back of GPU miners he told he no longer cares about well he already screwed up the relationship and there's no turning back now
This. Zooko’s statement has no meanining was my read. “A year from now” is forever in crypto. |
|
|
|
i have been running this FW on a Z9 batch 2 @650m for 17hrs to test it out, so far no errors here everything working as normal including temps.
Awesome -- hope it works out well for you! If not, you know where to look. :-) Jason 24hr using this firmaware with around 56k avg @650m and made the same as using the original firmware with turbo mode 46k avg @550m. something not right i guess, gonna try your firmware tonight at 12 and let it run for a day and see the difference. That's weird, regardless of the firmware -- are you on antpool, by chance? Jason |
|
|
|
i have been running this FW on a Z9 batch 2 @650m for 17hrs to test it out, so far no errors here everything working as normal including temps.
Awesome -- hope it works out well for you! If not, you know where to look. :-) Jason |
|
|
|
I decided to test the mini 1 last time before sending it to Bitmain. I tested 1 hashboard at a time and and it turned out 1 of the board was causing the PSU to trip. After opening it up, found the burned hashboard. Mini is hashing fine with 2 good boards now. 3 other Z9s and 2 A9s in the same server case are running fine.   Server Case, it is vented on top with a 1500 CFM QC-ES-1500 fan.  I don't think that was a heat problem It looks more of a surge or a short on the chip pins inorder for it to explode out and burn the board like that it would need a full short circuit or surge with good amperage. Heat will brown or discolor the board and the component just fails internally in most cases that I have seen in 30+ yrs of electronics Adding to this, were you trying to adjust chip voltages? -j |
|
|
|
when can we expect a paid firmware?
The paid update is going to need to have extensive "QA" performed on it since the impact of it failing will be large. I do not want to rush it out the door. I will update as soon as I can. Thank you, Jason |
|
|
|
|
Show Posts
