akadamson just cheated a few.
He extrapolated the results of mining short period (couple of hours) at latest difficulty at slush to a longer period (1 month, at more more little difficulties). This obviously created an error: luck of profit in slush.
Later on he calculates the real profit obtained mining at Kano and compares with the "extrapolated" profit in slush.... and the conclusion is:
Almost the same!! isn't it marvellous?
The worse thing here is that legendary one applaud this and even give merit...
I hope no new miners will come and trust this non sense without verifying the data!!! we are talking about loosing money!!!
I'd really like to know how I *cheated*. Let's work the numbers *another way*... Let's take the Minergain.com CDT from slush on their weekly numbers, that should remove the diff topic that you referenced (which is wrong BTW, but I'll leave that alone). Slush's CDT is 0.00008985 my calculated CDT for slush against my numbers was .00008703 Hmm... that's pretty dang close wouldn't you say? Enough that is would have changed the numbers *slighly* but not that much. So let's look at it again. @ 30Th/s that would be .0026955 BTC a day over 35 days, that would have been .094325 BTC Darn it, Kano still won.... hmmm |
|
|
|
|
Further to the above, I had provided my findings to Canaan and received this email in response.
e.g. there is an option to set a -2 to 1 voltage offset, this worked with the 741s, but with the 821, that option is there in the web interface, but the command that is actually used to start the miner,
*always* forces the 0 offset and ignores the value set in the web UI.
>> This option is for A741, sorry for we didn't delete this function from A821's firmware.
There is one other around the min fan speed, it defaults to 100%, what probably makes more sense is to set to that 50 by default. The outlet temperature is set to 70, where as on the 741 it was 90, because of this with the default min fan set to 100%, the fans run at 100% speed whether they need to or not.
>> We locked the fan speed in the A821 firmware since its chips cannot take high temperature.
Again, I wouldn't change that 70 on the temp... the fans shouldn't be a problem and it appears the undervolt will be coming out soon
|
|
|
|
I keep seeing all this about pool x, pool y, pool z vs here or maybe I should just say PPS vs here.
So after 35 days I figured it was time for a few real numbers.
I started here on the 18 (night) of January, with what was 15Th as the time, but that quickly doubled to 30 a day or 2 later and has been pretty flat since (just added 22 more, but I'll ignore that at the moment)
The other day when I brought up my 821's, I put them on slush for 2 hrs, just because I wanted to see what the daily CDT was there at my full hashrate. Then I moved them back to here.
So here is the basis to my numbers.
30Th/s for 35 days
Actual numbers from Kano based upon the above.
Let's start with my Kano experience.
I received .0976473 BTC in 35 days
this equals .0027899 per day
or .0000930 CDT
for the same amount of hash at slush, we'll work backwards
.00008703 CDT
.0026109 per day
or .0913815 BTC
You can say what you want and this is just a snapshot of 35 days, but even with the really ugly luck we've had, I've still made more here than the equivalent time at slush (who btw is averaging slightly *better* than 100% luck).
The upside that isn't in the above, if we find a block today, my numbers would go up again to a much larger amount of BTC than slush (e.g. .0045 (would be .007, but I'll factor out the .0027899 CDT) BTC to .002), if we improve on luck, we shatter the slush numbers. If I left today and went somewhere else, I'd still be paid for 5Nd ramp down along with where-ever else I go's take.
I just can't see any downside. Ok, except the fact that you don't have a little graphical ticker that is counting up based upon predicted BTC.... Oh but wait, it simply matches the number above, so what real benefit is it?
EDIT1: with 30Th/s this should work for anyone curious with a 14-15Th/s S9 or anything else because I took all the numbers down to CDT (Coin per Day per TeraHash/s)
Can I highlight one other thing. If you are getting into this to *make money*. Please understand, it took me 35 days @ 30Th/s to make *almost* .1 BTC, that means if I were a single S9, it would have taken approx 70 days to do the same. And that is only going to get harder with the growing network hash rate and diff changes. But most important. 70 * 10 is ~700 days to make 1BTC - that's a freaking long time (like almost 2 years) and why most people say if you *just* want to make money in this game *don't mine* go trade currencies! |
|
|
|
Thanks. Unfortunately I don't have a proper sound level meter, but I'm estimating that the noise level is at least 65dB (as advertised at www.canaan.io) , when the cooling fan speed is 100%. Most likely 1m distance is used for the advertised noise level measurement. These guys are a little quieter than the s9's...to my untrained and slightly deaf ears.  A little...  ... Sheesh the sorcerers are louder! lol |
|
|
|
Tried changing the frequency and so far, it doesn't work well with any setting other than 650. I tried 600, 625. 636, and 675. For each one I lt it try to hash for about 30 minutes and each time the overall hash rate barely got above 2TH. Setting it back to 650 and it went back to about 11TH immediately.
My guess is that the chips on the 821 are pretty much set to only work right at the 650 frequency. Of course, I could be doing something wrong so maybe someone else will have success.
So it looks like one setting that does work well is the "--avalon8-voltage-level". I set mine to -1 (range is supposed to be -15 to +15) and it is running cooler by 10 degrees and the fan is staying around 65-70% with a hash rate about 9-10TH. (Plus, it is much quieter now.) I also changed the target temp to 90 which makes the fan run more like the 741s. I'll let it run overnight and take a screenshot tomorrow. It will be interesting to see if there are power use differences when HagssFin retests.
Here is what I put in More Options: "--avalon8-voltage-level-offset -2 --avalon8-temp 90 --avalon8-voltage-level -1"
I'm more than a little leary of setting the temp to 90. Most chips today are rated for -20 to 80C, with a few exceptions that go to 105C (rare exceptions). And because *specifically* the developers at Canaan, changed the 90 which the 741's use to 70 in the code. I did see that overriding the voltage-level-offset made a difference, haven't tried the other, but for now, I'm going to leave mine running at a max temp of 70 and just work on keeping the inlet temp down. |
|
|
|
|
I keep seeing all this about pool x, pool y, pool z vs here or maybe I should just say PPS vs here.
So after 35 days I figured it was time for a few real numbers.
I started here on the 18 (night) of January, with what was 15Th as the time, but that quickly doubled to 30 a day or 2 later and has been pretty flat since (just added 22 more, but I'll ignore that at the moment)
The other day when I brought up my 821's, I put them on slush for 2 hrs, just because I wanted to see what the daily CDT was there at my full hashrate. Then I moved them back to here.
So here is the basis to my numbers.
30Th/s for 35 days
Actual numbers from Kano based upon the above.
Let's start with my Kano experience.
I received .0976473 BTC in 35 days
this equals .0027899 per day
or .0000930 CDT
for the same amount of hash at slush, we'll work backwards
.00008703 CDT
.0026109 per day
or .0913815 BTC
You can say what you want and this is just a snapshot of 35 days, but even with the really ugly luck we've had, I've still made more here than the equivalent time at slush (who btw is averaging slightly *better* than 100% luck).
The upside that isn't in the above, if we find a block today, my numbers would go up again to a much larger amount of BTC than slush (e.g. .0045 (would be .007, but I'll factor out the .0027899 CDT) BTC to .002), if we improve on luck, we shatter the slush numbers. If I left today and went somewhere else, I'd still be paid for 5Nd ramp down along with where-ever else I go's take.
I just can't see any downside. Ok, except the fact that you don't have a little graphical ticker that is counting up based upon predicted BTC.... Oh but wait, it simply matches the number above, so what real benefit is it?
EDIT1: with 30Th/s this should work for anyone curious with a 14-15Th/s S9 or anything else because I took all the numbers down to CDT (Coin per Day per TeraHash/s)
|
|
|
|
So at least for me a light bulb just went off.. I always thought about a block as a progress to completion, as such I always wondered what happened when a pool found a block, and what happened to all of the work that other pools/miners had done, were about to do. Then I found this... "Contrary to popular belief, mining is not something where there is progress. Each hash has the same probability of being a valid block hash. You could get lucky and find a valid hash with your next hash, or you could not. There is no progress that is made. When you mine on a lower difficulty, the target that your hash must be under is much higher than that of the networks. With enough hashes, someone will eventually find a hash that is lower than the network target, and that hash would also be lower than their pool target. The higher target (thus lower difficulty) is only for tracking who is actually doing work and how much work they are doing so that they can be paid appropriately." from here - https://captainaltcoin.com/what-is-pool-mining/So this then solves my problem with what happens when a block is found by another pool. It just an event, and we continue on trying to find another hash that meets the requirements of a *block* |
|
|
|
Not the same, kano is paying out of his pocket. We miners aren't paying for it via some of the block reward getting diverted to the 'winner'
(ps. I know the quote above isn't the target of my comments below) What is it with people... Do they not read? Sheesh The 10BTC is *NOT* coming from miners rewards, transaction fees, block fees, etc.... It's a completely separate source of BTC that the pool owner has and for some crazy reason (cuz, even with the drop in BTC price today, it's still *a lot of money*), is willing to contribute to actually help those that are mining on his pool. |
|
|
|
|
What we all need to remember is
a) the 10BTC incentive does *NOT* come from miners share, it's a separate, funded by separate means, set of funds
b) it doesn't *NOT* penalize the small guys any more than their small hashrate does
c) BUT WHAT IT DOES DO is get us back to the diff curve and hopefully ahead of it.
d) it will have a dilutive effect on your payments, but only because of the relationship of your shares to the pool shares
e) it will shorten the *5Nd* ramp time for everyone
f) and most important if *should* have the effect of causing more blocks to be found in the equivalent amount of time.
d) is really the only thing that might be of issue with some, but f) should more than make up for that.
Just my .05,
|
|
|
|
What results did you get?
So far it is running around 86% fan speed (it was 100% before) with chip temp at 67 C. The hash rate is staying between 10.5 -10.8TH - with the offset of 0, it was 11.2-11.5TH. Overall it is much quieter but I will probably keep working with it tomorrow to see that new fan ramping point can be adjusted and may the shutdown temp set to the standard 99. Screenshot after one hour with offset of -2 (sorry I don't have a before screenshot - if I have time I'll try to set it back tomorrow and let it run for an hour to directly compare): https://i.imgur.com/WRCu5Gf.jpgto get to the Avalon8 version of cgminer, make sure you select the avalon 8 branch in git. The shutdown temp on both that 741 and 821 is 105 in their respective versions of cgminer (you can see the 7's version by selecting the avalon7 branch in git) I found that the min fan value was picked up from the configuration screen, by default it was 100%, but if you changed that to 50%, you would get the 50-100 range in to the process listing of startup options. I had tried the -x values, but didn't even think to look at those being overriden. |
|
|
|
I posted a note about the voltage offset in the Avalon 8 official specs thread ( https://bitcointalk.org/index.php?topic=2616205.msg30730555#new). It looks like the values for voltage offset enter on the CGMiner Configuration page are being ignored and the offset and some other settings are getting applied to the command line of the cgminer. Try entering a value there other than zero and look in Processes - at the bottom of the list you should see the cgminer command line and it always has offset of 0 hardcoded. It looks like we can override if we use "More Options" with "--avalon8-voltage-level-offset -2" (figured that out by taking akadamson suggestion and then finding the developer documentation for cgminer on Canaan's github. For those who don't know or are new to this, the command line options can be found here (just look for the avalon7 options and maybe the options for avalon8 will be similar): https://github.com/Canaan-Creative/cgminer/blob/master/READMESo far it is running around 80% fan speed (it was 100% before) with chip temp at 67 C. The hash rate is staying between 10.5 -10.8TH - with the offset of 0, it was 11.2-11.5TH. Overall it is much quieter. I don't have a way to test the power usage so if someone else does, it would be a good thing to test. Here are the 8 options - Make sure you select the Avalon8 branch to see the version shipped with that firmware starting at line ~1500 https://github.com/Canaan-Creative/cgminer/blob/avalon8/cgminer.c |
|
|
|
Sheesh, fans set for 100% all the time in the 821 firmware... Anyone played with setting it back to something reasonable like 50% or 20, like I ran the 741's and let the temp circuit moderate the fan speed as needed to keep 90 degrees on the outlet? I guess I'll see what outlet temps I've got first, but my basement this time of year is pretty dang cold, and I'd just as soon trade off some noise if appropriate.
What's your 'minimum fan speed' setting in the Cgminer configuration? I'm guessing it is 100. Set it to 20 and see how the miner acts in your environment. Note that 20 is only the minimum value and miner will control the fan speed up or down in relation with the miner temperature. yes, by default the firmware has 100% in that min fan speed and I suspect I know why, the internal temp that they attempt to keep the outlet at is 70 and to do so, it appears to take 100% fan typically. You can put something like: --avalon8-fan 1-80 --avalon8-temp 90 in the 'More Options' box if you want to change the target temperature and fan speeds. Haven't played around with it too much but 80% fan speed still kept mine at a reasonable temperature so far. The difference between 80% and 100% fan was only a couple degrees after a few minutes. Yeah, but I wonder if they are *trying to keep* the outlet temp at 70. The 741 runs at approx 90 and thats what set value it has in the cgminer code. So someone must want the 821's to be at 70. I'd hate to adjust that and burn up an 821. I suspect the current/voltage changes pretty dramatically as the temps rise on the ASICs. And from others testing, the -1 undervolt has the opposite effect - hashrate and wattage goes UP, not the other way around... Be curious of others findings? Bit more info here after running these for a few days. a) I would change the "min fan speed" in the configuration to 50, from the 100 as the firmware came b) I would *NOT* change the max temp from the default of 70. I went back through all the git merges on the 821 code and that max temp was specifically changed from 90 which the 741's ran at to 70 and my guess is they want to keep the chips *COOL* and not let them get hot. The 105 number is still in there as the autoshutdown number, but that's a pretty scary temp and I would avoid at all costs letting it get much above 80 or 90 for any length of time. Because they are trying to keep the ASICS cool, if you shut one down that 50% min fan will still force enough air over them to cool them. The problem with aluminum is it works both ways, it will dissipate heat when used to cool, but if not cooling it will hold the heat and cook things. By lowering the min fan to 50, if your environment will allow for an inlet temp that can keep the outlet at 70, with a lower than 100% fan, the fan moderation will kick in and you'll save some noise. I noticed this today, my room temp was way below 30, and the outlet was 70, but the fan was at 85% which was way better than the default of 100 all the time. Just remember, you don't want them above 70 for any length of time (or so it would appear that Canaan doesn't) but if your environment will allow you can drop down the min fan. But leave it at something that would rapid cool the headsinks if the unit drops offline or is stopped. ps. this is why I pretty much go to an *unplug enet* when I want them to shut down gracefully. This keeps the rPi controller in the loop and once they run out of hash, they go to the min fan temp while they cool. Then I can unplug the I2C cable and turn them off. |
|
|
|
Do you think the reason it is taking longer for us to find blocks is because the number of bitcoin transactions are decreasing? Take a look at this chart - https://blockchain.info/charts/n-transactions#Or is this something that is unrelated, and just happens to be dropping at the same time? There was a burst in number of daily transactions in January just around the time we had a burst in finding blocks. It's taking longer, because the network difficulty continues to increase every 13 ish days, but our hashrate has stayed constant for the last little while. in fact, network difficulty is expect to go up 5% tonight overnight (US nite time) |
|
|
|
|
ok 22Th more is now online!
BLOCK ON!
|
|
|
|
Sheesh, fans set for 100% all the time in the 821 firmware... Anyone played with setting it back to something reasonable like 50% or 20, like I ran the 741's and let the temp circuit moderate the fan speed as needed to keep 90 degrees on the outlet? I guess I'll see what outlet temps I've got first, but my basement this time of year is pretty dang cold, and I'd just as soon trade off some noise if appropriate.
What's your 'minimum fan speed' setting in the Cgminer configuration? I'm guessing it is 100. Set it to 20 and see how the miner acts in your environment. Note that 20 is only the minimum value and miner will control the fan speed up or down in relation with the miner temperature. yes, by default the firmware has 100% in that min fan speed and I suspect I know why, the internal temp that they attempt to keep the outlet at is 70 and to do so, it appears to take 100% fan typically. You can put something like: --avalon8-fan 1-80 --avalon8-temp 90 in the 'More Options' box if you want to change the target temperature and fan speeds. Haven't played around with it too much but 80% fan speed still kept mine at a reasonable temperature so far. The difference between 80% and 100% fan was only a couple degrees after a few minutes. Yeah, but I wonder if they are *trying to keep* the outlet temp at 70. The 741 runs at approx 90 and thats what set value it has in the cgminer code. So someone must want the 821's to be at 70. I'd hate to adjust that and burn up an 821. I suspect the current/voltage changes pretty dramatically as the temps rise on the ASICs. And from others testing, the -1 undervolt has the opposite effect - hashrate and wattage goes UP, not the other way around... Be curious of others findings? |
|
|
|
Sheesh, fans set for 100% all the time in the 821 firmware... Anyone played with setting it back to something reasonable like 50% or 20, like I ran the 741's and let the temp circuit moderate the fan speed as needed to keep 90 degrees on the outlet? I guess I'll see what outlet temps I've got first, but my basement this time of year is pretty dang cold, and I'd just as soon trade off some noise if appropriate.
What's your 'minimum fan speed' setting in the Cgminer configuration? I'm guessing it is 100. Set it to 20 and see how the miner acts in your environment. Note that 20 is only the minimum value and miner will control the fan speed up or down in relation with the miner temperature. yes, by default the firmware has 100% in that min fan speed and I suspect I know why, the internal temp that they attempt to keep the outlet at is 70 and to do so, it appears to take 100% fan typically. |
|
|
|
Sheesh, fans set for 100% all the time in the 821 firmware... Anyone played with setting it back to something reasonable like 50% or 20, like I ran the 741's and let the temp circuit moderate the fan speed as needed to keep 90 degrees on the outlet? I guess I'll see what outlet temps I've got first, but my basement this time of year is pretty dang cold, and I'd just as soon trade off some noise if appropriate.
hmm, just looked at the cgminer code. The normal temp on the 821's is set to 70. And even in a cold room, the outlet temps at 100% fan are just at 70 so turning the fan min down isn't going to do anything.... curious |
|
|
|
|
Sheesh, fans set for 100% all the time in the 821 firmware... Anyone played with setting it back to something reasonable like 50% or 20, like I ran the 741's and let the temp circuit moderate the fan speed as needed to keep 90 degrees on the outlet? I guess I'll see what outlet temps I've got first, but my basement this time of year is pretty dang cold, and I'd just as soon trade off some noise if appropriate.
|
|
|
|
OK by me You can also ask me in PM how long they have been mining with the low hash rate and I'll check. Obviously I wont be giving out detailed answers, but I will let you know if they indeed are a longer term miner with a low hash rate. The Rewards page shows that clearly. cool.... will check with you before announcing. Nice job!... Looking forward to the winner! |
|
|
|
Signed up pointed my miners over lets hope they find a block here at some point they were on a solo mission for the past year . DrSha is in the building.. https://imgur.com/a/ac26SThanks for that info kano that cleared up the confusion surrounding that. wooo hooo all the best to the tiny guys (I'm the little guys with 40Th... lol) |
|
|
|
|
Show Posts
