Topic: DragonMint T1 16TH/S halongmining.com - page 32. (Read 87857 times)

They're designed to run at 70-75 degrees. What's yours running at?

this is the 4-15 firmware

https://download.halongmining.com/testing/t1_20180415_075807.swu

Where did you get a link/download for a new firmware?

There is an address that is accessible through public means! I hinted at it nearly a week ago in this thread. It doesn't give much information aside from potential upcoming models/names. The prominent figures are Yoshi and Drak, and probably other investors. Innosilicon confirmed their partnership with Halong. That, along with their ongoing deliveries, is legitimate enough for me. While I can see the perspective of people limited on where they can mine, I'm hopeful more pools adopt support for AB.

I'm personally a pretty happy customer so far with my experience with Halong. I'm eager to see what their next moves are, with regards to general releases and next iterations of hardware (and firmware!)

You choose to see what you want to see.

1. The lack of disclosure on the legality of the company is still a thing.
2. The fact that their miners are a rip off of bitmains design and Innos controllers shows they are not all that innovative.
3. The fact that you can literally only mine to 2 BITCOIN only pools was never disclosed. There are other coins a sha256 miner should be able to mine, but the fact that this miner was restricted to bitcoin only was not revealed until MONTHS after presale.

So let me get this straight....you are comparing an underclocked dragonmint against a full speed s9. Seems like a flawed test from the get go. I can probably go find 100 machines in my DC right now that will match that efficiency if i cherry pick them, and probably a lot more if I used a firmware that allowed me to underclock.

The real question is: what is preventing pools from adding the support for ASICboost? The official Core Bitcoin client always accepted blocks that were mined with boosting. The recent change was just to remove warning about unexpected forks or version numbers or some such. The required code change is pretty much trivial.

Personally, I see quite a few people felling ashamed that they gotten themselves drawn into the ASICboot propaganda and counter-propaganda. This just increases the inertia of the adoption of it.

Certainly, on the pool side the patents don't apply.

Edit: So you ninja-edited your post:


More or less standard operating procedure when developing a product that could be subject to contested patent. It appears that your familiarity with patent system is only at the head-end: how to get a patent and hang it on the wall. You don't appear to show any familiarity with the ass-end: how to extract the royalties from unwilling competitor and how to defend yourself against litigation trying to turn your precious framed certificate into a roll of toilet paper.

Samsung has 'been in the miner chips' biz for several years - not new. They made the 14nm chips for BW.com. What is new is they're now taking more clients since TSMC can't supply all needed.

What 'we' want is the the miners to work on ALL pools - not just ones that have their code tweaked to accept Little Dragon's AB design that Halong is using. Nothing at all was said about this until 3 months after sales started. Folks purchased them under the impression they would work on ALL pools - not just a very select few.

Then there is Halong and Little Dragon Technology LLC's (holder of the AB patent) intense secrecy as to who actually owns/runs them. Hell they won't even give a City/Provence/street address for contact. All they give is their Lawyers:
Little Dragon Technology LLC
c/o Getech Law LLC
203 N LaSalle, Suite 2100
Chicago, IL 60601
USA

This is called https://en.wikipedia.org/wiki/Blowback_(intelligence) , mostly related to all the propaganda about ASICboost in the past. And now even you increase the blowback but further spreading rumors about Samsung entering coin mining market, when they were just a contract foundry for a fab-less design boutique.

https://en.wikipedia.org/wiki/Live_by_the_sword,_die_by_the_sword
 
https://en.wiktionary.org/wiki/you_can%27t_make_an_omelette_without_breaking_eggs

Some people really went into alternate, schizoid mental state trying to comprehend the ASICboost propaganda and all the nonsensical related calculations of supposed profitability or supposed lack of profitability.

With all the superstitious thinking in the coin mining business all I can do is post the picture of a shrine to https://en.wikipedia.org/wiki/Saint_Barbara , patron saint of miners.

I don't get these HalongMining haters:

1) We have one of the world biggest chip manufacturers (Samsung) entering the miners race
2) 100% legal by Innosilicon confirming they are in partnership with Halong
3) Confirmed lowest W/GH on the market after the firmware update
4) Shipping date respected from Halong even for series 0 equipment
5) Price of  Bitmain gear drown down significantly because of healthy competition
6) Finally a realistic chance to break evil monopoly on mining gear

Instead of jumping up and down Halong did all this in a few months, I see constant nagging and whining and bad energy. What more do you want?

90.11W/TH on my test on efficiency settings. This is on 240v/MyRig PSU, on the beta firmware. Pretty good!

Okay...I will give it a try. What settings have you kept for the fan...high medium or low? Are you getting close to 16T for the efficiency setting? Thanks.

No mine run cool on efficiency setting.

Have you tried the 4-15 firmware update?

it has 3 speeds for settings and fans.

Is anyone having issues with the T1's running way hotter than the S9's? Any input? Thanks.

it has a hash rate average line  most of the T1's
on Efficiency   run from 13.6 to 14.1



 I multiple sample  it  and I compare to pool hash rate
here are the five at the solar array.  The 1 hour would be best to use  since the
1 wk is off due to less then  a week of run time
the 1 day is off  due to an internet drop during yesterdays spring storm


http://ckpool.org/users/16yLHLoeyuCLPMXkVpC3gyrRYvwRGwjKJr


here are  the five running from above  note I shifted them around


{ "workername": "16yLHLoeyuCLPMXkVpC3gyrRYvwRGwjKJr",
   "hashrate1m": "12.8T",
   "hashrate5m": "13.6T",
   "hashrate1hr": "13.7T",      This rounds at 1 digit  but is fairly close to the gui
   "hashrate1d": "35.4T",
   "hashrate7d": "49T",
   "lastshare": 1523970383,
   "shares": 11088010825,
   "bestshare": 65201213603.0,
   "lns": 11053409665.09229,
   "luck": 0.99,
   "herp": 10993567502.98039}

{  "workername": "16yLHLoeyuCLPMXkVpC3gyrRYvwRGwjKJr.1",
   "hashrate1m": "10.4T",
   "hashrate5m": "11T",
   "hashrate1hr": "13.2T",  Same here
   "hashrate1d": "12.7T",
   "hashrate7d": "7.02T",
   "lastshare": 1523970386,
   "shares": 1399594270,
   "bestshare": 1012819976.0,
   "lns": 1397664249.57471,
   "luck": 1.0,
   "herp": 1393773376.731752 },


{  "workername": "16yLHLoeyuCLPMXkVpC3gyrRYvwRGwjKJr.4",
   "hashrate1m": "16.7T",
   "hashrate5m": "14.8T",
   "hashrate1hr": "13.9T",
   "hashrate1d": "12.8T",
   "hashrate7d": "7.08T",
   "lastshare": 1523970386,
   "shares": 1412797275,
   "bestshare": 1129180623.0,
   "lns": 1410803253.588209,
   "luck": 1.0,
   "herp": 1413966576.42499 },

{ "workername": "16yLHLoeyuCLPMXkVpC3gyrRYvwRGwjKJr.6",
   "hashrate1m": "15T",
   "hashrate5m": "13.4T",
   "hashrate1hr": "13.4T",
   "hashrate1d": "12.7T",
   "hashrate7d": "5.98T",
   "lastshare": 1523970380,
   "shares": 1118274455,
   "bestshare": 619392421.0,
   "lns": 1118033669.763529,
   "luck": 1.0,
   "herp": 1119320405.892049 }

this has a dropped board will need to reset it
  {   "workername": "16yLHLoeyuCLPMXkVpC3gyrRYvwRGwjKJr.3",
   "hashrate1m": "10.3T",
   "hashrate5m": "9.36T",
   "hashrate1hr": "9.18T",
   "hashrate1d": "10.2T",
   "hashrate7d": "5.79T",
   "lastshare": 1523970370,
   "shares": 1139313707,
   "bestshare": 2769509710.0,
   "lns": 1138007124.306525,
   "luck": 1.0,
   "herp": 1141689357.810393},


All this is a work in progress.
I will be doing this with the avalon 841's  buysolar brings back from China

Using live hash rate value for power consumption calculations is not that great of an idea.

Does the T1 show an average hash rate value in the web status page?

For example, if I look at the live hash rate value (GHS5s) for Avalon841,
it shows a whopping big value of 32.5Th/s hash rate for two A841 units setup.
So about 16Th/s per A841 unit, for a short time period.

@ Thetaj

so 1 s-9  was a unicorn  unit.

I did have an S-7  that would clock higher then any other one.  It also did .234 watts a gh  by far the best of the 20+ s-7's I owned.

I ran the T1 over night

it is using the corsair Ax1500i

13720 gh

https://i.imgur.com/8OTiGXs.png

watts are 1318

that is 1318/13720 = 0.09606 a gh

but in video  hash jumped to 13920  that is

1318/13920 = 0.09468

final video using the corsair Ax1500i is here


Your video will be live at: https://youtu.be/2qbx9voYEQg

MY thanks to this loaner from a forum member.

He cherry picked his middle sample of five.

I will be going to the solar array and running  tests on the five units I have there.



note one of the five  units  dropped a board at the efficiency setting and will need a reboot and a re set to balanced setting

Alright boys, this is the ABSOLUTE BEST S9 I can find out of ~100 of them that I have.

I do have some of the older models that aren't Auto-tuned but they all turned out to be pretty bad Hashrate-wise so I didn't bother with them.

So I have 40 14th S9s. They are all ~October-December batch. Honestly I don't remember when each specific machine arrived (and am too lazy to look up the SN).

Full Disclosure: How I chose the machine
1. LEAST HWE/T (hardware error/time)
2. Highest Hashrate
http://i68.tinypic.com/97qtrr.jpg
http://i68.tinypic.com/30iz8gh.jpg
3. measured the wattage draw
4. Switched PSU to get absolute best Power-draw
http://i64.tinypic.com/2mfwx2r.jpg
5. All tests done on 230v/50hz


Here is the result
http://i63.tinypic.com/34rfyjb.jpg

After 30 minutes, I checked again to see if the power-draw creeped (because they usually creep in the s9s)
Sure enough, the power-draw creep up a little bit. Here is the result
http://i67.tinypic.com/2my22ky.jpg

Keep in mind: this is by no means the norm for S9s and I'm definitely not saying that every machine can do this......truth is they can't. S9s are notoriously finicky and the hashrate/power-draw varies wildly between machines/batches/chips etc etc. Out of 40 14TH, only 18 even consistently hash above 14th. and out of those 18, this one is the ONLY ONE that hash consistently over 14.5TH. So........this is pretty much as biased as you're ever gonna get with an S9.

So.....lets look at the numbers.

This machine hashes at 14.5TH and draws 1377W. That means the Efficiency is at ~95W/TH( 0.095W/GH)
In the very same environment, with the same PSU (who would have thought the same PSU had the best efficiency?) The Avalon 841s ALL consistently get around 13.5TH and draws 1230W with an efficiency of ~91W/TH (0.091W/GH)

SO.........IN CLOSING: I have absolutely no doubt that the S9s are probably NOT as efficient as the T1s. From what I have seen, the T1s can probably achieve higher efficiency than even the Avalon 841. Can Halong do consistency? I have absolutely no idea. But for now, We'd probably have to conclude that the T1 IS KING in efficiency for now. I already contacted Halong to see if they can sell me a batch

true for the 841 as I don't have them as of today.

I would be interested in seeing any s-9 under .10  I never got one to go below .101 and that was the older early batches that  allowed full frequency  adjustments.

Lets see.....

my 841s @-1 runs at ~13.5@1230w = ~91.1W/Th (or 0.0911w/gh if you old school)

Honestly I don't have any T1s so I don't know the power efficiency on my electric setup (I run 230v)

for the sake of THIS argument, I have MANY S9s (enough to make an unbiased test). So I will post some W/hash for you all for the 13.5 and the 14th ones.

No point in yelling without any actual data.