Pages:
Author

Topic: Avalon A10 series announced - page 4. (Read 4167 times)

legendary
Activity: 2030
Merit: 1569
CLEAN non GPL infringing code made in Rust lang
April 20, 2019, 08:54:20 AM
#75
Still... 56J/TH vs. BItmain's claim of ~40J/Th (29% less power).

Can't competitively buy anything compliments of the Trump Tax, and refuse to buy anything from Bitmain ever again, so guess I'll just have to wait for a power competitive product.  *sigh*

There is also MicroBT, and we have yet to see how much tweaking can the newer Canaan units take by downclocking. Price has always been a weakness for Canaan, but at least they are trying... I though Canaan was readying 7nm parts but their latest "A10" is still 16nm, hope is if they are brewing something new at 12nm or less like MicroBT is.
copper member
Activity: 658
Merit: 101
Math doesn't care what you believe.
April 19, 2019, 03:07:18 PM
#74
Still... 56J/TH vs. BItmain's claim of ~40J/Th (29% less power).

Can't competitively buy anything compliments of the Trump Tax, and refuse to buy anything from Bitmain ever again, so guess I'll just have to wait for a power competitive product.  *sigh*
legendary
Activity: 4256
Merit: 8551
'The right to privacy matters'
April 18, 2019, 01:48:44 PM
#73
Looks like canaan.io/shop updated.

So a 31th more efficient model
and a 33.5th less efficient model
full member
Activity: 1022
Merit: 221
We are not retail.
April 18, 2019, 01:16:49 PM
#72
Steve if you want me to "test" heat limits in my garage just let me know. Wink



Looks like canaan.io/shop updated.
legendary
Activity: 4592
Merit: 1851
Linux since 1997 RedHat 4
April 16, 2019, 09:00:56 PM
#71
Ok,  we only had an hour or two in the lab and the customer wanted to point the machines at their pool after we did our initial tests. We were in such a rush across beijing in traffic to get to the lab that I totally forgot about alerting you.

We were also surprised by getting 38TH as we predicted 37 based on the voltage setting of the PSU. There is even more headroom in frequency, ( say at least 33%) but we may hit cooling limits and we  hit the limits of the power strip when we pushed it beyond 2500.

I need to do some more structured tests and longer times at temperature, Needs some prep time and we will give you a heads up  when we head to the lab again.

I can probably find them (in the logs) via the diff they were mining at and hopefully a full hour ... will reply once I've done that.

Edit: didn't find any single a10's with 30-40k work diff - so I guess they were chained miners?
Next time just tell me the time (+timezone) you ran them and how you chained them (1, 2, 3 etc) and I should be able to find the stats
(the share logs are eternal)
If you give one a different worker name than the default (e.g. canaan.38) then I'll be able to find it even easier Smiley
member
Activity: 74
Merit: 76
April 16, 2019, 08:36:01 PM
#70
Ok,  we only had an hour or two in the lab and the customer wanted to point the machines at their pool after we did our initial tests. We were in such a rush across beijing in traffic to get to the lab that I totally forgot about alerting you.

We were also surprised by getting 38TH as we predicted 37 based on the voltage setting of the PSU. There is even more headroom in frequency, ( say at least 33%) but we may hit cooling limits and we  hit the limits of the power strip when we pushed it beyond 2500.

I need to do some more structured tests and longer times at temperature, Needs some prep time and we will give you a heads up  when we head to the lab again.
legendary
Activity: 4592
Merit: 1851
Linux since 1997 RedHat 4
April 16, 2019, 06:42:03 AM
#69
After seeing what Phillip was doing, I kinda pressed on engineering to make on the fly frequency changes a lot easier.  There is STILL a long way to go on what I would call proper specifications for systems: in particular I would like to see independent lab testing of machines and like to finally publish some hard data on how performance changes as a function of temperature. Who knows with a built in neural net asic maybe somebody will discover that the machine can learn to control itself. The clock cores already adaptively adjust  their clock rate as a function of DH.

Yeah I worked out with the A741 a couple of years ago how to do the frequency via the "More Options" field - since you can put any cgminer settings in there (rather than having to login and change the init script)
I started doing the A741 heater back in 2017 since it was less power than my home heater Smiley
https://bitcointalksearch.org/topic/m.19893916

As long as the options are in cgminer, it's not a problem to be able to change the settings, but of course a specific web page option to set it is more help for the less technical.
(and I failed dismally to get the A851 or A921 to do it last year when I got them from Canaan, since the stepping changed - so I didn't know til you released the code in Nov what the values where Tongue and by then it was summer again - winter coming up again soon, I'll get them out of the cupboard and try to get them to work as heaters also)

While we have a bunch of machines in test, Lily and I want to get test results from the first  Trial run. You should see those system hit the pool any day now.

The time I posted about it was to say "These really exist and they match the specs shown on the web sites" Smiley
https://bitcointalksearch.org/topic/m.50509226

If you have time, you can get one of the devs to message me per the usual channel, or send me a PM, to let me know when you do a faster test and I'll check out the individual miners and post the numbers.
(takes me a few minutes effort, I find a miner that was on before and after the hour, then do the calc on the shares it submitted for that hour, since their not individually reported on the pool, they all use the same worker name, which is ok of course)
legendary
Activity: 1554
Merit: 2037
April 16, 2019, 03:07:06 AM
#68
Nice to see the miner up and running.

I kept waiting for them to update the miner status page, or for cut to an hour or later. Either way definitely promising, a big plus with summer coming is seeing them actually push the miner to 40C.

Nope its 16nm  because the wafer cost per TH on yeilded die is lower on 16nm than 7nm.
No controller required in the RJ45  version, cgminer runs on the AI chip,  dual core  risc V.
I think the option with the non integrated PSU will probably rule the day, especially when folks figure out that if you have the right PSU, you can run it   above 31TH.. (31 TH is running about 1/2 the max clock) you'll take an effciiency hit that more than pays for itself at lower power prices.  
The purchasing decision and operating decisions become a bit more complex when you look at running the machine in max effciency mode ( say 24TH ) versus running it in max hash rate mode ( 38+ today) And how do I just put one number on the spec sheet? and then you have to factor in how efficiency changes with ambient temp.
Its almost like an AI problem. just kidding. But I do foresee customers who will now have fine grain control of the frequency that is tied to the actual operating conditions of their farm.

Thanks for clearing that up. Reading this it's hard to tell what the anticipation is for the miner or the fun/interest of being able to tweak and adjust for your needs. Note to self don't dump the PSU's yet.
member
Activity: 74
Merit: 76
April 16, 2019, 03:04:01 AM
#67
It was just a link to a youtube vid of the test at the enviromental lab. I emailed it to Matti.
its on my youtube channel.

Finalizing pricing now. It's rather difficult because most of the pricing is done on a  RMB per TH  basis in china. Irrespective of efficiency.

In short, the domestic market doesnt care about efficiency that much.

Lets  compare:

38  TH   @  2500  watts    .065J/GH    @ 3 cent power   =   $5.79  per day     2 years = 4226
31  TH   @  1800  watts    .058J/GH    @ 3 cent power   =    $4.89  per day.   2  years = 3569

Same machine:  same cost.  and everyone sells on RMB per TH.. well which   performance do you pick to price it?

Now compare M10

33  TH  @  1850 Watts    .056 J/GH    @ 3 cent power   = $5.26  Per day          2 years  = 3839

Now compare

33  TH  @  1850 Watts    .056 J/GH    @ 6 cent power   = $3.93  Per day          2 years  = 2868
38  TH   @  2500  watts    .065J/GH    @ 6 cent power   =   $3.99  per day       2 years =  2912

Of course guys here know how to do this math. The math on having a reliable system is harder. The math on how many machines actually deliver the promised spec ( or higher) is also harder to estimate. The math on reducing your number of network connections  because of daisy changing is probably easier.

So, its trivally true that the lower your power price the  more likely you are to trade efficiecny for hash rate. The vast major of huge customers are below  6 cents, all in. So they have to look at all the various modes the machine can run in. After seeing what Phillip was doing, I kinda pressed on engineering to make on the fly frequency changes a lot easier.  There is STILL a long way to go on what I would call proper specifications for systems: in particular I would like to see independent lab testing of machines and like to finally publish some hard data on how performance changes as a function of temperature. Who knows with a built in neural net asic maybe somebody will discover that the machine can learn to control itself. The clock cores already adaptively adjust  their clock rate as a function of DH.

Any way,   a few other updates.  machines will probably be built in a mix of with PSU  and W/O PSU. You can run the machine  with an integrated PSU ( it has voltage control over I2C)  or just attach your own PSU. And if you get it with the integrated PSU  you can always switch it.

Price on the website is  just a stand in price.  It will be below that once we start taking orders. While we have a bunch of machines in test, Lily and I want to get test results from the first  Trial run. You should see those system hit the pool any day now.
legendary
Activity: 2464
Merit: 1710
Electrical engineer. Mining since 2014.
April 16, 2019, 02:39:10 AM
#66
"The stuff" Steve mentioned sending earlier is video material from the A10 lab testing.

These videos are public at his YouTube channel,
so I'm comfortable sharing these here at the forum.

Really cool to see this process, and project going forward.

Enviromental test lab
https://youtu.be/17Yhi87A37k

A10 test lab
https://youtu.be/j74U08QD36Y
legendary
Activity: 4256
Merit: 8551
'The right to privacy matters'
April 15, 2019, 08:59:28 PM
#65
The mod here goes OCD on posts and edits and combines them all the time.
He gets paid by the forum for each change he does Tongue      << this is funny Grin
You have no choice but to put up with it.

The info you posted is probably all combined in your last post.

At mods  I just had to say I smiled a bit at the bolded line.

feel free to delete me Grin
legendary
Activity: 4592
Merit: 1851
Linux since 1997 RedHat 4
April 15, 2019, 08:57:22 PM
#64
I will email you some other stuff  that the mod deleted.

The mod here goes OCD on posts and edits and combines them all the time.
He gets paid by the forum for each change he does Tongue
You have no choice but to put up with it.

The info you posted is probably all combined in your last post.
member
Activity: 74
Merit: 76
April 15, 2019, 08:01:07 PM
#63
Thanks for sharing these details with us Steve!

I will email you some other stuff  that the mod deleted.
legendary
Activity: 2464
Merit: 1710
Electrical engineer. Mining since 2014.
April 15, 2019, 03:13:41 PM
#62
Thanks for sharing these details with us Steve!
legendary
Activity: 4256
Merit: 8551
'The right to privacy matters'
April 15, 2019, 12:43:54 PM
#61
... Its almost like an AI problem. just kidding. But I do foresee customers who will now have fine grain control of the frequency that is tied to the actual operating conditions of their farm.

Making this work well = key and very important to a lot of us.

yeah you will always have a guy with 10 + megawatts of 2 cent power that just runs the gear cranked.

But lots of us have multiple locations that need more efficient gear that is quiet. So having a unit that does both will be nice. I can't wait to get one to test out.
full member
Activity: 1022
Merit: 221
We are not retail.
April 15, 2019, 12:05:13 PM
#60
This is a nice luxury to have you here Steve. Thanks for the info.
member
Activity: 74
Merit: 76
April 15, 2019, 08:07:58 AM
#59
Have we worked out that these are standalone devices? i.e. if you just buy one A10 then you don't need anything else? Just http to the management interface on the miner itself?

Does that mean that if you have the choice of running them individually, each unit plugged into a switch ala most other miners out there, or you can just have one plugged into the network and all the others daisy chained off that one miner via rj45 cables?

A miner consists of 2 hash boards and a  MMU.  the PMU is gone. The MMU ( miner management unit ) has two  versions, and this will quickly transition to 1 version.

The first version uses the old style rasb pi  to AUC.  My best estimate is that only a few machines will be built in that style.

In parallel we are finalizing test on a new style MMU  that has  RJ45  instead of the old 5 pin connector. with this version, the PI will be gone, the AUC will be gone and machines will daisy chain over ethernet. no real limit on how many you can chain..

This MMU is up and working, final gerber released and so  mass pro will switch over  ASAP.  If you get a AUC  version, we will probably have a  replacement MMU ( with RJ45)  to upgrade. I ordered the parts to do this.



I know they might just want to sound "advanced" and all, but that description is kinda scary. AI System On Chip? What else does that thing do besides handling the Asics?

The  AI chip is here

https://kendryte.com/

https://kendryte.com/downloads/

it's not scary.



... and I checked one mining on the pool, it was getting 31.89TH/s for an hour of shares.
(advertised rate is 31+TH/s)

I was running at 38TH earlier today  at 40C in the enviromental test lab. fun times.



Hmmmm, that would be interesting to see them go back. Maybe they noticed MicroBT's results. I think it's more likely whoever made the product information page slipped up a little. In the past it's generally a copy past of older hardware then they edit it with the new specs. We'll just have to wait on Steven to come back and enlighten us a bit.

Biffa-

I think there will still be need for the controller, not positive. For the networking I believe that you will be using the rj45 cables in place of the AUC cables. So only 1 device whether that be the PI/Controller is connected to the router/switch, then the rest are connected in the same fashion as previous models just a different cable.

I could be wrong but that's my understanding of the info given. I'm interested to hear more from Steven on this, it's all pretty cool. Especially the option to have a built in PSU or not.

Nope its 16nm  because the wafer cost per TH on yeilded die is lower on 16nm than 7nm.

No controller required in the RJ45  version, cgminer runs on the AI chip,  dual core  risc V.

I think the option with the non integrated PSU will probably rule the day, especially when folks figure out that if you have the right PSU, you can run it   above 31TH.. (31 TH is running about 1/2 the max clock) you'll take an effciiency hit that more than pays for itself at lower power prices.  

The purchasing decision and operating decisions become a bit more complex when you look at running the machine in max effciency mode ( say 24TH ) versus running it in max hash rate mode ( 38+ today) And how do I just put one number on the spec sheet? and then you have to factor in how efficiency changes with ambient temp.

Its almost like an AI problem. just kidding. But I do foresee customers who will now have fine grain control of the frequency that is tied to the actual operating conditions of their farm.
legendary
Activity: 2030
Merit: 1569
CLEAN non GPL infringing code made in Rust lang
April 11, 2019, 01:07:41 PM
#58
Just noticed, this is using 16nm chips!

Quote from the product page:

240 x A3205 Third Gen 16nm ASIC

Very well, i hope, really that they also plan to come with a 12nm part and 7nm part. MicroBT is good competition but more is netter.

I'm not very sure $1200 for 31T will look very enticing vs, say. 33T for $785 (especially with that tax in the US), but lets see who makes the first independent review, maybe its quiet or can be made more efficient.



Model NameAvalonMiner 10
Hash Rate31TH/s,-5%~+5%
Power Consumption1736 Watts,-5%~+15%(On Wall)
(Minimum, Variant power consumption at different temperature and speed)
Power Efficiency56J/TH,-5%~+15%
Power Supply Input176~264V AC 50~60Hz 7.5~10A
Chips240 x A3205 Third Gen 16nm ASIC
Smart ControllerCanaan Kendryte K210 Artificial Intelligence SOC
ConnectionRJ45, Up to 254 Miners(theory) Daisy-Chain Connection
Operating Temperature-10℃~35℃
Air-intake Temperature-10℃~35℃
Cooling2 x 12038 FAN
Noise70db (Typical)
Net Dimensions190mm x 190mm x 292mm
Net Weight7.9kg

I know they might just want to sound "advanced" and all, but that description is kinda scary. AI System On Chip? What else does that thing do besides handling the Asics?

PS: Repost per request of moderator.
full member
Activity: 402
Merit: 116
April 09, 2019, 05:19:49 AM
#57
I might just get 12 of these guys instead of the S17, gonna wait and see for another month
legendary
Activity: 4256
Merit: 8551
'The right to privacy matters'
April 08, 2019, 09:40:58 PM
#56
A10 at stock spec is best over s15, m10, m10s. Pricing at 1.2k on canaan's site is pretty on par with the way this market is now. I believe the July release is the only barrier to making it a no brainer.

S15 is  60 high 50 low ------- cant buy it
m10 is 69 high 59 low ------- cant buy it
a1021 is supposed to be   56 high 50 low

I will get at least

 1 A1021
 1 s17  maybe the cheapest model
 1 m20 72 th on order

evaluate all three  and decide which one I will get a lot of.
Pages:
Jump to: