ANTMINER S2 Discussion and Support Thread

[1l1l11ll1l]

If anyone is wondering, I'm seeing 10A @ 208V for two S2 units on my PDU.

What PDU are you using?

AP7841

[1l1l11ll1l]

Why did I choose S2 over more S1? Well, this is the main reason:



4.5TH                 4.5TH                 6TH

[abracadabra]

1l1l11ll1l,

Does the top S2 in the stack have a dead display?

[robbcarpdm]

Why did I choose S2 over more S1? Well, this is the main reason:



4.5TH                 4.5TH                 6TH

Sweet setup bro!!!!!

Well if you underclock the left too you can get your density as the right for NO additional cost .....  just saying.

I would rather have numbers work because convenience aint going to pay you back in the end.  To each his own.....but I like your style if you can afford it

[dogie]

Well if you underclock the left too you can get your density as the right for NO additional cost .....  just saying.

I would rather have numbers work because convenience aint going to pay you back in the end.  To each his own.....but I like your style if you can afford it

No he wouldn't. He might get the same W/GH but that would make the already worse GH/U even worse.

[robbcarpdm]

Well if you underclock the left too you can get your density as the right for NO additional cost .....  just saying.

I would rather have numbers work because convenience aint going to pay you back in the end.  To each his own.....but I like your style if you can afford it

No he wouldn't. He might get the same W/GH but that would make the already worse GH/U even worse.

Well you cant use the same frames ...you would have to mod the blade rack ...

Your right though the heatsinks will not allow you to get identical footprint.

Geeze.....I was just making a point....lmao

[kano]

Has anyone tried the new firmware, yet?  Any idea what it's supposed to fix?

https://github.com/AntMiner/AntGen1/tree/master/firmware_S2

Just upgraded to the new firmware. Don't see any differences, still a very high discard rate...

Discard rate?!?

Post the numbers.

Discarded in cgminer is pretty much a meaningless number.

The 2 numbers that matter are the hash rate, and the pool reject rate.

[1l1l11ll1l]

1l1l11ll1l,

Does the top S2 in the stack have a dead display?

Not dead, just dim.

[robbcarpdm]

OKay I update the firmware.

Given it more time here with Slush the rej rate is dropping as the numbers are not increasing as fast as the start up.

Not sure why its so much better at Slush .....will try Eligius and report back...... still 15% but its declining as the accepts go up.

[1l1l11ll1l]

Well if you underclock the left too you can get your density as the right for NO additional cost .....  just saying.

I would rather have numbers work because convenience aint going to pay you back in the end.  To each his own.....but I like your style if you can afford it

No he wouldn't. He might get the same W/GH but that would make the already worse GH/U even worse.

Footprint is irrelevant, only power usage. I pay $50/mo per rack + Power. So if I underclock the S1's and managed to get the same W/GH then they'd only cost me < $100/mo more than the S2's

[robbcarpdm]

For me I dont have your constraints so I would rather have 40% more hashing power for the same dollar and deal with the power under clock later

[dalenorman2005]

Great news!  Got all 3 S2's today.  First one's setup and running, after reseating the main board and reconnecting the ribbon cable to the display.  One board was crooked but not too bad. 

Others have reported major grief after changing to a subnet other than 192.168.1.x, so before turning it on, i connected a 2nd router to the primary router just to create a 192.168.1.x subnet. 

So all i've done is configure my pool, and it's working....

two questions:

-what's the best way to shut it down?  there is no "shut down" option in the web gui.  there's a reboot, but no shutdown.  just pull the plug?  was it just exceptionally bad luck for the other guy who shut down and couldn't boot again?

-is DHCP to a subnet other than 192.168.1.x working?  i'd hate to loose contact with it over something so trivial as a changed IP.

[jelin1984]

Is noisy ??

[abracadabra]

1l1l11ll1l,

Does the top S2 in the stack have a dead display?

Not dead, just dim.

Kinda bummed about my dead displays 

[innoman]

Discard rate?!?

Post the numbers.

Discarded in cgminer is pretty much a meaningless number.

The 2 numbers that matter are the hash rate, and the pool reject rate.

https://dl.dropboxusercontent.com/u/19658540/images/antminers2rejects.png

Hashrate is decent, though it sometimes spends a lot more time around 950s than it should.  

[kano]

Discard rate?!?

Post the numbers.

Discarded in cgminer is pretty much a meaningless number.

The 2 numbers that matter are the hash rate, and the pool reject rate.

https://dl.dropboxusercontent.com/u/19658540/images/antminers2rejects.png

Hashrate is decent, though it sometimes spends a lot more time around 950s than it should.  

OK yep your DiffR (cgminer Rejects) are high indeed.
DiffA: 524000
DiffR: 196882

So basically, 27.3% of your submitted shares are rejected by the pool.
OUCH!
Me thinks the driver needs some skilled work ...

Edit: though doing the actual accepted hash rate calculation:

35m 51s = 2151s
524000 * 2^32 / 2151 = 1.046TH/s - which is of course OK

So yeah it's most likely the driver doing something wrong with it's work difficulty calculation.

[btc_jumpnrl]

Before I head to the gym, here's the quick video of the initial boot with sound so you can hear the fans. http://youtu.be/sQYtbtsyKdU

I hope you dont sleep with that thing on

well...I'll give it a night and see how it goes...


My discards are also crazy high, haven't updated yet though. Slush hashrate is showing about 700GH/s

[kano]

Before I head to the gym, here's the quick video of the initial boot with sound so you can hear the fans. http://youtu.be/sQYtbtsyKdU

I hope you dont sleep with that thing on

well...I'll give it a night and see how it goes...


My discards are also crazy high, haven't updated yet though. Slush hashrate is showing about 700GH/s
https://i.imgur.com/9eQIZmU.png?1

DiffA: 3089511
DiffR: 2048684

Yep crazy high again.
39.9% reject rate.

However, the DiffA tells a different story:

Elapsed is 3hr 44min 38s = 13478 seconds

So hash rate for that time is:
3089511 * 2^32 / 13478 = 984.5 GH/s

So it looks like the driver is throwing shares at the pool that are below diff.

Still needs some skilled work on it though

[robbcarpdm]

Discard rate?!?

Post the numbers.

Discarded in cgminer is pretty much a meaningless number.

The 2 numbers that matter are the hash rate, and the pool reject rate.

https://dl.dropboxusercontent.com/u/19658540/images/antminers2rejects.png

Hashrate is decent, though it sometimes spends a lot more time around 950s than it should.  

OK yep your DiffR (cgminer Rejects) are high indeed.
DiffA: 524000
DiffR: 196882

So basically, 27.3% of your submitted shares are rejected by the pool.
OUCH!
Me thinks the driver needs some skilled work ...

Okay after someone pointed out my algebra mistakes....the Pools are reading the correct hash as I forgot about a missing miner I have.

So perhaps the API is not reporting correctly.....my understanding is the only thing that matters is hashing as seen by the pool and if you have it there you have all you need regardless of other stats.

So on this one I can officially call this one a PEBKAC and its me  =)

[kano]

...
Okay after someone pointed out my algebra mistakes....the Pools are reading the correct hash as I forgot about a missing miner I have.

So perhaps the API is not reporting correctly.....my understanding is the only thing that matters is hashing as seen by the pool and if you have it there you have all you need regardless of other stats.

So on this one I can officially call this one a PEBKAC and its me  =)

I added an edit to my post.
https://bitcointalk.org/index.php?topic=515448.msg6076528#msg6076528

Edit: though doing the actual accepted hash rate calculation:

35m 51s = 2151s
524000 * 2^32 / 2151 = 1.046TH/s - which is of course OK

So yeah it's most likely the driver doing something wrong with it's work difficulty calculation.