ice_chill (OP)
|
|
October 13, 2012, 06:04:37 PM Last edit: October 13, 2012, 07:23:37 PM by ice_chill |
|
Hi, it seems the new Little Single (30Ghash) is better then the old SC Single 60Ghash, here's why:
1. Better value for money: 1 x 60Ghs Single = $1299, 2 x 30Ghs Singles ($649*2=$1298) 2. Better redundancy: If 1 x 60Ghs Single fails then you are down by 60Ghs, if one of the two 30Ghs Singles fails, then you are down by only 30Ghs 3. Even more redundancy: with 2 x Little SC Singles you get 2 power bricks.
I am personally disappointed that BFL now allows to upgrade the Japaleno orders to Little Singles, but does not allow the Standard Single to be changed into Little Singles orders.
----------------------------------- Edited the numbers value mistake -----------------------------------
|
|
|
|
ElectricMucus
Legendary
Offline
Activity: 1666
Merit: 1057
Marketing manager - GO MP
|
|
October 13, 2012, 06:07:51 PM |
|
That depends, if you consider the possibility of their devices being vaporware, not living up to specs or even a scam it would be best to wait.
|
|
|
|
aTg
Legendary
Offline
Activity: 1358
Merit: 1000
|
|
October 13, 2012, 06:13:26 PM Last edit: October 13, 2012, 06:26:20 PM by aTg |
|
No longer jalapeno? Soon orders you can upgrade to the new generation of ASICs 10THs if you wait a year or any such offering, this seems like the shopping channel.
|
|
|
|
Fjordbit
|
|
October 13, 2012, 06:29:34 PM |
|
1. Better value for money: 1 x 60Ghs Single = $1299, 2 x 30Ghs Singles ($699*2=$1298)
I don't think this is valid because you have to also include the cost of the hub. 2 Little singles will take up 2 ports vs just one. If you have two free ports on your computer, then yeah, but I personally don't. Another aspect that you haven't factored in here is power usage, including loss in the power brick. It might be half, but likely two littles will use more power. 2. Better redundancy: If 1 x 60Ghs Single fails then you are down by 60Ghs, if one of the two 30Ghs Singles fails, then you are down by only 30Ghs 3. Even more redundancy: with 2 x Little SC Singles you get 2 power bricks.
This is a good point. There is also an annoyance of how many devices you have. If you are just getting one sc single vs two little singles, then this makes sense, but if you are getting 10 scs single vs 20 little singles, that can get to be a bit much.
|
|
|
|
pieppiep
|
|
October 13, 2012, 07:19:51 PM |
|
My calculator says $699*2=$1398
/edit Oh, $699 must be $649, so the $1298 is correct.
|
|
|
|
ice_chill (OP)
|
|
October 13, 2012, 07:22:37 PM |
|
My calculator says $699*2=$1398
/edit Oh, $699 must be $649, so the $1298 is correct.
Sorry my mistake, the SC LITTLE costs $649, so the thread is still on topic, sorry for numbers mistake.
|
|
|
|
pieppiep
|
|
October 13, 2012, 08:00:47 PM |
|
2. Better redundancy: If 1 x 60Ghs Single fails then you are down by 60Ghs, if one of the two 30Ghs Singles fails, then you are down by only 30Ghs
If a device has 10% downtime, the 1 x 60GH/s gives you 54GH/s effective, the 2 x 30GH/s gives you 54GH/s effective, so no difference in there. If a device has a chance of a failure capable of bringing the host system down it also brings other mining devices down on that way. So if the chance of such type of failure is the same on a 30 vs 60 GH/s device, then a single 60GH/s is better than 2 times 30GH/s. If the devices are build with chips capable of doing 1.5GH/s they need 40 chips for the 60GH/s and 20 for the 30GH/s device. If a single chip failure can bring the entire device, the 2 times 30 GH/s is better. If the devices are build with smaller boards containing 5 chips of 1.5GH/s and a single chip failure would bring such a board down, if won't matter if you have 60 or 2 times 30 devices. So, my conclusion is I don't know. I need to know more about the setup and possible risks to make a good judgement.
|
|
|
|
ice_chill (OP)
|
|
October 13, 2012, 08:05:15 PM |
|
By redundancy I meant that if a device fails independently and has to be sent back to BFL for repair.
|
|
|
|
creativex
|
|
October 13, 2012, 08:20:51 PM |
|
...but just like a RAID0 array 2x30 has a greater chance of failure than 1x60.
|
|
|
|
crazyates
Legendary
Offline
Activity: 952
Merit: 1000
|
|
October 13, 2012, 08:25:04 PM |
|
Hi, it seems the new Little Single (30Ghash) is better then the old SC Single 60Ghash, here's why:
1. Better value for money: 1 x 60Ghs Single = $1299, 2 x 30Ghs Singles ($649*2=$1298) 2. Better redundancy: If 1 x 60Ghs Single fails then you are down by 60Ghs, if one of the two 30Ghs Singles fails, then you are down by only 30Ghs 3. Even more redundancy: with 2 x Little SC Singles you get 2 power bricks.
I am personally disappointed that BFL now allows to upgrade the Japaleno orders to Little Singles, but does not allow the Standard Single to be changed into Little Singles orders.
----------------------------------- Edited the numbers value mistake -----------------------------------
You are so full of fail. 1) 2 Little Singles vs 1 SC Single is the difference of $1! One freaking dollar over a 1.3K USD investment? Who cares? 2) More parts =/= more redundancy. It just means more points of failure. 3) Same thing. If BFL has a 2% failure rates for their PSUs (I just made that number up), then you're twice as likely to get one that fails. And as someone else said, you're also now looking at added costs of USB hubs, and space constraints.
|
|
|
|
ice_chill (OP)
|
|
October 13, 2012, 08:55:44 PM |
|
...but just like a RAID0 array 2x30 has a greater chance of failure than 1x60.
Having 1 Single fail does not put the other one out of action, so it is more like RAID 1 and not RAID 0.
|
|
|
|
scrybe
|
|
October 14, 2012, 04:54:11 AM |
|
...but just like a RAID0 array 2x30 has a greater chance of failure than 1x60.
Having 1 Single fail does not put the other one out of action, so it is more like RAID 1 and not RAID 0. You should not mix in RAID terminology here, it is there for different reasons. RAID 1 is mirroring of data for protection, so the equivalent would be more like having 2 little singles doing exactly the same work in lockstep to prevent errors. This is a lot more like a farm of servers (like a bunch of web servers) There is a point where having more units of a smaller size can indeed reduce your risk that you will stop earning altogether, but it increases the risk that you will have a failure at the same time. Given the declining returns on mining, this could indeed be worth a few BTC if you get to keep mining instead of taking a week off. But as you add more and more units you have a larger and larger power overhead (it's more than double the power draw due to redundant VRM's) as well as the space and maintenance issues that come with more devices. The point is that there is a trade-off. if you are only buying one SC Single, then getting 2 little ones might make sense, but don't double your unit count if you are getting several SC Singles. This is not a new decision matrix for anyone who has had to choose between 2 (or 20) 8-socket servers and 8 (or 80) 2-socket ones.
|
"...as simple as possible, but no simpler" -AE BTC/TRC/FRC: 1ScrybeSNcjqgpPeYNgvdxANArqoC6i5u Ripple:rf9gutfmGB8CH39W2PCeRbLWMKRauYyVfx LTC:LadmiD6tXq7gFZvMibhFUZegUHKXgbu1Gb
|
|
|
scrybe
|
|
October 14, 2012, 04:56:02 AM |
|
Hi, it seems the new Little Single (30Ghash) is better then the old SC Single 60Ghash, here's why:
1. Better value for money: 1 x 60Ghs Single = $1299, 2 x 30Ghs Singles ($649*2=$1298) 2. Better redundancy: If 1 x 60Ghs Single fails then you are down by 60Ghs, if one of the two 30Ghs Singles fails, then you are down by only 30Ghs 3. Even more redundancy: with 2 x Little SC Singles you get 2 power bricks.
I am personally disappointed that BFL now allows to upgrade the Japaleno orders to Little Singles, but does not allow the Standard Single to be changed into Little Singles orders.
----------------------------------- Edited the numbers value mistake -----------------------------------
You are so full of fail. 1) 2 Little Singles vs 1 SC Single is the difference of $1! One freaking dollar over a 1.3K USD investment? Who cares? 2) More parts =/= more redundancy. It just means more points of failure. 3) Same thing. If BFL has a 2% failure rates for their PSUs (I just made that number up), then you're twice as likely to get one that fails. And as someone else said, you're also now looking at added costs of USB hubs, and space constraints. I posit that if the difficulty is increasing and you are out 30Gh/s for 1 week now and 1 week 6 months from now it will actually work out in your favor to have the partial outage rather than a full one. I agree that it is not worth the hassle though.
|
"...as simple as possible, but no simpler" -AE BTC/TRC/FRC: 1ScrybeSNcjqgpPeYNgvdxANArqoC6i5u Ripple:rf9gutfmGB8CH39W2PCeRbLWMKRauYyVfx LTC:LadmiD6tXq7gFZvMibhFUZegUHKXgbu1Gb
|
|
|
VeeMiner
|
|
October 14, 2012, 04:23:50 PM |
|
Hi, it seems the new Little Single (30Ghash) is better then the old SC Single 60Ghash, here's why:
1. Better value for money: 1 x 60Ghs Single = $1299, 2 x 30Ghs Singles ($649*2=$1298) 2. Better redundancy: If 1 x 60Ghs Single fails then you are down by 60Ghs, if one of the two 30Ghs Singles fails, then you are down by only 30Ghs 3. Even more redundancy: with 2 x Little SC Singles you get 2 power bricks.
I am personally disappointed that BFL now allows to upgrade the Japaleno orders to Little Singles, but does not allow the Standard Single to be changed into Little Singles orders.
----------------------------------- Edited the numbers value mistake -----------------------------------
You are so full of fail. 1) 2 Little Singles vs 1 SC Single is the difference of $1! One freaking dollar over a 1.3K USD investment? Who cares? 2) More parts =/= more redundancy. It just means more points of failure. 3) Same thing. If BFL has a 2% failure rates for their PSUs (I just made that number up), then you're twice as likely to get one that fails. And as someone else said, you're also now looking at added costs of USB hubs, and space constraints. +1
|
|
|
|
HDSolar
|
|
October 14, 2012, 04:44:58 PM |
|
Hi, it seems the new Little Single (30Ghash) is better then the old SC Single 60Ghash, here's why:
1. Better value for money: 1 x 60Ghs Single = $1299, 2 x 30Ghs Singles ($649*2=$1298) 2. Better redundancy: If 1 x 60Ghs Single fails then you are down by 60Ghs, if one of the two 30Ghs Singles fails, then you are down by only 30Ghs 3. Even more redundancy: with 2 x Little SC Singles you get 2 power bricks.
I am personally disappointed that BFL now allows to upgrade the Japaleno orders to Little Singles, but does not allow the Standard Single to be changed into Little Singles orders.
----------------------------------- Edited the numbers value mistake -----------------------------------
You are so full of fail. 1) 2 Little Singles vs 1 SC Single is the difference of $1! One freaking dollar over a 1.3K USD investment? Who cares? 2) More parts =/= more redundancy. It just means more points of failure. 3) Same thing. If BFL has a 2% failure rates for their PSUs (I just made that number up), then you're twice as likely to get one that fails. And as someone else said, you're also now looking at added costs of USB hubs, and space constraints. By the way didn't BFL just come out and offer lifetime warranties on their gear? So the fail part of your analysis is not a very big deal? Also with the bricks, saving power is a big deal and cost item so would you not want to do something like cablez anyways to drop the brick from the device? Also having one power source and one USB item is less cost power and easier to manage so to save a buck I am not really seeing it. I think this kind of analysis needs to happen but this is what I am seeing.
|
|
|
|
bce
|
|
October 14, 2012, 08:32:23 PM Last edit: October 14, 2012, 09:24:41 PM by bce |
|
It does make sense that the effect of a failure of one Little Single would be of less consequence than a failure of a Single, although worrying about two Little Singles could be more trouble than it's worth. Asics should be highly reliable, so variance reduction may just not be worth it.
A bigger question - If these ASIC chips are highly overclockable, might air cooling using the reference heat pipe cooler favor the Little Single (as there are would only be 4 chips to cool)? I remember seeing in a post someplace that the Mini Single will use the same heat sink / fan setup as the Single. However, given the released design previews, water cooling kits may be made available as an option for future overclocking. In the case of a water-cooling setup, the long term advantage goes back to the SC Single.
I'd rather have an SC Single, due to simplicity of setup and long term efficiency- as scrybe said, redundant VRMs = waste. For me, Little Single is also a waste of potential for a beautifully designed board (that has potential to be used in a water-cooled setup).
|
|
|
|
MrTeal
Legendary
Offline
Activity: 1274
Merit: 1004
|
|
October 14, 2012, 09:41:30 PM |
|
With two little singles the cost of a unit going down is half that of a big single. However, the chance of a unit going down is twice as high as with a big single. As other have said, the net result is the same.
The only real benefit that I can see is that little single is going to use the same chassis and heatsink as the big single. Because of that it should run cooler and if BFL does allow the user to adjust the clock it will have more overclocking headroom.
|
|
|
|
bce
|
|
October 15, 2012, 09:36:49 AM |
|
Nevermind what I said before about this topic- It's likely the case that ice chill just wants some free stuff
|
|
|
|
thatbluedude
|
|
October 15, 2012, 06:14:32 PM |
|
Has BFL ever officially stated the small singles power consumption? Because I don't remember seeing official numbers like with the normal single
|
|
|
|
ice_chill (OP)
|
|
October 15, 2012, 06:24:27 PM |
|
Has BFL ever officially stated the small singles power consumption? Because I don't remember seeing official numbers like with the normal single
All their products have an advertised power consumption of 1watt per Ghash.
|
|
|
|
|