Glasswalker
|
|
August 19, 2012, 04:21:13 AM |
|
I should chime in and clarify too. The current released hashvoodoo bitstream (and all future ones) do in fact hash a full nonce range on a single chip.
|
|
|
|
LazyOtto
|
|
August 19, 2012, 04:25:03 AM |
|
I should chime in and clarify too. The current released hashvoodoo bitstream (and all future ones) do in fact hash a full nonce range on a single chip.
That seems very 'clean/simple' to me. One USB port per hash producer. Especially if the auto-clocking / optimizing will work per FPGA. Assurance that whatever the device can do, it will do the best it can.
|
|
|
|
LazyOtto
|
|
August 19, 2012, 04:30:49 AM |
|
Of course, the major consideration / wild-card here is if BFL delivers on their promises and starts shipping ASICs, in volume, in November. If so, none of the current FPGA based products will ever break even on their up front capital expense if buying now. An interesting 'bet'. Personally, I'm hoping BFL fails. (Such a sad attitude.) Just because of what I consider their unethical practices to date have been. -- Oh, full disclosure, I own one BFL single. Which is operating at 800 MH/s. And have a pre-order in for two CM's with an estimated delivery date in September. It is really on the razor's edge whether it will be to my advantage/profit/break-even to pay for those two or not.
|
|
|
|
Lethos
|
|
August 19, 2012, 07:20:27 AM |
|
So if Glasswalker's bitstream does a full nonce range on a single FPGA, yes indeed it would be --icarus-options :1:1 or even --icarus-options :1 (both mean the same thing)
I only went off what you said earlier in the thread. Not like I really went into finding out what it did. It has worked beautifully since I done the change. I will update and try :1:1 next time. No worries - just thought I should explain it Trying out a :1:1 setting now Btw Glasswalker, they all decided to disable themselves like a domino effect around 6am this morning, so they last around 36 hours. Strangely no Hardware errors, but I did have to cold restart them to get them mining again. (thankfully now they are perma-flash). Updated the icarus options, So lets see if that changes anything.
|
|
|
|
yohan (OP)
|
|
August 19, 2012, 11:15:46 AM |
|
State of the Nation Time
I have not done one of these for a while so probably time to do one now.
Bitstreams as you all probably know are being done by Glasswalker amd Makomk and we are doing our best to keep up with them to test and support them as much as we can. Our own bitstream development isn't stopped altogether but as these guys are doing so well it is better we put as much of our available resource into supporting them as best we can.
Shipments are going fairly smoothly and about half of September's promised deliveries have been shipped already. We will be asking for the balance to confirm during this coming week as we have to decide both on how much extra silicon to buy and if we can accept any more orders in the near term. Obviously a lot of you if the later shipment promise customers are debating whether you want to go ahead because of ***. If you are uncertain it is better for us that you tell us sooner that later. It does hurt us if we get left with a lot of silicon or boards that are not wanted for whatever reason and if we lose money on this project we won't be doing anything for Bitcoin in the future.
If *** actual produce what they promise it will change the bitcoin market somewhat but not necessarily so much that a CM1 board isn't viable still. There is much more to this that just hashing rates. It may change the return period but my guess is that it will still make a profit. Personnally I still don't think *** have done their own customers much good which is very silly and no good to anyone. I also think the real people to suffer will be GPU miners. GPUs miners banking on good resale values of GPUs might be very dissappointed as well in the next few months. My crystal ball shows resale values dropping through the floor for GPUs if *** are successful.
Some of you are asking how much we have lost to *** and it's difficult to be totally accurate as we are still getting new orders. We always expected a certain amount of drop off anyway with a pre-order system. People do change their mind over a couple of months. I'd probably say 25% as a rough number which sounds a lot but even after that we are still running at x5 the numbers of orders we expected when we started this project. The project has given us a few cashflow issues but that has been managed. We have done projects bigger than this before and that experience has been invaluable in coping with the demand and what goes with that.
Going back to the technical we are now understand the CM1 pretty well and the fixes that the successful bitstreams use are working well with that. Remembering that we are learning about Bitcoin as we go we have learned a lot in the last 16 weeks since we announced the CM1 concept and started the design and manufacture of this product. We believe the "noise" issues that have to worked around in the bitstreams are 80% down to the Spartan-6 packaging which something that we can't do anything about. This 80% is peculiar to the way that hashing causes very large current surges. The 20% part that we can change in PCB design will be incorporated into what we do next as a Bitcoin mining product. The whole "noise" subject is a combination of factors and is not simply one thing or another. It is a combination of various things. On the difference between the early pre 0051 boards and the later post 0050 boards and we think the difference is actually very small and that the better pre-0051 are as good as the lower to mid performance boards post 0050 boards. It's a tolerance band overlap. We also seem now to have bitstreams that can make the worst pre-0051 boards work. We have a "very bad" board working here now as a test platform to help prove out things.
We are getting very stable performance now out of production boards and all production boards get test mined for 1 or 2 days before they ship. Nobody should see an underperforming board now unless there is something peculiar to there setup. We can't cover all possibilities of software, network, or mining pool in our testing but unless a board meets our internal performance marker it doesn't ship to a customer.
During September we will try improve documentation on the various aspects of CM1.
For the future we are looking at what we might do as products for Bitcoin mining and we slowly putting together concepts and seeing what we might do with those. We won't release any information on these until at least October for obvious reasons. We will evaluate if they can compete with *** solutions and if we can compete they will probably be announced. If not some of those ideas will appear in our other product areas and won't be wasted time.
We have been trialling Bitcoin payments for a few weeks and we are ready to start accepting payment this way in a limited fashion. Our biggest concern now in this is that transactions can't be traced as such and we don't want to get into "I sent money" and we say we didn't get it syndrome so we will limit direct taking of Bitcoins to less than GBP £1k equivalent. For the existing pre-order price we will also add a risk/conversion margin of 15% based on Intersango rates. Once we have more experience of this method we may revise the limits or margins but that is what we will offer for now.
There is a second way that Bitcoins can be used to pay us and this way came out of our investigations into using Bitcoins. This method uses Intersango as Bitcoin conversion, other exchanges may be able to do this as well, but Intersango is UK based as that is an advantage in this method. It works on the basis that the customer themselves hold a GBP/Bitcoin account/s at Intersango and convert Bitcoins into GBP. Once in GBP a transfer to our GBP bank account can be done on smaller amounts in less than one banking day and probably 3 days on larger amounts. We are awaiting confirmation of the "faster payments" limit for the 1 day process but we have already done up to several hundred pounds so far as a trial and that works well. The money transfer from Intersango to us should be fully traceable as it is simply a conventional UK bank transfer. Charges for the bank transfer we believe as follows £0-100 -0 £0.20, £100-200 - £1.00, £200+ - free. You can all argue about whether the Intersango converstion rates are competitive and the same for ther commission but we have not found them them to be too bad in what we have done ourselves. We may also ask you add a marker in this sort of transfer and that will simply be a variation in the number of pennies you pay us so we will vary the amount by £0.01 to £0.99. This means we unlikely to get identical payments that we can't 100% determine from whom they came from. All we see is it's from "Intersango" when it arrives in our bank account. We will be interested to how all of you get on with this method. We think this hybrid payment system is very straight forward and provides the tracability that avoid disagreements about whether payments have actually been made. As far as we can tell, and it's a lot more we can find on some exchange operations, Intersango has been good in it's dealings with customers. If you have any experience on that, good or bad, we would be interested in the feedback. This is still an area that we are learning about. The advantage with Intersango being in the UK we do understand what reglatory and business norms apply to businesses operating in our particular backyard.
|
|
|
|
Isokivi
|
|
August 19, 2012, 11:38:39 AM |
|
I did some math when the exchange rate was 12$ and for me (at 0,17$/KWH) about 40million difficulty is the point where I have fairly expensive paperweights, to put it the other way around if the network hashrate goes up 18x.
|
Bitcoin trinkets now on my online store: btc trinkets.com <- Bitcoin Tiepins, cufflinks, lapel pins, keychains, card holders and challenge coins.
|
|
|
yohan (OP)
|
|
August 19, 2012, 12:58:52 PM Last edit: August 19, 2012, 06:02:01 PM by yohan |
|
I did some math when the exchange rate was 12$ and for me (at 0,17$/KWH) about 40million difficulty is the point where I have fairly expensive paperweights, to put it the other way around if the network hashrate goes up 18x.
That's a fairly much seems to be on the worst end of most peoples estimates but everyone has their own numbers. I can only say we don't think it will be anything like that but we could be wrong. Of course if *** actually deliver that 3.5GH/s rate, the price is $149, it actually appears in October in numbers maybe it will be that. I'm not going to try and disagree on something I can at best only guess at. There are still more questions than answers here but only waiting for October and maybe a lot longer will you find out. And when we get there maybe one set of followers will say "I told you so". We will carry on delivering what we promised as best we can and make a better evaluation in October of what return can be had. Remember the GPU rigs have a 5-10X power usage for the similar hashing levels to CM1 rig so by your numbers how many GPU miners will be left by say December. I doubt many if *** are successful.
|
|
|
|
Isokivi
|
|
August 19, 2012, 01:23:46 PM |
|
I did some math when the exchange rate was 12$ and for me (at 0,17$/KWH) about 40million difficulty is the point where I have fairly expensive paperweights, to put it the other way around if the network hashrate goes up 18x.
That's a fairly much seems to be on the worst end of most peoples estimates but everyone has their own numbers. I can only say we don't think it will be anything like that but we could be wrong. Of course if *** actually deliver that 3.5GH/s rate, the price is $149, it actually appears in October in numbers maybe it will be that. I'm not going to try and disagree on something I can at best only guess at. There are still more questions than answers here but only waiting for October and maybe a lot longer will you find out. And when we get there maybe one set of followers will say "I told you so". We will carry on delivering what we promised as best we can and make a better evaluation in October of what return can be had. Remember the GPU rigs have a 5-10X power usage for the similar hashing levels to CM1 rig so by you numbers how many GPU miners will be left by say December. I doubt many if *** are successful. I for one dont believe *** will be able to deliver everything they've promised, especially the 3,5ghs trough usb-power seems far fetched. My gpu's can take a difficulty of 6,5 million at an exchange rate of 12$ before I need to shutdown. So I'd guesstimate my gpu's go on sale as soon as first reports of ***-products mining.
|
Bitcoin trinkets now on my online store: btc trinkets.com <- Bitcoin Tiepins, cufflinks, lapel pins, keychains, card holders and challenge coins.
|
|
|
Glasswalker
|
|
August 19, 2012, 02:22:48 PM |
|
Btw Glasswalker, they all decided to disable themselves like a domino effect around 6am this morning, so they last around 36 hours. Strangely no Hardware errors, but I did have to cold restart them to get them mining again. (thankfully now they are perma-flash).
Updated the icarus options, So lets see if that changes anything.
This was just this one board right? Did anything else happen around that time? Major change in environmental? (heat, power fluctuation, and so on). How is it connected through usb? (hubs/powered hubs) and so on. For it to run rock solid for 36 hours then up and puke sounds more to be related to something off the cairnsmore board (environmental, or infrastructure) unless the board itself is seriously faulty. Let me know, we'll see if we can diagnose.
|
|
|
|
Lethos
|
|
August 19, 2012, 02:46:01 PM |
|
Btw Glasswalker, they all decided to disable themselves like a domino effect around 6am this morning, so they last around 36 hours. Strangely no Hardware errors, but I did have to cold restart them to get them mining again. (thankfully now they are perma-flash).
Updated the icarus options, So lets see if that changes anything.
This was just this one board right? Did anything else happen around that time? Major change in environmental? (heat, power fluctuation, and so on). How is it connected through usb? (hubs/powered hubs) and so on. For it to run rock solid for 36 hours then up and puke sounds more to be related to something off the cairnsmore board (environmental, or infrastructure) unless the board itself is seriously faulty. Let me know, we'll see if we can diagnose. Nope, both boards did it. they both collapsed 30seconds apart. They are unpowered hubs, I could try having them powered separately. I don't think their was any major environmental issue, but I can't rule it out, I wasn't awake at 6am, just found what was left in the logs. Any suggestions on how to diagnose it, I've not looked into cgminer more verbose commands, could it tell me more why they became disabled?
|
|
|
|
Glasswalker
|
|
August 19, 2012, 03:30:44 PM |
|
When they disabled, did they disappear from the host? (the com ports disappeared?). Or did they simply stop mining? Did invalids shoot up? Were the LEDs doing anything wonky on the boards when it occurred? Can you give any further details on the exact behaviour you saw?
|
|
|
|
Lethos
|
|
August 19, 2012, 03:52:11 PM |
|
When they disabled, did they disappear from the host? (the com ports disappeared?). Or did they simply stop mining? Did invalids shoot up? Were the LEDs doing anything wonky on the boards when it occurred? Can you give any further details on the exact behaviour you saw?
It did appear as if the com ports disappeared. Doing a quick usb unplug and plug back in, didn't bring them back. Power cycle was needed. Invalids didn't shoot up, seemed pretty normal and low. So very low rejects and no hardware errors. The LEDs were doing what they usually do when it's not mining when I found it, blinking a steady red and yellow. If it happens again I'll try provide more details.
|
|
|
|
Glasswalker
|
|
August 19, 2012, 04:24:31 PM |
|
With the COM ports disappearing, that sounds like either a hub, cabling, motherboard port, driver, or FTDI issue. (with the last being the least likely, but still possible).
it could be due to power surging, a tiny variation in power supply (a minor surge on your power lines) combined with a power spike on the board or something might have thrown the FTDI into a wierd state.
Or cabling/hub being slightly out of whack could throw the FTDI into a wierd state too (which wouldn't be corrected without a power cycle.)
The fact the red led was still flashing, implies the 4 worker FPGAs were continuing to work just fine, but since there was no com port there was no way to give them more work.
Is this windows or linux? If it's linux try monitoring dmesg and/or kernel logs. Also if you can run udevd in verbose logging mode it could help show any wierdness on the USB bus that preceeds (and leads up to) this issue.
|
|
|
|
Lethos
|
|
August 19, 2012, 05:59:51 PM |
|
With the COM ports disappearing, that sounds like either a hub, cabling, motherboard port, driver, or FTDI issue. (with the last being the least likely, but still possible).
it could be due to power surging, a tiny variation in power supply (a minor surge on your power lines) combined with a power spike on the board or something might have thrown the FTDI into a wierd state.
Or cabling/hub being slightly out of whack could throw the FTDI into a wierd state too (which wouldn't be corrected without a power cycle.)
The fact the red led was still flashing, implies the 4 worker FPGAs were continuing to work just fine, but since there was no com port there was no way to give them more work.
Is this windows or linux? If it's linux try monitoring dmesg and/or kernel logs. Also if you can run udevd in verbose logging mode it could help show any wierdness on the USB bus that preceeds (and leads up to) this issue.
All very possible. I'm going to start by removing the hub from the equation.
|
|
|
|
rjk
Sr. Member
Offline
Activity: 448
Merit: 250
1ngldh
|
|
August 19, 2012, 11:19:55 PM |
|
Yohan, I'd be very very wary of using Intersango right now. Although it is technically a separate company, its owners are being sued for their dealings with an unrelated matter (mismanagement of Bitcoinica resulting in funds loss), and Who Knows what might happen in the future although they were legally and technically separate companies. I understand that their UK incorporation holds much weight with your decision, but still can't imagine that Intersango will escape unscathed from the fray.
I believe there might be one other exchange that is incorporated in the UK, but I am not sure of that and I don't remember the name of the one I am thinking of either. At least one merchant service provider (Bit-pay) offers a GBP cashout from their service, although I don't know who (if anyone) is using it or how well it works. (I would assume it has no major hassles).
|
|
|
|
zefir
Donator
Hero Member
Offline
Activity: 919
Merit: 1000
|
|
August 20, 2012, 07:46:55 AM |
|
[...]
As far as we can tell, and it's a lot more we can find on some exchange operations, Intersango has been good in it's dealings with customers. If you have any experience on that, good or bad, we would be interested in the feedback. This is still an area that we are learning about. The advantage with Intersango being in the UK we do understand what reglatory and business norms apply to businesses operating in our particular backyard.
Yohan, is there a reason why you would not use existing merchant payment solution like provided by bit-pay and MtGox merchant solutions? AFAIK, they offer ready-to-use fully embedded services that ensure you to get the exact amount of money donominated in fiat for a convenient fee (less then CC companies usually charge). There is neither need for you to take currency risks at all nor for adding a margin on top of your prices. From my personal experience with Intersango, I'd strongly suggest you to check their TOS and ask the staff explicitly if that would be possible. I tried to use Intersango for the payments of my boards around two months ago. At that times I could withdraw funds only to my own bank account, they lack the required licenses to wire money to a third party. Maybe they got the related licenses meanwhile, otherwise your outlined plan to accept BTC payments via Intersango as conversion proxy won't work. Another fact to consider is the low trading volume there. It took me almost a week to sell 1k BTC. Finally, as rjk mentioned, the management group behind Intersango lost quite some trust and reputation after the Bitcoinica debacle.
|
|
|
|
gigantic
Member
Offline
Activity: 89
Merit: 10
|
|
August 20, 2012, 07:55:31 AM |
|
not only after what happened with bitcoinica,
i heard from a friend that intersango UK metro bank account was frozen several weeks ago, all the persons that deposit funds were stuck for some time, eventually, after some time, metro bank finished the investigation and the account went back to normal, tough so, many people waited an unusual time for their funds, i think they lost many costumers because of that.
|
|
|
|
LazyOtto
|
|
August 20, 2012, 08:13:10 AM |
|
Merely the substantial entanglement with Bitcoinica is sufficient to seriously taint Intersango.
I suggest you do a lot of homework before making that decision.
|
|
|
|
spiccioli
Legendary
Offline
Activity: 1379
Merit: 1003
nec sine labore
|
|
August 20, 2012, 09:15:45 AM |
|
makomk's dcmwd2 190MH/s to all but ICA 9 which is running the 180 one. cgminer version 2.6.4 - Started: [2012-08-14 22:15:50] -------------------------------------------------------------------------------- (5s):6563.1 (avg):7560.4 Mh/s | Q:178541 A:836721 R:1225 HW:0 E:469% U:104.9/m TQ: 0 ST: 21 SS: 21 DW: 48618 NB: 866 LW: 1543751 GF: 804 RF: 7 Connected to http://eu.ozco.in with LP as user Block: 000004758cbe0866a5d4e5764c436bf1... Started: [10:58:16] -------------------------------------------------------------------------------- [P]ool management [ S ]ettings [D]isplay options [Q]uit ICA 0: | 379.4/378.4Mh/s | A:41893 R:59 HW: 313 U: 5.25/m ICA 1: | 379.8/379.3Mh/s | A:42538 R:58 HW:7132 U: 5.33/m ICA 2: | 379.7/379.0Mh/s | A:42259 R:73 HW: 188 U: 5.30/m ICA 3: | 379.8/376.9Mh/s | A:41751 R:58 HW:7764 U: 5.24/m ICA 4: | 379.7/379.2Mh/s | A:42173 R:58 HW: 121 U: 5.29/m ICA 5: | 370.4/371.6Mh/s | A:40264 R:59 HW:7081 U: 5.05/m ICA 6: | 379.7/379.3Mh/s | A:42004 R:64 HW: 115 U: 5.27/m ICA 7: | 379.8/378.3Mh/s | A:41816 R:63 HW: 388 U: 5.24/m ICA 8: | 379.6/377.8Mh/s | A:41966 R:56 HW: 400 U: 5.26/m ICA 9: | 372.1/371.8Mh/s | A:40069 R:55 HW: 41 U: 5.03/m ICA 10: | 379.8/377.4Mh/s | A:41831 R:51 HW: 495 U: 5.25/m ICA 11: | 379.8/379.7Mh/s | A:42137 R:73 HW: 38 U: 5.28/m ICA 12: | 379.7/379.4Mh/s | A:41812 R:62 HW: 100 U: 5.24/m ICA 13: | 379.7/379.4Mh/s | A:42089 R:52 HW: 131 U: 5.28/m ICA 14: | 379.7/376.4Mh/s | A:41550 R:63 HW: 661 U: 5.21/m ICA 15: | 379.6/379.6Mh/s | A:42103 R:60 HW: 61 U: 5.28/m ICA 16: | 379.5/379.2Mh/s | A:42174 R:62 HW: 124 U: 5.29/m ICA 17: | 379.8/379.5Mh/s | A:42206 R:61 HW: 87 U: 5.29/m ICA 18: | 379.8/378.8Mh/s | A:41777 R:73 HW: 205 U: 5.24/m ICA 19: | 379.5/379.5Mh/s | A:42309 R:65 HW:7080 U: 5.31/m --------------------------------------------------------------------------------
Note that boards with HW: > 7074 are showing some strange error since all of them, at a certain point in time, went from a normal error count to 7074 and from them kept increasing it. spiccioli.
|
|
|
|
yohan (OP)
|
|
August 20, 2012, 09:22:25 AM |
|
[...]
As far as we can tell, and it's a lot more we can find on some exchange operations, Intersango has been good in it's dealings with customers. If you have any experience on that, good or bad, we would be interested in the feedback. This is still an area that we are learning about. The advantage with Intersango being in the UK we do understand what reglatory and business norms apply to businesses operating in our particular backyard.
Yohan, is there a reason why you would not use existing merchant payment solution like provided by bit-pay and MtGox merchant solutions? AFAIK, they offer ready-to-use fully embedded services that ensure you to get the exact amount of money donominated in fiat for a convenient fee (less then CC companies usually charge). There is neither need for you to take currency risks at all nor for adding a margin on top of your prices. From my personal experience with Intersango, I'd strongly suggest you to check their TOS and ask the staff explicitly if that would be possible. I tried to use Intersango for the payments of my boards around two months ago. At that times I could withdraw funds only to my own bank account, they lack the required licenses to wire money to a third party. Maybe they got the related licenses meanwhile, otherwise your outlined plan to accept BTC payments via Intersango as conversion proxy won't work. Another fact to consider is the low trading volume there. It took me almost a week to sell 1k BTC. Finally, as rjk mentioned, the management group behind Intersango lost quite some trust and reputation after the Bitcoinica debacle. The biggest problem we have with choosing an exchange is that there is virtually no exchange of any size that hasn't had some sort of problem either with security or time to pay out going by forum postings. This partly of why we won't take large payments directly in Bitcoins as yet and our aim initially will be to convert to GBP/USD as fast as we can on receiving Bitcoins as a risk reduction strategy. In time that might be able to change because we are not operating on a wafer thin margin, as we are currently, or maybe even some of our suppliers take Bitcoins that mitigates the conversion risk a little. The latter I think is probably a few years away so basically we are left with the currency conversion risk that we have to manage. The other side of taking this mitigation is that by say converting funds every day we could have very high bank transfer fees on a foreign based exchange e.g. MTgox. MTgox would be the logical place for exchange based on volume but when we looked for the sort of information we look for as general business due diligence it's either not available or simple somewhere we don't know where to look for MTgox. I can say the same about a lot of exchanges. Ther are a lot nice looking websites with very little information about the owners of them. Overall we can see that Bitcoin is a good thing and there is a lot of nice things in it. From a business perspective the biggest issue is the lack of traceability. Again on the forum there are enough arguements about people having sent Bitcoins and the other side saying they didn't get them for us to be cautious. If we have missed something there do tell us. We are still learning about Bitcoins and things surrounding Bitcoins so there may be something here we don't know. Many of these sorts of problems with dissappear as Bitcoin gets bigger. At the moment it's a chicken and egg situation in that Bitcoin needs more conventional businesses to take Bitcoins but conversely they either won't, or will only in limited way as we do, because other companies that trade with them don't and they have the same exchange rate risk we have. On the second option we are open to people using other exchanges and sending us the money that way but the transfer fees are always going to have to be paid. If the exchange is UK and possibly EU based (using SEPA) it will probably be viable for smaller amounts. What we don't want to do is running a lot of accounts wih different exchanges ourselves as that is a huge admin issue. Using merchant services like BitPay is a possibility but we simply are not ready to do that yet. We can run live currency conversion on our USD and Euro pricing that runs on our existing webshop but we don't because it is an administration nightmare because we have to account for everything in GBP to our taxation authorities. We simplify that by usually only changing USD/Euro exchange pricing conversion rates every 1-3 months and sometiimes longer. Doing the same in Bitcoins simply won't work unless we put a huge risk margin on Bitcoins which I am sure none of you want.
|
|
|
|
|