 Show Posts
|
|
Pages: [1] 2 3 4 5 »
|
|
I have 3x Minirig SC's. Interested?
|
|
|
|
private message from NLA You need to update your BFL SC Wait List thread!
Sorry if I am not Quick enough for you, I do this in my spare time, Please make your feelings public than private messaging this. Kind regards squall1066 Ok! You need to update your BFL SC Wait List thread!  |
|
|
|
|
10/19/2012, 11113, N, 0, 0, 0, 1, NLA
|
|
|
|
The BFL wait list thread has just under 50TH/s of orders. I think someone else around here claimed that Inaba (or BFL_Josh) said they were closing in on 90TH/s of preorders. The bASIC is closing in in 35TH/s of preorders. The Avalon preorders were around 18TH/s.
For my personal calculations, I'm assuming a 200TH/s network by the end of 2012, and then increasing by 20% per month after that. You can do the math to figure out the rest.
Well, if you're assuming a 200TH/s network by the end of 2012, what are you assuming the true overall networking hash rate is right now to correspond to current difficulty levels, so I can do some simple math and multiply the current difficulty to get expected profit by the end of 2012? |
|
|
|
I posed this question to #bitcoin, and they weren't very responsive, so here's a chatlog of my thoughts that I'd like to get some feedback on. [16:35:02] <NLA> maybe I'm just not seeing it
[16:35:10] <NLA> but has anyone done some calculations
[16:35:20] <NLA> based on the BFL ASIC waiting thread
[16:35:30] <NLA> as to what the project difficulty will jump to
[16:35:39] <NLA> in October/November/December
[16:36:04] <NLA> and how that will affect how much that will affect the profitability of the different BFL ASIC
[16:36:49] <NLA> ie. at the current moment, the BFG of the bunch, the SC Minirig, estimated at 1500GH/s by BFL, would currently earn $180,000/month at current rates
[16:37:02] <NLA> obviously the difficulty is going to spike up when BFL starts delivering preorders
[16:37:21] <NLA> and likely small miners are going to be forced out due to the difficulty spike
[16:37:31] <NLA> but it seems like a lot of threads are ending there and focusing on the little guy
[16:38:06] <NLA> what I'm interested in is what happens to the big guys? does $180,000/month today become $30,000/month almost overnight? or $15,000? or $1,000?
[16:41:22] <NLA> Take for instance http://blockchain.info/charts/hash-rate and https://bitcointalk.org/index.php?topic=89685.0
[16:41:42] <NLA> if the current overall network hashing rate is estimated at 25TH/s
[16:41:57] <NLA> and based on the data from the wait list
[16:42:11] <NLA> which will likely not cover all the pre-orders made to BFL
[16:42:33] <NLA> of a total increase of 48.339TH/s when everyone receives their units and flips the switch on
[16:43:22] <NLA> thats almost a 3x increase in total hashing rate, and therefore a linear increase of 3x difficulty to accomodate the increased hashing rate
[16:43:50] <NLA> Consulting a simple profit calculator http://www.alloscomp.com/bitcoin/calculator
[16:44:31] <NLA> and multiplying the difficulty by 3x and plugging in 1500GH/s
[16:44:48] <NLA> results in an expected return of $59,924.51/month
[16:45:17] <NLA> which is significantly down from $180,000 but is still reasonable
[16:45:44] <NLA> but if everyone quits mining because the difficulty is so high
[16:45:56] <NLA> and lets pretend the exchange rate of bitcoin falls to $5/coin
[16:46:22] <NLA> that results in an expected return of $25,024.91/month
[16:47:40] <NLA> and I could what-if further along all day and make wild assumptions, but has anyone actually thought about the implications for the big guy, the guy that dropped $30K on a Minirig SC? Thoughts, anyone? Am I overlooking some possibilities here? Is it possible for Minirig SC owners to net $180,000 for a month or two before, say, the block worth is cut in half, or the network collapses, or is it possible that almost upon arrival that Minirig SC owners will net $10,000 first month, $8,000 next, etc., and never earn back their initial investment? |
|
|
|
|
I tried to do 7 with Windows, and it turns out that you need a motherboard with some PCI latency setting to get it working right (and that the motherboard I had did not support said setting). Expect to do 5 unless you're lucky and have a motherboard with a BIOS that supports it.
|
|
|
|
|
Is it possible to point particular GPU's in a single instance of cgminer to different pools, rather than having to do so with multiple cgminer instances? For instance, say I want to launch just one cgminer instance, and point one GPU explicitly at Eclipse, one at Bitminter, and one at Slush, all with different login credentials?
|
|
|
|
|
Wait, maybe I'm not understanding here. Are you saying that with Ubuntu you're using more power than with Windows 7? Sounds good to me.
|
|
|
|
- Support .bat as well as .lnk for launching cgminer?
It should be supported now. Let me know if it does not work for you. Lnk is just an example of how to wrap a long command line, set working directory etc.
Ah, I just tried it and it works. I swear I tried it before and it didn't work, must be my mistake. - Support attempting to reset network adapters before resorting to rebooting Windows?
Not sure how this would help. Would need more info on the kind of problems you are experiencing. I could enable/disable network devices, but if your router is a toast, it is kinda beyond akbash reach.
Well, its an oddball issue with Windows Vista and 7 (maybe XP?) that some people are aware of, and some people aren't, so I'm not surprised you haven't run into it before. Microsoft reps have stated they've fixed the issue before, etc., but the problem still exists: the network adapter just stops working, and network access is lost unless the network adapter is disabled and re-enabled in Windows. After that, it works again just fine (until it stops working again). So it would be a good check to try disabling/re-enabling all network adapters before restarting Windows. I believe this can be done with a simple command in the command prompt, so you should be able to implement this fairly easily. - Enable/Disable email status updates option, instead of "set it to infinity minutes" work-around?
Already done in version 1.0.6
Ah, I see it. It was kinda hidden in this big list of very-long-named options. Why not shorten the listen down to more simple terms, ie. STRING "email-address", what email address to send to BOOL "email-status", INT "email-status-interval", to email regular status updates BOOL "email-OS-reboot", to email on OS reboot BOOL "email-miner-reboot", to email on miner reboot If email-address is empty or not valid, ignore values of other options. Simple, to-the-point, conveys meaning, easy-to-read, takes up little space. Same goes for all the other options in the config file. - Email a portion of the akbash log as an attachment in emails?
How many "last bytes"? Configurable? The whole log? Attachments, means MIME (multipart) etc. Not sure if I want to get into this.
Maybe last 2000 bytes sent only when a serious problem occurs? Not sure, have to think about this one some more.
However many entries since the last one or two times that akbash was started, up to X number of kilobytes. Also, while we're at it, it would be great if akbash could have the option of splitting up the log files each time akbash is started and starts logging (ie. akbash_log.001.txt, akbash_log.002.txt, etc). Maybe even include the last Y number of logs with the email, maybe as a ZIP? == BUG ==Also, I've found a bug with akbash: will improperly detect GPU usage (somehow), misinterpret as 0% and restart the miner. I know it's a bug because I'm monitoring GPU usage with MSI Afterburner, and because the miner hash rate is fine all throughout the timeout period. |
|
|
|
Program seems to work alright, but every now and then I do experience some crashing that akbash doesn't help with. But otherwise the program is fantastic. Are you in the market for some critique/feature requests? - Yahoo is terrible (no forwarding/pop features for new users, messages occasionally make it through but many do not), can you add proper support for GMail?
- Include FAQ with akbash download, as a readme (which it is)?
- Support .bat as well as .lnk for launching cgminer?
- Support attempting to reset network adapters before resorting to rebooting Windows?
- Enable/Disable email status updates option, instead of "set it to infinity minutes" work-around?
- Open source!
 - Make the akbash output text more human readable?
- Email a portion of the akbash log as an attachment in emails?
Hell, with just better email support I'd donate some coin your way. |
|
|
|
|
I'm interested in maximizing my profits, and I could care less how it happens. I would be willing to sacrifice some fun features for more profit. In general, will I earn just as much, slightly less, or slightly more money if I mine as DGM versus PPS? This is mainly what I'm concerned about. Once I know that, I can decide whether to jump to some of these promotional pools I'm seeing advertised (110%, ProjectX), or if I should go with ABCPool because their PPS fee is only 1.5% (fairly low it seems). Or does the +10% of the 110% pool outweigh that low 1.5% PPS fee of ABCPool?
I'm hoping some people with experience can point me in the right direction.
|
|
|
|
Topic says it all really. Right now I have two rigs running with 7x 7970's, and I've been mining with EclipseMC as DGM, but I'm wondering if maybe there's a better pool out there that I could apply my 5GH/s to? I enjoy EclipseMC's perks, such as - Google Voice-compatible SMS messaging (!)
PaypalDwolla instant cash-out option- Fancy added-security authentication options
- Fairly high pool mining speed, for fairly quick block solving (and thus fairly quick payout via DGM)
Any opinions? Also, at the moment, I'm mining as DGM, but I'm wondering if with 5GH/s I should switch over to PPS or some other variant. My (likely wrong) understanding of DGM is that you're paid a set amount based on your hash rate, not so much based on the number of shares you contribute in a round.. but wouldn't that mean then that I would want to switch to as fast a pool offering DGM as possible? And I haven't done too much investigating into whether I would make more as PPS than DGM, but can anyone shed some expert opinion into this? I'm just a hardware guy at heart, trying to earn some moneys. Halp? |
|
|
|
A mixture of good ideas and basic troubleshooting: - Do a clean re-install of Windows. It could be that simple!
- If you don't want to re-install Windows, do a clean re-install of your drivers with ATIman, here: http://atiman.go-board.net/
- Your RAM might be bad. Run Memtest86+ overnight and make sure the RAM is alright. http://www.memtest.org/
- When you mine, do you mine one card at a time (for testing) or all at once? Instead of mining with all cards and causing BSoD, try mining with single cards at a time. See if the 3rd card alone causes the BSoD, or if the 5970's have something to do with it.
- Are you using a custom BIOS on any of your cards, by chance? Undervolting/overvolting? Underclocking/overclocking?
- Test the 3 cards you intend for this rig in a completely different rig, to make sure that the issue is definitely card related, and not related to motherboard, RAM, hard drive, or PSU.
|
|
|
|
You cannot run 7970s with SDK 2.5; if it's working with that combination it's only because cgminer has cached the .bin files.
Odd.. I wonder where "Lions" came from, then? I had a similar problem in widows 7 after adding a fifth gpu, 4 gpus ran great, 5 sent cgminer into a tail spin with the hash rates, but worked fine under ubuntu, I figured out it though I had to set the gpu threads to 1 then got full steady hash rates on all 5 gpus
Reducing the number of GPU thread doesn't reduced hash rate? I could not get the 4th to work until I increased the PCIE latency in the bios from 64 to 96.
Sadly my BIOs has no options to change PCI latency.  |
|
|
|
Thank you very much!
Donations appreciated!  Just did a clean uninstall/re-install with ATIman and installed 11.12 + the 2.5 SDK, the hash rate is terrible even though I'm getting 99-100% GPU usage (~482MH/s on stock settings with SDK 2.5, whereas SDK 2.7 did 550+MH/s). Also, "cgminer -n" identified my 7970's as "Lions" and not "Tahiti", so I'm guessing 2.6 and 2.7 are the only SDK's that will work properly with the 7970.. trying again with 2.6 before I return to 2.7.. EDIT: Just finished the clean 11.12 + SDK 2.6 install, and oddly enough it seems to not be crashing. Sometimes I can only run one GPU for mining because running all GPU's ends up with a cgminer crash. Currently have very reasonable CPU usage from cgminer (5-12%). GPU usage is still all over the place when mining with multiple GPU's.. some GPU's are hitting 93-99%, some are in the 80's.. its just a mess. Intensity at 14. |
|
|
|
Ignoring the gpu usage reported, what hashrates are you getting? Max you should get about .6MH per Mhz of engine clock speed.
Well, it's hard to ignore the reported GPU usage, since I've been noticing that low GPU usage corresponds to low hashrate, and high GPU usage corresponds to high hashrate. So while the cards might be capable of hitting 550MH/s at stock settings, when I run all the GPU's in tandem I might only be hitting 450-480MH/s per GPU. I'm noticing about .5MH/s per MHz. Also, I just used ATIman to do a clean uninstall and re-install of 11.12 and the 2.6 SDK, and cgminer is crashing when I choose a single GPU or all GPU's for mining. Do I need to compile for Windows from the latest commit? And by the way, ckolivas, do you have any recommendations for Linux (Ubuntu) drivers + SDK for a rig with 7x 7970's? I've been reading around that some people like 12.2 + 2.6 for 2x 7970's, but is it different somehow with 7x 7970's? I forget if AMD's Linux drivers have a GPU limitation like they do in Windows. How does one edit the BIOS of 7970? RBE doesn't support it The old fashioned way, with a hex editor. RBE should be able to support 7970 bios editing, but the dev seems to believe that there are new security mechanisms to prevent BIOS editing.. which there aren't. This post over on the OCN forums outlines how to edit the 7970 bios to change clocks and voltages: http://www.overclock.net/t/1196856/official-amd-radeon-hd-7950-7970-owners-thread/4220#post_16496161 |
|
|
|
Odd, I swear when I was using cgminer 2.4.1 it said the intensity limit was 10. Right now I'm using the 11.12 drivers, which I believe are the oldest drivers with 7970 support, so I'm not sure what to say. One GPU going, perfectly fine. Many GPU's running, GPU usage is between 90-95%, so I'm missing out on a good bit of performance between all the GPU's. |
|
|
|
|
WELL, oddly enough, it turns out the new cgminer (2.4.3) has higher intensity limits than the version I was using (2.4.1), up to 14 from 10. My GPU usage is now between 91-95%.. which is better, but still not the 99% I'm aiming for.
|
|
|
|
Currently testing my 7970's in Windows before I flash the BIOS to the settings I want (voltage, clocks, for maximum performance) and switch over to mining in Ubuntu, and I'm testing with AMD's 2.7 APP SDK. When I'm mining with just one or two cards, I hit 99% GPU usage and I get all the performance out my 7970's; when I'm mining with five cards, GPU usage wildly fluctuates between 70-86% and I net substantially lower hashrates. What gives? Less GPU utilization equates to lower hash rates! I'm suspecting that either the SDK or mining in Windows itself is to blame. Specs: 5x 7970's: 2 Diamonds, 3 Sapphires Windows 7 Ultimate x64 AMD 11.12 Drivers + AMD APP SDK 2.7 cgminer: process priority high, CPU usage between 6-15%, poclbm kernel, 2 threads, intensity 10, powertune 20, mining with EclipseMC - Any suggestions?
- Any evidence to the contrary that SDK 2.7 works just fine? Is it just me having the issue?
- Do the Linux and Windows SDK's have the same performance, ie. if SDK 2.7 is netting 70-86% GPU usage in Windows, it will be the same in Linux?
- What is the best SDK to use with 7970's in Linux? In Windows?
|
|
|
|
One last thing I remember is maybe playing with the PnP OS option if it exists: See this KB article.Don't remember seeing this option in the BIOS, but will check again! Thanks for the heads-up! Do yourself a favour and configure it on a linux machine.
Use kernel parameters to disable ACPI and self-assign IRQ's to each GPU (ACPI can stuffup horribly for some reason with so many GPU's)
Ive had 8 GPU setup like this temporarily but it really puts alot of strain on the board and I could literally feel the slots heatup with all 8 running thus I dont really support this for a stable longterm setup.
I did attempt to set everything up in Linux initially (Ubuntu 12.04) but I remember having problems getting it set up properly with the drivers and having the cards show up as being available for mining. I suppose I could sink some time into it and use Ubuntu 10.10 like I used to. Hm. Also, if I set everything up with Linux, I doubt I could underclock the RAM as much as I do in Windows (thanks to a "bug" in the beta versions of Afterburner). I guess I could BIOS edit since the 7970's don't do hash checking like the 6XXX series and below did (not a lot of people know this.. including the guy that makes RBE, sadly), but thats very uncharted territory at this point, and not many people have attempted it. Also, I've set the SATA type to AHCI in the BIOS as opposed to ACPI.. is that fine? Or should I configure things under IDE? Go in your BIOS and disable EVERY non-essential piece of hardware that you wont be using.
Did that. Disabled everything but USB, then disabled USB and gave that a try. Didn't help. Also, using 16x > 16x risers may be causing a problem, because your board is trying to negotiate to 16x speeds. Some boards, when negotiating to max slot speed, will turn off other slots entirely in order to have the full amount of bandwidth available (PCI-E lanes are not unlimited). Try using 1x risers across the whole shebang. If you have limited 1x risers, use the 16x risers in your PCI > PCI-E adapters since they are always 1x speed.
Seems reasonable. When all cards are plugged in via risers, I noticed that the two cards that weren't available for mining were (incidentally?) the two 16x connected cards. Could I just cut the part of the cable that makes the whole thing a 16x cable so that there are only enough connected wires for it to function as a 1x to 16x, and use the cable as a 1x to 16x cable? I'm willing to sacrifice a cable to make this all work! For science! Also make sure your risers are not defective.
The risers aren't defective; I unplugged two 1x to 16x cables from two other cards, and after a reboot all of the remaining cards were recognized and functional (including the 16x cards, which were not functional before). Powered shouldn't matter if you are doing underclock/volt. Also, I already have experience with 6 cards on 1 board as well as using PCI > PCI-E adapters (I'm using one in a 4 PCI-E slot board now, now have 5 slots).
But I'm not underclocking/undervolting, quite the opposite, I'm majorly overclocking and majorly overvolting (1231mV, 1270MHz core). Oh, almost forgot. You say you are using 2 PSU's. If you can, power on the PSU that ISN'T powering the motherboard FIRST, then the one that does SECOND.
Way ahead of you; I'm using this bad-boy:  |
|
|
|
|
