How can you load 2 GPU on a single 6 pin connector? That isn't possible. It will melt. Do you think you're smarter than electronic engineers at work? Lol. I saw the dividers 3 PCI connector on the riser card. It's nonsense. Such equipment will not work for a long time. I am opposed to expanding the capabilities of the equipment by such methods.
It will only load the PCIe socket draw of 2 GPUs, not the whole GPU, will be on 1 PCIe 6 Pin connector so 150W @12V, I'm quite happy to incorporate 1 per GPU but people will just use splitters anyway which are far more dangerous than a PCB based solution. Each contact of PCIe connector is rated at 9A, equating to 27A total draw limit for the connector or a 324W @12V limit, the PCIe standard is very conservative in this matter. 3 slots is not enough space for cooling.
Advanced GPUs all use 2.5 - 3 slots of space so it would only work for blower style cooling solutions.
With a spacing 4 - 6 slots and some provisions for compartmentalizing the GPUs so exhaust air is not cross-fed, it will be much more useful.
Cooling is the big problem to solve and a solution that takes this into account will be worth more on the market.
4-6 slot spacing isn't really possible for a single PCB solution, it would be far too big. Exhaust fans and sufficient air extraction are the best way of preventing warm exhaust air from getting to any other GPUs. Intelligent purchasing of GPUs for the particular situation will always be a factor with noise to cooling capability the main factor, and I'd argue that GPUs with horizontally aligned heatsinks are sufficient as well as blower style to keep decent airflow at low noise levels, however testing would obviously be necessary. However, only the most extreme coolers tend to go over 2 slots these days. It could possibly be increased to 3.5 slot but would no longer neatly fit behind 3 120mm fans. it would fit if you use 120mm fans to a 140mm shroud gpu 140mm 120mm fan>2 gpus behind each shroud https://www.amazon.com/Bgears-Cooling-Fan-Adapter-140mm-Black/dp/B0043GMY1U/ref=sr_1_3?now this means 3 x 140mm = 420mm Plus a little more but 420 mm = 16.535 inches so say the full build is 451mm that = 17.756 inches which means it fits in a standard rack that extra spacing allows 1080ti's at 200 to 225 watts which = beast a 19 inch rack mount with 6 1080ti's that can do 1200 to 1350 watts is a monster the fan adapter will actually cut fan noise down you could use a lessor delta like this one https://www.newegg.com/Product/Product.aspx?Item=N82E16835213015if you used blower 1080ti's like this one https://www.newegg.com/Product/Product.aspx?Item=9SIAAZU6HW6961 |
|
|
|
Still waiting on shipping and invoice for my demo. Maybe the few days I hope  |
|
|
|
I think I read somewhere that Bitmain probably made more then Nvidia last year.
Rumored, but no HARD numbers are available for Bitmain. Given they probably have a lot higher margin on their sales, their profits being higher than Nvidia in 2017 would be no shock. I think Bitmain is buying the chips from TSMC so they are probably getting good deals on those and I would assume they have higher profits that Nvidia does. --ypsi they where hold a ton of coins at the start of 2017 more then 40000 btc so it was 900 a coin in jan 2017 it was 19000 a coin in dec 2017 and they added 40000 BCH which was worth 1500 in Dec 2017 so 40000 x 900 = 36,000,000 40000 x 20500 = 820,000,000 so just on the holding of coins for certain addresses they have that is a 750million profit not counting anything sold. BTW they have sold off coins in these addresses So between Mt Gox and Bitmain coin sales this is the price drop for the market. and a 2000usd s-9 costs .25 btc in march a 3000 usd s-9 cost .15 btc in dec so dropping price on coins gets them more coins on sales. |
|
|
|
If it is to believed pool hashrate is correct, then Bminer does overstate hashrate in console by a big margin...    11 038 in console, but just 10 094 on the pool... I can tell from my tests so far that Bminer is a little better then dstm. And I can tell from my tests that Bminer reports more hash then dstm 1294 for 2 1080ti's = Bminer 1257 for 2 1080ti's = Dstm these numbers are what smos reports and what can be seen if I attach a monitor to either pc so 1294 /1257 = 1.0294 that means Bminer is 2.94% faster then Dstm on identical machines with the same settings nicehash has reported $5.37 in earning for dstm and $5.48 in earns for bminer so 548/537 = 1.0204 which is 2.04 % faster for bminer so 2.94 vs 2.04 both show Bminer to be faster and the variance between 2.94 and 2.04 is small enough to be luck. Bottom line shows that bminer would earn more. I will continue testing but so far bminer shows better hashrates at the pc and the pool then dstm. I still need to test power draw those test will be next week. |
|
|
|
1st 24 hour test shows bminer is a clear winner 5% on the plus  new test I flipped the pc's so if the pc was the reason for success and not the bminer this will show the totals to get closer but if the bminer was the reason and not the pc the totals will get further apart in the favor of bminer. be back in a day!  |
|
|
|
https://chainradar.com/xmr/chartIn situations like this I think it is helpful to look at this from a purely economic standpoint: what are the incentives to Bitmain/Baikal/etc. using their own miners before selling them vs. the costs. Since these are foreign companies with near-zero liability to, say, customers in the US or EU, they don't have to worry about individuals claiming they were sent a used or defective product; big farms, perhaps, but not individuals. Conversely, stealth mining the shit out of a high value coin like XMR with your hot new ASIC miner would be all but irresistible so there is a huge incentive to do so. Frankly, the incentives grossly outweigh the costs, and so you are left with trusting companies like Bitmain and Baikal to act ethically. You can take that bet if you want, but I certainly won't. The "Vega are huge monero miners" news got widespread in November - which is when Vega supply dried up VERY quickly and the hashrate climb started. Then the "3'd party" cards hit in Febuary or so, likely feeding the second hashrate surge 'till the price spike collapsed and hashrate went flat. I doubt Bitmain had a LOT of miners running in that timeframe, though I'm sure they had SOME "engineering sample" type miners running to test out. Yeah, I definitely remember seeing all the Vega pumping, and I just as distinctly recall them disappearing from Newegg's stock in a matter of days... and ever since then they have only been sporadically available, and at ~2x MSRP (Newegg's price, not some janky reseller's). No doubt some of the increase in hashrate is from the unprecedented success of compromising servers into XMR mining botnets, but come on, average network hashrate just 3 months ago was around 200 MH/s and now it is 1 GH/s. If half of that jump was from Vegas, and each Vega does 2 kH/s, then AMD must have sold ~400000 Vegas in less than 3 months. That doesn't sound realistic to me, as that would normally be a fantastic number of sales for the entire lifecycle (~2 years) of a flagship "gaming" GPU. Then again, I can't say I would be surprised to be proven utterly and totally wrong about this. Crypto kinda defies logic (just like dotcoms did 20 years ago). 400,000 vegas is 240,000,000 = 600 x 400,000 I would think that is more then they sold. so maybe 10 tester prototypes then a small 100 unit run that is 110 which is the same as 20 x 110 = 2200 vegas A good way to understand how many are mining as I type is list the hash rate every day til the al- gore - rhythm is altered If it drops off by 20% odds are bitmain mined more then 110 units. |
|
|
|
19 hours into first part of the test. and bminer is 5% more earnings. now this is one of six parts to do a decent test. part 1 24 hour test ------------- 19 hours done part 2 24 hour test flip the software part 3 20 min test check power use on rig 1 with bminer part 4 20 min test check power use on rig 1 with dstm part 5 20 min test check power use on rig 2 with bminer part 6 20 min test check power use on rig 2 with dstm  |
|
|
|
I think it said that it was NAND flash so I assumed it was more the SSD side of things but I'm no expert. Oh may not be much of an issue. |
|
|
|
Are you trolling or you just didn't read my message? It's not a matter of naive or not naive, math doesn't lie, hashrate can't be hidden. You repeat yet again that Sia had a "steady rise in hashrate" "months before bitmain...". THAT IS NOT TRUE  From: https://siastats.info/miningA3s batch 1 sold on Jan 18th. So, tell me, how in hell bimain "mined with them for months before selling them"? The entire Sia network hashrate was equivalent to only a few A3s before end of Jan. That chart does seem to prove your point w/r/t SIA, but in the case of XMR it really looks like some front-running has been going on: https://chainradar.com/xmr/chartIn situations like this I think it is helpful to look at this from a purely economic standpoint: what are the incentives to Bitmain/Baikal/etc. using their own miners before selling them vs. the costs. Since these are foreign companies with near-zero liability to, say, customers in the US or EU, they don't have to worry about individuals claiming they were sent a used or defective product; big farms, perhaps, but not individuals. Conversely, stealth mining the shit out of a high value coin like XMR with your hot new ASIC miner would be all but irresistible so there is a huge incentive to do so. Frankly, the incentives grossly outweigh the costs, and so you are left with trusting companies like Bitmain and Baikal to act ethically. You can take that bet if you want, but I certainly won't. Bitmain made so much money last year they simply don't know what to do this year. So they figured WTF why not attack gpu coins. The next few months will be very interesting to say the least. |
|
|
|
Scorpio, pools do not know your hashrate, they only estimate it.
Hi Bimmber would you please explain to us how the pool works? usually how pool estimate the hashrate? well you don't need to know that and it varies pool to pool. But it is a good question. So here is an attempt to explain it. They all count your good shares submitted to them. They then devide them by a time limit and they say a rate. That is only an estimate because you can toss lots of (7's) in a give time frame and look like you have a higher rate then you do. Of lots of good shares in hour 1 = high hash followed by no shares in hour 2 = low hash the reality is your tosses of the dice per hour are your true hash rate but pools only count your good shares per hour thus it is an estimate. |
|
|
|
This tests don't have power consumption comparision.
yes they do wait for screen shot all gpus are set to 155 watts all settings are the same, but for the actual power drawn I need to stop at friends office and put the same meter on each one. that will be done in time as I have to use the same k-watt meter as no k-watt meters are identical. so to test power would interrupt this stage of the test. I SEE THAT bminer is reading higher cores with the same settings look at screen shot . this indicates how higher hash may be acheived. it may also mean the higher you do tdp and set the core clocks there will be a difference between the 2 softwares.
Ie in a closed case at 155 watts a 1080ti bminer wins
in an open case at 205 watts a 1080ti dstm wins
Much more complex to test for but I can do it with the two identical pc's if I can cool the cards off
settings for bminer 5.5 https://i.imgur.com/7IPdxeA.pngsettings for dstm 6.0 https://i.imgur.com/X3SWUr2.png |
|
|
|
They're going to be bloody expensive paperweights real fast.
I doubt even the FIRST adopters will manage to make back half their price.
Seems like Bitmain didn't think THIS move through at all.
Bitmain and baikal knows exactly what they're doing. They got the rug pulled out from under them with the algo change, making their ASICs worthless. All they can do besides recycling them for precious metals is dump them on unsuspecting, uninformed miners. This is 100% a malicious move on both their parts. If the algo switch hadn't been announced, these would still be humming away in their farms making them huge profits. It's sad that they'll get away with selling some of these to people with more money than brains. they think they are smart but in the long term this is a big miscalculation based on thier overwhelming greed. they have basically exposed the fact to even the nonbelivers that they mine for months with thier asics before releasing them. They have shown withoute a doubt that they are a bad actor in the mining space , heck I think this might eveb give sia and other devs the ammunition they needed to push thier forks though. I hope this is the info people needed. I’ve been saying bitmain is bad actor since s5 sales. Most people don’t like to admit when their wrong. But those who don’t see this coming is also not good for this space. Maybe is time to go back and read thru some old threads when people defended bitmain with every word they spoke. Once there was a time in crypto when we were proactive not reactive. I miss those times I used to like bitmain did a lot of purchases from them. As time went on I see them as simply too big and as too greedy. Once in a while I still think they have a master plan but reality is that plan does not include me. We are in very turbulent times. For crypto coins . If all coins die but BTC and BCH it will mean only really big owners of gear and all of us in the small to medium mines will be out of the game. Hobby -------- will stay around Small --------- will die Medium ------- will die Large ---------- may merge into Huge Huge ---------- 3 to 5 companies Massive ------- 1 to 2 companies I am in the small size not the medium I am over 10k and under 100k total gear coin cash My hope is rolling algo's for multiple coins a zec that changes algos a monero that changes algos this keeps gpus alive. gpu miners need understand that 110% tdp and lots of rmas hurt the go mining indsutry as gpu companies have no interest in dealing with 2x 3x or 4x the rma rates I will continue to hold me last 24 1080ti's I will not load up on asics I have 6 or 7 from sidekick and 8 from jestanop all usb sticks. I will still get the 1 dragon mint I will still get the bitmain L3+ I may get a few avalon's as I hope they will become leaders in servicing their gear. Oh I am going to be doing a new thread. I have the two identical omen dual 1080ti pc's I will be doing software tests as with 2 identical machines I can fairly compare zec programs 1st test will be bminer 5.5 dstm 6.0 |
|
|
|
They're going to be bloody expensive paperweights real fast.
I doubt even the FIRST adopters will manage to make back half their price.
Seems like Bitmain didn't think THIS move through at all.
Bitmain and baikal knows exactly what they're doing. They got the rug pulled out from under them with the algo change, making their ASICs worthless. All they can do besides recycling them for precious metals is dump them on unsuspecting, uninformed miners. This is 100% a malicious move on both their parts. If the algo switch hadn't been announced, these would still be humming away in their farms making them huge profits. It's sad that they'll get away with selling some of these to people with more money than brains. they think they are smart but in the long term this is a big miscalculation based on thier overwhelming greed. they have basically exposed the fact to even the nonbelivers that they mine for months with thier asics before releasing them. They have shown withoute a doubt that they are a bad actor in the mining space , heck I think this might eveb give sia and other devs the ammunition they needed to push thier forks though. They're 100% dumb, but I doubt they miscalculated. They have been able to mine YUGE sums of money with these in the last months. They have for sure covered their R&D and production costs on these ASICs. Their cash cow is now dead, so they try pass these on to us. ASIC companies are absolutely, 100% of the time bad actors. Does that mean you can't make money with some ASICs? No. But when they decide to sell ASICs to the public, they have calculated that they've been milked as much as possible and then throw us the scraps to make even more money off of our backs. Fuck ASICs, and fuck bitmain in baikal in particular. yeah I only have a few asics and after this im resolved to stop feeding the pig Im out , idgaf how much they claim we can make Im onky buying gpus , every dev out there needs to fork the fk out of thier coins to get them out of the business of dumping used gear on the masses keep your useless toastes baikal and bitmain and fuck off Same here. My Titan is on life support, after I pull it's plug I'll have rid myself of all my ASICs. Only decentralized GPU minable coins for this guy from here on out. Hopefully Monero inspires other coin devs to shove a cylindrical cheese grater up bitmain & companies assholes. any small to middle guy should side with gpu's Basically my single biggest problem with asic's = real warranties with the solar array I can fix power cost to always make some money unless the gear breaks. An evga 1080ti with a 3 year warranty = a certain profit an asic with a 90 day shit warranty = certain uncertainty |
|
|
|
It looks like it has a static price value for BTC, but that skews it because the electric bill is due monthly
No, it's gross mining revenue in USD. Basically (block reward)x(market price)/difficulty at a constant hashrate, using published historical data. Electric bill is not in it. Did your OP involve utility or other costs at all? It looked to me like you were simply lamenting low gross revenue. Please clarify if I misread it. I fully admit your chart is accurate . I see a gross revenue chart that is being used in a misleading manner, since you left out the gross costs chart, but if you want to say that gross revenue has dropped per th okay yeah fine To not show that cost per th and power per th has decreased enough to fully offset the revenue per th is a joke Once you do that net revenue shows up. Frankly gross revenue means nothing to me Now the fact that net revenue which is what you are leaving out is clearly what matters to me. Proper way is to show that is gross revenue decreased per th and show that gross expenses decreased per Th. So I fully admit your chart is accurateI don't bother most of the time but the accountant in me hate to see gross revenue without gross expense and of course net profit.  That is why I said your chart is a joke. < and that is not what I should have said although I did say I did it as a shock value statement. To get attention Below is what I could have said: I stand corrected your perfectly accurate chart on gross revenue was used in a manner that is a joke. Since you left out the charts showing cost and net revenue many people without a background in accounting will be mislead and not realize net profits have not dropped.That is a more accurate description on my part. But it lacks the shock value of my first statement. Hopefully people read this post and realize what I am talking about. |
|
|
|
There is nothing impressive, difference between them two is very very small as you can see from Nicehash screenshot. Even bminer can handle higher OC, shows more sol/s in console, in reality in doesn`t translate in much higher valid share count on pool side.
I realize I can do lots of tests here to show many software programs against each other and how they stack up. Having the 2 identical machines purchased on the same day and shipped to my house on the same day allow for a chance to fairly compare the various programs used on smos simple mining test has just started about 9 hours. bminer is better so far. but that could simply be one omen dual 1080ti is better. after 24 hours we switch software and see if the better rates follow bminer to second omen   |
|
|
|
They're going to be bloody expensive paperweights real fast.
I doubt even the FIRST adopters will manage to make back half their price.
Seems like Bitmain didn't think THIS move through at all.
Bitmain and baikal knows exactly what they're doing. They got the rug pulled out from under them with the algo change, making their ASICs worthless. All they can do besides recycling them for precious metals is dump them on unsuspecting, uninformed miners. This is 100% a malicious move on both their parts. If the algo switch hadn't been announced, these would still be humming away in their farms making them huge profits. It's sad that they'll get away with selling some of these to people with more money than brains. hold onto your 1080ti's they are 11gb maybe that would be good for next algo. if the next algo allows for only 10gb of ram the 1080ti's could go to the moon. |
|
|
|
|
war
asic vs gpu
just remember asic = shit service when gear breaks.
gpu = real warranties when gear breaks.
|
|
|
|
thats pretty much the same so far but over a week we should get a meaningful number. All bets are off !
yeah it is a fair test as far as I can tell. |
|
|
|
|
Show Posts
