Thanks a bunch, i tried google, bing, and stackexchange extensivly over the past 24 hours to look for a windows version with no resolve, it would be awesome if such a thing was out there for us lay-persons.
You know, to get better help you should describe the tools you are planning on using to analyze the data once you get it. What kind of software you already have and are familiar with? What kind of hardware do you have available? Have you ever actually processed some really big data-sets on your own, without using external database? In my previous company we had certain class of prospective customers that are just not worth dealing with, not even worth the salesperson time spend talking. Somewhat contrived example to maintain privacy: a department has site license for a non-current version of SPSS for 32-bit Windows, and always inputted the data either in the old dBaseIII+ format (2GB limitation) or used Oracle ODBC driver for 32-bit Windows (with the actual data on the mainframe/cluster). This is the type that isn't worth helping, even though they have budget to theoretically buy a working solution, because they don't understand their own tools and they've been doing bullshit GIGO for years. Any money gained isn't worth the time it would take to support them. |
|
|
|
Yes, that's fine for my purposes. As I told the other person above, this is a learning exercise to explore what Bitcoin is capable of.
Understood. The point of using Bitcoin is that we want its blockchain to know the program and the inputs and verify the output, then release funds depending on the output. It isn't enough that both you and I can run the program off the blockchain, because one of us could lie and say "hey when I run this program it says I won, so give me the money." Trustless here means you don't have to trust your partner to be honest and pay you if he loses. The blockchain resolves any dispute about who the winner is, and causes funds to be paid to them.
OK. So we have fairly unsophisticated but way too large (for current Bitcoin blocks) program to verify the result of the game. We could say it is doable with a future soft fork. The only remaining problem is how to verifiably enter the moves of the players into that program. Do you want to do that (1) one by one (2) in pairs (3) entire game in one fell swoop? I guess we now have a "trustless move After the soft-fork gets implemented the whole operation would be: <white-pubkey> <black-pubkey> OP_OFFICIATE_CHESS_GAME which pops two public keys of the stack and pushed one of three numbers {0,1/2,1} back onto the stack. So you are back into square one of Bitcoin: at least 50%+ of miners have to have the correct implementation of OP_OFFICIATE_CHESS_GAME soft-forked operator. It isn't a technical problem but a political one. More precisesly: it isn't 50% of all miners, but of those miners that implement and mine OP_OFFICIATE_CHESS_GAME a majority (>50%) have to use the correct implementation. |
|
|
|
The smart contract would not be generating moves. Just officiating a human vs. human match. I have in mind something like in the article I linked to. Checking whether a move is valid is far easier than generating good moves.
I'm sorry, I misunderstood your intentions. Timing moves is tricky, however it may be possible to play long games where the time per move is measured in blocks. If the time is long enough (for instance, 100 blocks per move), the variance of block creation may not be an issue. However since the goal is not to post every move on the chain (as mentioned in the original post, this is super inefficient), there may be some way to play with faster time controls with all verification done off-chain or over a Lightning channel.
But judging the game isn't just verification of the moves. Most of the effort in judging actually goes to verifying the fairness, meaning that any of the sides isn't using any disallowed resources. So it would not be "human vs. human" but more like "correspondence game with unlimited computing resources and unlimited consultation". I mention zero knowledge proofs because those are solutions that Greg mentioned already exist (ZKCP) or are eventually coming (SNARKS). You're right that it seems like zero knowledge shouldn't be needed for this. If anyone knows how to trustlessly verify an arbitrary computation without it, I'm interested to know how.
Is that a trick question? If you know the computation program and the input state then just rerun that computation as many times as you want until the you are convinced. Do you have some strange definition of "trustlessness" in your mind? Something like "left hand doesn't trust right hand?" I think chess programming is just a nice way to really learn a new programming languages. It is non-trivial yet small enough that interesting and sensible programs can be written by a single programmer. Other than that what is the point? It really is "play to play" not "play to win". |
|
|
|
If this is not possible, what is keeping it from being feasible?
Competitive chess playing is very resource intensive: fastest multicore processors, gigabytes of hash tables for middle game, hundreds of megabytes for opening books, tens or hundreds of gigabytes for endgame tablebases (6- or 7-piece respectively). You would have to come up with a different chess game scoring system that doesn't take account of time. Also, chess has no hidden state, so what would be benefit of using the zero-knowlegde system in the game? |
|
|
|
I think that someone could make money by buying a few dozen servers distributed across the globe and selling GRE-tunnel-based DDoS protection from SYN floods and maybe also bandwidth leeching (by tracking when new IPs start using way more traffic than anyone else), ideally with anycast IP addresses to distribute traffic among the firewall servers. I think that you could do it largely with standard iptables rules, though it'd be very complicated. If I was setting up a service like this, I would oversell like crazy -- each site is only actually DDoSed a very small percentage of time, so you only need enough ordinary capacity to protect against one or two active attacks --, but then have some sort of backup plan to add more servers in an emergency (maybe by spinning up EC2/DigitalOcean/Vultr instances, which are expensive compared to a dedicated server but quickly available in case more capacity is needed now).
Anycast to distribute state-full traffic? Anycast only really works with stateless/connectionless services like DNS over UDP. Anything else requires a modified client side to recover the hidden state. And in addition to the above modifying the routing rules after the DDoS started to add more firewall servers? Guaranteed failure because it will prolong the instability and limited availability. "standard iptables rules, though it'd be very complicated" - this claim is such a deep bullshit, that I can't believe a sane person with IT knowledge would utter it. What about the state of the TCP/IP socket required to track sequence numbers? To me it seems like you've talked to too many professional bullshit salesmen in the DDoS mitigation industry and they successfully managed to turn your brain to mush to prepare you for closing a sale. Four days ago you had a generally correct idea. Within AWS the GRE tunnels are not required because EC2 offers a private LAN segment for free to allow connections between instances spawned from the same account. Maybe just get some sleep and then implement it yourself. |
|
|
|
|
The whole idea is ambiguous and not well-defined. The problem centers around what it means to "verify zk-SNARK". There are two possible interpretations:
1) evaluate a given zk-SNARK to verify that the input data satisfies the particular condition.
2) verify that the given zk-SNARK actually tests for the condition stated by whoever had given it to you.
If you re-read the linked paper paying attention to trust boundaries, you'll find that it doesn't really apply to the trust-less cryptocurrency model.
What is missing is a distinction between two processes:
1) generate a zk-SNARK for a particular condition without knowing your resultant zk-SNARK in advance
2) knowing a particular zk-SNARK perform a regeneration of it to verify that it was generated properly and tests for the condition that it was supposed to test, not something else
Only if you'll find process (2) that takes substantially less energy than process (1) the "generation of zk-SNARKs" would be anything suitable for trust-less mining.
|
|
|
|
If anyone is an actual expert in Linux networking (ie. the term "GRE tunnel" is familiar to you), I could use your help in figuring some of this stuff out.
I'll be glad to help. My testing computer & router farm is temporarily in storage due to moves, but my brain is available. The further discussion probably has to switch to private messages for security reasons. |
|
|
|
Location - London, I doubt any mining farm could be located there
I don't know anything about this particular company. But I do happen to have some fresh information about the real-estate/property developers in Greater London area. There is plenty of freshly build/renovated residential and commercial locations that are looking for a buyer without any current tenants. Developers are again resorting to having fake occupancy by letting trustable people to live and use the facilities free of rent&utilities. Conditions typically are: be personally known to trusted people in the construction business, be able to move out on a very short notice (few days at most), be trusted and capable of leaving the property in the "as new" condition, don't look and behave like a chav. The holy grail of the Bitcoin mining is free electricity. Paradoxically, it is easiest to get it in the areas that normally have very high rents and costs of utilities. You won't get it on the regular commercial/residential terms, only as a temporary fallback financed by the developers/bond holders/speculators. Certainly, you won't find them advertised anywhere. From my broader experience I'm pretty sure that such conditions aren't limited to London, UK. So keep your eyes open for opportunities like that in the city near you. Again, the above was a general information, not specific to any particular offer discussed above. |
|
|
|
My guess (hope) is that Chip Temp is the highest reported on the board. Given the simplicity of on-die chip sensing and the space afforded @ 16nm node it would be shear stupidity not to have each report its temperature... Hell, TSMC probably provided that IP block "free" of charge.
Unfortunately it isn't that simple. There are two issues: 1) lack of calibration. Performing a proper calibration of the internal temperature sensor would be probably prohibitively expensive. Uncalibrated sensor will have good differential accuracy (is the temp higher or lower?) but bad absolute accuracy (is it 100C or 120C?). 2) internal switching noise. The mining chips have super-high levels of internal simultaneous switching noise, beyond what is tolerated by the standard library IP blocks. Spondoolies tried to work around it by configuring that temperature sensor cell to provide intentionally coarse measurements; IIRC the step was 5C. The other way to work around those limitations is to use external measurement chip (like LM75) with external sensing diode that is internal to the mining chip. This setup is much less sensitive to the manufacturing process variations. Is that or anything over 100 maybe 105C 'safe'.... Maybe.
The determination of true maximum operating temperature would be too time consuming and too expensive. There published values are just very approximate goals that are used as an input to the electromigration models in the CAD software. And again the normal operating points of the mining chip are outside of the ranges where those models have any accuracy. So it is back to keeping the fingers crossed. There's no time for the proper qualification procedures and fault analysis. In addition to the above the maximum safe operating temperature would be dependent on the core supply voltage. For higher voltages the temperatures would be lower. My textbooks are in storage, I can't look them up now. But anyone interested could dig around the web and try to find some school problems like: the capacitor has plates at a distance of 10nm, the voltage difference between plates is 1V. What is the equivalent pressure in PSI required to keep those plates apart? The results are frighteningly high. Modern processed utilize pre-stressed silicon to partially deal with those electrostatic forces. |
|
|
|
|
By the way: Google has announced that starting sometime in January 2017 it will start penalizing in their search results pages that require horizontal scrolling to read the text. To avoid penalization pages should properly reformat to be readable on most desktop and mobile screens.
|
|
|
|
|
I confirm that "Satoshigames.io" advertisement creates same problem for Safari users on Mac OSX (older versions).
|
|
|
|
Thanks, graymatter, for a very thoughtful reply. I will quote it for my future reference. I guess we are most familiar with the opposite corners of the data center business space. You seem to be more of a greenfield guy hiring a temporary workforce. I'm more of the brownfield guy preferring to work with employees, both part-time and full-time, and maybe hiring contractors also keeping their own permanent workforce. I'm particularly glad that you've mentioned that I neglected mentioning fire protection when using improvised fixtures for racking. This is again probably issue of working at different corners in the same large business space. You seem to have experience with dry climates whereas I'm familiar with moderate ones, where the main danger to the cardboard and wood is moisture, rot and other living things, not fire. As a mostly-brownfield person I'm mostly familiar with too sensitive and too wet fire protection, completely unlike the case in Thailand. The two things I completely don't buy in your scenario is "10 minute MTTR" and "1 person working 20hr/week". To me they look like clear warning signs. I've seen many facilities making similar claims, they were seriously under-maintained. This is one point of the checklist that our buyout team uses: carefully inspect the facilities for the sign of rushed maintenance work: dropped washers from the rack-mount screws, whole rack-mount screw kits still in their plastic baggies, entire cable assemblies dropped on the floor and replaced. It is particularly important with the rack-mount case designed like yours, which has contiguous bottom, which make it additional "floor" to drop things on. The room with long aisles like in one of your original photos is obviously unmaintainable by single person, it is a physical impossibility to replace equipment there by one person. It requires a work-team, probably equipped with walkie-talkies to actually productively work while the facility operates. Harmonics don't really effect ASIC and solid state systems. For example, the only concern you need to have with harmonics and vibration is with spinning disks (magnetic disks) in a traditional DC environment. Bitcoin miners have none of these problems. As for the rest of it, I won't even bother addressing as Helmholtz resonance is completely irrelevant to anything really... No part of the HVAC system has empty cavities, the hot isle exhaust most of the air out of the facility except for what we wish to recirculate over winter..
There is also plenty of space in the hot and cold isles to handle the volume of air that is passing. We're not using high pressure 36'' (3ft by 3ft) ducts, these are 6ft by 14ft ducts...
The design of the PSU to be above the miner is actually imperative, as bitmain supplies such short DC supply cords. The entire weight of the PSU is supported by two rails that the PSU slides into, and the PSU itself is simply screwed into the front to allow someone to unplug / re-plug the power cord if they need to simply power cycle a miner without having to go into the hot isle...
"The corrugated cardboard + plastic foam + particle board can direct the air around metal warehouse" - yeah and cardboard and plastic have one thing in common, they are both combustible. Regardless if I were stupid enough to do it this way and risk burning thousands of dollars worth of equipment to a crisp, the firecode does not allow me to, nor does the facility itself, so both 'cheap' solutions are irrelevant. If you want to look at what happens when people do this, lookup spoondoolies-tech fire hong kong.
As for multi crew... you obviously are not a very organized person... All S9's we know the hashrate, operation, and the location of every miner in the facility... We keep a database online (intranet) that shows when an S9 goes down, or is hashing slower. A single user can walk to the floor, find the miner, and take it back to the NOC for servicing, often times about a 10 minute operation to swap bad parts and RMA.
As for deployment, yeah, we bring in a team of temps that we pay about 10 dollars an hour to 'install' cable' and do basic stuff. The core team supervises, does final install, and logs the position of the equipment with a tablet on wifi (temporary wifi) pressing the broadcast button one rack at a time. Takes a little while, but once mapped, 1 person working 20 hours a week @ the facility can service 2500 S9 units easily. Its called being organized...
To give you some numbers, the total cost of the solution, compared to the cost of the miners / equipment purchased was 2%. Meaning, 2% of our total cost that we spent on the entire deployment FOR JUST THE MINERS, was the cost of the shelves... The shipping cost was approximately 8% of the total order...
Estimated lifetime for 16nm equipment I give 18 months, before 10nm readily available. With our cost of power 99% lower than the rest of the US, if we're not profitable, then no miners are going to be. at 2.5c per KW, there are only a few places in the world that are cheaper, and nothing else in the US. Technically our old spoondoolies tech miners are still profitable on our current rates, but it would be a waste of space for us to keep them.
Final thought, these things take 5 minutes to install, work flawlessly, and will accommodate future generations of bit-main miners. They work with the S1,S3,S5,S7, and S9... My guess is, they'll work with the S11 as well... How much time labor and money do you spend trying to get the combustible cardboard and plastic up? I've done it before in a cheap facility and its a lot more work than you think.
Obviously you'll go with the combustible datacenter in a chicken coop option, but some of us prefer making money hand over fist and being cheap doing it and not have to worry about 30% of our equipment failing in the first year... I feel many other people are in the same position we are when dealing with multi-million dollar deployments of bitcoin miners. For about a total cost of ownership of 2% greater than fire...
The opposite of your neat greenfield build out isn't a chicken coop. It is a brownfield deployment opportunity like the classical https://en.wikipedia.org/wiki/One_Wilshire . I have no information on how good is your estimate of 18 months of useful life of Antmier S9. I take it on face value and make a comment: the brownfield opportunities frequently come with electricity already prepaid for some time, typically 6 to 12 months. So your "lower than 99% of the USA" has to be taken with qualifiers: only for multi-year contracts, only for multi-megawatts deployments and only for full construction/renovations. If somebody can live with 1 to 2 years contracts, accept smaller deployments (of the size of supermarket or shopping mall food court) and promise to not make permanent alterations to the facilities; then the electricity price can be zero because it was already paid by the previous contract or the bond issue financing. Thanks again for your thoughtful reply. |
|
|
|
It is working as it should, may i ask if some friends of you have tried to connect?
My Team and friends seem to have no problem when trying to access the site.
Also it doesnt make any sense to me why you should be blocked, maybe you are using a blocked ip-range. (we blocked some spam related ip ranges)
But you can also visit us via smartphone/tablet as we are optimized for that too.
Not even close. But your misfeature made me check out how professionals do the "browsercheck". I checked my bookmarks for the sites that use similar technology. One related to Bitcoin is http://chain.so , that has block explorers for several mainnets and testnets. It is handled by Cloudflare, either by the free level or the cheapest level, because it uses the SSL certificate issued by Cloudflare. The block explorers are very AJAX-y, but do have some limited usefulness even with Javascript disabled. I checked it on all available browsers, both with Javascript enabled and disabled, both through the clearnet and through Tor. I also had tested it over the weekend through a cellular Internet provider that uses heavy IPv4 sharing on the public side, so that I would use an IPv4 shared with many other customers. When Javascript is disabled Cloudflare displays: Please turn JavaScript on and reload the page.
DDoS protection by CloudFlare
Ray ID: XXXXXXXXXXXXXXXX
The page reloads itself without the need to do anything, in particular site continues to work (in their limited capability) with Javascript disabled. When Javascript is enabled Cloudflare displays: Checking your browser before accessing chain.so.
This process is automatic. Your browser will redirect to your requested content shortly.
Please allow up to 5 seconds…
DDoS protection by CloudFlare
Ray ID: XXXXXXXXXXXXXXXX
and again everything refreshes itself without the need to do anything. So that about sums up the difference between professionals and random code monkeys. Cloudflare's free account probably has some non-obvious limitations, but I see absolutely no reason to use 1net.tech DDoS protection. It would seem particularly effective way to choke the traffic to one's site. |
|
|
|
Maybe you first ask some other guys before posting things like this. This is not an error, you are just not using javascript i guess.
Our browsercheck need javascript to run. Even TOR got javascript installed as default today.
Wrong guess. I have Javascript enabled in all 4 browsers and all of them are correctly passing browser checks and captchas served by the likes of Cloudflare and others. My guess is that you're a beginning programmer/system administrator working alone, without backup of a test team or even a friend that you could bounce your ideas off in a friendly way. That is why you are so angry: hard work and no results, not even the foggiest idea of what is wrong in your setup. |
|
|
|
Yes indeed, you're not goin throu our custombrowsercheck. This can happen if you block services such as javascript or you visit our browser from a ip very often or under other circumstances, we wont fully describe how it works becaus of security purposes.
//I maybe overestimated your English
I didn't get this joke, as i told, english is not my foreign language so please excuse the bad grammar.
Also please don't be disrespectful with no reason.
Didn't get a joke, but managed to get offended? Another red flag for scammer. They are usually humorless but always demand "respect" (the last word pronounced with Marlon Brando's accent in The Godfather). I'm pretty sure this business is like dry gangrene: it is self-limiting by cutting off the blood/traffic flow to the infected/protected site. But typically they would try to restart it under different name to disassociate themselves from the original failure. For the future reference I preserve the "custom browsercheck" in case anyone wants to investigate. I downloaded exactly the same, non-working, page using 4 different browsers and several IPv4 addresses, both through clear-net and through Tor. I don't have time to debug it, I just noticed a hotlink to an image hosted by imgur.com . I just X-ed out the tracking cookie. <!DOCTYPE html><html><head><title>...</title><meta charset="UTF-8"/><script type
="text/javascript">!function(window){function success(e){var a=new Date;a.setTim
e(a.getTime()+26784e5),document.cookie="1net=XXXXXXXXXXXXXXXX-XXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXX-"+e+";path=/;expires="+a.toGMTString(),window.location.re
load()}var HashCat=new function(){this.sha="function SHA1(s){function U(a,b,c){w
hile(0<c--)a.push(b)}function L(a,b){return(a<<b)|(a>>>(32-b))}function P(a,b,c)
{return a^b^c}function A(a,b){var c=(b&0xFFFF)+(a&0xFFFF),d=(b>>>16)+(a>>>16)+(c
>>>16);return((d&0xFFFF)<<16)|(c&0xFFFF)}var B=\"0123456789abcdef\";return(funct
ion(a){var c=[],d=a.length*4,e;for(var i=0;i<d;i++){e=a[i>>2]>>((3-(i%4))*8);c.p
ush(B.charAt((e>>4)&0xF)+B.charAt(e&0xF))}return c.join('')}((function(a,b){var
c,d,e,f,g,h=a.length,v=0x67452301,w=0xefcdab89,x=0x98badcfe,y=0x10325476,z=0xc3d
2e1f0,M=[];U(M,0x5a827999,20);U(M,0x6ed9eba1,20);U(M,0x8f1bbcdc,20);U(M,0xca62c1
d6,20);a[b>>5]|=0x80<<(24-(b%32));a[(((b+65)>>9)<<4)+15]=b;for(var i=0;i<h;i+=16
){c=v;d=w;e=x;f=y;g=z;for(var j=0,O=[];j<80;j++){O[j]=j<16?a[j+i]:L(O[j-3]^O[j-8
]^O[j-14]^O[j-16],1);var k=(function(a,b,c,d,e){var f=(e&0xFFFF)+(a&0xFFFF)+(b&0
xFFFF)+(c&0xFFFF)+(d&0xFFFF),g=(e>>>16)+(a>>>16)+(b>>>16)+(c>>>16)+(d>>>16)+(f>>
>16);return((g&0xFFFF)<<16)|(f&0xFFFF)})(j<20?(function(t,a,b){return(t&a)^(~t&b
)}(d,e,f)):j<40?P(d,e,f):j<60?(function(t,a,b){return(t&a)^(t&b)^(a&b)}(d,e,f)):
P(d,e,f),g,M[j],O[j],L(c,5));g=f;f=e;e=L(d,30);d=c;c=k}v=A(v,c);w=A(w,d);x=A(x,e
);y=A(y,f);z=A(z,g)}return[v,w,x,y,z]}((function(t){var a=[],b=255,c=t.length*8;
for(var i=0;i<c;i+=8){a[i>>5]|=(t.charCodeAt(i/8)&b)<<(24-(i%32))}return a}(s)).
slice(),s.length*8))))}; ",this.worker=this.sha+"onmessage=function(e){var orig=
e.data;while(true){var rand=String(Math.floor(Math.random()*100000))+String(Math
.floor(Math.random()*100000));var hash=SHA1(orig+rand).substr(0,4);if(parseInt(h
ash,16)==0){postMessage(rand);break}}}",this.createWorker=function(e){if("undefi
ned"!=typeof BlobBuilder){var a=new BlobBuilder;a.append(this.worker);var r=wind
ow.URL.createObjectURL(a.getBlob())}else{if("undefined"==typeof Blob)return!1;va
r a=new Blob([this.worker],{type:"text/javascript"}),r=window.URL.createObjectUR
L(a)}var t=new Worker(r);return t.onmessage=e,[r,t]},this.process=function(sessi
on,form){if("undefined"==typeof Worker||"undefined"==typeof BlobBuilder&&"undefi
ned"==typeof Blob)for(eval(this.sha);;){var rand=String(Math.floor(1e5*Math.rand
om()))+String(Math.floor(1e5*Math.random())),hash=SHA1(session+rand).substr(0,4)
;if(0==parseInt(hash,16))return setInterval(function(){success(rand)},10)}else{v
ar data=this.createWorker(function(e){success(e.data),worker.terminate()}),worke
rURL=data[0],worker=data[1];worker.postMessage(session)}}},data="XXXXXXXXXXXXXXX
X-XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX".split("-");setTimeout(function(){Has
hCat.process(data[0])},1e3)}(window);</script><style type="text/css">*{margin: 0
;padding: 0;}html, body{background: #2E2C2A url(https://i.imgur.com/BvHgDTO.png)
;font-family: "Helvetica Neue", "Arial", "Verdana", sans-serif;color: #FFF;font-
size: 130%;}h1, p{font-weight: 200 !important;}a{color: #EEE;font-size: 10pt;}#c
ontent{position: absolute;top: 50%;left: 50%;width: 700px;height: 400px;margin:
-100px 0 0 -350px;text-align: center;}</style></head><body><div id="content"><h1
>Hold on just a second...</h1><p>We're just making sure you're not a robot, give
it a few seconds...<br/><a href="http://1net.tech">1net Anti DDoS Proxy (Beta)<
/a></p></div></body></html>
|
|
|
|
1net Anti DDoS Proxy (Beta) is our service, we guarantee for your security. http://1net.tech/ <- if you visit the link, wich is displayed in the Custombrowsercheck then you will see, it refers to 3st.me We have our own servers, for sure dedicated. (no virtual servers bullshit) And one of the best protections available. I told you that you are clueless code monkey. I maybe overestimated your English. You seem to write clearly, but have comprehension problem when reading. When some visit your sites all they get is a static unmoving webpage. There's no way to see what is behind. Your DDoS is indeed very efficient. Like the proverbial protection against sexually transmitted diseases: - Drink glass of water. - Before or after? - Instead. Yep, THE best protection. |
|
|
|
// We are not using CloudFlare. We got our own firewall and custom browserchecks as you can see whilst your first visit and sometimes when it appears randomly checkin if you are a robot or not.
Don't flatter yourself. You don't know what you are doing technically. I was intrigued about "custom browserchecks". So now I have 3 different browsers open on my Macintosh, each displaying static text white on black: Hold on just a second...
We're just making sure you're not a robot, give it a few seconds...
1net Anti DDoS Proxy (Beta)
with the last line linking to http://1net.tech also displaying exactly the same white on black screen. Looks like clueless code monkey directing the development. But at least writing a very serviceable English. |
|
|
|
LMAO - you have obviously never managed a large scale mining operation, this mine when fully upgraded will house several thousand S9's. Now, to put that in perspective, the heat load and CFM requirements are massive. I can stick a bunch of S9's on a shelf, but the blow back from the hot isle due to positive pressure is a nightmare.
Remember, every open penetration that does not flow through equipment means additional strain on the CFM capacity from your system. It also represents a loss of temperature delta between the cold isle and hot isle. Having lets say, several cubic inches of open space does not seem like much until you multiple that by a thousand. You are now talking about just having 10 cabs completely open wasting conditioned cold air to hot air.
In addition, the S9's do not screw in to the shelf, they sit on it. The only thing screwed in is the PSU so you can easily reboot the system from the cold isle... 95% of the time its not the PSU we have an issue with. We have ~ a 70 degree cold isle supply with a 100+ hot isle exhaust and a 30 degree delta. If you aren't an engineer, and have never worked in a pressurized / or data-center environment you have no idea what you are talking about, this isn't over engineering, its engineering to actually make the system work.
You are 100% correct saying that I've never managed a coin mining operation. I have completely different background. In particular I'm not a mechanical engineer, but I have worked with them. I also don't normally do "operations," but in my company there's a team that specializes in acquiring bankrupt data processing operations, so I've seen quite a few and helped restart them under new management. I didn't made myself clear in my previous posts. I don't consider your design over-engineering, it is something akin to mis-engineering or show-off-engineering. It is for designed for "lookists" and quick sales, not for the cost-effective operations. From the mechanical point of view, you've only partially addressed the air flow and the MTTR issue. Your design seems to put the whole weight of the power supplies on the edge of the sheet metal. The better design would be to place the weight of the inside devices on a sort of ledge that is at least 1cm wide and lined with soft rubber. (Sorry, I'm not an ME, I lack proper vocabulary.) Basically your design seriously ignores the vibration and resonances caused by the high speed air flow. The resonances are particularly important when nearly the whole facility will be filled out with exactly identical equipment. Your design also wastes a lot of sheet metal for the bottom of the case (and maybe sides). The Bitmain miners are already quite similar to the Helmholtz resonator ( https://en.wikipedia.org/wiki/Helmholtz_resonance). The multi-mega-watts facility will be like a mad-scientist punk-rock one-note-organ or maybe one-note-straight-flute ensamble. People already are complaining that they are loud, the vibration of the whole facility will badly affect it mechanically. The other MTTR issue is with the facility shown in the photo. The very long aisles, over 20 racks long, may look impressive on the photo. They are very detrimental to proper operations: it requires multi-person crews for maintenance. If people working alone are forced to work in it they will either skip important work steps or will spend more time walking around than productively working. I've seen what an experienced water-damage contractor could improvise in hours to direct dry-air flow around the equipment to baffle and cool it. The corrugated cardboard + plastic foam + particle board can direct the air around metal warehouse shelving no worse than dedicated metal racks and cases. It doesn't look good, but operates well, is much cheaper and has no lead time for ordering. I wouldn't advocate that for permanent facility, but what is going to be the expected operational lifetime of the particular models of coin miners? |
|
|
|
|
Show Posts
