I really do like the idea of Vertcoin. Still, two things come to my mind:
1.) Claiming something is "ASIC proof". No algorithm is that. You can always develop an ASIC. GPUs are ASICs, after all. Take the design of a GPU, strip everything not needed for Vertcoin mining and that's your VTC-ASIC. Still, of course, that ASIC would be way more complex than any SHA/BTC ASIC and would require loads of fast memory, thus making it very expensive (compared to SHA and even scrypt/LTC ASICS). But this immediately leads to my second point.
2.) How bad are ASICs after all? With SHA, there is a low entry level for developing an ASIC. With VTC, this "entry barrier" would be much higher. Imagine VTC would be as popular as BTC is now. I'm going to guarantee you that at least one person/group would be into developing a VTC ASIC. But that's the problem: Only few entities would have the funds to do so, much fewer than with Bitcoin. After all, this will lead to more centralization, not less.
I don't like that "hobby mining" is gone in the Bitcoin world, I don't like how much current ASICs are overpriced.
But I still think the Bitcoin network is better off with all the ASICs than the VTC network would be.
No worries. The coin is software. ASICs are hardware. A simple algorithm tweak proposed by the VTC developers and agreed upon by the community will shake off an ASIC. Even a change to N-factor schedule would wreak havoc on ASIC makers' business plans. As long as Vertans say ASICs are not welcome here, we'll be able to preserve the 'hobby mining' angle for a good while. Edit: BTW I don't like the asics at all, mostly because of the incredible disruption they caused. There are not many companies that can make asics this complex, and when they do have working units they extensivly 'test' them before they ship them out. That's what I think BFL and such did. HinnomTX, ShadesOfMarble does have a point though, IMO. Asics (and FPGAs) are sort of in between hardware and software. The GPU (the core itself) is basically an ASIC. With lots of functionality that's not required for mining. So if you could strip the unneeded parts from a graphicscard's GPU you could create an mining-specific PCB with a higher efficiency and probably lower power requirements. And you could change the N-factor just as easily. I think that would would require an incredible amount of engineering though, you'd be doing AMD's and Nvidia's work basically. To me that doesn't sound feasible. Unless AMD and Nvidia themselves would start producing those miners. Then again, the downside of anything application specific is that they're completely useless if this whole *coin-world collapses. I don't think that will happen, but for a company this is probably a big risk. This is my whole point. If there is enough money in it, some entity will do it. But only companies with a very good funding will be able to do it. Fast forward to 2017, everybody is using Vertcoin now because they thought it will free the cryptocurrency world from "those ASIC companies". But, boom! AMD just developed an ASIC miner, because their GPUs are really good at mining anyway, so they just removed everything not needed for mining, no video output etc, high power VRMs for insane clock speeds, loads of high-speed memory (because they are such a big company and buy tons of it they get it cheaper than anyone else), so they run 10-100x more efficient (both MH/s/J and USD/(Mh/s)) than the GPUs they sell. So now all the hashing power basically lays in the hand of one company. Because the design of the ASIC is very complex, no other company competes with AMD. What now? With SHA, you could even go the "ultra cheap" route and do a hardcopy ASIC with existing (open source) HDL code. How many SHA ASICs do we have? A dozen? They all compete and are at least an oligopoly. Anyway: The statement "No more ASICs" ist just not true. It should say "Currently no ASICs".The thing is, if the community doesn't want ASICS, they will never mine successfully for long. And here we go : https://bitcointalk.org/index.php?topic=421921.msg7087970#msg7087970 |
|
|
|
|
Lass es bleiben. Selbst wenn du Strom und Hardware geschenkt bekommst, ist die Abnutzung der Hardware größer als deine Einnahmen.
|
|
|
|
which for SP10 means getting rid of the cases Simply removing the case will destroy the airflow completely. |
|
|
|
There are blade design differences bearing differences motor differences etc. There are 40mm fans ranging from 10 dB to 100 dB. Again if someone wants to bother and look at what fan and what specs. they currently are, they can then source their own "better" or "quieter" equivalent. A fan spinning at 6000~9000 rpm can never be quiet, no matter the bearing or blade design. It's impossible. At this level, you don't hear any noise from the ball bearing, you just hear the airflow. At this fan speeds, you will NEVER achieve laminar flow. The bearing does matter if you look for fans spinning at 300~500 rpm. Then you even prefer a sleeve bearing over ball bearing, because sleeve bearing is more quiet. Again, at this level you only hear the mechanics of the fan, not the actual airflow. But for fans with many thousand rpm, ball bearing is the only option and turbulent airflow will always be noisy. If you want a quiet SP10, go for watercooling. |
|
|
|
To put it bluntly these fans are probably the cheapest available to do the job. "Loud" does not mean "cheap", I think that's a common misconception  |
|
|
|
Having gone through this whole BFL and Avalon (and partially BitFury) mess I takes nothing as granted  |
|
|
|
Web UI screenshots can be found here: http://imgur.com/a/H3dee - Nice, clean and functional. The SP10 is now running for over two weeks without the slightest problem. It's just delivering rockstable 1,4 TH/s.If I compare that to my BitFury units, which cry for attention all the time.... |
|
|
|
What ambient temperature would Sp-10 tolerate? Up to 35°C |
|
|
|
It is impressive but it's a linear scale which always looks impressive if you plot exponential growth  http://bitcoin.sipa.be/speed-ever.pngThis carries much more information and is no less impressive. (Straight line with exponential axis = exponential growth) |
|
|
|
there is so much scope for profit outside of mining bitcoin
Like cracking passwords which use sha(sha()) as hashing function? |
|
|
|
This is the future but I am not sure if mineral oil will be the mass deployment coolant. It is still pretty messy to service. I am working on a project with another coolant that is a little less messy. Regardless this is cool and green tech. I love it.
It should be possible to deploy immersion cooling at large scale, as shown here: https://bitcointalk.org/index.php?topic=346134.0 |
|
|
|
Could we get a video of the fan noise. Start video, turn unit own, struggle to hear you talk over noise That should be possible -- I now have some more numbers to present: Mode (Web UI) | Performance | In/Out Temp | Power (Wall)
------------------------------------------------+-------------+---------------+---------------
Idle | 0 Gh/s | N/A | 35 W
1.00Th / ~720W / ~quiet / 1,39 GH/s/W | 1139 Gh/s | 29 °C / 67 °C | 780- 820 W
1.35Th / ~1100W / ~quiet / 1,23 GH/s/W | 1341 Gh/s | 29 °C / 82 °C | 1150-1200 W
1.43Th / ~1350W / normal / 1,06 GH/s/W | 1412 Gh/s | 29 °C / 79 °C | 1340-1380 W
1.47Th / ~1370W / turbo fans / 1,07 GH/s/W | 1456 Gh/s | 29 °C / 76 °C | 1380 W
------------------------------------------------+-------------+---------------+---------------
The power measurements were done on a 230 V circuit. The numbers for every mode except "turbo" and "idle" fluctuated, so I included the lowest and highest reading. The "average" power draw is somewhere in between. In normal mode, the unit hits the advertised 1.4 TH/s exactly. "Turbo" mode does not give much more performance (about 3,1%), but decreases the temperature by 3°C. The lowest performance mode also gives the lowest temperature, whereas the other quiet mode gives the highest temperature (3°C more than "normal") and a 5% reduction in performance compared to the standard mode. So, as the "quiet" mode does not deserve its name (I would rather call the mode "efficient", as that is what is is - more energy-efficient than standard/turbo) - my recommendation is to either run this unit using the (slowest) 1Th-setting, which will give the best performance per Watt and also a low temperature. Otherwise, one should choose "turbo" mode over "normal" mode because you get both a lower temperature and a bit better performance at only marginal increased power consumption. |
|
|
|
Hey guys, it would be nice if you could keep this all time favourite discussion "mining vs. buying BTC" out of this thread I would like to have this thread dedicated to my SP10 review and questions regarding this unit/review. (I'm not saying either option is better. Everyone has to do due diligence. Then, buy mining gear or buy BTC or buy nothing.) |
|
|
|
|
This is not going to reduce the loudness by a single dB, sadly.
|
|
|
|
I have a fuw idea's
Care to share? |
|
|
|
Seriously, though, I'd still see if a cross flow fan setup would work.
How would that look like? Currently, cool air is sucked in at the front and blown out of the case at the back by a battery of fans (see http://imgur.com/a/kASCu#2 ). So the air is moving in a "straight line", I really don't believe it does get any more efficient than that. Also, the PSU adds some noise. Having such a small form factor (it's really small considering it delivers up to 1.4 TH/s) just limits what you can do in terms of cooling. |
|
|
|
So I take it is loud. LOL. After reading through the comments other people, who received their units earlier, have made, I was expecting something, but my expectations were greatly exceeded why simple put better quality fan
Well, this is a 1U device, and power consumption is above 1 kW. Looking at the webinterface, there is a temperature reading: Temp Front / Back 28 °C / 82 °C I have not opened the unit yet (will do that, but there are already pictures of an open SP10), but I guess the thermal management is well-engineered. So I don't see any way to make this unit more quiet, at least not using air cooling. You have to dissipate the heat somehow. The problem isn't that the fans are of bad quality, but that they are spinning at (estimated) 4000-6000 rpm. Even if the fans had loud/bad bearing, air movement will always be magnitudes louder at this level. Even if Spondoolies-Tech manage to build in a better (read: more quiet) fan management, this unit will never be "quiet". |
|
|
|
Initial post:The SP10 offers the following modes of operation (first three columns taken from the webinterface, last column added by me): 1.00Th / ~720W / ~quiet / 1,39 GH/s/W
1.35Th / ~1100W / ~quiet / 1,23 GH/s/W
1.43Th / ~1350W / normal / 1,06 GH/s/W
1.47Th / ~1370W / turbo fans / 1,07 GH/s/W
As you can see, the slowest (or "underclocked") mode is most efficient, whereas the two fastest modes offer about 20% less performance per Watt. Update 24.04.14: Mode (Web UI) | Performance | In/Out Temp | Power (Wall)
------------------------------------------------+-------------+---------------+---------------
Idle | 0 Gh/s | N/A | 35 W
1.00Th / ~720W / ~quiet / 1,39 GH/s/W | 1139 Gh/s | 29 °C / 67 °C | 780- 820 W
1.35Th / ~1100W / ~quiet / 1,23 GH/s/W | 1341 Gh/s | 29 °C / 82 °C | 1150-1200 W
1.43Th / ~1350W / normal / 1,06 GH/s/W | 1412 Gh/s | 29 °C / 79 °C | 1340-1380 W
1.47Th / ~1370W / turbo fans / 1,07 GH/s/W | 1456 Gh/s | 29 °C / 76 °C | 1380 W
------------------------------------------------+-------------+---------------+---------------
The power measurements were done on a 230 V circuit. The numbers for every mode except "turbo" and "idle" fluctuated, so I included the lowest and highest reading. The "average" power draw is somewhere in between. In normal mode, the unit hits the advertised 1.4 TH/s exactly. "Turbo" mode does not give much more performance (about 3,1%), but decreases the temperature by 3°C. The lowest performance mode also gives the lowest temperature, whereas the other quiet mode gives the highest temperature (3°C more than "normal") and a 5% reduction in performance compared to the standard mode. So, as the "quiet" mode does not deserve its name (I would rather call the mode "efficient", as that is what is is - more energy-efficient than standard/turbo) - my recommendation is to either run this unit using the (slowest) 1Th-setting, which will give the best performance per Watt and also a low temperature. Otherwise, one should choose "turbo" mode over "normal" mode because you get both a lower temperature and a bit better performance at only marginal increased power consumption. Update 26.04.14:More pictures: http://imgur.com/a/afpMF#0
|
|
|
|
|
Show Posts
