Is that thing rack mounted? Does it make more than it uses in electricity?
Yes the case is primarily rack mounted with bolt on rails but the front handles also unbolt for a desktop or tower like usage.
In theory it should generate about 3 times the cost of electricity (not including hardware investment) but I've yet to prove that. I was hoping for 120 MH/s but they aren't performing quite as predicted but may improve with driver updates etc.
As Ethereum has come back down recently ($20 > $12) that margin has lessened and I guess there may only be a few months left of profitability for small miners. Assuming you're just mining to convert to money straight away.
Any plans for the system once you stop mining?
Assuming it survives :) I bought X99 bits rather than go as cheap as poss as my main gaming system is also X99 so may be a good source of parts. I did get lucky on the Gigabyte Gaming 5P mobo though as it's the same as my newly bought one but 2nd hand via Ebay. Similar for the Xeon engineering 8/16 core CPU. Much less than retail yet plenty powerful enough and very low power. Also has AES NI instructions so can also mine minor alt coins at the same time which may contribute to the running costs.
The headline parts of course are the GPUs so unless some other alt coin comes along to mine I'll have 3 fairly new GPUs to sell on and one can go in my race sim PC to replace the aging GTX 670. Though that's still good enough at the mo.
I don't think I'll sell the case though as it's very versatile. Really good airflow and surprisingly quiet for what's in it.
I think my motherboard will be fine...
https://www.asus.com/Motherboards/Z97WS/
...even with two of them :P
Update OP. Please try to stay on topic guys.
Alright, so basically, I've become an electrical engineer overnight reading every fucking resource imaginable and here's my conclusion.
Is it out of spec? yes
will it kill your motherboard? no
will two kill your motherboard? unlikely
will two heavily overclocked kill your motherboard? if you motherboard is a potato maybe but still unlikely.
The motherboard is responsible for limiting how much power the card can draw from it, if the motherboard allows it to draw more it's either within that motherboards capabilities or it's a piece of incorrectly designed shit and that's on the motherboard manufacturer. The card is drawing out of spec because the motherboard gives it the go ahead. All claims of dead motherboards thus far have remained unproven additonally SIG gave this card the go ahead. This is a mountain out of a mole hill
5 Likes
I agree with you. However this is a legitimate concern, even though, I do agree that people are making it worse than it seems, but people are reporting some issues. However this pretty normal for a GPU and honestly I'd give it a month or two for everything to settle down to see if it's just older/crap motherboards giving out or if it's something more serious. Thanks for doing the research. Would you mind adding links so I can put this in the OP?
Base specification
http://composter.com.ua/documents/PCI_Express_Base_Specification_Revision_3.0.pdf
PC perspective testing
http://www.pcper.com/reviews/Graphics-Cards/Power-Consumption-Concerns-Radeon-RX-480
Jay and Barnecles talk about it briefly in the beginning of this episode. (Jay barely seems to know more than we do so take it with some salt)
AMD reddit (trying to find specific posts will update once/if located)
https://www.reddit.com/r/Amd/?count=26&before=t3_4qtl5t
Overclock.net thread (lots of valuable info here, one of the posts even goes on to explain that the sum of the ratings of each pin in the connector adds up to like 150w)
http://www.overclock.net/t/1604477/reddit-rx-480-fails-pci-e-specification
Really not a fan of this video. Seems very dismissive with little evidence for his position short of anecdotes and I think PcPer pretty much debunked the claim of drawing a parallel to the 960. I like Adored but he seems to be in full damage control mode.
While I don't think we should jump to conclusions this should be investigated further. Especially since someone today posted, with proof, that their PCIE slot died using a 480.
1 Like
He does make some valuable points though about pci-e only cards drawing a lot of power.
I hope very much that this is just a problem with reference cards, and a small one at that. Otherwise AMD could sink even further.
Although this does make me wonder, what will the 490 and whatever card above that do? Draw more? Less maybe? Could the board partners just make theirs with more power cables and negate this entirely?
Again though the 480 does it on stock, those don't, and as the PcPer article established that the average matters more. Momentary spikes aren't as big of a deal.
Eitherway it did seem very dismissive. Especially when he brought up reddit.
He has a point though, reddit is full of half informed keyboard warriors... I just think the video is worth taking into consideration along side other sources, not necessarily by itself.
This could probably fixed with a BIOS or driver update to shift more load to the PCIE 6 pin. An 8 pin would solve this problem entirely.
490 won't have this issue because it is an entirely different board design and architecture.
1 Like
Very true but dismissing all of them out of hand when many post very good evidence is not a great move.
Agreed we need multiple opinions and investigation. Although after I watched his video about power supplies I greatly question what he actually knows....
Oh good. So me getting an AMD card isn't entirely out of the realm of possibility.
I really want to get a top end card this year, along with a nice 1440p monitor.
We could ask @Fouquin to overclock the shit out of the 480 in the shittiest motherboard he has to see what happens, for science of course.
3 Likes
No he doesn't... PCper said so...
What?
So Tomshardware is wrong in their reviews, because PCPer says so? PCPer, the guys, that said "Get Nvidia for future proofing"... I don't like PCPer... you may have caught on that.
I would be interrested to see more, but this is just another point of view, and as much as i love AMD if the fucked up, it is fine to be punished.
I haven't seen that... May i see that? Could be true, but this seems pretty weird already. For 2 days 10-15 extra watts of power have destroyed his PCI-E slot.