Posts: 8
Threads: 2
Joined: Oct 2016
10-26-2016, 12:38 AM
(This post was last modified: 10-26-2016, 12:42 AM by demonachizer.)
I can't find any good hashcat benchmarks for the R9 Nano so I am trying to decide which I would be better off with. I have a power limit that allows me to use a single P100 or dual R9 Nanos and would like to make the right choice.
I am inclined to think the P100 is obviously the better but I am not sure if just comparing SPFP or FP64 TFLOPS is the way to go for this application.
Any ideas?
It is possible that I might be able to squeeze in two P100s on the system and then the answer is obvious (I think) but I think that there might not be much wattage overhead.
EDIT: I know that there might be better approaches for the same cost or same power etc. but I only have the choice between the two and they are being provided gratis so I have to stick with one or the other.
Posts: 413
Threads: 2
Joined: Dec 2015
Why would you be buying a P100? For something other than hashcat? Also the common suggestion has been "don't buy AMD for a new rig". Buying a 1080 or maybe a few 1080's would be best as far as I'm aware.
Posts: 8
Threads: 2
Joined: Oct 2016
10-26-2016, 01:10 AM
(This post was last modified: 10-26-2016, 01:10 AM by demonachizer.)
(10-26-2016, 12:42 AM)Chick3nman Wrote: Why would you be buying a P100? For something other than hashcat? Also the common suggestion has been "don't buy AMD for a new rig". Buying a 1080 or maybe a few 1080's would be best as far as I'm aware.
I am not purchasing any of this hardware it is being loaned to me gratis. My choice unfortunately is either of them not anything else. I can either use CPU or choose from those two cards. It is a weird situation I know but unfortunately those are the parameters I must work with.
Posts: 413
Threads: 2
Joined: Dec 2015
I'm inclined to say that the P100 is the better option as it looks like it will perform a fair bit better by looking at its raw specs however I'm not sure it's 100% supported and I haven't seen anyone post benchmarks for it yet.
Posts: 2,936
Threads: 12
Joined: May 2012
The P100 is a much better option, and yes it definitely should be supported.
Posts: 8
Threads: 2
Joined: Oct 2016
Does it make any difference that the P100 can run at half precision at 18.7 TFLOPS?
Posts: 347
Threads: 3
Joined: May 2010
10-26-2016, 11:37 AM
(This post was last modified: 10-26-2016, 11:38 AM by Xanadrel.)
wikipedia Wrote:In computing, FLOPS or flops (an acronym for FLoating-point Operations Per Second) is a measure of computer performance, useful in fields of scientific calculations that make heavy use of floating-point calculations.
Guess what, hashing algos are integer-based, so FLOPS means close to nothing here.
Posts: 8
Threads: 2
Joined: Oct 2016
10-26-2016, 04:47 PM
(This post was last modified: 10-26-2016, 04:47 PM by demonachizer.)
This was my inclination initially but I saw a bunch of talk about the SPFP of consumer cards on these forums. Thanks for the clarification.
