Stories
Slash Boxes
Comments

SoylentNews is people

posted by on Tuesday June 06 2017, @10:38AM   Printer-friendly
from the so-many-triangles-wasted dept.

AMD Radeon RX 500 series graphics cards, particularly the RX 580 and 570, have been out of stock for weeks now owing to the cryptocurrency mining craze. The market had shifted away from GPU mining a couple of years back after several China based companies launched specialized ASICs that were much faster and more power efficient at resolving the block chain equations necessary to mine Bitcoin and Litecoin, the Gold and Silver of cryptocurrencies.

However GPU mining has seen a massive resurgence over the past little while due to the rising popularity of ASIC resistent coins. Chief among which is Etherium which has seen its price more than triple in a matter of months. AMD's Add-In-Board partners have caught on to this and have already started directly advertising to miners.


Original Submission

 
This discussion has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
  • (Score: 0) by Anonymous Coward on Tuesday June 06 2017, @11:00PM

    by Anonymous Coward on Tuesday June 06 2017, @11:00PM (#521639)

    Moreover, there is no such thing as "resistance" to ASICs; sure, the algorithm used by Ethereum means that the dedicated ASIC hardware is more expensive to make than similar hardware for Bitcoin, but guess what? That just means that only the large, deep-pocketed organizations will be able to afford such hardware—and such hardware is generally orders of magnitude more performant than GPU-based systems.

    Well, not exactly. It is not necessarily the case that you can throw money at an ASIC designer to get something that gives orders of magnitude more performance than, say, an off-the-shelf programmable component. Getting good performance with an ASIC requires good utilization of the physical resources on your chip, but also you need to keep your signal paths short otherwise you can't run it very fast. This normally means pipelining and not all problems are amenable to this.

    Depending on the problem your super-optimized custom ASIC might just end up looking a lot like a bunch of microcontrollers thrown on a chip with some memory attached. Which starts to look a lot like a GPU. At which point you will probably lose to the guy who decided not to spend any money on ASIC development and instead used that money to buy more off-the-shelf GPUs.