Title: estimating/calculating hash rates for FPGA chips Post by: CanaryInTheMine on August 03, 2011, 01:28:27 AM I seem to remember seeing "back of the napkin" calculations regarding arriving at what a hashing rate would be of a particular fpga/chip. but I can't find that post anymore...
Can someone point me in the right direction/location for this please? Title: Re: estimating/calculating hash rates for FPGA chips Post by: fpgaminer on August 03, 2011, 09:09:20 AM https://en.bitcoin.it/wiki/Mining_hardware_comparison#FPGA_Devices (https://en.bitcoin.it/wiki/Mining_hardware_comparison#FPGA_Devices)
Title: Re: estimating/calculating hash rates for FPGA chips Post by: CanaryInTheMine on August 03, 2011, 03:04:21 PM https://en.bitcoin.it/wiki/Mining_hardware_comparison#FPGA_Devices (https://en.bitcoin.it/wiki/Mining_hardware_comparison#FPGA_Devices) :) thanks! I'm very familiar with this page I was thinking along these lines: if I take x number of some fpga chips, assume a bunch of other things, here's how to do a rough estimate on the expected hashing rate.... That's what I'm after. I can't find that post... Title: Re: estimating/calculating hash rates for FPGA chips Post by: fpgaminer on August 04, 2011, 01:12:59 AM Quote if I take x number of some fpga chips, assume a bunch of other things, here's how to do a rough estimate on the expected hashing rate.... That's what I'm after. I can't find that post... It's easiest to do with Altera Cyclone chips, and roughly works out to:1,000 LEs = 1MH/s Could be more in some cases, less in others, but that's a fairly good baseline. For Spartan-6 it's roughly: 1,000 LEs = 0.5MH/s For high-performance families like Virtex or Stratix you should multiply by 3 or 4. |