We would need design and production of FPGA or ASIC units optimized to run just a specific OpenCL program as fast as possible.
This OpenCL program calculates cryptographic hash functions and has a benchmark report that displays how many hash calculations are made per second.
Your FPGA/ASIC should calculate at least 20 billion hashes per second while executing our OpenCL kernel, and should consume less than 1500W.
The software in question is here: [url removed, login to view]
and the OpenCL kernel that needs to be executed in hardware by your device is the /zcash/gpu/[url removed, login to view] source file.
We would prefer that you provide OpenCL drivers for Windows 32-bit , but we can discuss the possibility to switch to Linux.
The device interface can be USB or PCI-e or other , there are no constraints for that.
Important: Our budget will vary depending on the actual performance of your device.
Benchmark is measured by the application itself and verified by a server, and we will test together the actual performance achieved before releasing our payments.
Errata corridge: "20 billion hashes per seconds" was wrong. The requirement is 50000 hashes calculations per second
Alternative: if the task described above is hard to accomplish, maybe you can do the following task which may be easier:
Implement a powerful FPGA/ASIC board running the software described in this board: https://forum.decred.org/threads/fpga-bitstreams.827/page-2 which uses an FPGA bitstream that can be found here: https://github.com/sprocket-fpga/Decred_LX9
What we expect is:
- you should provide me with that same software but running on Windows
- you should make an FPGA/ASIC board (and drivers) that calculates 100+ GH/s with that software ( 100 billion hashes/second), while consuming less than 1500W
The bounty we will release , as said above, depends on the actual performance obtained.
5 freelancers are bidding on average €4433 for this job
I am a reverse engineer. I know about assmembly and c++. I had developed ethereum miner using opencl. If you trust me, i will do my best. best regard.