|
Title: scripting language-dependent hashrate - C++ faster in theory than Python?(polcb) Post by: jgm_coin on April 12, 2013, 02:14:45 AM I'm an interdisciplinary grad student recent exposed to coding by random placement with a CS major in grad housing and an interest in writing [R] for scientific programming.
I've always learned that C is the fastest (albeit less readabiliable [play on accountabilybuddy]). So should CGminer be fundamentally faster than the python-based mining? or is most of the compute carried elsewhere? Discuss. Title: Re: scripting language-dependent hashrate - C++ faster in theory than Python?(polcb) Post by: crazyates on April 12, 2013, 03:53:26 AM One of the advantages of CGMiner is that it's written in C. This isn't just a theoretical advantage, as it uses very little CPU, even when running many GH/s. Avalon uses a modified version of CGMiner for their products, and it runs over 65GH/s on a tiny little chip with barely 64MB of RAM.
Title: Re: scripting language-dependent hashrate - C++ faster in theory than Python?(polcb) Post by: Marrs on April 12, 2013, 04:45:55 PM I'm an interdisciplinary grad student recent exposed to coding by random placement with a CS major in grad housing and an interest in writing [R] for scientific programming. I've always learned that C is the fastest (albeit less readabiliable [play on accountabilybuddy]). So should CGminer be fundamentally faster than the python-based mining? or is most of the compute carried elsewhere? Discuss. There isn't much to discuss. A well-optimized C based CPU miner should be faster than a Python based CPU miner. However, CPU mining is mostly obsolete (due to reward being lower than power cost). With GPUs or FPGAs or ASICs, all of the Real Work is being done elsewhere, and the C or Python (or Java in the case of Bitminter's custom client) miner code is just feeding work to the hardware that does the calcs and relaying results back to the pool, so there's pretty much no difference in hashing speed. |