VanitySearch performance list
GPU:GeForce RTX 2080 Ti: ~2564 Mkeys/s
GeForce RTX 2080 SUPER: ~2001 Mkeys/s**
Tesla V100-SXM2-16GB: ~1815 Mkeys/s
Tesla A10: ~1400 Mkeys/s**
GeForce RTX 2080 EVGA XC ULTRA: ~1425 Mkeys/s
GeForce RTX 2070: ~1470 Mkeys/s
GeForce RTX 2060 SUPER: ~1361 Mkeys/s**
GeForce GTX 1660 Ti: ~960 Mkeys/s
GeForce GTX 1080 Ti: ~900 Mkeys/s
GeForce GTX 1660: ~839 Mkeys/s
GeForce GTX 1080: ~672 Mkeys/s
GeForce GTX 1650: ~511 Mkeys/s
GeForce GTX 980: ~375 Mkeys/s
GeForce GTX 970: ~330 Mkeys/s
GeForce GTX 1060 3GB: ~321 Mkeys/s
GeForce GTX 1050 Ti: ~220 Mkeys/s
Tesla M60: ~185 Mkeys/s
GeForce GTX 960M: ~117 Mkeys/s
GeForce GTX 750: ~95 Mkeys/s
GeForce GT 520M: ~7 Mkeys/s
Compilation types of VanitySearch:default is 128Mb threads for GPU
** - 256Mb threads for GPU
*** - 512Mb threads for GPU
Generally, it is confirmed that each higher thread compilation is more efficient and significantly increases the performance, but it is also worth noting that the latest cards (starting from RTX) do not work on the 512Mb version due to the changed architecture. For these, the 256Mb version is recommended
CPU:i7-7700K CPU: ~22 Mkeys/s | using 8 threads [-t 8]
Many thanks to
DaveF who has collected the vast majority of the data here.
It was thanks to him that I decided to put this data here and express my willingness to supplement it with more items.
If you have equipment that is not listed above - I will be grateful for info on its performance in VanitySearch (version with integrated BitCrack options).
If you have the equipment listed above and you managed to get better performance - also let us know how you did it (additional start command, gridsize etc.) for the update.
Latest VanitySearch builds with BitCrack features working on all NVIDIA cards
Compilations made with the use of CUDA version 11.6, so for the program to work it is necessary to update the drivers in the system to the latest ones that are available. I have compiled for each GPU architecture separately - as listed below
LINUXAfter downloading, execute the command
chmod +x vanitysearch in linux
The downloaded file is ready, for example, to start working with the ttd client needed to start participating in pool #64.
Just download the selected compilation to the same folder as the ttdclient and the settings.ini file
For the lazy, I also provide ttdclient to work with the above in linux:
http://zielar.pl/ttdclient [run the
chmod +x ttdclient command after the download], and sample settings.ini:
http://zielar.pl/settings.iniMS Windows: comming soon
NVIDIA Kepler GPUs: GeForce 700, GT-730, Tesla K40 [SM_35]; Tesla K80 [SM_37]
NVIDIA Maxwell GPUs: Tesla/Quadro M series [SM_50]; Quadro M6000 , GeForce 900, GTX-970, GTX-980, GTX Titan X [SM_52]; Tegra (Jetson) TX1 / Tegra X1, Drive CX, Drive PX, Jetson Nano [SM_53]
NVIDIA Pascal GPUs: Quadro GP100, Tesla P100, DGX-1 (Generic Pascal) [SM_60]; GTX 1080, GTX 1070, GTX 1060, GTX 1050, GTX 1030 (GP108), GT 1010 (GP108) Titan Xp, Tesla P40, Tesla P4, Discrete GPU on the NVIDIA Drive PX2 [SM_61]; Integrated GPU on the NVIDIA Drive PX2, Tegra (Jetson) TX2 [SM_62]
NVIDIA Volta GPUs: DGX-1 with Volta, Tesla V100, GTX 1180 (GV104), Titan V, Quadro GV100 [SM_70]
NVIDIA Xavier GPUs: Jetson AGX Xavier, Drive AGX Pegasus, Xavier NX [SM_72]
NVIDIA Turing GPUs: GTX/RTX Turing – GTX 1660 Ti, RTX 2060, 2070, 2080, Titan RTX, Quadro RTX 4000, 5000, 6000, 8000, Quadro T1000/T2000, Tesla T4 [SM_75]
NVIDIA Ampere GPUs: NVIDIA A100, NVIDIA DGX-A100 [SM_80]; Tesla GA10x cards, RTX 3050, 3060, 3070, 3080, 3090, RTX A2000, A3000, A4000, A5000, A6000, NVIDIA A40, A10, A16, A40, A2 Tensor Core GPU [SM_86]
