1,746
edits
m (1) add note linking to a benchmark site and info about GPUs under NVIDIA section. Cleanup) |
mNo edit summary |
||
| Line 105: | Line 105: | ||
== NVIDIA == | == NVIDIA == | ||
''April 2021 note:'' Check this a GPU benchmark list like this for a '''rough''' idea of performance: [https://benchmarks.ul.com/compare/best-gpus Direct link] ([http://web.archive.org/web/20210412095421/https://benchmarks.ul.com/compare/best-gpus April 2021 archive]). If you are selecting a laptop find out what your GPU is - ask or check the manufacturers website page for the model of the laptop. Laptop GPUs: Laptops with NVIDIA 9xx series GPUs often have the mobile "M" variant. This is a lot slower than the normal version e.g. a GTX 960M is slower than a GTX 960. NVIDIA10xx series and later use desktop equivalents in all laptops or close to it. There are also Q-MAX variants that are only slightly slower than normal versions - e.g. a GTX 1060 Q-MAX is only slightly slower than a GTX 1060 - see benchmarks. | ''April 2021 note:'' Check this a GPU benchmark list like this for a '''rough''' idea of performance: [https://benchmarks.ul.com/compare/best-gpus Direct link] ([http://web.archive.org/web/20210412095421/https://benchmarks.ul.com/compare/best-gpus April 2021 archive]). If you are selecting a laptop find out what your GPU is - ask or check the manufacturers website page for the model of the laptop. Laptop GPUs: Laptops with NVIDIA 9xx series GPUs often have the mobile "M" variant. This is a lot slower than the normal version e.g. a GTX 960M is slower than a GTX 960. NVIDIA10xx series and later use desktop equivalents in all laptops or close to it. There are also Q-MAX variants that are only slightly slower than normal versions - e.g. a GTX 1060 Q-MAX is only slightly slower than a GTX 1060 - see benchmarks. (April 2021) | ||
Stay away from nVidia GPUs with a low second digit (x20, x40). Higher 2nd digit means more CUDA cores. | Stay away from nVidia GPUs with a low second digit (x20, x40). Higher 2nd digit means more CUDA cores. | ||
edits