Not a truly "great" deal, but it's the lowest I've seen from a US based seller recently. I paid $175 after shipping and taxes.
Nvidia Tesla P40 24gb for $149.99 +sht
- Thread starter bwahaha
- Start date
Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.
Not a truly "great" deal, but it's the lowest I've seen from a US based seller recently. I paid $175 after shipping and taxes.
They meet the VRAM requirements to run open models and will be slow, for any serious ML work you'd want a 3090 imo
The P100 is probably the only one of that generation worth getting at today's prices. The real place it shines is float 64 workloads in scientific computing- consumer cards simply take too much of a performance hit where the P100 is only 1/2 the 32bit.
Not just that, the P100 has FP16 which GP102/the P40 doesn't. Only 16GB VRAM though
it's the most VRAM per dollar right now, they're somewhat popular for running LLMs locally
I dunno about you but I'm not buying anything with sht in the box.
Can this do transcoding for emby/jellyfin/plex media streamers? Or encoding to H.264/H.265/AV1?
nVidia Hardware Transcoding Calculator for Plex Estimates
NVENC - Wikipedia
en.wikipedia.org
Last edited:
Video Encode and Decode Support Matrix
Find the related video encoding and decoding support for all NVIDIA GPU products.
Thats very power hungry transcoding card. (It ain't worth it.)
(p100 and p40)
(here's some comparison last 2 on the bottom of the list) (2080 is faster in ai)
P40 & P100 can only interest you if you plan on playing with nv vgpu (but if you have commercial or performance requirements in vgpu ~> go with A5000 instead)
(p100 and p40)
(here's some comparison last 2 on the bottom of the list) (2080 is faster in ai)
P40 & P100 can only interest you if you plan on playing with nv vgpu (but if you have commercial or performance requirements in vgpu ~> go with A5000 instead)
Last edited:
Agreed, but if you need something for more than just transcoding it fills a niche for the current price.
I just bought one from that listing. Gonna run some local models and maybe transcode some stuff. I also see people running two. Seems you want the vram to actually run the llms. Make sure you have power sorted. The EPS 12 connector makes things complicated.