This collaborative space allows users to contribute additional information, tips, and insights to enhance the original deal post. Feel free to share your knowledge and help fellow shoppers make informed decisions.
70W - can power it easily right thru PCI-E, no external power required.
Low bandwidth 288GB/s or so - not terrible, similar to current AMD Ryzen AI chip .
16GB is nice
CUDA core count is low, I think it was in the 4000'?
Kind of low powered but nice for a companion card to the main card or a super low power system that you want to add some local AI capabilties
If you got the room find a deal for 5080 ($1000 possible) or def 5070 TI as you will run much much faster.
comparison table in the picture below. You'd get this for niche slim profile and low power draw for light AI/Pro work. Otherwise you can stick to consumer cards that are better bang for the specs for almost all other use-cases.
This is tough at the current market price, but one note is this can be fitted in a single slot aftermarket copper cooler to use in low profile servers like minisforum ms-a2. The previous gen ada 2000e was single slot and 50w (i use one in an n5 pro with similar space constraint), those are impossible to find and this is a nice uplift.
Very niche and was a lot more attractive at 699 months ago
Join The Conversation
Share information with the community. Please follow our Community Guidelines and be kind!
3 Comments
Sign up for a Slickdeals account to remove this ad.
70W - can power it easily right thru PCI-E, no external power required.
Low bandwidth 288GB/s or so - not terrible, similar to current AMD Ryzen AI chip .
16GB is nice
CUDA core count is low, I think it was in the 4000'?
Kind of low powered but nice for a companion card to the main card or a super low power system that you want to add some local AI capabilties
If you got the room find a deal for 5080 ($1000 possible) or def 5070 TI as you will run much much faster.
comparison table in the picture below. You'd get this for niche slim profile and low power draw for light AI/Pro work. Otherwise you can stick to consumer cards that are better bang for the specs for almost all other use-cases.
Very niche and was a lot more attractive at 699 months ago
Join The Conversation
Share information with the community. Please follow our Community Guidelines and be kind!