10-18-2026, 07:11 PM
I tested this tonight. On Threadripper box with llama.cpp, I saw basically the same behavior. The surprise was that memory pressure showed up before throughput collapsed.
|
RTX 5090 local inference on M3 Max
|
|
« Next Oldest | Next Newest »
|
| Messages In This Thread |
|
RTX 5090 local inference on M3 Max - by thetarunner - 10-19-2026, 05:32 AM
Re: RTX 5090 local inference on M3 Max - by ggufwire22 - 10-18-2026, 06:00 PM
Re: RTX 5090 local inference on M3 Max - by rasterhall - 10-18-2026, 07:11 PM
Re: RTX 5090 local inference on M3 Max - by thetarunner - 10-18-2026, 09:46 PM
Re: RTX 5090 local inference on M3 Max - by fusedhall - 10-18-2026, 10:36 PM
Re: RTX 5090 local inference on M3 Max - by torchshed - 10-19-2026, 03:33 AM
Re: RTX 5090 local inference on M3 Max - by dockkite69 - 10-19-2026, 03:54 AM
Re: RTX 5090 local inference on M3 Max - by signaltrack44 - 10-19-2026, 04:52 AM
Re: RTX 5090 local inference on M3 Max - by batchwire - 10-19-2026, 05:36 AM
|
Users browsing this thread: 1 Guest(s)