Dual DGX Sparks- 40tk/s single 1M ; 350 tk/s agg. - Deepseek V4 Flash (vs RTX Pro 6000 vs Mac M2 Ultra 192)
COVERAGE [1]
-
Dual DGX Sparks- 40tk/s single 1M ; 350 tk/s agg. - Deepseek V4 Flash (vs RTX Pro 6000 vs Mac M2 Ultra 192)
<!-- SC_OFF --><div class="md"><p>First of all shout out to Aiden/Antirez & geniuses at the Nvidia community threads. I'm merely claude-vibing off of their works.</p> <p>That a said, i thought i'd share recipes & learnings & benchmarks so far on running big MOE models…