Palit rtx 5070 ti gamingpro-s running a zotac apocalypse OC bios, 350w power limit, deshrouded and running 2 arctic p12 pros whirling at 3000rpm, a single 140mm nzxt fan at 900rpm and a p9 pwm blowing on the back of the gpu pcb at about 900rpm (best fans i managed to pull from what i had) ambient temps were about 10~ degrees. It ran steady 3290-3320mhz throughout the whole benchmark, i have my LOD at +3 and texture filtering quality at high performance. Oh and im also using sodimm to udimm adapters running the cheapest 8gb samsung sodimm ram for laptop with 4 sticks, although im not sure how much ram and cpu affect this benchmark, probably not much if any.

I knew i had a very lucky bin of a card, and i just saw this post https://www.tomshardware.com/pc-components/gpus/frankenstein-rtx-5070-ti-with-an-rtx-2080-ti-pcb-breaks-world-record-with-extreme-modding-card-was-damaged-salvaged-with-amd-donor-parts-and-lots-of-soldered-wires-and-tape And decided to try my luck against this fellow hardcore overclocker, goes to show that silicon lottery can still beat hardcore hardware modding.