GameUP

GameUP’s Advice Engine uses your PC, game and measurements to pick the next setting to test, then verifies it with a same-scene retest. Steam + Microsoft Store.

[GameUP @GameUpOffi]

@xotic23fn @FortniteStatus @FNCompetitive That rules out the original 60-second pattern. For the random ~1s freezes, I’d mark the fight moments in the trace and compare GPU busy, CPU per-core and disk/VRAM activity. If one signal jumps each time, that’s much better evidence for Epic.

[GameUP @GameUpOffi]

GameUP 2.1 is live on Steam. New captures start clean, same-scene comparisons require an explicit reference, and the Advice Engine can step back when evidence is weak. Creator Mode also adds local branding to the Full overlay. https://t.co/wep7w5NAwu

[GameUP @GameUpOffi]

@Slowww74 @FortniteStatus Fair—stop testing. Try one practical fix: Epic Launcher → Fortnite → Options → enable Pre-download Streamed Assets, then restart. That can remove cosmetic-streaming hitches. If it still drops below the 120 cap, I wouldn’t chase registry tweaks; report it to Epic.

[GameUP @GameUpOffi]

@Slowww74 @FortniteStatus Thanks—the 120 cap is on, but you're still at ~97–104 FPS with only 34–37% GPU use and a steady ~2130 MHz clock. That points away from GPU saturation. Enable frametime plus CPU per-core use in the OSD and capture the exact hitch; total CPU % can hide one maxed game thread.

[GameUP @GameUpOffi]

@Slowww74 @FortniteStatus It doesn’t need to hit 20 to feel bad—the jump from 200 FPS (~5 ms) to 80–90 (~11–12.5 ms) is a big frametime step. Try a 120 FPS cap on the same route. If the lag disappears, unstable headroom is likely; if not, capture frametime + CPU/GPU clocks at the drop.

[GameUP @GameUpOffi]

@Slowww74 @FortniteStatus Got it. The HML is the right kind of log, but I can’t read its graph directly from Drive here. Could you export or screenshot CPU per-core, GPU clock/power and frametime around one drop? A 10–20 s window is enough.

[GameUP @GameUpOffi]

@BloobyVT That makes the patched in-game limiter the strongest suspect. I’d leave RTSS at the same cap for now and repeat one Training route after the next hotfix. I wouldn’t keep changing unrelated graphics settings if RTSS stays flat.

[GameUP @GameUpOffi]

@BloobyVT At that point I’d keep RTSS as the control and send both frametime captures with the bug report. Since restarts, shader warm-up and presets don’t change it, the in-game limiter vs RTSS A/B is much stronger than another settings sweep.

[GameUP @GameUpOffi]

@BloobyVT That really looks like the in-game limiter. I’d try its cap 2–3 FPS below refresh, then replay the same Training sequence without alt-tabbing. If the spikes stay while RTSS remains flat, that’s a pretty clean A/B.

[GameUP @GameUpOffi]

@Nosejak A new GPU might not fix bad frame pacing. Try a 60/90 FPS cap and replay the same swing route twice; if the frametime spikes land in the same spots while GPU use isn’t pinned, it’s probably streaming/CPU-side rather than raw GPU load.

[GameUP @GameUpOffi]

@WiiNK_64 55→45 swings are where I’d stop chasing average FPS: use one short training-mode sequence, change one setting, then compare frametime/1% lows. I’m the GameUP dev; we just added a Tōkon profile: https://t.co/pcUE1UUAKj

[GameUP @GameUpOffi]

@Slowww74 @FortniteStatus Thanks—these show ~88–111 FPS, not the 40 FPS instant. GPU use is only ~37% with normal temps/clocks, so the 4050 isn’t maxed here. Record GPU power/clock and CPU per-core graphs through the drop; that can separate a game-thread/streaming spike from a laptop power limit.

[GameUP @GameUpOffi]

@GalvanicSquare @GreasyPixel Yeah—the hard part is changing five things and not knowing what helped. I built GameUP around a loop: use PC/game/measurements to pick the next setting, rerun the same scene, then compare frametime and 1% lows. https://t.co/G3qWpV2PNE

[GameUP @GameUpOffi]

@LonginusOffcial @una_panesilla Los mínimos indican que abre, no frametime estable. La CPU maneja draw calls, IA, físicas y streaming; bajar gráficos descarga sobre todo la GPU. Repite la ruta limitada a 60 FPS: si el pico cae siempre en el mismo punto, parece más streaming/motor que tu 7600X.

[GameUP @GameUpOffi]

@LonginusOffcial @una_panesilla Para salir de esa paranoia: repite 3 veces la misma ruta de 60 s mirando frametime y uso de GPU; luego baja solo la resolución. Si los tirones no cambian y la GPU no va al 95–99 %, apunta más a CPU/streaming/motor. Trabajo en GameUP, pero RTSS + CapFrameX sirven.

[GameUP @GameUpOffi]

@Tenvid_san @Roltas01 Si ya estás en bajo/720p, probaría un límite estable de 30 o 40 FPS y repetiría la misma zona mirando temperaturas y uso de CPU/GPU. Si sigue igual, puede ser compilación de shaders o throttling, no calidad gráfica. Cambia una cosa cada vez. ¿Qué GPU lleva el portátil?

[GameUP @GameUpOffi]

@LegionzGaming That RT-shadow lead is worth an A/B: replay the same route on the current driver with only RT shadows off. If it still crashes, then test the older driver—don’t change both at once. Count crashes + watch frametime. That one-change/retest loop is what I build GameUP around.

[GameUP @GameUpOffi]

GameUP is now on Microsoft Store.

Benchmark a repeatable scene, let the Advice Engine pick the next setting to test, change one thing, then replay the same scene and keep only what the retest proves.

What game/settings pair would you test first?
https://t.co/cy8ZocafFR

[GameUP @GameUpOffi]

@PJam06 @ArcSystemWorksU Does it tank in an offline training stage too? Repeat 30s at native resolution and 70% scale while watching frametime + GPU use. If the dips land in the same spots and GPU use stays low, that points to game/CPU-side pacing. I work on GameUP; PresentMon is enough for this A/B.