{ "meta": { "title": "TechFuelHQ GPU Undervolt Settings and Attributed Results", "description": "GPU undervolt and power-limit results, one row per setting a source measured: the stock and tuned figures, the source's author and publication date, and the passage the numbers come from. One row is TechFuelHQ's own RTX 5080 stock baseline. Workloads differ between rows, so no row predicts another card.", "license": "CC BY 4.0", "schema_version": 2, "updated": "2026-09-28", "contribute": "A GitHub account is required to submit a pull request to https://github.com/iBlessi/gpu-undervolt-db. Email contact@techfuelhq.com as an alternative. Include stock and tuned measurements from the same workload and a source; leave unmeasured values null. A setting nobody measured is not a row.", "fields": { "id": "unique slug for the row (gpu-board-setting)", "gpu": "canonical card model, e.g. 'RTX 5080' or 'RX 9070 XT'", "vendor": "NVIDIA or AMD", "board": "Board the source tested, or '' if the source does not name it", "tbp_watts": "the reference total board power in watts for the card as tested; null where the row's board is a partner model whose own rating the source does not state", "method": "Tuning control: MSI Afterburner V/F curve, AMD Adrenalin voltage offset or maximum voltage, or a power limit alone", "setting": "Settings as the source states them; not a stability guarantee", "voltage_mv": "NVIDIA curve target voltage, or AMD RDNA3 absolute max voltage, in mV (nullable)", "offset_mv": "AMD negative voltage offset in mV, RDNA4 (nullable)", "clock_mhz": "target/held core clock in MHz (nullable)", "power_limit_pct": "power limit applied; percent of stock for NVIDIA (e.g. 90) or delta for AMD (e.g. -10 or +15); null if untouched", "power_saved_w": "Stock minus tuned board power in watts for the stated workload; null when the source gives no matched pair or the tune drew more (see stock_result and tuned_result)", "temp_drop_c": "Stock minus tuned temperature in C for the stated workload; null when the source gives no matched pair", "fps_delta_pct": "Performance change versus stock in percent, rounded from the source's stock and tuned figures where both are given, otherwise the source's reported change: positive gain, zero parity, negative loss; null when not reported", "stability_hours": "Reported stable test duration in hours; null means no duration established", "stock_result": "The stock figures the source measured, with the workload", "tuned_result": "The tuned figures the source measured in the same workload", "test_method": "Workload, whose card it was, and the limits of the measurement", "source_type": "editorial-firstparty is TechFuelHQ's own bench; community-verified is an owner's report on their own card; review-measured is a publication's or reviewer's test; TechFuelHQ reproduced none of the external rows", "source_url": "Where the figures were published", "source_url_2": "A second source for the same test, or ''", "source_publisher": "Publication or forum that carries the source", "source_author": "Author, poster or tester the source names", "source_title": "Title of the source page or thread", "source_date": "YYYY-MM-DD the source was published or posted", "source_quote": "A short verbatim passage from the source describing its result, setting or measurement method", "source_read": "YYYY-MM-DD TechFuelHQ last read the source", "submitter": "Who compiled or submitted the row, with original attribution", "date": "YYYY-MM-DD the row was added or corrected; not the experiment date", "flagged": "true if the outlier-flag script flagged this row (do not set by hand)", "flag_reason": "why it was flagged (set by the script)" }, "repository": "https://github.com/iBlessi/gpu-undervolt-db" }, "results": [ { "id": "rtx-5080-rog-astral-oc-stock-baseline", "gpu": "RTX 5080", "vendor": "NVIDIA", "board": "ASUS ROG Astral OC", "tbp_watts": 400, "method": "None: stock baseline, Astral OC Default profile", "setting": "Stock, no undervolt applied", "voltage_mv": null, "offset_mv": null, "clock_mhz": 2784, "power_limit_pct": null, "power_saved_w": null, "temp_drop_c": null, "fps_delta_pct": null, "stability_hours": null, "stock_result": "Time Spy Graphics 32,350; 2,784 MHz sustained in Time Spy; 266-340 W gaming, 360 W maximum under stress; 63-72 °C core while gaming", "tuned_result": "—", "test_method": "My own card on its stock profile, 3DMark and HWiNFO64, June 8-11, 2026. It is one card and stock only. The same dataset's 271 W gaming and 400 W stress figures are its manual-overclock profile, not stock.", "source_type": "editorial-firstparty", "source_url": "/data/gpu/rog-astral-rtx-5080-oc-2026-06-09/", "source_url_2": "", "source_publisher": "TechFuelHQ", "source_author": "LK Wood IV", "source_title": "ASUS ROG Astral RTX 5080 OC first-party bench dataset", "source_date": "2026-06-11", "source_quote": "Thermals and board power captured with HWiNFO64 during sustained-load runs.", "source_read": "2026-09-26", "submitter": "TechFuelHQ (first-party bench, stock profile)", "date": "2026-09-28", "flagged": false, "flag_reason": "" }, { "id": "rtx-5070-ti-935mv-2955mhz", "gpu": "RTX 5070 Ti", "vendor": "NVIDIA", "board": "ASUS Prime OC", "tbp_watts": 300, "method": "MSI Afterburner V/F curve", "setting": "935 mV @ 2955 MHz (+450 MHz core offset), +1500 memory", "voltage_mv": 935, "offset_mv": null, "clock_mhz": 2955, "power_limit_pct": null, "power_saved_w": 50, "temp_drop_c": null, "fps_delta_pct": 4, "stability_hours": null, "stock_result": "300 W in 3DMark Steel Nomad; fans near 50%", "tuned_result": "250 W in Steel Nomad, 170-200 W in most games; about +4% performance; fans 35-45%", "test_method": "The author's own Asus Prime RTX 5070 Ti OC. The 50 W is Steel Nomad alone; the author puts the saving at 50-70 W across games and benchmarks. No stability duration given.", "source_type": "community-verified", "source_url": "https://www.esportstales.com/tech-tips/how-to-undervolt-the-5070-ti-with-msi-afterburner", "source_url_2": "", "source_publisher": "Esports Tales", "source_author": "Vincenzo Milella", "source_title": "How to undervolt the 5070 Ti with MSI Afterburner", "source_date": "2025-03-14", "source_quote": "Power consumption: 250W in 3DMark Steel Nomad. (300W at default)", "source_read": "2026-09-26", "submitter": "TechFuelHQ (from the Esports Tales author's report)", "date": "2026-09-26", "flagged": false, "flag_reason": "" }, { "id": "rx-9070-xt-sapphire-nitro-plus-neg100mv", "gpu": "RX 9070 XT", "vendor": "AMD", "board": "Sapphire NITRO+", "tbp_watts": null, "method": "AMD Adrenalin voltage offset", "setting": "-100 mV offset", "voltage_mv": null, "offset_mv": -100, "clock_mhz": null, "power_limit_pct": null, "power_saved_w": null, "temp_drop_c": null, "fps_delta_pct": 10, "stability_hours": null, "stock_result": "58 FPS average, Cyberpunk 2077 high, RT medium", "tuned_result": "64 FPS average", "test_method": "The author's own Sapphire NITRO+ RX 9070 XT, one game. Power limit, board power and temperatures for this setting are not given, and neither is a stability duration.", "source_type": "community-verified", "source_url": "https://www.xda-developers.com/i-undervolted-my-gpu-and-got-the-upgrade-i-actually-wanted/", "source_url_2": "", "source_publisher": "XDA", "source_author": "Ty Sherback", "source_title": "I undervolted my GPU and got the upgrade I actually wanted", "source_date": "2026-01-17", "source_quote": "with a -100mv offset, I saw my average FPS jump from 58 to 64", "source_read": "2026-09-26", "submitter": "TechFuelHQ (from the XDA author's report)", "date": "2026-09-26", "flagged": false, "flag_reason": "" }, { "id": "rx-9070-xt-red-devil-neg170mv-110pt", "gpu": "RX 9070 XT", "vendor": "AMD", "board": "PowerColor Red Devil", "tbp_watts": null, "method": "AMD Adrenalin voltage offset and power limit", "setting": "-170 mV offset, 110% power target", "voltage_mv": null, "offset_mv": -170, "clock_mhz": 3360, "power_limit_pct": 10, "power_saved_w": null, "temp_drop_c": null, "fps_delta_pct": 10, "stability_hours": null, "stock_result": "60 FPS, 2.90 GHz, 339 W maximum; Cyberpunk 2077 4K Ultra, no RT", "tuned_result": "66 FPS, 3.36 GHz, 358 W maximum", "test_method": "Der8auer's test of a PowerColor Red Devil, as Tom's Hardware tabulates it. A performance tune: it draws more power than stock, not less. Maximum power only; no temperatures or stability duration.", "source_type": "review-measured", "source_url": "https://www.tomshardware.com/pc-components/gpus/undervolted-rx-9070-xt-beats-rtx-5080-rx-9070-and-9070-xt-models-with-heavy-coolers-have-massive-oc-headroom", "source_url_2": "", "source_publisher": "Tom's Hardware", "source_author": "Sunny Grimm, reporting Der8auer's test", "source_title": "Undervolted RX 9070 XT beats RTX 5080 - RX 9070 and 9070 XT models with heavy coolers have massive OC headroom", "source_date": "2025-03-11", "source_quote": "RX 9070 XT (OC) (source: Der8auer) 66 3.36 358 -170 110%", "source_read": "2026-09-26", "submitter": "TechFuelHQ (Der8auer's result as tabulated by Tom's Hardware)", "date": "2026-09-26", "flagged": false, "flag_reason": "" }, { "id": "rx-9070-neg100mv-30pl-efficiency", "gpu": "RX 9070", "vendor": "AMD", "board": "ASRock Steel Legend", "tbp_watts": null, "method": "AMD Adrenalin voltage offset and power limit", "setting": "-100 mV offset, -30% power limit", "voltage_mv": null, "offset_mv": -100, "clock_mhz": 2400, "power_limit_pct": -30, "power_saved_w": 73, "temp_drop_c": null, "fps_delta_pct": -7, "stability_hours": null, "stock_result": "54 FPS at 244 W, 2.6 GHz; Cyberpunk 2077 at 4K", "tuned_result": "50 FPS at 171 W, 2.4 GHz", "test_method": "Alva Jonathan's ASRock Steel Legend, as Club386 reports it. Notebookcheck calls the drop about 10%; 50 against 54 FPS is 7.4%. No stability duration given.", "source_type": "review-measured", "source_url": "https://www.club386.com/amd-radeon-rx-9070-maintains-93-of-its-performance-at-171w/", "source_url_2": "https://www.notebookcheck.net/Radeon-RX-9070-undervolting-yields-surprising-performance-still-beats-RTX-4070-Ti-Super.977181.0.html", "source_publisher": "Club386", "source_author": "Fahd Temsamani, reporting Alva Jonathan's test", "source_title": "AMD Radeon RX 9070 maintains 93% of its performance at 171W", "source_date": "2025-03-11", "source_quote": "This resulted in 50fps at 171W – down from 244W", "source_read": "2026-09-26", "submitter": "TechFuelHQ (Alva Jonathan's result as reported by Club386 and Notebookcheck)", "date": "2026-09-26", "flagged": false, "flag_reason": "" }, { "id": "rx-9070-neg125mv-plus10pl-performance", "gpu": "RX 9070", "vendor": "AMD", "board": "ASRock Steel Legend", "tbp_watts": null, "method": "AMD Adrenalin voltage offset and power limit", "setting": "-125 mV offset, 110% power target", "voltage_mv": null, "offset_mv": -125, "clock_mhz": 3000, "power_limit_pct": 10, "power_saved_w": null, "temp_drop_c": null, "fps_delta_pct": 11, "stability_hours": null, "stock_result": "54 FPS, 2.60 GHz, 244 W maximum; Cyberpunk 2077 4K Ultra, no RT", "tuned_result": "60 FPS, 3.00 GHz, 270 W maximum", "test_method": "The same tester and card as the -30% row, as Tom's Hardware tabulates it; Club386 gives the same -125 mV and 60 FPS. Notebookcheck gives this tune as -100 mV. A performance tune: it draws more power than stock.", "source_type": "review-measured", "source_url": "https://www.tomshardware.com/pc-components/gpus/undervolted-rx-9070-xt-beats-rtx-5080-rx-9070-and-9070-xt-models-with-heavy-coolers-have-massive-oc-headroom", "source_url_2": "https://www.club386.com/amd-radeon-rx-9070-maintains-93-of-its-performance-at-171w/", "source_publisher": "Tom's Hardware", "source_author": "Sunny Grimm, reporting Alva Jonathan's test", "source_title": "Undervolted RX 9070 XT beats RTX 5080 - RX 9070 and 9070 XT models with heavy coolers have massive OC headroom", "source_date": "2025-03-11", "source_quote": "RX 9070 (OC) (source: Alva Jonathan) 60 3.00 270 -125 110%", "source_read": "2026-09-26", "submitter": "TechFuelHQ (Alva Jonathan's result as tabulated by Tom's Hardware)", "date": "2026-09-26", "flagged": false, "flag_reason": "" }, { "id": "rx-7900-xtx-1100mv-max", "gpu": "RX 7900 XTX", "vendor": "AMD", "board": "", "tbp_watts": null, "method": "AMD Adrenalin voltage slider (RDNA3 maximum voltage)", "setting": "1100 mV on the voltage slider, driver 25.3.1", "voltage_mv": 1100, "offset_mv": null, "clock_mhz": null, "power_limit_pct": null, "power_saved_w": null, "temp_drop_c": null, "fps_delta_pct": null, "stability_hours": null, "stock_result": "300-400 W, hot spot in the 80-90s °C, Topaz Video AI upscaling (Adrenalin readout)", "tuned_result": "200-something W, hot spot 65 °C; Topaz's Rhea model still 8-10 fps", "test_method": "One owner's forum report from Topaz Video AI upscaling, not games; the watts and temperatures are Adrenalin's readout, given as ranges. The poster sets normal settings for games because they crash on this tune.", "source_type": "community-verified", "source_url": "https://community.topazlabs.com/t/7900-xtx-under-volting-suscess-3-400-watts-down-to-200/89312", "source_url_2": "", "source_publisher": "Topaz Community forum", "source_author": "j.mcleod", "source_title": "7900 xtx under volting suscess, 3-400 watts down to 200+", "source_date": "2025-04-05", "source_quote": "I set the mV slider to 1100", "source_read": "2026-09-26", "submitter": "TechFuelHQ (from the forum poster's report)", "date": "2026-09-26", "flagged": false, "flag_reason": "" }, { "id": "rtx-4090-985mv-2800mhz", "gpu": "RTX 4090", "vendor": "NVIDIA", "board": "MSI Suprim X", "tbp_watts": 450, "method": "MSI Afterburner V/F curve", "setting": "985 mV @ 2800 MHz", "voltage_mv": 985, "offset_mv": null, "clock_mhz": 2800, "power_limit_pct": null, "power_saved_w": 40, "temp_drop_c": null, "fps_delta_pct": -1, "stability_hours": null, "stock_result": "Cyberpunk 2077 74.1 FPS average at 398 W average; Port Royal 12,162", "tuned_result": "73.4 FPS at 358 W; Port Royal 12,072", "test_method": "The poster's own MSI RTX 4090 Suprim X, averages of three to six runs per profile. One Port Royal run returned no score and the poster had a system freeze the day before, so they considered adding 5 mV.", "source_type": "community-verified", "source_url": "https://forum.level1techs.com/t/undervolting-my-rtx-4090/190315", "source_url_2": "", "source_publisher": "Level1Techs forum", "source_author": "Azulath", "source_title": "Undervolting my RTX 4090", "source_date": "2022-10-26", "source_quote": "2800Mhz @985mV", "source_read": "2026-09-26", "submitter": "TechFuelHQ (from the forum poster's tables)", "date": "2026-09-26", "flagged": false, "flag_reason": "" }, { "id": "rtx-4090-60pct-power-limit", "gpu": "RTX 4090", "vendor": "NVIDIA", "board": "", "tbp_watts": 450, "method": "Power limit only", "setting": "Power limit 60% (270 W of the 450 W default)", "voltage_mv": null, "offset_mv": null, "clock_mhz": null, "power_limit_pct": 60, "power_saved_w": null, "temp_drop_c": null, "fps_delta_pct": -8, "stability_hours": null, "stock_result": "100% power limit (450 W); TechPowerUp's own stock gaming draw is around 340 W", "tuned_result": "About 8% slower at 4K, averaged across five AAA games at maximum settings", "test_method": "Quasar Zone's test as TechPowerUp reports it. A power cap is not a curve undervolt; the tuned board power is not given. The same test's 850 mV undervolt is described only as similar, with no number.", "source_type": "review-measured", "source_url": "https://www.techpowerup.com/301649/nvidia-geforce-rtx-4090-with-nearly-half-its-power-limit-and-undervolting-loses-just-8-performance", "source_url_2": "", "source_publisher": "TechPowerUp", "source_author": "btarunr, reporting Quasar Zone's test", "source_title": "NVIDIA GeForce RTX 4090 with Nearly Half its Power-limit and Undervolting Loses Just 8% Performance", "source_date": "2022-11-30", "source_quote": "lose just around 8% of performance at 4K UHD, averaged across five AAA games", "source_read": "2026-09-26", "submitter": "TechFuelHQ (Quasar Zone's result as reported by TechPowerUp)", "date": "2026-09-26", "flagged": false, "flag_reason": "" } ] }