- Joined
- Jun 6, 2002
i will need to go back and run the benchmarks again with screen grabs of temp/TDP. what i think is interesting people might have already have know? NV spec TDP/Stock clock is no the same as what the card runs. I just threw in the EVGA 3070 XC3 Ultra, no stock speed listed on their site but Boost speed of 1770mhz. Techpower up reports stock 1500mhz boost 1725mhz. https://www.techpowerup.com/gpu-specs/geforce-rtx-3070.c3674
when i first ran superposition bench MSI AB showed it would be pulling up to 240watts for short periods. it was always around 230 watts, the Boost speed showing was just under 2ghz, some times as low as 1870mhz. temps were in check around 72c with fans only hitting 80% speed. i figured why not lets see what happens right? use the MSI AB power curve, find the point where it hits 1.8ghz. Set that point to the rest of the graph, so there is still the step up points from low to 1.8ghz. the board power never gets over 200watts, mostly sits around 190watts just limiting the gpu clock speed. FPS in superposition bench was not effected much ,think it lost 1 or 2 frames on the minimums. it lowered the the score by about 300 points vs other evga 3070's listed. not sure why when you search to compare on their site it lists by brand? is this what people have been talking bout basically limiting power and keep the same performance? i kinda got the impression some where just limiting % of power in msi ab vs just limiting core clock. To go from 100% do what you want board settings using around 230watts and about 2ghz gpu speed. to just limiting gpu speed to 1.8ghz, with no power limit in msi AB under 200watts i like! does this actually mean then, that the 3060TI is pulling maybe 220watts to get close to the 3070 numbers?
really wish i had boards to test this out. would this be a new way to test GPU power efficiency? Limit either board power to a set wattage or do it by gpu speed till it hits target TDP(i assume this is total card power draw)
when i first ran superposition bench MSI AB showed it would be pulling up to 240watts for short periods. it was always around 230 watts, the Boost speed showing was just under 2ghz, some times as low as 1870mhz. temps were in check around 72c with fans only hitting 80% speed. i figured why not lets see what happens right? use the MSI AB power curve, find the point where it hits 1.8ghz. Set that point to the rest of the graph, so there is still the step up points from low to 1.8ghz. the board power never gets over 200watts, mostly sits around 190watts just limiting the gpu clock speed. FPS in superposition bench was not effected much ,think it lost 1 or 2 frames on the minimums. it lowered the the score by about 300 points vs other evga 3070's listed. not sure why when you search to compare on their site it lists by brand? is this what people have been talking bout basically limiting power and keep the same performance? i kinda got the impression some where just limiting % of power in msi ab vs just limiting core clock. To go from 100% do what you want board settings using around 230watts and about 2ghz gpu speed. to just limiting gpu speed to 1.8ghz, with no power limit in msi AB under 200watts i like! does this actually mean then, that the 3060TI is pulling maybe 220watts to get close to the 3070 numbers?
really wish i had boards to test this out. would this be a new way to test GPU power efficiency? Limit either board power to a set wattage or do it by gpu speed till it hits target TDP(i assume this is total card power draw)