450w and rx Vega 56?

[PLOB]

I have an SF450 80+ gold and an r5 1400, would i be fine use a Vega 56 if i undervolted it a little?

[InfinityJosh]

引用自 Omega：

With my Vega 64 and i7 5820k I see peeks of 550w being pulled from the wall.

It may be a difference due to different voltage :100v vs 230v.

[Monk]

More likely the hugely power hungry HEDT chip vs the new mainstream King.

If I was to run the test with my 7900x (I won't, that would take 4 days to do tearing down and rebuilding / testing the loop lol) I would have a huge power draw also from the cpu.

The crazy vega power figures came from unlocked crazy bios settings pulling enormous voltage, in other words, not a normal use case.

[Omega]

引用自 Monk：

Why, that's a ridiculous use case, I guess the 9900k is just Way more efficient than we all thought, boosting all cores to 4.8 it seems, so far, heaven (1080p), superposition (1080p extreme and 4k optimised) and firestrike all come in around 350w with only the extreme final test of firestrike pulling just 400w momentarily from the wall.

so on a evga supernova g2 1000w (huge overkill but I already had it and only paid £80 for it new) which is around 90% efficient to make life easy, that means my system, with a 9900k boosting to 4.8 all cores (according to hw monitor thought it was only meant to boost to 4.7, hmm) and a stock vega 64 is using between 315w to 360w under benchmark load.

Remember that HEDT chips use ALOT more power, my 7900x with sli 1080ti overclocked uses 450w idle and 8-900w under load.

So unless OP goes messing with custom bios I suspect he would be fine even overclocking the vega 56 and r3 1400 on his psu.

Edit.

These benchmark tests are far more realistic than running two torture tests similtaneously.

Doesn't matter, the machine is capable of pulling it therefor the PSU should be capable of supplying it.

My 550w max power draw is messured during normal daily use of my machine, gaming, browsing, some other stuff. The PSU is supplying around 500w when it peaks while on average it's supplying around 450w while gaming. My idle power usage is below 100w.

I wouldn't consider a benchmark to be a piece of software that is capable or supplying realistic usage scenarios to check total system power usage. It's way too perfect and controlled and can not be compared to the random chaos in a game. You look at the sky in Battlefield for 1 seconds and BOOM 300fps, 4 CPU cores maxed, GPU maxed while in a benchmarkt test #1 runs the GPU maxed and the CPU is barely utilized and in test #2 the CPU is maxed but the GPU is being held back.

[Monk]

Your power draw is entirely from your cpu not the gpu, the Op is using an r3 1400 that barely wants 80w of power.

Dual torture test power draw is hugely outside the realm of normal use, plus, we're both using much more powerful cpu's and higher power draw gpu's.

Honestly, I dislike using prime and furmark, they are unrealistic and I am yet to run into issues from my overclock none of which have ever been tested on prime or furmark, I find the various benchmarks I use do an excellent job of stressing components beyond what they will encounter in day to day use without subjecting them to a load that CAN kill them even at stock, which is why you have never seen me recommend them in any real capacity.