Vega FE Hybrid Mod - Passing the 1080 but 400W+ Power Draw

[Hunter259]

It's great to see the card performing so much better when it is able to keep it's clocks high but sweet lord does this thing pull a fuckton of power and this is without PCIE power draw testing.

 

Written Section : http://www.gamersnexus.net/guides/2986-amd-vega-hybrid-mod-results-overclocking-liquid-vs-air

[Billy_Mays]

The ghetto mounting solution reminds me of a go kart that I made it looked ugly but it worked (little kid bike on a 47" long peice of 2x4)

[JediFragger]

Sweet Jesus....

[goodtofufriday]

400+ is fine me with. lets get it going!

[PCGuy_5960]

Performance makes more sense now, but 400W power draw is A LOT!

[VagabondWraith]

1 minute ago, PCGuy_5960 said:

Performance makes more sense now, but 400W power draw is A LOT!

7900X + Vega = your own personal sauna.  

[MorganO]

4 minutes ago, PCGuy_5960 said:

Performance makes more sense now, but 400W power draw is A LOT!

how do you measure how much power its using? 

can i use this measurement in HWINFO 64

(cpu)http://prntscr.com/fvebgr

(gpu)http://prntscr.com/fvebtl

[PCGuy_5960]

Just now, KOMTechAndGaming said:

how do you measure how much power its using? 

Idk, ask GN?

[Hunter259]

Just now, KOMTechAndGaming said:

how do you measure how much power its using? 

can i use this measurement in HWINFO 64

(cpu)http://prntscr.com/fvebgr

(gpu)http://prntscr.com/fvebtl

If you had watched the video you see they were using an amperage clamp. Given that the PCIE connector only uses 12v you can multiply your Volts * Read Amps to get your wattage which was around 400W. The + in my title was due to not knowing the read out for the PCIE interface itself.

[MorganO]

4 minutes ago, PCGuy_5960 said:

Idk, ask GN?

lol

 

4 minutes ago, Hunter259 said:

If you had watched the video you see they were using an amperage clamp. Given that the PCIE connector only uses 12v you can multiply your Volts * Read Amps to get your wattage which was around 400W. The + in my title was due to not knowing the read out for the PCIE interface itself.

havent watched it yet

my question is can i use said sensors in the screenshots that are highlight to get power usage or not even if its a rough estimation

[zMeul]

13 minutes ago, KOMTechAndGaming said:

how do you measure how much power its using? 

can i use this measurement in HWINFO 64

(cpu)http://prntscr.com/fvebgr

(gpu)http://prntscr.com/fvebtl

that only shows you the GPU power draw and not the board power draw, there are other things that need power on a video card

plus the FETs are not 100% efficient

 

for total power draw I suspect north of 450W - that's just stupid on so many levels

[MorganO]

1 minute ago, zMeul said:

that only shows you the GPU power draw and not the board power draw, there are other things that need power on a video card

plus the FETs are not 100% efficient

so its a ball park figure (im not looking for 100% accourate power measurements) but for the cpu i guess its mostly accurate or is it the same?

[zMeul]

4 minutes ago, KOMTechAndGaming said:

so its a ball park figure (im not looking for 100% accourate power measurements) but for the cpu i guess its mostly accurate or is it the same?

the CPU is a different thing

 

ps: don't make the confusion of the GPU being the video card

[Drak3]

22 minutes ago, VagabondWraith said:

7900X + Vega = your own personal sauna.  

Sun.

 

What you said was just ridiculous. You'd need to constantly dump water on them to keep the steam up, and there is a thread in the news section about electronics and water mixing.

 

Plus, with how hot they're getting, those heat sinks are glowing.

[MorganO]

1 minute ago, zMeul said:

the CPU is a different thing

 

ps: don't make the confusion of the GPU being the video card

kk gpu video card samething in my head

 

are the figures provided by hwinfo ball park figures

[Belgarathian]

2 of these in CF would give my AX1200i a reason to exist. 

[zMeul]

6 minutes ago, KOMTechAndGaming said:

are the figures provided by hwinfo ball park figures

for GPU ballparking board power draw, they aren't - you also need to add the VRAM power and the efficiency of the FETs

[MorganO]

Just now, zMeul said:

for GPU ballparking board power draw, they aren't - you also need to add the VRAM power and the efficiency of the FETs

ok, so cpu i can ballpark the power usage, but graphics card cant because  the variables involved? correct?

[zMeul]

12 minutes ago, KOMTechAndGaming said:

ok, so cpu i can ballpark the power usage, but graphics card cant because  the variables involved? correct?

no, it's because a video card has a lot more components attached to it

[Arika]

Holy shit. Surely there Had to be something AMD could have done to make it more power efficient.

 

It more sounds like they did testing, found its performance was close enough to their intended target and said "perfect"

[potoooooooo]

6 minutes ago, Sierra Fox said:

Holy shit. Surely there Had to be something AMD could have done to make it more power efficient.

 

It more sounds like they did testing, found its performance was close enough to their intended target and said "perfect"

If it were that easy, surely they would have done it.

 

Designing chips isn't as easy as changing some sliders and letting the autorouter plug away.

[mr moose]

I haven't read the article yet, but this quote suggests they need to redo their math:
 

Quote

though is done at the cost of ~33A from the wall over 12V PSU power.

 

 

33 amps from the wall is 3300watts.   And not even possible in most cases.  33 amps on the 12v side of the PSU is a different story though. 

[Hunter259]

1 minute ago, mr moose said:

I haven't read the article yet, but this quote suggests they need to redo their math:
 

 

33 amps from the wall is 3300watts.   And not even possible in most cases.  33 amps on the 12v side of the PSU is a different story though. 

It's bad wording not bad math. 33*12 = 396W.

[mr moose]

Just now, Hunter259 said:

It's bad wording not bad math. 33*12 = 396W.

I understand the math.  but that bad wording suggests the author may not understand the core principles,  as such it warrants that either the reader be cautious of implications or that someone redo the math to ensure it is right.

[Drak3]

6 minutes ago, mr moose said:

33 amps from the wall is 3300watts.   And not even possible in most cases.

It just takes a 40A breaker, running off of 2 20A circuits, or 3 15A circuits. You'll need multiple PSUs though.