World’s Fastest Gaming CPU Could Be Holding Back the NVIDIA RTX 5090 at 4K

All very interesting.

Anyone have a spare $2k and an extra, hmm, how much would my energy bill go up? 7900 xtx to 5090?

I asked copilot. Not sure it got the idle and light to medium power usage right though. Anyone have a better model? Or perhaps a better question (that makes me think of the Hitchhikers guide to the Galaxy).

Asking CoPilot:

what is the tdp of the rx 7900 xtx?

The AMD Radeon RX 7900 XTX has a maximum power draw (TDP) of 355 watts. However, some variants like the Sapphire 7900 XTX Nitro+ have a higher TDP of 420 watts.
Is there anything specific you'd like to know about this graphics card?

Assuming 350, and that I am in Reno Nevada and game 4 hours a day and have a light to medium load 10 hours on a computer with a 670e chipset and ryzen 7950x3d cpu, just changing the gpu from the 7900 xtx to an NVidia 5090, please estimate for me, how much I can expect my energy bill to increase?

Sure, let's break it down step by step:
  1. Power Consumption:
    • AMD RX 7900 XTX: 350 watts
    • NVIDIA RTX 5090: 575 watts
  2. Daily Usage:
    • Gaming: 4 hours/day
    • Light to Medium Load: 10 hours/day
  3. Electricity Cost in Reno, NV: 11.05 cents/kWh
  4. Monthly Usage Calculation:
    • Gaming: 4 hours/day * 30 days * 350 watts = 42 kWh/month
    • Light to Medium Load: 10 hours/day * 30 days * 350 watts = 105 kWh/month
    • Total for RX 7900 XTX: 147 kWh/month
    • Gaming: 4 hours/day * 30 days * 575 watts = 69 kWh/month
    • Light to Medium Load: 10 hours/day * 30 days * 575 watts = 172.5 kWh/month
    • Total for RTX 5090: 241.5 kWh/month
  5. Monthly Cost Calculation:
    • RX 7900 XTX: 147 kWh * 0.1105 = $16.23/month
    • RTX 5090: 241.5 kWh * 0.1105 = $26.68/month
Estimated Increase: $26.68 - $16.23 = $10.45/month
So, you can expect your energy bill to increase by approximately $10.45 per month if you switch from the RX 7900 XTX to the NVIDIA RTX 5090.
 
I envy your 4hr a day availability. I consider it a good week if I get 4hrs total on a weekend.

Haven't crunched those numbers but given daily usage it seems plausible. Thankfully we were able to get solar installed a few years back so electricity isn't normally and issue any more. Paid off a while back and now I just smile when the bill comes. BTW my 4090 pulls between 420-520 when I do game, and sometimes slightly higher but pretty rare there, with most of the titles I like to play at 4K, so not that far off from a 5090.

Edit: Side note, the 4080 SUPER I have is probably the most energy efficient card I have ever seen. Between ~260W upto ~330 (rare and extreme cases) for 4K w/ DLSS quality at 60-120 FPS. It's almost more fun watching the TDP than gaming.
 
How do you check your watts?

I can check at the PSU level, using ICUE, to see how much it is taking from the wall and providing to my system, I dont get any kind of sub system breakdown though. Can Windows 11 see that? I dont see anything in the Resource monitor, or is there an App?

Thanks :)

Oh me at light use:

1738282586100.png
Me with POE 2 running - moderate to heavy use:

1738282769907.png

I've seen it get up to 600-650 but just spikes not sustained - that includes the 7950x3d, 3 NVME, a Sata, etc.
 
Elf_Boy said:
How do you check your watts?
Click to expand...
Mostly just by MSI AB but occasionally I'll take a look at what the UPS says for actual from the wall draw. It's rare when my 4090 rig pulls more than 600W but I think the most I've seen has been upwards of 750ish with real-world gaming and not just benchmarking to see something maxed out.
 
Elf_Boy said:
I asked copilot. Not sure it got the idle and light to medium power usage right though. Anyone have a better model? Or perhaps a better question (that makes me think of the Hitchhikers guide to the Galaxy).
Click to expand...
Pretty interesting results. I'd say it's not bad as just a differential between two cards. But it's assuming full TDP for "light to medium use" for your cards. Not sure what the idle draw on these cards are - that number is harder to find sometimes.
 
Small correction:

Per TechSpot:​


RE: Marvel’s Spider-Man 2 Remastered: “At 4K, the CPU bottleneck is mostly removed, but the RTX 5090 still appears slightly limited, averaging 212 fps. As a result, the RTX 5090 is just 26% faster than the RTX 4090.”
Click to expand...

TechSpot/HUB tested Spider-Man, not Spider-Man 2. ;)
 
Brian_B said:
Pretty interesting results. I'd say it's not bad as just a differential between two cards. But it's assuming full TDP for "light to medium use" for your cards. Not sure what the idle draw on these cards are - that number is harder to find sometimes.
Click to expand...
j2c and hardware nexus use devices to specifically measure power draw and have posted numbers.

Still seems off that ltt did their I initial review with 7800x3ds.
 
Grimlakin said:
Still seems off that ltt did their I initial review with 7800x3ds.
Click to expand...
They used those because they have a dozen of them they sourced to be similar in performence to allow parallel testing.
 
Grimlakin said:
Is it physx causing that? :)
Click to expand...
Funny you should mention that. I was doing some DLSS 4 transformer model ray reconstruction testing this weekend with CP2077 and Hogwarts and noticed NV listed it in their copyright notices before one of those games. It was a blink and miss it moment but I was a little surprised.
 
You must log in or register to reply here.
Become a Patron!
Back
Top