Power consumption during gaming
Let’s move on to the power consumption in games, where the TBP of 450 watts is put into perspective very quickly, depending on the resolution and feature set of course. In QHD, the new card is even more frugal on average than a GeForce RTX 3080 10GB, which may certainly also be due to the CPU limiting, but it also indicates an enormously increased efficiency. It is and remains a tangible surprise.
Even in Ultra-HD without CPU limit, it is only just behind the GeForce RTX 3090, which is again astonishing. However, I’ll come to the efficiency analysis across all cards and resolutions later. First of all, we are talking about the wattage figures, and they are more than just pleasing.
If you use the supersampling helpers, the picture is reversed because the actually powerful CPU is slightly limited again. But we can also state here that the new GeForce RTX 4090 is the measure of all things when you put everything in relation to the achieved gaming performance.
Power consumption in factory state as summary
Yes, I did run Furmark at just under 600 watts with the Power Target set to maximum, but that’s basically pointless and purely for the gallery. But again, I briefly checked the load peaks and found nothing that would contradict all the other measurements and statements. What is a real step backwards from my point of view is the very high idle power consumption, which rarely falls below 35 watts even when the GPU is heavily downclocked. However, this is a special problem of the X670E motherboards, which is supposed to be fixed by a new UEFI or will be soon. PCIe problems during idle (fallback) are suspected.
Load sharing between PCIe slot and external 12VHPWR connector
NVIDIA has connected all 10 phases or the controlled 20 control loops to the 12VHPWR, which makes sense that way. The PCI-SIG says in the PCIe 5.0 specification that the card may consume a maximum of 600 watts of power in the sum of all 12 connections. Although this is slightly exceeded by a few watts in extreme cases, it is still quite accurate within the permissible tolerances. Instead of the maximum possible 5.5 A, it is only 1.2 A in the measurements, even in extreme cases, which is about 14 watts. The card should generate various extra-low voltages here, which hardly change as a load even when overclocking. The main load is therefore almost exclusively via the 12+4 pin connector (12VHPWR), which got up to 70 °C hot after about 1 hour of full load during operation with the 4-pin adapter (4x 6+2) on the card. However, the circuit board where the connection is soldered has already reached a temperature of more than 60 °C. This solution doesn’t really work, but at 450 to 500 watts you’re still in the pain-free range.
- 1 - Introduction, technical data and technology
- 2 - Test system in igor'sLAB MIFCOM-PC
- 3 - Teardown: PCB, components and cooler
- 4 - Gaming Performance WQHD (2560 x 1440 Pixels)
- 5 - Gaming Performance UHD (3840 x 2160 Pixels)
- 6 - Gaming Performance UHD + DLSS/FSR/XeSS (3840 x 2160 Pixels)
- 7 - DLSS 3.0 and the longest bars
- 8 - NVIDIA Reflex and Latency
- 9 - Workstation graphics and rendering
- 10 - Power consumption and load sharing
- 11 - Load peaks, capping and power supply recommendation
- 12 - Temperatures, clock rate, OC, fans and noise
- 13 - Summary and Conclusion
545 Antworten
Kommentar
Lade neue Kommentare
Urgestein
Veteran
Urgestein
Urgestein
Urgestein
Urgestein
Veteran
Urgestein
Veteran
Urgestein
Mitglied
Veteran
Urgestein
Veteran
Urgestein
Urgestein
Mitglied
Mitglied
Veteran
Alle Kommentare lesen unter igor´sLAB Community →