Articles: Cases/PSU

Bookmark and Share

Pages: [ 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 ]

High-End Gaming PC 2

The most serious gaming station is the same as in the previous section but the graphics card is replaced with a dual-chip ASUS ENGTX295 (i.e. GeForce GTX 295). Here is the full configuration:

  • Processor: Intel Core i7-920 (2.66GHz)
  • Mainboard: Gigabyte GA-EX58-UD3R (iX58 chipset)
  • System memory: 3x1GB Samsung (PC3-10666, 1333MHz, CL9)
  • Hard disk drive: 1000GB Seagate Barracuda 7200.11 ST31000333AS
  • Graphics card: PCI-E 1792MB ASUS ENGTX295/2DI
  • Optical drive: DVD±RW Optiarc AD-7201S
  • System case: IN-WIN IW-J614TA F430 (550W)

I installed Microsoft Windows Vista Home Premium SP1 (32-bit) and necessary drivers on the PC.

It is easy to see the moment the ACPI driver is loaded and the CPU’s power-saving technologies are enabled: 15 seconds into the test (the 150 mark of the X-axis). It’s different with the graphics card, though. The consumption on one of its power connectors dropped 30 seconds into the test, but the +3.3V consumption grew up at the same moment. This must be only due to the GTX 295 because the previous system, which only had a different graphics card, did not have such a change in the graph. The power draw on both of the card’s additional power connectors grew up 40 seconds into the test. The power consumption of the mainboard grew up, too, and this addition can only be attributed to the graphics card consuming from the PCI Express slot.

Thus, you should not hope for the GTX 295 to be comparable to single-chip cards in terms of power consumption even when displaying Windows’ Desktop. For more details about that, refer to our graphics card articles.

3DMark06 cannot ensure a constant high load for a modern gaming PC: the power consumption of the graphics card and CPU is varying wildly.

Well, we’ve got FurMark for drawing pretty-looking graphs. Take note of the growth of the power consumption throughout the test: it is due to the CPU getting hotter.

Prime95 makes the CPU consume more than a hundred watts like in the previous configuration. The slanting graph is again explained by the temperature rise: the higher the temperature, the higher the power consumption of an electronic chip.

Take note that the graphics card takes about 3A from the additional power connectors when displaying the Desktop. And the mainboard and drives consume about 5A more from the +12V rail. For comparison, these numbers were 2A and 4A with the previous configuration that had a different graphics card.

When launched simultaneously, FurMark and Prime95 produce a familiar picture: the CPU is overloaded and cannot feed data to the graphics card quickly enough.

To check out how this might affect the “wall outlet” measurement, I took the PM-300 wattmeter mentioned at the beginning of the review. It reported a maximum of 490W. Considering the 90% efficiency of the PSU, it means that the computer consumed 441W from the PSU. But my testing tool reports a maximum consumption of somewhat higher than 500W. This big difference is due to the wattmeter’s reporting an average rather than maximum value when the power consumption is so fluctuating.

Of course, my testing tool can also calculate the average that is indicative of the system’s heat dissipation and of your electricity bill. But you want to know the maximum power draw in order to choose an appropriate power supply.

It is still unclear who needs those 1000W power supplies because a 750W unit is quite enough even for this very top-end configuration. A 1000W PSU is two times the required wattage, which is obviously redundant.

Pages: [ 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 ]


Comments currently: 0

Add your Comment