Three Aftermarket Graphics Coolers On GeForce GTX 480

Test System And Benchmarks

We’re testing idle and load temperatures, in addition to noise levels. The graphics load we’re using is the brutal FurMark stress test at 8x AA.

We’re also testing overclocking ability by seeing how far we can push our stock-voltage GeForce GTX 480 when each of these coolers is used.

All sound and noise tests were recorded in an open testbed. Results will always change on a per-case basis as every type of case has it's own unique airflow. Noise results were recorded with a decibel meter positioned two inches above the graphics card.

Swipe to scroll horizontally
Test System
MotherboardAsus M4A785TD-V EVO Socket AM3, AMD 785G, BIOS 0410
ProcessorPhenom II X4 970 3.5 GHz, Quad-Core, 6 MB L3 Cache
CPU CoolerCooler Master Hyper TX3
MemoryCrucial DDR3-1333 Dual-Channel 2 x 2048 MB, 669 MHz,  CAS 9-9-9-24-1T
GraphicsGeForce GTX 480 700/1401 MHz GPU/Shaders, 924 MHz GDDR5 Memory***all clock rates set to reference specifications for the purposes of benchmarking***
Hard DriveWestern Digital Caviar Black 1000 MB 7200 RPM, 32 MB Cache SATA 3Gb/s
Software and Drivers
Operating SystemMicrosoft Windows 7 x64
DirectX VersionDirectX 11
Graphics DriversGeForce 258.96
Swipe to scroll horizontally
Benchmark Configuration
Synthetic
FurMarkVersion: 1.6.5, Stability Test - Xtreme Burning Mode, 8xAA
  • Poisoner
    You just can't mess with the Accelero.
    Reply
  • AMW1011
    Wow, I can't believe the Accelero and the Zalman were basically neck and neck, with no tangible victory in cooling. I always thought the Accelero would be better, but Zalman actually pulled this one off.

    However, what would be AMAZING, is to have a follow-up to see which of these two solutions work best in SLI, assuming EITHER work well in SLI. I think that is the big question, especially since we never expected bad temps.

    Also can you please add the voltages used for each overclock? It might give people, especially stock GTX 480 owners, a better idea of what these can handle, since cards will always vary when overclocking.
    Holy ***, stock voltages! That is crazy! Please post up your VID (stock voltage) for us. Either you have a golden GTX 480, or the VID is a lot higher than it needs to be, which is pretty common.

    Great article, I love to see these kind of side articles/reviews, and it was well done.
    Reply
  • gabwerkz
    You mean Zalman VF3000F, not Zotac VF3000F on page 7.
    Reply
  • Nerdbox87
    Given that this entire article is based on cooling you really should have included all GPU temps - as vram / vrm cooling may be the difference between the Accelero and the Zalman (as I know my Accelero Twin Turbo on a 5850 struggled with vram temps in Furmark)

    On this my guess would be the Zalman would smash it
    Reply
  • joytech22
    How was the 480 at just under 60C when my 470's are always idling at 86C in a Antec twelve hundred..?

    I mean sure this case isn't the best cooling case but there's never really much hot air in the case and with a sidepanel fan blowing on the two cards i just cant see how.

    Unless.. you guy's test in a cool room at about 15-20C or the fan speeds are maxed 100% of the time.
    Reply
  • anacandor
    joytech22my 470's are always idling at 86C
    Well, there's your praablem.

    Reply
  • dvijaydev46
    Oh, Zalman's cooler looks weaker than Accelero but the later actually outperforms a bit. That's amazing.
    Reply
  • Th-z
    Don, what is your view on how to apply thermal paste. There are numbers of ways people talk about: a pea in the middle, several peas, line in the middle, several lines, and old school spread method. You use spread method when you apply the RAM/VRM sink on Zalman, I presume you use the same method on other parts, too. Maybe Tom can do a comparison test on different ways of applying thermal paste.
    Reply
  • compton
    These look good and would help me out quite a bit with a 465/470, but I wish I could get them in a design that would help maintain the existing footprint. My GF100 doesn't run quite as hot as some reference card designs, but that sure as hell doesn't mean that I'm happy with the leaf blower lurking in my case. Maybe I find a way to make the Zalman work.
    Reply
  • avatar_raq
    I think the Zalman one wins here. It performs slightly better, costs slightly less, weighs less and is shorter than the Arctic cooler. As for noise they're trading blows. Being shorter, the VF3000 does not increase the length of the GTX 480, unlike the accelero which may theoretically cause incompatibility with some cases. Being lighter, it strains the PCB much less. The fact it performed so close to the Arctic cooler with only 2 fans surprised me.
    Reply