[TechSpot and Tom's] Crysis 3 GPU and CPU benchmarks

[Solomutt]

I'm still trying to understand why people keep saying that screen resolution is what makes more VRAM advantageous. This will be true as soon as all textures come from system memory, and not card memory. Which, in the case of discrete cards, is never.....

 

[blackened23]

I'm still trying to understand why people keep saying that screen resolution is what makes more VRAM advantageous. This will be true as soon as all textures come from system memory, and not card memory. Which, in the case of discrete cards, is never.....

You truly have absolutely no idea what you're talking about. Game asset and especially anti aliasing causes VRAM use to multiply as resolutions are increased. Pixel count increases as resolution increases - That means game asset VRAM use increases. This also applies to anti aliasing, the pixels for which AA must be applied increases exponentially with higher resolutions. It's like you have a completely absurd assumption that pixel count remains the same as resolution increases, which is not the case.

 

[escrow4]

Gainward Phantom 680 2GB + i7 3770 (stock) + 16GB 1600MHz DDR3, all settings at high, textures very high, TXAA medium = average 45FPS, frequently peaks at 55-60+ FPS in outdoor levels. All at 1920 x 1200.

All settings at very high, AA/AF disabled at 1920x1200 = laggy 35-40 FPS. I'm too used to 50+. Dropped settings down a notch. I highly recommend not buying until this is reduced - its 50% shorter than Crysis 2 and I think it needs some optimisation patches. Full credit to raping CPU's though.

 

[chimaxi83]

I'm still trying to understand why people keep saying that screen resolution is what makes more VRAM advantageous. This will be true as soon as all textures come from system memory, and not card memory. Which, in the case of discrete cards, is never.....

Listen to Blackened. You're clueless dude.