MSI GTX 660 GAMING 2 GB

[W1zzard]

MSI's new GTX 660 GAMING seeks to introduce the popular MSI GAMING brand to the GTX 600 Series. The card, which comes overclocked out of the box, was very quiet in our testing and is available online at an affordable $210.

Show full review

 

[Wark0]

Hi,

MSI GeForce GTX 660 Twin Frozr 2 GB was reviewed one year ago using 306.23 Beta drivers
MSI GTX 660 GAMING 2 GB review is done with 320.18 WHQL drivers

As you know, drivers can have an impact on GPU performance and btw power consumption in a game (you are using "a stressful game" for noise). In this review you directly compare fan noise between the two cards but with such a driver difference this comparison is pointless, isn't it ?

Power consumption is in fact higher on Crysis 2 on both card with the same PCB, 110W for the MSI GeForce GTX 660 Twin Frozr 2 GB and 119W for the MSI GTX 660 GAMING 2 GB, so a higher fan noise is logical, but could be more related to the power consumption difference than cooling. This power consumption difference can be related to the driver and/or simply GPU variance.

 

[erixx]

SKU... were did i read that last time? Mmmm Yeah, we all want a nice computer full of nice SKU's boy! I love my Es-kai-jus!

 

[W1zzard]

In this review you directly compare fan noise between the two cards but with such a driver difference this comparison is pointless, isn't it ?

all data that I have suggests that fan noise and power consumption stays almost constant between drivers. any software optimizations have a minimal impact that is below the measuring accuracy for all related tests.

feel free to prove me wrong. do some temperature testing (game, not furmark) on old driver, new driver and old driver again

SKU... were did i read that last time? Mmmm Yeah, we all want a nice computer full of nice SKU's boy! I love my Es-kai-jus!

industry people pronounce it skew, not es-kai-jus

 

[Wark0]

all data that I have suggests that fan noise and power consumption stays constant between drivers. any software optimizations have a minimal impact that is below the measuring accuracy for all related tests.

feel free to prove me wrong. do some temperature testing (game, not furmark) on old driver, new driver and old driver again

Sure. For example on Radeon 7970 GHz with Catalyst 12.8, 172 watts under BF3, with Catalyst 13.6b 195 watts. Of course it depends on the game, the gpu and the driver.

 

[W1zzard]

check gpu temperature, fan speed is based on temperature, not on power draw

 

[Wark0]

You said that power consumption stays constant between drivers, my numbers show that it's not always the case. With higher power draw you know that you temperature and/or fan speed will be necessarily higher ...

 

[W1zzard]

You said that power consumption stays constant between drivers, my numbers show that it's not always the case. With higher power draw you know that you temperature and/or fan speed will be necessarily higher ...

what I mean was that it stays constant enough to not have any measurable effect on heat/noise. can you check gpuz for me? how much higher is temperature and fan speed with those 2 drivers?

 

[Wark0]

Nope, i've done this test a few months ago, the card is not here at this time. But with such a difference even if it's stable with 13% - 23W more power draw their must be a difference. I do agree that for 2-3W difference it will not be necessarily the case.

 

[W1zzard]

just do it with your current card

 

[Casecutter]

W1zzard on page 4 you said... "NVIDIA's new GK106 processor is produced on a 28 nm at TSMC".

I suppose that a simple scripted over-site from like last years "Gigabyte GTX 660 Windforce". There nothing different on this after 12 months. Well there is... this thing is expensive! I mean a two Heat-pipe cooler vs. and 4-Pipe found on the "Windforce" and the Gigabyte provided cooling the voltage regulation circuitry, and included good Samsung memory.

Given the base price on what appears to be a lower quality/construction card, and only reduced like 15% from a year ago! I see the better buy was the Gigabyte... a year ago for full MSRP. Today spending +$200 for a GK106 is not any value proposition. You would've been better off getting the Gigabyte at MSRP 12 months ago, it would end up like $3.30 a month depreciation all along receiving the enjoyment. Given today's market this has no real standing unless you can get it for $160-170.

The Gigabyte Windforce got a 8.9 (fair enough at the time), now this MSI is 8.7 - I don't see it? The PowerColor HD 7870 Devil got an 8.5 it wasn't much value, but at least it gave some stuff for $20 more like the backing-plate and software voltage control/monitoring. It also was straddled with the Elpida same memory chips, but offers 5-7% faster performance.

I said with the Devil "Regrettably just too late to have any merit" that holds true for this MSI Gaming, but given the content/construction MSI provides it's even less of a value.