- Aug 2, 2017
- 2
- 0
- 510
While dealing with many other various issues today, I decided to do some more in-depth benchmarking of my system.
(Asus Z170-A i3 6100, Zotac GTX960 4GB, 8GB DDR4 2400, EVGA 500W PSU, and a 480GB SSD + 1TB HDD) (Running latest version of Windows 10, and a 1440x900 monitor)
The major issue I noticed was when I ran the 3D portion of Passmark, comparing it to the average GTX 960 marks. The three most notable tests were the GPU compute test, DX11, and DX12. My (slightly) overclocked 960 crushed most 960s in DX12 and compute, scoring around the 80th and 90th percentile of all 960s.
However quite the opposite was the case when running the DX11 benchmark, when I scored 72-77FPS. The overall average 960 score was 106. BUT, 81% of the scores are above this mark, peaking at 114-116FPS, which 18.5% of all 960s landed in. After a bit of painstaking hovering over tiny bars, and simple addition, I found that 90% of all GTX 960s performed better than mine in DX11.
What's the deal? Is this just a crazy driver issue?
Extra stuff- My FireStrike score was 6296, compared to the 7640 that is their official 960 score.
Newest versions of nvidia drivers and MSI Afterburner were installed today. (actually caused a BSOD that I had to fix. problem hasn't returned thankfully.)
Reverting my 960 to stock clocks only made my score go down. (+150 on core and +50 on memory, over stock Zotac 960 clock. [the non AMP edition])
According to (http://www.tomshardware.com/forum/id-3192807/ultimate-bottlenecking-guide.html) an Intel i3 6100 shouldn't bottleneck a 960 at all. (especially in a GPU benchmark, I would assume)
Finally, possibly the most telltale sign of something, would be that MSI AB reported about 50-55% Frame buffer During the DX12 test, while peaking at 90-96% during the DX11 test.
(Asus Z170-A i3 6100, Zotac GTX960 4GB, 8GB DDR4 2400, EVGA 500W PSU, and a 480GB SSD + 1TB HDD) (Running latest version of Windows 10, and a 1440x900 monitor)
The major issue I noticed was when I ran the 3D portion of Passmark, comparing it to the average GTX 960 marks. The three most notable tests were the GPU compute test, DX11, and DX12. My (slightly) overclocked 960 crushed most 960s in DX12 and compute, scoring around the 80th and 90th percentile of all 960s.
However quite the opposite was the case when running the DX11 benchmark, when I scored 72-77FPS. The overall average 960 score was 106. BUT, 81% of the scores are above this mark, peaking at 114-116FPS, which 18.5% of all 960s landed in. After a bit of painstaking hovering over tiny bars, and simple addition, I found that 90% of all GTX 960s performed better than mine in DX11.
What's the deal? Is this just a crazy driver issue?
Extra stuff- My FireStrike score was 6296, compared to the 7640 that is their official 960 score.
Newest versions of nvidia drivers and MSI Afterburner were installed today. (actually caused a BSOD that I had to fix. problem hasn't returned thankfully.)
Reverting my 960 to stock clocks only made my score go down. (+150 on core and +50 on memory, over stock Zotac 960 clock. [the non AMP edition])
According to (http://www.tomshardware.com/forum/id-3192807/ultimate-bottlenecking-guide.html) an Intel i3 6100 shouldn't bottleneck a 960 at all. (especially in a GPU benchmark, I would assume)
Finally, possibly the most telltale sign of something, would be that MSI AB reported about 50-55% Frame buffer During the DX12 test, while peaking at 90-96% during the DX11 test.
Facebook
Twitter
Instagram