Dolphin, the GameCube and Wii emulator - Forums

Full Version: [Unofficial] New Dolphin 5.0 CPU benchmark - results automatically updated!
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Did anyone figure out why the Core2Quads were performing so much better than Core-is at the same clockspeed? I see more Core2s performing almost as good as and occasionally better than Sandy Bridge at similar clockspeed in the benchmarks.
I'm honestly not sure. They shouldn't be at all. My only thought is user inaccuracies as the only verification we can give is a screenshot of the completed POVRay run.
Both me and Nintendo Maniac conducted tests with Core2 and Core-i systems (eg. Sandy Bridge), and Core2 were scoring unusually well. My Sandy Bridge was doing poorly against Core2s at the same clockspeed.

foxyshadis

I listed the CPU frequency based on the maximum Turbo boost as shown in CPU-Z (which it sustained about 50% of the time, dropping 100MHz from time to time), if that's incorrect let me know.
(03-19-2017, 02:04 PM)Helios Wrote: [ -> ]I'm honestly not sure. They shouldn't be at all. My only thought is user inaccuracies as the only verification we can give is a screenshot of the completed POVRay run.

Remember, this uses luabench, not POVRay like the older Dolphin 4.x benchmark used. For reference, running the older POVRay benchmark in Dolphin 5.0 does not result in this performance discrepancy.

I unfortunately have no desktop Core 2 Duo systems nor any mobile Core 2 Duo systems with discrete GPUs. Therefore, due to a GPU bottleneck, I am unable to test the performance difference in actual games.

For all we know, it may simply be a quirk with luabench.
Submitted the results for my Xeon E3 1230 V5.

Tested at 3.6 ghz (Highest speed it can maintain across all cores):
https://drive.google.com/open?id=0B0tJ49...UlQQlVhSFU

Tested at 3.0 ghz:
https://drive.google.com/open?id=0B0tJ49...XA0bDlZNWM

Tested at 2.0 ghz:
https://drive.google.com/open?id=0B0tJ49...2EycGJ1aEk

Tested at 1.0 ghz (yes, really):
https://drive.google.com/open?id=0B0tJ49...VNMcEhUMnM
Hey, I ordered an Open Box Supermicro A1SRi-2758F from Newegg because I had been considering getting one for a while and it was cheap.

I'm eventually going to be using it as a storage server, but for a little while I'm going to use it as a (Windows 10) desktop. Would you folks like for me to benchmark Dolphin on the Atom CPU?

I'm pretty sure these industry-grade chips won't want me to adjust their clock speed, but I could be mistaken. How would I look into changing that to benchmark at different speeds?
Probably in the BIOS like most motherboards
(04-14-2017, 06:48 AM)KHg8m3r Wrote: [ -> ]Probably in the BIOS like most motherboards

That's the first place I'm going to check (when my RAM comes in), but on my last server board with two E5-2670's, there wasn't anywhere for me to adjust clock speeds that I saw. I then kind of took for granted that you just aren't ordinarily allowed to fool with that stuff on 'server' products.
(03-24-2017, 07:07 PM)Nintendo Maniac 64 Wrote: [ -> ]
Nintendo Maniac 64 Wrote:Remember, this uses luabench, not POVRay like the older Dolphin 4.x benchmark used.  For reference, running the older POVRay benchmark in Dolphin 5.0 does not result in this performance discrepancy.

I unfortunately have no desktop Core 2 Duo systems nor any mobile Core 2 Duo systems with discrete GPUs.  Therefore, due to a GPU bottleneck, I am unable to test the performance difference in actual games.

For all we know, it may simply be a quirk with luabench.




your in luck i have a system with an 8600 gt and a E6550 core 2 duo i can run twilight princess (GC) its most likely gonna run like garbo but i can test it for you i also have a c2q q6600 i can test as well and a pentium d830 and a pentium E5700 its gonna take alot of time as i dont have windows installed on the box.
Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14