...why the hell would you think I meant that? No no no no no, I've built enough computers to know that when you change motherboards windows goes berserk. It's not worth it. The problem was because my DVD drive failed, and I just spent over $400 on parts so I didn't want to spend more to buy a new one. Trying to get my Windows 7 disc onto a USB key, getting it on, and then having the activator nag me. Ugh. It's an official freakin tool just for that! *facepalm* Whatever. So I had to call and all that fun. So it's all sorted out.
Now that that is taken care of, I'll run some benchmarks when I can stop playing Skyward Sword
.
"activator" .... pirate alarm
the normal usb installation works like a charm
my pc got from the beginning no DVD-Drive...
now 2years without and i dont miss it...
there where 2-3 cases where i needed one...
network dvd drive via notebook
Alright guys. Here's some benchmarks at stock clocks. I don't want to overclock it just yet, you know, I want some chance to enjoy my new system, and to settle in a bit. Plus... the AS5 needs 200 hours to really cure... Bah. But here we go. Here's the system specs.
Core i5 3750K @ stock everything
Asus P8Z77-V LX
Crucial Ballistax DDR3-1600 4GBx2
Geforce 275GTX overclocked to 743-1282-1605
Now the results. Note that I'm going to include some graphics related stuff as well, cause they kinda go hand in hand. But I'll skip the boring stuff. Still, just read the bolded parts for the CPU specific results.
Cinemark 11.5 (64bit)
OpenGL - 58.47fps
CPU - 6.89pts
CPU (single core) - 1.76pts
MP Ratio - 3.91x
Passmark7 Performance Test
CPUmark - 8887.1
3D Graphics Mark - 2567.3
3D Mark Vantage (more a GPU test but I might as well include it)
3DMark Score - P15446
GPU Score - 12454
CPU Score - 55343
CPU Test 1 - 2993.04 plans/s
CPU Test 2 - 190.50 steps/s
There we go. That's all the benchmarkers I have. I'd like to try to benchmark Dolphin in some way. I have Mario Galaxy, maybe that would do?
So guys, any other benchmarks you guys want me to try at stock? (trial or free please). Any ideas for benchmarking Dolphin?
Yeh, benchmark Super Mario Galaxy at the observatory. And NSMBW at World 1-1 with EFB to RAM and Accurate texture cache set to safest using DX9. Native resolution and non-scaled EFB to rule out the GTX275. Turn off the framelimiter too.
Quote:CPU - 6.89pts
CPU (single core) - 1.76pts
Nice, that looks like a small improvement over the 2500K.
2500K @ 3.8GHz (to match the 3570K)
CPU: 6
Single Core: 1.53
2500K @ 4.5GHz
CPU: 7.08
Single Core: 1.80
2500K @ 5GHz
CPU: 7.76
Single Core: 2.01
That's not a small improvement dude. My Core 2 Duo got 1.69 in the full Cinebench test, and .87 in the single core test. OVERCLOCKED from 2.66ghz to 3.2ghz, the maximum stable overclock I could get. So from Sandy to Ivy is a full Core 2 Duo overclocked core. So yea, that's my perspective
. Unfortunately I didn't turn off the overclock on my old machine and run any of these benchmarks (doh) so I can't really give a stock to stock comparison. Oh well.
Oh, and the 3570K is 3.4ghz stock clock, btw. Not 3.8. Of course, turbo was still on (up to 3.8ghz), is that what you used?
Quote:That's not a small improvement dude.
Pretty small if you ask me. You've got 0.89 more in the multithreaded test and 0.23 in the single threaded test, not the values you listed lol.
Not saying your CPU is bad ofcourse. Just not worth it upgrading from a Sandy Bridge
.
Update your sig btw.
Quote:Oh, and the 3570K is 3.4ghz stock clock, btw. Not 3.8. Of course, turbo was still on (up to 3.8ghz), is that what you used?
Yeah, your CPU Turbo Boosts when its on load, so basically the test was at 3.8GHz. Unless you turn off Turbo Boost, it's never going to run your benchmarks at stock.
The difference between the 2500K and 3570K in terms of clockspeed is 100MHz. 3.3/3.4 @ stock and 3.7/3.8 @ Turbo Boost respectively.
Quote:Update your sig btw.
DOH. Thanks for pointing it out.
Anyway, I finished up my testing on Mario Galaxy. Everything is done at the observatory on a new game save, at the spot where you wake up after loading a save, and given a few seconds to stabilize. Here's the settings and results. Again, all the CPU settings are stock, including turbo. Framelimiter is always off, vsync off, etc etc etc. All testing was done in 3.0-600.
Results are underlined. What I was aiming to test is in bold.
DX11
Internal Resolution: 1x Native
Scaled EFB Copy: Off
EFB to Texture
Tex Cache - Fast
DSP HLE
Result - 68fps
DX9
Internal Resolution: 1x Native
Scaled EFB Copy: Off
EFB to Texture
Tex Cache - Fast
DSP HLE
Result - 71fps
DX9
Internal Resolution: 1x Native
Scaled EFB Copy: Off
EFB to Ram (not cached)
Tex Cache - Fast
DSP HLE
Result - 66-67fps
DX9
Internal Resolution: 1x Native
Scaled EFB Copy: Off
EFB to Ram (cached)
Tex Cache - Fast
DSP HLE
Result - 66-67fps
DX9
Internal Resolution: 1x Native
Scaled EFB Copy: Off
EFB to Texture
Tex Cache - Safe
DSP HLE
Result - 66-67fps
DX9
Internal Resolution: 1x Native
Scaled EFB Copy: Off
EFB to Texture
Tex Cache - Fast
DSP LLE (on thread)
Result - 66fps
DX9
Internal Resolution: 1x Native
Scaled EFB Copy: Off
EFB to Ram (cached)
Tex Cache - Safe
DSP LLE (on thread)
Result - 56-57fps
DX9
Internal Resolution: 3x
Scaled EFB Copy: On
EFB to Ram (cached)
Tex Cache - Safe
DSP LLE (on thread)
Result - 52-53fps
DX9
Internal Resolution: 3x
Scaled EFB Copy: On
EFB to Texture
Tex Cache - Fast
DSP LLE (on thread)
Result - 64fps
(05-08-2012, 11:29 AM)MaJoR Wrote: [ -> ]DX9
Internal Resolution: 1x Native
Scaled EFB Copy: Off
EFB to Ram (cached)
Tex Cache - Safe
DSP LLE (on thread)
Result - 56-57fps
With this configuration, I get the same speed on a Sandy Bridge 2600K running at 4.5Ghz. I wonder how much faster it would be if you overclocked your CPU to the same speed. I think RAM bandwidth is the bottleneck with these settings.
Hey MaJoR have you tried Super Mario Galaxy with the official 3.0 release? You should be able to get really good speeds with that version. When I tried the 3.0-631 dirty version I got real real slowdowns.
But the 3.0 version OH BABY, it'll fly...