Dolphin, the GameCube and Wii emulator - Forums

Quote:I checked the gaming performance benchmarks on both AMD CPUs before I made that statement. So, if it doesn't really gain much on the gaming side/aspect, what make you think it would do better on emulation? Also, there are some extra data to look at:

http://forums.pcsx2.net/Thread-CPU-Bench...d-on-FFX-2

57.66 FPS - SLUS 20672 - AMD Phenom II X4 955 - 3.8 GHz OC -Ryner Lute
56.64 FPS - SLUS 20672 - AMD Athlon II X2 250 - 3.8 GHz OC - denimu

54.05 FPS - SLUS 20672 - AMD Phenom II X2 555 Black Edition - 3.6 GHz OC -denimu
53.24 FPS - SLUS 20672 - AMD Athlon II x4 640 - 3.6 GHz OC - Spikexp

40.46 FPS - SLUS 20672 - AMD Athlon II X2 245 - 2.9 GHz Stock - CKL
40.20 FPS - SLUS 20672 - AMD Phenom II X6 1055T - 2.80 GHz Stock (Turbo Core off) - ilovejedd

So, where is this day and night differences that you stated? I stated that he should save up and get something better. I think it's you that have an reading comprehending issue.

Here's my example of day and night differences should look like:

49.08 FPS - SLUS 20672 - Intel Core i7 920 - 2.66 GHz Stock - Master_DX
48.78 FPS - SLUS 20672 - Intel Core 2 Quad Q9550 - 2.83 GHz Stock - tuanming
48.41 FPS - SLUS 20672 - Intel Core 2 Duo E8400 - 3.00 GHz Stock - antespo
48.34 FPS - SLUS 20672 - AMD Phenom II X4 945 - 3.3 GHz OC - Abelus
48.05 FPS - SLUS 20672 - AMD Phenom II X6 1055T - 3.33 GHz OC - ilovejedd (Turbo Core off) - CPU-Z
47.69 FPS - SLUS 20672 - AMD Phenom II X4 955 - 3.2 GHz Stock- Ryner Lute

AMD CPUs clocked at over 3GHz still could not manage to beat a lowered clock C2Q. Let play a bit fair and say bump the clock speed of the Q9550 to 3.2GHz and see what would happen, shall we?

56.44 FPS - SLUS 20672 - Intel Core i5 650 - 3.2 GHz Stock (Turbo Boost on) - ilovejedd
55.17 FPS - SLUS 20672 - Intel Core 2 Quad Q9550 - 3.2 GHz OC - tuanming

First of all gaming performance != pcsx2 software rendering performance. Both games and emulators are usually heavily impacted by cache size since they need to make frequent large block memory read/write operations. I know for a fact that some games on dolphin are very heavily impacted by cache size. An E5200 @ 3.6GHz will perform worse in SSBM than an E8400 @ 3.0GHz. Most of the people who complain about performance have cpus with a small cache.

Second of all you've shown only one benchmark from one software application. What you've displayed here is a software rendering benchmark for pcsx2 where performance will instead be dictated by the cpus simd performance. Game benchmarks on tomshardware/anandtech/guru3d/any other hardware review site clearly show that a phenom II clocked at 3.0GHz will significantly outperform an athlon II clocked at 3.0GHz in any memory/cpu intensive game. The performance lead of a phenom II in these applications is usually about 20% at the same clock rate.

Third of all you're comparing athlon II and phenom II (K11 vs. K11). Not Athlon X2 and phenom II (K8 vs. K11), which is what you made your original statement about.
Here is where an athlon X2 ranks on that chart:

Quote:33.40 FPS - SLUS 20672 - AMD Athlon 64 X2 5200+ - 3.06 Ghz OC - Rezard - CPU-Z Screenshot
29.20 FPS - SLUS 20672 - AMD Athlon 64 X2 5200+ - 2.7 Ghz Stock - Rezard - CPU-Z Screenshot
27.63 FPS - SLUS 20672 - Intel Core 2 Duo E6300 - 1.86 GHz Stock - tuanming
27.30 FPS - SLUS 20672 - AMD Athlon 64 X2 5200+ - 2.7 GHz Stock- Abelus
26.19 FPS - SLUS 20672 - Intel Core 2 Duo E7200 - 1.6 GHz Underclocking - pcsx2fan
24.54 FPS - SLUS 20672 - Intel Pentium Dual-Core E2160 - 1.8 GHz Stock - Shadow Lady
23.38 FPS - SLUS 20672 - Intel Core 2 Duo T5200 - 1.5 GHz Stock - prafull
21.23 FPS - SLUS 20672 - AMD Athlon X2 QL-64 - 2.1 GHz Stock - Rezard - CPU-Z Screenshot
20.03 FPS - SLUS 20672 - AMD Athlon 64 X2 4200+ - 2.2 GHz Stock - abdo123 - CPU-Z Validation

That's right, HALF the performance of an athlon II/phenom II at the same clock rate. Now that's what I call a night and day difference.

Please think before you post.

Quote:and we don't know if it wasn't video card that bottlenecked those systems.

Yes we do. That benchmark was done using software rendering.

Quote:Also all AMD CPU there were forced to SSE2 and Phenom II's are capable of SSE4.

SSE4a, not SSE4.1. SSE4a doesn't really have any useful instructions for software rendering.

Quote:If you want to bash AMD so badly, you'll better change target to video cards, because it will be true that geforce -> radeons because AMD drivers suck. Old game (Ragnarok Online), GeForce 6600 in linux with wine vs HD5870 native. Full speed on 6600 and less than 10 fps on ATI at same scene. Fail.

He's not talking about graphics cards, please don't change the subject.

Yes, any Phenom II would outperform an Athlon 64 X2. Once again, don't do the mistake of comparing the clock. The only situation where clock comparisons make sense is when comparing 2 CPUs from the exact same architecture.
So a 955 seems good to me. However if you can get a good Phenom II X3 or X2, you'd save money without losing performance (and 3 cores may be helpful with LLE on thread. It helps a lot with my i5).