(05-15-2011, 05:56 AM)Rezard Wrote: [ -> ]Have you heard of AMD's Bulldozer? It has the instruction sets that it has lacked, and a few other unique changes. Maybe you could look to those for PCSX2 salvation in June (if you're an AMD fan). 
Bulldozer looks like its going to be a Multi Threaded monster of a CPU for commercial release games and apps.
But for PCSX2 i doubt its going to match the 2500K & 2600K SandyBridge CPU's, seeing as PCSX2 only really makes use of 2 cores, and from what i've seen of leaked performance figures, Bulldozer doesnt seem to be as fast clock for clock, core for core as Sandybridge.
They only seem to be 20% or so faster than Phenom II's at the same clock speed, which is a good improvement, but still WAY short of the 40% speed advantage the Sandybridge CPU's have over the Phenom II's at the same clock speed today.
So i think Intel will still be the CPU of choice for just about all emulators out there for the forseeable future.
Of course if PCSX2 was to go quad core or better......
...
Yeah, AMD is more focused on the aggregate market, so they're not going to be really trying to beat out Intel clock-for-clock. I've even read that their arch will be less effective when utilizing less cores, which is not gonna help there.
AMD has done some things differently, though. We're all curious to see what they pull out of their you know whats.
Shadow Lady:
Unfortunately our CPU benchmark thread uses an FMV, so the all important VU recompiled code performance isn't measured.
Have you tried switching from microVU to superVU in that FMV? it's a 10FPS difference here (not using speedhacks), the performance difference is similar to what you get in VU heavy games, seems to be enough to consider that as performance test for PCSX2, and even on other PC apps it's always been the same with comparable performance between Phenom II and Core 2 at same clocks, only the higher (hax) Core 2 Quad Q9xxx and i3/5/7 being ahead by a good amount
(05-15-2011, 02:11 PM)DaMan Wrote: [ -> ]MS VC doesn't even do auto vectorizing but doesn't enabling SSE on the compiler do FP operations as a scalar SSE operation so it can use a XMM register instead of GPR or something like that? This is how 64bit "obsoletes" x87. GCC has an option to use both x87&SSE so it can use both sets of registers.
i believe enabling SSE optimizations on msvc2008 tries to use SSE where it can, but it still uses x87 FPU operations for some stuff iirc. its most likely similar to the GCC mode you mentioned.
and i don't have experience with 64bit compilers so not sure what they try to do. since on 64bit architectures SSE has 8 additional XMM regs, its probably even more advantageous to use SSE than the x87 FPU.
Intel compiler is far more aggressive with SSE
I am playing Smackdown vs Raw 2011 using PCSX2.It is running at 60fps when i am playing with single player mode.But when i m playing 6man tag match,royal rumble my fps going down to 35fps.Can i know why?
no may not. (in this thread)
The Intel Core 2 Q9xxx series is a Core i3/i5/i7 in a disguise, at least that is what the data are saying with pcsx2.