If I have money I would buy Intel,but I am not rich so I stick with AMD

I needed to record videos from my games, AMD just dont have the ability to do it like intel do, that, and some other reasons, is which i op't for intel this time round.

Hi im new to forums, been keeping up with this sit for going on 2 years i think?

Anywho i was scanning on through this topic and noticed your mention on i7 im not to sure how good it is, i just recently built a rig and started back testing on beta on this thing, went from first gen duo Core 1 core 2 threads
to
i7920 2.67 ghz, Vista Ultimate X64bit
X58 evga Mobo, Evga gtx 280 SSC,

o and azzure to your comment, i do like your point ^^ but i think it varys on your type of person, Personally id rather save and have the best,

I prefer Intel over AMD.
From personal experience i've seen that Intel outperforms AMD.

Well Core i7 is probably useless for PCSX2 but in future probably its nice. PCSX2 can only use all 8 threads with GSdx 0.1.12 graphic plugin when you select software rendering and select threads for it (1-7 for core i7)

(12-25-2008 11:55 AM)FireFlower Wrote:  Well Core i7 is probably useless for PCSX2 but in future probably its nice. PCSX2 can only use all 8 threads with GSdx 0.1.12 graphic plugin when you select software rendering and select threads for it (1-7 for core i7)

oohh i dint know that ^^

(12-22-2008 11:09 AM)FireFlower Wrote:  Well Azure I could explain why Intel processors have been faster since C2D came out. Times were different on AMD Athlon XP vs. Intel Pentium 4. Pentium 4 processors were CISC based meaning (Complex Instruction Set Computer). Cisc uses very complex technology to run even simple task thats why its very slow with basic use. AMD Athlon XP in otherhand used RISC (Reduced Instruction Set Computer) where tasks are done with very simple pieces of code is pretty fast with normal use.

Okay then Intel releases Core 2 processors. These processors are RISC/CISC based where simple codes are divided to RISC operations and CISC operations are very complex code. This boosted up perfomance so much but AMD is doing this too but with different architecture they haven't yet managed to get same perfomance with same clocks.

Yup AMD is foor poor and Intel is for those who want the best =). AMD generate a bit too much heat and don't overclock so well.

Thats not true, CISC/RISC is more about ISAs, and obviusly, the uArch tradeoffs of implementing them (In x86 case, Decoding penalties, and some executing ones, like MUL being slower than in, I.E,Alpha, the need of a Microcode Engine, and so on).
Since the P6 (Pemtium Pro) and K5, AMD/Intel processors turns the Instructions lines prefetched from the cache, to RISC-like uOps, wich are handled with more ease (Neither of those RISC-Like instructions were compatible in each processor, AMD implemented them sligthy better) and are much more suitable for OoOE. PIV bad performance, comes for a bad uArch, based on a speed-demon desings, whereas K7 is based upon a Brainac desing, wich was a proven way to archieve good performance, while there were less Speed-demons desings.
Now Core 2 keeps the same x86 ISA, but with a much better uArch (Way more) and still now, the complex instructions are converted into uOps trough the Microcode Engine.
However, there is no way that the Athlon XP, using a RISC ISA, were able to execute the same code as the P4, using a CISC ISA.
Obviusly this is much more complicated, but i wouldn't explain it here, and more likely i would not be able

EDIT: On Topic, looking at Intel Roadmaps (Sandy Bridge, Westmere) AMD can take a leap on the race, if Bulldozer turns out as a revolutionary desing, and knowing that AMD have the best engineers in the CPU world (They have them, but they dont have the money for long and expensive projects) this is possible, implementing some of the things EV-8 was about to do, Sync+Register renaming OoOE, a powerful Branch predictor, and a whole 8-issue wide uArch. If this rolled out in 2001 as planed, it would still kick ass on today days, including Power6, Rock, I7 and K10

Well probably my studies have still long way to go =) Well basically new CPU's are more and more like RISC type.

I think RISC/CISC have nothing to do in this. They are used to classify ISA's. It doesn't matter how the CPU implements it. x86 is CISC, and has always been CISC, from the 486 to i7. It doesn't matter whether those complex instructions are really implemented in circuitry or in microcode, as long as they are provided, the CPU is CISC. RISC CPUs are things like MIPS, IBM PowerPC (used in older Macs), or Sun's SPARC (used in servers).

(12-26-2008 02:27 PM)FireFlower Wrote:  Well probably my studies have still long way to go =) Well basically new CPU's are more and more like RISC type.

Internally this may be true, but they present a CISC ISA at their boundary. That's all that really matters, whether it handles the instructions natively or uses an engine to translate to RISC-style operations is just a design choice.

I love RISC instruction sets as much as anyone, but CISC ISA's like x86 on consumer desktops has too much momentum to stop. It's a boon for people like me that have to write assembly, we don't have to waste time reimplementing complex operations since you can do damn near anything with 5 or 6 ops. SIMD instructions allow a somewhat RISC-like design while still keeping all of the addressing modes that I couldn't live without.