Dolphin, the GameCube and Wii emulator - Forums

(05-12-2013, 12:40 PM)omega_rugal Wrote: [ -> ]Sorry to bump but as i was reading this thread i found this...

Quote:3. Even if it didn't, historically Intel has never produced a cpu
series that was slower than its predecessor (despite the pentium 4
haters who would have you believe otherwise).

So you think P4 Williamette was a GOOD processor in someway?

I never said that. I said that Intel has never produced a cpu that was slower than its predecessor.

The main reason that people didn't like Willamette was because it was:

1. Outrageously expensive
2. Improved performance very little over its predecessor compared to previous leaps (pentium to pentium II, pentium II to pentium III, etc.). Little did they know that this new slower improvement curve of 10-20% per year in singlethreaded performance would continue for the next 10 years with northwood and conroe being the only exceptions.
3. Intel's Willamette chipsets only supported rdram. Which was also outrageously expensive.
4. Its performance improvements over its predecessor were inconsistent. It ranged anywhere from -15% to +70% depending on the application despite the 50% increase in clock rate. And people weren't used to that concept at the time. On average I would say most applications were around 20-25% faster than on coppermine.
5. The motherboards were also expensive and choices for chipsets and manufacturers were more limited than the other platforms.
6. It requires more cooling and requires good case ventilation to prevent throttling from overheating. A concept that people weren't used to at the time.
7. It was harder to OC.
8. Due to its small increase in performance and large increase in cost and power consumption it offered less performance per watt and performance per dollar than coppermine or thunderbird/palomino.
9. The internet acts as a sort of "drama amplifier". Especially for hardware and software products. Products that are slightly below average become widely known as "the worst piece of shit ever invented" and products that are slightly above average become widely known as "the greatest thing since sliced bread".
10. Power consumption was higher than coppermine or thunderbird/palomino.

Reason number 8 is a good summary reason. Notice that nowhere in there did I mention it being slower than coppermine, in the vast majority of applications it wasn't. Reason number 4 elaborates on this.

And yes it was good "in a way". It was faster on average than any other x86 cpu at the time. I think if they had marketed it as an enthusiast product rather than a mainstream product it might have done better. The concept that this thing would replace the pentium III completely really scared people.

Also I should point out that while I believe that I have done sufficient research to draw these conclusions I wasn't interested in microprocessors at that time since I would have been around 10 years old. So obviously none of these conclusions were reached from first hand experience.

Quote:3. Intel's Willamette chipsets only supported rdram. Which was also outrageously expensive.

Mmm no. there was a version compatible with pc-100... which ran even slower.

Quote:It was faster on average than any other x86 cpu at the time

Nope, only if you compare it with a PIII coppermine at 600 Mhz it could win, anything between that and 900 MHz was a tie and the Tualatin definitly mopped the floor with it.

Quote:So obviously none of these conclusions were reached from first hand experience.

i was 19, but i`ve been in this "business" since 1998... and yeah i tested this piece of crap when it first launched.

i still even them, a RAMBUS at 1.4 Ghz and PC-100 at 1.8 GHz, more or less, i can`t remember them exactly.

omega_rugal Wrote:Nope, only if you compare it with a PIII coppermine at 600 Mhz it could win, anything between that and 900 MHz was a tie and the Tualatin definitly mopped the floor with it.

Yeah.....no: http://www.tomshardware.com/reviews/intel,264-19.html

There is no question that it beats any pentium III in the 600-900 MHz range.

omega_rugal Wrote:Mmm no. there was a version compatible with pc-100... which ran even slower.

Are you sure it was an Intel chipset not a 3rd party chipset and that it used socket 423/478?

I thought Intel signed an exclusivity agreement with rambus that ran from 2000-2002.

Quote:Are you sure it was an Intel chipset not a 3rd party chipset and that it used socket 423/478?

Intel 845 Clarkdale supports pc-133 (just like the board i threw away, from which i got the cpu)

and willamette uses socket 423, socket 478 was for NorthWood, don`t mix them.

Quote:There is no question that it beats any pentium III in the 600-900 MHz range.

Using Rambus, yeah. too bad it was way too expensive back in those days.

from emulators.com/Pentium4

"As Tom's Hardware site documented last month, the Pentium 4 lost miserably against the AMD Athlon at MPEG video encoding. Only after Intel engineers personally modified the code did the Pentium 4 suddenly win the benchmarks. A side effect of this is that the benchmarks on the Athlon improved considerably as well, indicating that the code was very poorly written in the first place."

wanna know the true story?

http://www.emulators.com/pentium4.htm

careful, very long article.

omega_rugal Wrote:and willamette uses socket 423, socket 478 was for NorthWood, don`t mix them.

I know. I was including all 2002 motherboards since that was the last year of the contract.

omega_rugal Wrote:Using Rambus, yeah. too bad it was way too expensive back in those days.

I don't see how this invalidates my statement. I never mentioned the price being good or even reasonable.

omega_rugal Wrote:"As Tom's Hardware site documented last month, the Pentium 4 lost miserably against the AMD Athlon at MPEG video encoding. Only after Intel engineers personally modified the code did the Pentium 4 suddenly win the benchmarks. A side effect of this is that the benchmarks on the Athlon improved considerably as well, indicating that the code was very poorly written in the first place."

Once again I don't see how this invalidates my statement. It was a poorly optimized application, so they optimized it. Your implication is the equivalent of "but it was optimized to run well so it doesn't count!". That's ridiculous. It ran faster, the why is irrelevant to the claim. Some developers optimize their applications well, this doesn't make their applications any less valid as a benchmark. And on top of that it's also not the only application that ran faster.

omega_rugal Wrote:wanna know the true story?

http://www.emulators.com/pentium4.htm

careful, very long article.

That's very interesting but I don't see how that invalidates my statement. Unless you're talking about the follow up articles, which I don't really have time to read.

The fact is there are plenty of benchmarks from that era floating around the web from hardware review organizations. Some are faster on the P4 compared to the alternatives of that time, some are slower. It depends on the type of application you're planning on running. However there is no question that a P4 willamette 1.5GHz will completely destroy a 600-900 MHz PIII in every or nearly every application. If you want to refute that claim please show me some reviews backed up with a wide range of benchmarking data that have concluded otherwise. From what I can tell the follow up articles here are more of a technical analysis of the architecture at a low level.

Also since this is now getting grossly off-topic I would advise that we continue this via PM if you're ok with that. Otherwise SS may send us a friendly "stop that right now" message.

gone pm

Has this issue been resolved or a work around been found?

my i7 with a hd 7750 are preforming far worse then expected.


even 1.5x isnt smooth.

What game? And also tell us what "far worse you expected" means XD

Have you checked to see if it's throttling?

Super Mario Galxy 1 and 2

Skyward Sword

New Super Mario



Expected was at least full 1080p (set in game res to auto per screen?) or at least 2x. Lowest setting AA or at least 100% capped fps (60) with 0 aa. and the best i can to match all other settings to what other players have mentioned in these forums. Am i missing some blatant checkbox that everyone but me knows about? I spent at least 3 weeks searching the dolphin-emu.org forums before developing the never to ask..

thanks, seriously, thanks -