This is just another way of saying that the Pentium 4 is broken except for multimedia, which is pretty much what I have been saying all along. The Athlon has all-around good performance, and if you look at price/performance ratios, the Athlon totally wins.
The only apps that the consumer needs that demand raw CPU power are multimedia apps. You do not need a several GHz processor to run business apps. You need it for: Gaming (ala Quake III), Encoding (ala FlasK), Decoding, etc.
More technical arguments about emulators.com page
on
Pentium 4 Under Linux
·
· Score: 1
"The restrictions on ordering of instructions is again something of a compiler issue, IMHO. For example, if you're dealing with a RISC processor, ordering of instructions is very important. It's something the compiler can do before hand - the job doesn't need to be done by the CPU. Leaving it to the compiler saves transistors which can then be used by something else, like the SSE2 units which the article glosses over.
Again, with lack of execution units he's focusing primarily on the weak FPU, and ignores the very fast SSE stuff. With the release of ICL5 which is smart enough to parallelize loops for SSE2 by itself, there's no excuse for that.
The fourth point was the small instruction cache. Intel doesn't use a normal instruction cache on the P4, it uses what it calls a trace cache. P2, P3, P4, Athlon, they all decode instructions into smaller micro-ops, as you know. Unlike the other instructions, the P4 doesn't cache x86 instructions at all. It caches the decoded micro-ops in the trace cache instead, saving the job (and several pipeline stages) of decoding instructions. The theory is that because the P4 works, for the most part, on the level of micro-ops instead of normal instructions as earlier instructions are, it doesn't need as much cache."
The Pentium 4's L1 cache is significantly different than the Athlon's. The Pentium 4's is tracecache.
The Intel engineers looked at the pros and cons and decided on the lowest figure for L1 cache. There's no reason why they wouldn't include the extra space if there was a noticable performance delta with it.
Sometimes you just gotta think things through logically...
BTW, SDRAM support will be here in 1-2 months. From the preliminary benchmarks, RDRAM/is/ better than SDRAM now that t he frontsidebus is fast enough for the extra bandwidth to matter.
The emulators.com guy is just pissed off because the Pentium 4's core doesn't work as well with emulators than the P6 core did. It's more for multimedia, not for heavy logic programs like emulators are.
Intel is Nvidia, and AMD is ATI. ATI has very promising upcoming chips and some alternative solutions to fixing the problem, just like AMD. Intel and Nvidia are both the dominant makers, but ATi/AMD are gaining on them.
Anyway, AMD is adopting SSE/SSE2 now too, so why WOULDN'T you optimize for it? You're not just optimizing for the Pentium 4, you're optimizing for all future Intel 32-bit processors, and probably upcoming AMD 64-bit processors too.
I really wish people would stop linking to the emulators.com article, it's akin to linking to the Weekly World News.
I'm not going to waste my time explaining everything about the article, but I'll summarize it: Who do you think knows more about how to design a processor: A guy who makes emulators for a living, or Intel?
Example: The L1 cache thing. Did you know if the L1 cache on the Pentium 4 was increased, the latency also increases? Did you know that the higher latency would hurt performance more than the additional cache? Probably not, but then again, neither did this emulators.com guy. Why? Because he designs emulators for a living, not microprocessors.
Actually, as much as you'd like to think that, no console maker makes money on the box itself. They all make money off software designed for it. Sony and Nintendo all lose money on the hardware.
If MS can flood the market with XBoxes, then more people will buy more titles for the Xbox, and the company would make more money.
All console makers use the same strategy, except MS is fortunate enough to have huge cash reserves to splash the name all over the place.
In the free market, business DO fail, peple DO lose their jobs. Market conditions change. Live with it.
MS made things more convenient for the consumer by bundling components which are necessary for the operation of a modern computer (thus OPERATING system), and unfortunately for Quarterdeck, TrumpetWinSock, Stac, (and some might argue Netscape), they were simply inconvenient. You either used the free bundle MS gave you, which worked fine in most cases, or you could acquire (or sometimes BUY) a 3rd party program to do the exact same thing. Why would you do that?
If anything, this should encourage innovation, because the competitors must add features WORTH switching to, and not giving us the exact same features MS is giving. How does this stifle innovation in any way?
The price of MS' consumer OS' have consistently been around the same mark, despite the fact that their features and bundled software have expanded dramatically.
It increases MS' costs, but historically this doesn't add on to the OS' price.
I also had many problems in Linux with 0.7, including a few times where I had to kill the process, and also some problems with 0.7 and 0.8 in Win9x that took down the whole OS with it.
The OS didn't do anything wrong, Mozilla's buttons simply didn't work. I clicked them, the animation worked, but nothing else did. Moments later, the Task Manager reports it's not responding.
I used Mozilla 0.9 for about 20 seconds in Windows XP Beta 1. Then...*CRASH*
It prompted me about setting it for the default browser or something, and none of the buttons worked. A few seconds later, it just stopped responding. Had to kill the process.
Nice going, Mozilla team.
Is this some sort of record?
I think he's going deeper down. The operating systems the people used were very easy to use, which allowed things like AOL and Netscape to take off and take advantage of the WWW and internet.
I never said that Linux was more unstable, I said that Cold Fusion + Win2K was more stable than when they used it with Linux (which, I stand corrected, was Solaris before. My bad)
I use my webcam to be triggered by motion, and find it fascinating to see what the maid does in my room.
Apparently she likes some of my music.
Re:1.7 GHz is a lot like a 1.2GHz Athlon
on
What 1.7Ghz Is Like
·
· Score: 1
How do you explain the FlasK benchmarks from Toms Hardware, then?
A nearly 4x increase by adding SSE2...
Re:1.7 GHz is a lot like a 1.2GHz Athlon
on
What 1.7Ghz Is Like
·
· Score: 1
Of course, we all really need the 1.33GHz Athlon to be faster in those all-important MS Office and other critical business applications!
It should also be noted that the Pentium 4 scores better in 3DMark 2001, as well.
It all boils down to this: The Athlon 1.33GHz may be faster in yesterday's apps in general...but why do we need the old apps to run so damn fast? They're still/more/ than playable/runable on the Pentium 4 systems. The difference you get with the Pentium 4 is that future SSE2-coded applications will become mainstream shortly (AMD licensed it, as well) as a replacement to the x87 FPU, and performance will take off for future apps, where it is actually needed.
Slashdot Mirror
The only apps that the consumer needs that demand raw CPU power are multimedia apps. You do not need a several GHz processor to run business apps. You need it for: Gaming (ala Quake III), Encoding (ala FlasK), Decoding, etc.
Again, with lack of execution units he's focusing primarily on the weak FPU, and ignores the very fast SSE stuff. With the release of ICL5 which is smart enough to parallelize loops for SSE2 by itself, there's no excuse for that.
The fourth point was the small instruction cache. Intel doesn't use a normal instruction cache on the P4, it uses what it calls a trace cache. P2, P3, P4, Athlon, they all decode instructions into smaller micro-ops, as you know. Unlike the other instructions, the P4 doesn't cache x86 instructions at all. It caches the decoded micro-ops in the trace cache instead, saving the job (and several pipeline stages) of decoding instructions. The theory is that because the P4 works, for the most part, on the level of micro-ops instead of normal instructions as earlier instructions are, it doesn't need as much cache."
The Intel engineers looked at the pros and cons and decided on the lowest figure for L1 cache. There's no reason why they wouldn't include the extra space if there was a noticable performance delta with it.
Sometimes you just gotta think things through logically...
BTW, SDRAM support will be here in 1-2 months. From the preliminary benchmarks, RDRAM /is/ better than SDRAM now that t he frontsidebus is fast enough for the extra bandwidth to matter.
The emulators.com guy is just pissed off because the Pentium 4's core doesn't work as well with emulators than the P6 core did. It's more for multimedia, not for heavy logic programs like emulators are.
Intel is Nvidia, and AMD is ATI. ATI has very promising upcoming chips and some alternative solutions to fixing the problem, just like AMD. Intel and Nvidia are both the dominant makers, but ATi/AMD are gaining on them.
Anyway, AMD is adopting SSE/SSE2 now too, so why WOULDN'T you optimize for it? You're not just optimizing for the Pentium 4, you're optimizing for all future Intel 32-bit processors, and probably upcoming AMD 64-bit processors too.
I'm not going to waste my time explaining everything about the article, but I'll summarize it: Who do you think knows more about how to design a processor: A guy who makes emulators for a living, or Intel?
Example: The L1 cache thing. Did you know if the L1 cache on the Pentium 4 was increased, the latency also increases? Did you know that the higher latency would hurt performance more than the additional cache? Probably not, but then again, neither did this emulators.com guy. Why? Because he designs emulators for a living, not microprocessors.
Does anyone else find that funny?
Anyway, I didn't even know there was a 1.2GHz Pentium 4.
That would work, except sometimes it's nice to play with no lag. :)
The reason is your browser on Linux doesn't have 128-bit encryption. If you did, it would work.
If MS can flood the market with XBoxes, then more people will buy more titles for the Xbox, and the company would make more money.
All console makers use the same strategy, except MS is fortunate enough to have huge cash reserves to splash the name all over the place.
Besides, that's a pretty lame response. Loki isn't a game platform, Xbox is. You're comparing apples to oranges.
You really don't need the X-Box to play TuxRacer.
X-Box is a gaming console. Linux is...not. Linux on Dreamcast is just so wildly useful, just as it will be for the XBox...
People have been saying this for years. It's a fad. Do you REALLY care what powers your toaster?
We seem to be getting on just fine without them.
In the free market, business DO fail, peple DO lose their jobs. Market conditions change. Live with it.
MS made things more convenient for the consumer by bundling components which are necessary for the operation of a modern computer (thus OPERATING system), and unfortunately for Quarterdeck, TrumpetWinSock, Stac, (and some might argue Netscape), they were simply inconvenient. You either used the free bundle MS gave you, which worked fine in most cases, or you could acquire (or sometimes BUY) a 3rd party program to do the exact same thing. Why would you do that?
If anything, this should encourage innovation, because the competitors must add features WORTH switching to, and not giving us the exact same features MS is giving. How does this stifle innovation in any way?
It increases MS' costs, but historically this doesn't add on to the OS' price.
It's almost cheaper to just buy a mag. Haaaah.
Couldn't be more wrong. Every single app except Mozilla functions properly.
Same problem in WinMe. Go figure.
I also had many problems in Linux with 0.7, including a few times where I had to kill the process, and also some problems with 0.7 and 0.8 in Win9x that took down the whole OS with it.
The OS didn't do anything wrong, Mozilla's buttons simply didn't work. I clicked them, the animation worked, but nothing else did. Moments later, the Task Manager reports it's not responding.
I used Mozilla 0.9 for about 20 seconds in Windows XP Beta 1. Then...*CRASH* It prompted me about setting it for the default browser or something, and none of the buttons worked. A few seconds later, it just stopped responding. Had to kill the process. Nice going, Mozilla team. Is this some sort of record?
I think he's going deeper down. The operating systems the people used were very easy to use, which allowed things like AOL and Netscape to take off and take advantage of the WWW and internet.
I never said that Linux was more unstable, I said that Cold Fusion + Win2K was more stable than when they used it with Linux (which, I stand corrected, was Solaris before. My bad)
Apparently she likes some of my music.
How do you explain the FlasK benchmarks from Toms Hardware, then? A nearly 4x increase by adding SSE2...
It should also be noted that the Pentium 4 scores better in 3DMark 2001, as well.
It all boils down to this: The Athlon 1.33GHz may be faster in yesterday's apps in general...but why do we need the old apps to run so damn fast? They're still /more/ than playable/runable on the Pentium 4 systems. The difference you get with the Pentium 4 is that future SSE2-coded applications will become mainstream shortly (AMD licensed it, as well) as a replacement to the x87 FPU, and performance will take off for future apps, where it is actually needed.
Just so you know, James Gosling is the inventor of Java...