P4 @ 1.7Ghz ROCKS!!!

THG showed the two flagship products (AMD TBird 1333, Intel P4 1700) neck and neck overall, although occasionally one or the other lost out in a benchmarks badly.

I think the P4 is becoming a serious contender at long last. What with the awful performance of the VIA KT266 and ALi's MaGiK, combined with the difficulties in finding 760 systems, RAMBUS and 850 are looking more like a competitor rather than an underdog!

I'm still not happy that Northwood will change the socket format. However, it will be smaller, faster, cheaper and cooler, and will certainly get people's attention. Intel may be onto a winner.

The P4 core can theoretically clock much higher than the Thunderbird's, or indeed the Palomino's. I still think the prices are a <i>little</i> too high though! I guess you pay a premium for thermal protection and SSE2.



~ I know the human being and fish can coexist peacefully ~

hehehe, step out of your little box! AMD is also a well know brand name.

"AMD...you are the weakest link, good bye!"

It probably doesn't have 10% of the recognition that Intel does though. I could ask my Mum if she'd heard of Intel and she'd most likely say yes, she's not going to be placing AMD in a hurry.



-* This Space For Rent *-
email for application details

Amdmeltdown is right, it will be very interesting to see the next generation palomino (i think its the hammer?) after it is "perfected" like the northwood should be. If the northwood is the true p4 then this will be a great battle of the cpu's...that sounded dumb but hey...

next palomino will be the thoroughbred.


<font color=red>"My name is Ozymandias, King of Kings:
Look on my works, ye Mighty, and dispair!"</font color=red>

Enthusiast Intel fans are all waiting either for the price drops on the 29th or for Northwood. We'll hear all about personal benchmark scores at that time.

-Raystonn

-- The center of your digital world --

What will all of you AMD lovers start to criticize about the p4 when its priced comparable to the Athlon? Which will happen soon.

AMD lovers will be able to talk about the price of memory. RDRAM is usually going to be more expensive than SDRAM & DDR RAM.

"And I m waiting for Intel to fail like they did with the 1.133."

It won't happen. The P4 is at the very beginning of its lifecycle. Those kinds of issues pop up at the end of the lifecycle of a CPU, when you start really overclocking the core.

-Raystonn

-- The center of your digital world --

Well, let's see...

If you go back and look at the benchmarks from this site comparing the 1.7 and 1.5 GHz P4 to the 1.2 and 1.3 GHz Athlon T-Bird, you see some interesting trends:

1) The P4 only shows superiority in 2 areas: video-based applications, and memory bandwidth. Note, of course, that despite having a data bandwidth channel nearly 3 times that of the Athlon (800 vs. 266 MHz), at best the P4 could only pull 2 to 2.4 times the memory through (see the Sandra2001 Memory benchmarks). Whether that's a latency issue or not is immaterial to the fact that, to all appearances, the P4 architecture (or possibly the RAM itself) isn't able to deliver all the advertised bandwidth to the PC.

2) While the P4 1.7 won more benchmarks than the Athlon 1.3, it did NOT stomp the Athlon. Winning 17 of the 28 benchmarks is just below 61% of them. That may be good enough for an election, but not for computers. More importantly, though, is that 7 of the benchmarks were from games, and involved conditions where the game card (in this case, a GeForce 2 Ultra -- surely nothing to sneer at) is the limiting factor, not the CPU. Without those benchmarks, the P4 only wins 12 out of 21, and only improves its win percentage to 62%.

3) The benchmarks, of course, are simply raw numbers. However, it's interesting to note what happens when you start anaylizing them. For example, if you scale the benchmarks to get the theoretical performance for a 1 GHz processor of that type (ie. multiply the benchmark by 1000 and divide by its CPU rating), the Athlon 1.2 GHz takes the top spot, including all the game benchmarks, with 13 of the 21 benchmarks (62%). Athlon 1.3 GHz took 6 of them (28.5%), with the P4 1.5 GHz taking the last 2 (9.5%) -- and those only being the Sandra2001 Memory benchmarks. In other words, the ONLY reason the P4 wins the benchmarks it does is because it has more CPU cycles per second. Scale an Athlon up to meet it, though, and the P4 will lose. In this case, less CAN be more.

4) Price comparisons aren't that great, either. Granted, Intel announced price reductions to their OEM's this week. Based on performance per dollar, the P4 1.5 GHz (NOT the 1.7) is a clear winner, with 16 of the 21 benchmarks locked in (76%). However, of those 16, the P4 1.7 GHz model comes in second place only 7 times to its slower sibling. The rest of the time, the Athlons take a close 2nd. More importantly, these are the announced price breaks to the OEM's, and don't reflect street prices. In fact, if you check for the latest prices on the street, you'll see that the OEM's are still selling them for more money -- nearly $400 for a 1.5 P4 and nearly $600 for the 1.7 -- and you don't necessarily get any RAM with that, either. Based on current street prices, the Athlons are the clear winners for performance-per-dollar still. And since Intel announced their price drops because the P4 sales were drooping thanks to Athlon sales, do you really think AMD will sit back and not announce any price breaks of their own?

5) And let's not forget the upgrade paths. AMD has already committed to having their next CPU's use the same Socket A technology. That means that they will be backwards compatible with the current Athlon motherboards, which is great for those people who like to build their own systems, or prefere to upgrade piece by piece as needed instead of buying a brand-new system every time they want to "upgrade". Intel, on the other hand, has already said the upcoming P4's and their next designs will NOT use the P4's socket design, even though they just made the change from the P3 to the P4. So, for an upgrader, buying a P4 is NOT smart, since any future CPU purchases from Intel will require a new motherboard as well.

6) If you look at the benchmarks, they tell the tale: if you're doing a lot of multimedia work, or like playing your 3D games at really low resolutions, then a P4 might be better than an Athlon, depending on the game or application. However, if you analyze the original benchmarks again, despite having a 400 MHz advantage over the Athlon 1.3, the P4 1.7 only shows that advantage in a CPU-intensive benchmark once: the Flask MPEG encoder. The best it gets beyond that (save in the memory benchmarks -- which reflect the FSB and RAM speeds, NOT CPU speed) is only 16% better prformance. And so what if your computer can get nearly 230 FPS in Q3A, as opposed to nearly 198? You don't buy a super-fast processor to get great 3D performance at high resolutions, you buy a great 3D card. It goes back to the price thing, too. If I'm only getting a 16% increase in performance for a chip that's rated nearly 31% faster, and I have to pay nearly twice as much for the chip, it's not worth it.

Show me a P4 rated at the same speed as an Athlon, that costs only 5-10% more than the Athlon, and delivers 10-20% more performance than the Athlon, and you can say that the P4 kicks Athlon butt. Since you can't, stop the cheerleading and only quote facts.

another well thought out intellectual argument...
what more could we expect from the intellectual giant
amdmeltdown?
oh deary deary me

ThePoo!

i just wish you would say something constructive.
hell flame amd all you like, but back it up with data, facts, reviews AND refrences etc

p.s.
"you know i mean"?
is there meant to be another word in there? it doesnt make much sense

ThePoo!

<<i just wish you would say something constructive.
hell flame amd all you like, but back it up with data, facts, reviews AND refrences etc>>

Besides being mentally retarted I think he's too lazy to do that.

I'm not sure which sucks more - the P4 or you. It's truly a toss up, pun intended.

Intel does tend to use the Microsoft theory on quality control for release - use the public, there's more of them!

Actually I quite admire all manufacturers trying to release products into the PC world. The sheer size and volume of interoperability issues are incredible. I don't happen to like companies that release unready products to the masses just to get market position - VIA, Intel, Microsoft etc.

-* This Space For Rent *-
email for application details

"Whenever there is a problem, there will be an <b> immediate </b> recall".

Like with the P60 FPU glitch ? not quite immediate, was it. Also, to me quality is about creating products that dont require recalls. Im sure it can happen to any company, but the P60/820/P3 1.13 recalls are not exactly signs of devotion to quality to me.. rather time to market issues.

The P60 glitch was maybe a storm is a glass of water, and granted, very hard to detect.. but the 820 was clearly rushed to market (or how else must I feel about a disabled ram slot ? How hard is it to insert 3 sticks and see if it works ? And then there was this MTH issue.. not too hard to test either) The 1.13 was an even more obvious example.

Intel seems to value its reputation for quality only when it can afford the time. WHen its under pressure, it seems to quickly forget about decent quality control.