Aces High Bulletin Board
General Forums => Hardware and Software => Topic started by: bloom25 on July 12, 2001, 02:52:00 AM
-
You guys remember when the Celerons first came out and they outperformed the top end P2s? I guess Intel must have learned something from that. Take a look at this:
http://www.theinquirer.net/11070112.htm (http://www.theinquirer.net/11070112.htm)
What do you figure Intel refuses to release these until they are obsolete as well? They are out in Japan though. I guess the Japaneese are not stupid enough to buy P4s because of nightmares involving golf clubs, vaccuum cleaners, and blue mimes.
I sure hope this doesn't backfire too badly on Intel. I actually am getting kind of tired of AMD's continual delaying of their newer stuff because of Intel's bungling. Unfortunately for AMD their stock holders are not happy either, the other day I think AMD stock was down over $7.00 a share! :eek:
Consumers generally come out ahead when these two are fighting each other by releasing chips at higher and higher clock speeds. Unfortunately for us, both Intel and AMD are afraid to let this continue. Intel and AMD have both absorbed heavy losses to their stock prices recently as a result of price competition. When have you EVER been able to buy the most powerful PC processor for less than $200! (Tbird 1.4 GHz being the most powerful right now.)
What do the rest of you think?
-
It is all confusing to those of us who don't follow this stuff closely. When ya decide to upgrade to a faster system, heheeeee, by the time the parts arrive at your house, the system is dated!! Geez, this stuff is changing fast!
-
BUY LOW! SELL HIGH!
xBAT
-
It also means the 1.4 Gig Tualatin will likely lay the smack down on a T-bird doesn't it?
Actually it would be good if Inten stumbled a bit more and AMD had a chance to crank a bit of profit for 6 months or so. I'm sure I don't have to point out how awesome it is to have good competition in the CPU space for a change. :) We've been kicking the toejam out of Moore's law in terms of raw Mhz lately.. we're likely right on in terms of computing power though.
-
I was wondering that myself Leph, but the p6 core will never be as powerful as the k7 core at the same clockspeed. The k7 is just plain and simple much more powerful. (If you want to know why, you might want to look at www.arstechnica.com (http://www.arstechnica.com) .) All this goes to show is just how bad the p4 really is.
(k7 features over the p6 core include fully pipelined FPU. 9 Instructions in flight at one time vs 5. Larger L1 caches. 72 entry vs 40 entry TLB. Better branch prediction, and above all else the EV6 bus protocol.)
I wish AMD would release the palomino now. They could, but they are waiting on Intel to release faster chips before doing so.
-
Yep, I've read through that before.
The P4 is marketing driven, that much is obvious. They crippled it to allow it to scale to higher Mhz because that's how the marketing works moreso than with real performance. I think the market is a bit more educated these days at the top end, the enthusiast market... but it remains to be seen if that will influence the mainstream purchases. Doubtful that it would IMHO. The fact is that most folks will compare the numbers, so it's most important to have a higher number. Silly, but that's the way things work.
I agree that on a per Mhz basis the K7 is still superior... although not by tons in real performance. The trick is that the Tualatin will be "pretty close" in real world stuff, and should scale up much better than the old P3's.
Anyway, I just like to see good competition taking place. I hope AMD can make some decent profits, and fund taking the fight to Intel for the long term. A good 2 party battle has been great for us consumers. :D
-
There really isn't much crippling involved with the P4. Its more of a direction. The P-4 will still run your games plenty fast, but it is actually tailored towards multi-media type applications. The main roadblock for highlighting its capabilities is a classic story... nobody <not many> has used the optimized features yet.
The same thing was true when the P-3 was released... the P-2 was actually faster on the benchmarks present at the time. It wasn't until people started using all of the features the P-3 introduced (many used in that wonderful AMD processor) that it finally started blowing the doors off of the P-2.
As far as the article goes... I wouldn't give it much credibility. I haven't seen much in regards to us making a 1.4 MHz P-3. I do know some 1.266 and 1.3 MHz stuff may be coming... but we definately aren't "withholding" any 1.4 MHz P3 processors right now. Don't get too hung up on hypothetical strategy... especially when its actually more mythological.
AKDejaVu
-
Hmmm... my local shops are listing 1 Gig and 1.2 Gig Palomino's in stock.
You mean they are just not shipping the faster ones?
http://www.mysterybyte.com/pricelst.htm#114 (http://www.mysterybyte.com/pricelst.htm#114)
I notice he lists them as "Server Processors", what are these Xeon type deals with craploads of cache?
-
OMG... your one of THEM!! The evil empire!
xBAT
-
We are not evil... nor are we good.
Think of us more as a non-discriminating soul vacuum :D
AKDejaVu
-
Also... something on a slightly different subject..
AMD has recently announced that they expect an opperating loss next quarter. They blame Intel for driving down the price of processors.
I still can't help but laugh out loud at that.
AKDejaVu
-
Originally posted by Lephturn:
I notice he lists them as "Server Processors", what are these Xeon type deals with craploads of cache?
I dunno, but Holy toejame!...
INTEL P3 1GHz 133FSB 256K Cache, OEM $358.00
I just bought the same, only in a retail box for less than $200.
SOB
-
Originally posted by SOB:
I dunno, but Holy toejame!...
INTEL P3 1GHz 133FSB 256K Cache, OEM $358.00
I just bought the same, only in a retail box for less than $200.
SOB
Must be Canadian!
-
ahh, cc, Monopoly money. That explains it! :D
-
Deja, did you mistype or did you actually think the p3 benched worse than the p2? You should know that the p2 and p3 (at least the Katimi (sp) core) were totally identical except for SSE 1 support and the processor serial number. They benchmarked identically unless SSE was used.
I also find it hard to believe the P4 is specifically designed for video, etc. That's just putting a positive spin on the fact that the pipeline is very long, so the applications that get the least performance hit are going to be those that "behave well" with regards to branch prediction. Anyone wan't to guess what applications those would be? :D Besides some SSE 2 instructions, I personally can't find a single architectural feature of the Willamette (current P4s) core that is "optimized for video." Nah, it's just that in general video applications perform the same algorithms time after time and generally the conditional statements are very easy to predict.
On the other hand the p4 does have a few very excellent features, but the current clock speeds the chip is running at prevents them from helping the performance much. (I am of course speaking of the ALUs running at 2x speed. As clock speed increases the integer performance difference between p4 and Athlon should shrink up some more.) Right now there is hardly anything the p4 does better than the Athlon, even if the p4 has a 200 - 400 Mhz clock speed advantage.
Personally I'd say if AMD falls behind, which I don't see happening for a while, AND once the 478 pin socket and the new .13 micron P4s (I think Northwood is the name being used) are in full scale use the p4 may eventually be OK in my book. Right now in it's current form it just stinks IMHO. :(
I'm really unhappy with both AMD and Intel right now, but at least I don't feel like I've violated some ethical code in telling someone the Athlon is better for the money than the p4. I'd have a pretty hard time telling ANYONE to buy a p4, mainly because of no upgrade path for all those who bought systems with Socket 423 and Rambus ram.
-
Good post bloom25 ... Thanks for all the updates you give us.
sprint
*MOL*
-
Deja, did you mistype or did you actually think the p3 benched worse than the p2? You should know that the p2 and p3 (at least the Katimi (sp) core) were totally identical except for SSE 1 support and the processor serial number. They benchmarked identically unless SSE was used.
No mistype. Come on bloom.. use your memory... remember the articles. I do.. I read them every day.
The uproar when the P3 was released pointed to the fact that the P3 lost to the P2 in some tests (not all... not many... but some). Much like the stir that occured when the K6-2 managed to beat the P-3 in one type of test and was dubbed a "better choice for office applications" by many authors as a result.
Remember bloom... remember?
I also find it hard to believe the P4 is specifically designed for video, etc.
Sorry to hear that. Of course, I said multimedia.. not just specifically video. Besides.. you believing one way or the other doesn't really mean its not true. Kinda equates to "feel".
I personally can't find a single architectural feature of the Willamette (current P4s) core that is "optimized for video."
Are you quoting me? I can't see where I said what you put quotes around.
Nah, it's just that in general video applications perform the same algorithms time after time and generally the conditional statements are very easy to predict.
LOL! I find this statement rather funny. It doesn't matter how fast they do it if you think its easy?
On the other hand the p4 does have a few very excellent features, but the current clock speeds the chip is running at prevents them from helping the performance much. (I am of course speaking of the ALUs running at 2x speed. As clock speed increases the integer performance difference between p4 and Athlon should shrink up some more.)
Yep.. the fireball core is pretty spiffy. Its not really the main advantage of the chip, however. The main advantage is that the P-4 has room for development whereas the P-3 is at the end of its life.
Right now there is hardly anything the p4 does better than the Athlon, even if the p4 has a 200 - 400 Mhz clock speed advantage.
OK.. but I fail to see what this is in something adressed to me. I did not compare the P4 (or P3) to the Athlon in any way.
As a matter of fact, you'll not find me ever making any serious processor recomendations to people. I have not knocked the Athlon processor and have not argued with anybody recomending them. Its a good processor. Maybe you are reading more into what I posted than what is actually there.
I'm really unhappy with both AMD and Intel right now, but at least I don't feel like I've violated some ethical code in telling someone the Athlon is better for the money than the p4. I'd have a pretty hard time telling ANYONE to buy a p4, mainly because of no upgrade path for all those who bought systems with Socket 423 and Rambus ram.
I simply don't tell people what to buy. That way I don't feel guilty when someone's computer won't boot because of a faulty chipset... faulty power supply... lack of support for a video chipset... or whatever reason.
To be honest... there is so little difference between the top end processors anymore that people aren't able to notice. Programs are created to tell them that something is .2% faster in a certain test because somehow... that has become important. The truth is... it doesn't matter.
Go with the reliable motherboards and chipsets. The processors are no longer the main cost nor the main concern. Of course... many would have you think so... but really... where are you going to pick up the most speed for normal use or even more advanced use?
AKDejaVu
-
Originally posted by AKDejaVu:
Go with the reliable motherboards and chipsets. The processors are no longer the main cost nor the main concern. Of course... many would have you think so... but really... where are you going to pick up the most speed for normal use or even more advanced use?
AKDejaVu
hence the reason i buy intel. I'll wait about 6 months on a P4 tho :D
<S!> Hamish
P.S. now i know who to call wheni am lookin to upgrade my puter again :D
-
Whoa, AK. Hold it there, in no way was a responding directly to your post there. :) (Except for the very first part.)
I still stand by my early p3 = p2 performance though. :) If benchmarks were showing otherwise, they did them wrong. ;)
You are totally correct about the k62 IMO, the p3 is far superior to it. THe k62 was just a very inexpensive processor that in the end started the price wars when Intel responded with the Celeron.
When I said "video", that could for the most part be taken as multimedia, and I wasn't quoting you, I was quoting an Intel press release some time back.
The conditional statements I was speaking of are the if, for loop, while loop, etc statements in programs. All modern PC processors have a branch predictor unit that attempts to guess the outcome and begin execution of the following instructions prior to them actually being needed. Most multimedia ( :) ) applications, say for example an Mpeg 4 encoder, are going to be performing the same series of calculations over and over again.
Please don't take anything I post about Intel as being an attack against you, they aren't in ANY way. On the contrary it's good to hear from someone with first-hand knowledge from within the company. :)
(BTW: I'll find some P3 vs P2 articles for you. :) )
-
I think Toms pretty much spells it out on the second page.
http://www6.tomshardware.com/cpu/99q1/990223/index.html (http://www6.tomshardware.com/cpu/99q1/990223/index.html)
Benchmarks:
http://www6.tomshardware.com/cpu/99q1/990223/index-04.html (http://www6.tomshardware.com/cpu/99q1/990223/index-04.html)
The P2 450 and P3 450 score identically most of the time.
[ 07-14-2001: Message edited by: bloom25 ]
-
I must be really bored today, I just keep posting and posting ... ;)
AKDejaVU, I think you are totally correct when you say the the top end processors are within .2 percent of each other. I admit it is kind of silly to compare benchmarks that are showing a .1 % perfomance increase in some synthetic test between a 1.4 Tbird and 1.8 P4. Neither top of the line CPU today really destroys the other in a particular area where the average user is going to notice it. (With a couple exceptions: The p4 encodes Mpeg 4 quite a bit faster than the Athlon, and the Athlon is quite a bit faster in FPU intensive engineering software. )
As for chipset reliability, Intel has had a pretty good record compared to say, SiS and VIA. Via has gotten better recently, and for the most part Intel has stayed about the same. All things being equal, the MB and it's chipset play the biggest role in stability. I'd say all the current chipsets for P3 (except i820), Athlon, and P4 are pretty good. If you buy a good motherboard, an AMD system is going to be just as stable as an Intel system.
-
Toms is biased. He has the inate abbility to mix facts with spin to prove his point. He has been "anything but Intel" since the mid 90's.
Other than that.. most of the hardware sites have been trying to make something of nothing. I do read them occasionally to try to find out about latest hardware releases... but the tests they do on processors are so abstract that they really hold little relevance to virtually anyone that is stopping by the site.
Of course, people will disagree... they'll set up benchmark tests so that people can measure their computer's peformance against others. These tests must exist because most people wouldn't really notice the difference during actual gameplay. Yet somehow, the abillity to get 115 FPS in AH vs 110 FPS has become important.
Sites like Tom's are to blame. Almost directly.
The "I must hate everything that is Intel because they are so big" mentallity is finally toning down. Intel is actually grateful to AMD for that. I would sit back at work and read article's on Tom's in regards to how we were threatening motherboard manufacturers, pc manufacturers, doing deceitful things with our processors... whatever. Any move Intel made was spun into. Of course, all this was going on during an Anti-Trust investigation by the U.S. Government... the ideal time to start strongarming people.
At Intel, almost down to the man, we believe AMD has a better designed chip. They took 75% of our design and were able to improve on it due to the aquisition of another chip manufacturer's design team and came out with an excellent product.
At Intel, almost down to the man, we know that AMD will never have the manufacturing capabilities (neither in quanity or quality) that we have. Nobody will even aproach us there for some time. Nobody can make chips as for the low $$$ we can.
We're working to completely change architecture to meet demand. I don't know exactly what direction that is going, but I have a tendancy to think that making things specifically for benchmark tests is going to go by the wayside. Its encouraging that Intel has moved from the "lets just make the same old chip smaller and faster" policy. But lets face it.. until last year there really was no need to do so.
The strange thing is the "until last year" part. These same arguments have been raging even when Intel had a far superior product. Yet, people still recommended AMD chips (K5 or K6) citing some article that someone wrote somewhere... when in fact... those chips simply sucked.
With an argument that is pretending to be about the "here and now"... I find it ironic that so many are harboring 8 year grudges against Intel and still proclaiming to be unbiased.<not a reflection on you Bloom.. just a general statement>.
...ah... enough of the soap box.
Glad we weren't talking about overclocking... I woulda gone on for years ;)
AKDejaVu
-
LOL, yeah I think Tom has finally flipped. :D That was just somewhere I knew would have some p2 vs early p3 benchmarks.
AMD did hire the designers who previously came up with the Alpha from Digital (Compaq now). That's why the K7 cores use the EV6 bus. (Though there are rumors flying that Intel is considering purchasing this part of what used to be Digital. I'm sure AMD would really love to pay royalties to Intel for the use of the EV6 bus! :D )
As for your above post, I agree with you 100%.
Edit: Should I mention that of the 4 PCs I own, 3 of them are Intel? ;) For it's time the p6 core was great, it's time has passed though. I own a Tbird 700 (soon to be 1200), P3 450, Celeron 466, P1 100.
[ 07-14-2001: Message edited by: bloom25 ]
-
Though there are rumors flying that Intel is considering purchasing this part of what used to be Digital. I'm sure AMD would really love to pay royalties to Intel for the use of the EV6 bus!
I hadn't heard about a purchase of the Alpha technology. I'd only heard that Compaq is no longer going to be making them, and they are licensing Alpha technology to Intel.
I don't think Intel would be allowed to buy the Alpha. The US Government would definately block that one. But it wouldn't come to that... the government's desire to nail Intel with anti-trust charges pales in comparison to Intel's desire to avoid being charged ;)
AKDejaVu