Slashdot Mirror
Intel Discontinues Extreme Edition P4
bizpile writes "X-bit Labs reports that Intel is stopping production of its Extreme Edition Pentium 4s. The company said in its statement sent to clients, 'Market demand for the Intel Pentium 4 processor Extreme Edition supporting Hyper-Threading technology 3.20GHz with 800MHz processor system bus in mPGA478 packaging has shifted to higher performance Intel processors.'"
Since the arrival of Doom 3 I think we know to where the REAL market demand shifted.
Market demand has shifted to higher performance, and cheaper AMD processors ;-)
-- duh
Listen Rummy. I'm gonna say it plain and simple. Where'd you pinch the hooch? Is some blind tiger jerking suds on the side?
Market demand for the Intel Pentium 4 processor Extreme Edition supporting Hyper-Threading technology 3.20GHz with 800MHz processor system bus in mPGA478 packaging has shifted to higher performance Intel processors.
Which is That..... and we made it especially for dell and we dont like dell anymore...
From the statement in the article, it sounds like they're just discontinuing the 3.20Ghz, socket 478 version of the chip, not the whole P4 EE line.
Why would I need a 2 meg l3 cache on a gaming processor that only increases performance by 1-5%? Combine that with extrordinary cost, cooling measures, the size of the proc itself, and power consumption and failure to sell is predictable.
Candy-Coated Knowledge
You mean everyone has been heading for the less expensive, better performing AMD chips, from which you are now copying instruction sets.
Great ideas often receive violent opposition from mediocre minds. - Albert Einstein
Start a new paragraph for christs sake.
Saying Java is nice because it works on all OS's is like saying that anal sex is nice because it works on all genders.
Intel needs to wake up to reality and stop regurtiation of the old wayward windbag speech, we here time and time again.
It's All Politics
Must be shifting to the new Sun technology that got double posted...
Bye!
SeqBox
Honestly, I think that the only thing that as far as cost and performance ratios go, AMD has the upper hand. People who keep up with the industry are (I assume) fairly well aware of this fact.
From what I can tell, intel's only remaining advantages are in niche markets (not consumer desktops), and the fact that most people buying consumer-level desktops haven't even heard of AMD. I doubt that AMD will be able to overthrow intel's brand-recognition supremacy, but intel will be facing some tough decisions if they do...
"Operating systems suck: you're better off using only the BIOS" --trainsaw.com
I assume you mean like a Sparc or Power, right? I'm happily an intel shareholder, but seriously, AMD's just as much a "real" cpu as any other x86-braindamage.
HyperThreading simply gives better multi-tasking, and games don't require that.
:)
But for real work - I prefer real CPU.
lol. Moderators, do your worst!
There's no mention whatever of the whole Extreme Edition line being stopped, in fact they recently said they would be making further new ones in the near future... This is mentioned (with new FSB and clock speeds) here and here and here, for instance - and all quite recently.
> thus I use Intel. I tested AMD once.
Yup, you're right, them XScales sure beat those K5s
While you're at it, you'd better check your Microsoft using your Norton, I think your Adobe just got trashed.
Advanced users are users too!
But for real work, I prefer real computer.
"VAX - when you care enough to steal the very best"
> While you're at it, you'd better check your Microsoft using your Norton, I think your Adobe just got trashed.
Norton? That performance hog? pfft. Symantec Antivirus Corporate Edition - Now there's a nice light-weight AV client.
Anyway, windows sucks by nature for multi-tasking due to its awful memory management, so I'm not even going there. 1gb RAM and it eats into swap on first chance. Only way to avoid it is disable swap completely.
Madcat.
...everyone could take one look at the specs and see that it wouldn't sell well if at all. Intel isn't that stupid either. But it managed to squeeze the little extra on the performance graphs comparing the "best" AMD vs Intel processors, cost and other things be damned.
As processor speeds exceed what most people use them for (multi-GHz machine to check mail and surf web, sigh), it is all about perception. Most people would be happy with both AMD and Intel running their box. Of course the slightly more tech-savvy saw what was going on, but the average Joe doesn't know.
To him, Intel is still the strong market leader and AMD the outsider. Intel fumbled once with the Athlon being the fastest thing around, they're not making the same mistake twice. If they showed signs of weakness, it could cost them vastly more in "mainstream" P4 sales than keeping a EE line to put on charts.
Kjella
Live today, because you never know what tomorrow brings
Market demand for the Intel Pentium 4 processor Extreme Edition supporting Hyper-Threading technology 3.20GHz with 800MHz processor system bus in mPGA478 packaging has shifted to higher performance Intel processors.
Try saying that with a mouth full of cheese three times fast...
And tomorrow the stock exchange will be the human race
All insdustry watchers, maybe except Toms Hardware, saw the EE as a trick to get the performance crown from AMD again. At (literally) any price. I don't know if it sold; you must really want to have the ultimate performance at any price to buy one.
Now that they've released faster processors that are up to par on performance with AMD (and removed the GHz speeds from the processor names) there is no need for the extreme edition any more. So now they don't have to sell server chips to make up for bad performance on the PC front.
Anyway, windows sucks by nature for multi-tasking due to its awful memory management, so I'm not even going there. 1gb RAM and it eats into swap on first chance. Only way to avoid it is disable swap completely.
So you prefer keeping unused portions of kernel in the system memory? Good memory management tends to use swap to page out unused pages so more physical memory is left free for the applications that really need it.
Hyperthreading is a neat hack, nothing more. It seems designed exclusively to fool non-techies into thinking that they have a 2 or 4 way system when in fact they have half the number of actual processors, and additional really crippled ones.
That combined with the price means my last purchase was a pair of dual opteron workstations.
Martin Brooks / Slayer99 #linux / UIN 2178117
SPARC? You have to be joking. Intel's Itanium 2 is twice as fast as a SPARC, and POWER5 is faster again.
P4 and Opteron would also make a meal of it.
So you prefer keeping unused portions of kernel in the system memory?
I'm pretty sure Windows, like most operating systems, has a non pageable kernel. So.. yes.
try rendering a heavy scene with mentalray 3.3 on a dual xeon, then on a dual opteron, and tell me which one is for gaming and which one is for serious work...
Well, actually Windows does page out unused portions of kernel.
Actually, Windows 2000/XP have executive paging enabled by default (swapping out the kernel), and has to be disabled through a registry tweak or a program that peforms said registry tweak.
If there was a joke in there somewhere, sorry.
It would be cool if it didn't suck.
We've got some people where I work who come from a proprietary unix background and therefore don't really know anything about PC's and PC hardware. I've had to explain hyperthreading to a couple of them on more than one occassion. They just can't get past the fact that "top" says there are 4 cpus and refuse to believe me when I say there are only two. These same people are also hold on to the illusion that their proprietary Sparc and SGI workstations are somehow architecturally superior to a standard PC. That was certainly true ten or fifteen years ago, but not today. Nowadays you're paying a lot more money for a lot less performance when you buy a Sun or SGI system. The only reason to buy one of those is if you have software that won't run on a standard system, and even then you're often better off switching to some other software package, especially in the long run.
if you want to spot me some hardware.
Fujitsu has good SPARC processors, though, and with 2048 processors they can beat systems with 1936 Itaniums (on the top-500 supercomputer lists); so you're making the same falacy of dismissing an brand based purely on its brand name.
Joking aside:
unlike the parent implied, you can do real work on an AMD chip;
unlike you implied, you can do real work on a SPARC based chip; and
unlike I implied (but thankfully for my stock), you can do real work on an Itanium.
P4's however are another issue altogether; hence this article about INTC dropping them. :)
What? You have to be kidding... You're talking out of your ass, buddy. Smells like astroturfing to me.
Slashdot is proof that Sturgeon's Law applies to mankind.
My God, the bloated POS "microkernel" on that monkey-created operating system must be fuckin' huge to justify the added complexity of a pageable kernel....
Hmm, no it seems that executive paging only pages out the NT executive (ie services on top of the microkernel). The kernel itself isn't pageable.
If there was an attempt to show of your coolness in there somewhere, sorry.
Fijitsu has better SPARCs than Sun, but the SPARC64 V isn't that great...
SPECfp 2000
2161 - Itanium 2 6MB 1500MHz
1642 - POWER4+ 1700MHz
1548 - P4 EE 3400MHz
1505 - Opteron 2200MHz
1096 - SPARC64 V 1350MHz
SPECint 2000
1666 - P4 EE 3400MHz
1408 - Itanium 2 6MB 1500MHz
1405 - Opteron 2200MHz
1077 - POWER4+ 1700MHz
776 - SPARC64 V 1350MHz
So basically you would use a SPARC if you are insane, have it given to you, or have legacy systems.
I don't know about the Fujitsu supercomputer - it could be an Itanium 2 system. Look at the SPARC64 system at number 22.
You're right. I mean, you'd think they'd both at least try to standardize on the same instruction set or something. Oh, wait.
I thought they shifted to the higher performing Athlon 64 FX.
If someone says he and his monkey have nothing to hide, they almost certainly do.
no you won't
Which would you prefer, inactive code sitting in main memory, or more space for buffers?
It would seem that as "Market Demand has shifted towards higher performance" means that only the 3.2 has been discontinued, I mean, if you were gonna go all out and buy an EE, you would buy the top of the line, not the bottom.
AMD is still more consistant...they only make 1 A64-FX at a time...the current one is the 53 based of top of the line Opteron tech, tweaked to run 1 proc as fast as they can ... sure you can get 51's but AMD is no longer making them. Like you pointed out, what good are 3 versions of an "Extreme" processor? [Who buys an "Extreme" in medium?] It's a case where Intel's marketing "3" of everything "just because" fails miserably.
You make it sound like the whole "Extreme Edition" processor line is cancled. This is not true. They are just dropping the oldest processor in the lineup, which was a little too slow to be considered "ultimate performance" like the rest of the lineup.
Intel still has a couple of advantages, and one of them is the heat problem that comes with AMDs. OK, in a normal case you'd never notice because you basicly have the room to fit in airconditioning if you wanted to. In a 1U case however, you have to stick to about 1cm (that's less than half an inch, for those who don't know the metric system) for a fan.
On a 1U with an intel on the motherboard, I've rarely had any cooling problems. On a 1U with an AMD I've on regular basis had problems. The money customers save on buying an AMD, they'll have to spend on buying a 2U instead of a 1U.
When it comes to desktop PCs I've found that home-consumers these days are sold a lot of AMD as well, although that may be because I know a lot of gamers. But the truth is that in local computer shops (which are horribly expensive, btw) I see a lot more AMD machines than I see intel. Of course, the brand machines will have an intel, but IMHO brand machines for home use are consumer-stupidity (unless it's warranty you're after).
After AMD whipped their butts in Doom 3 performance with much cheaper A64 parts, they threw in the towel.
Hail Eris, full of mischief...
E pluribus sanguinem
Why would it need swap for that? The kernel pages aren't going to be dirty -- they're executable, read only.
Good memory management doesn't write clean pages to swap. Since those pages haven't changed since they were pulled from disk, they can just as easily be pulled from disk as from swap.
Also, good memory management only swaps when more space is needed by active applications than is currently available. If I have gobs of RAM to spare, it shouldn't swap at all.
I don't know what windows is compulsively swapping, but compared to linux on the same machine, windows seemed to swap constantly. I have a lot of RAM in there, so I turned off swapping in windows, and haven't had any problems (but then again, I only use it to play games...).
Honest, I'm thinking about getting a new cpu/motherboard. As many people are touting the cheapness and goodness of AMD vs Intel, I took a gander at newegg, so check prices on the Athlon FX 64, and I don't see any that are any cheaper than the Intel 2.8 ghz. Really, are they less expensive? What about motherboards? Is there a good, credible comparison out there between the AMD and Intel of similar power? Really, I'm curious.
"Would it kill you to put down the toilet seat?" -- Maya Angelou
actually intel fanboys are the real sheep.
it's time to let go, your mighty god intel has shit all over you while charging you an arm, a leg, and your soul.
...between Dell's theme music and Intel's jingle? Not too damn many I suspect...
"...Well, there's egg and bacon; egg sausage and bacon; egg and spam; egg bacon and spam; egg bacon sausage and spam..."
They are not getting rid of it totally - only the socket 478 version of the chip.
The best sites (IMHO) to check for this sort of information are anandtech.com and tomshardware.com. Based on the information they (and others) have presented, I've decided to purchase an AMD 3500+ Athlon 64 next week.
Now this was a difficult decision for me because my main box does an awful lot of stuff - it's rarely just sitting there. If I'm not gaming or surfing, then I might be rendering or producing a DVD. Render times for this stuff can be annoying.
It's not that Intel couldn't do the job - in fact, there are places where Intel beats AMD (DiVX encoding and Lightwave rendering comes to mind). But it's close - and here's the kicker: So far, most of these are all just 32-bit software comparisons. Wait for the 64-bit stuff to show up! If I got an Intel, I wouldn't even have that future option.
"...Well, there's egg and bacon; egg sausage and bacon; egg and spam; egg bacon and spam; egg bacon sausage and spam..."
> PC's are just toys
Tell that to Google.
It's good to see they are finally ditching the Intel Pentium 4 "Expensive Edition".
This article is misleading. Intel is just discontinuing the Socket 478 versions of the P4EE. The Socket 775 Versions will still exist.
Kristopher
So when I was in the market I went for the EE. Probably 70% of that decision was just 'I'm dropping big cash, which I normally don't do - let's go all the way this time', 25% was 'I've seen the horrors of small data cache firsthand and I need a security blanky to forget' and 5% was 'That additional few percentage points really matter!'. So, yeah, it's a little on the silly side. But, I need to feel comfortable after the PS2 :)
"Improves your internet experience", indeed...
Wow, you really are a pretentious, arrogant arsehole.
Way to completely miss the point.
THe Athlon FX line is not the series of CPUs that are being talked about as the cheap Intel killers. The standard Athlon 64s are what you want to look for. Myself, I am going to build an Athlon 64 3000+ system, and the CPU is going to cost me about $170. Rougly the same as a 2.6-2.8 GHz Pentium 4, but the A64 kills the P4 performance wise.
Say you have a CD with a 600 MB file on it and you copy it to your HD. Windows will use 600MB caching the disk into memory, and it will swap out whatever hasn't been touched during the file copy.
--
WHO ATE MY BREAKFAST PANTS?
Say what?! Have you seen the difference in the wattage an Athlon 64 uses compared to an equivalent Prescott?!
Swapping the data you're trying to copy during a copy operation is idiotic, so I must be misinterpreting what you are trying to say.
If I copy a 600MB file from a cd onto a HD, here's what I think *should* happen:
1. Allocate the first page for the new file. Setup the destination file structures in the filesystem and mark the page as filesystem cache.
2. Start reading file from cd into the memory allocated for the file (this will mark that page as dirty).
3. After finishing that page, allocate another page and update the filesystem file structure accordingly. Continue reading.
4. Repeat 3 until entire file has been read. Note that allocating the new pages may cause some of the first pages to be written to disk (sync-ed not swapped), but in most cases should not cause any swapping since filesystem buffers are a lower priority than all but the oldest application pages.
5. When the entire file has been read into memory (as part of the filesystem cache), the copy is said to be complete.
6. After a certain period of time, or due to demand for more virtual memory or filesystem cache or a 'sync' system call, the file will be completely written to disk.
You probably meant something like what I said, though my system doesn't end up with the entire disk cached into memory. It also is less likely to swap out application data that will need to be swapped in once the copy completes. Swapping application data to perform a file copy is >2 times slower (2 writes and probably 1 read) than a cached write that I described.
That said, when I use the 'cp' command to copy a file on Linux, I'm not entirely sure what the kernel does. I know that 'cp' will open the source O_RDONLY and then open the destination O_WRONLY, and then reads from the source into its buffer and writes from its buffer to the destination (with some special cases for sparse files).
I'm not sure if they do a special case optimization for this in the kernel or not, but if they don't, then it would read the source into the filesystem cache, and copy part of it into 'cp''s buffer, switch to usermode, execute a couple lines, copy part of 'cp''s buffer into a separate filesystem cache entry, and repeat until done. (this would strain the filesystem cache, and cause lots of context switching)
Of course since moving and copying files is such a frequent activity, I would expect them to either special-case it out the wazoo (who knows?), or have a syscall specifically for copies and moves (which Linux doesn't).
Intel still has a couple of advantages, and one of them is the heat problem that comes with AMDs. OK, in a normal case you'd never notice because you basicly have the room to fit in airconditioning if you wanted to. In a 1U case however, you have to stick to about 1cm (that's less than half an inch, for those who don't know the metric system) for a fan.
On a 1U with an intel on the motherboard, I've rarely had any cooling problems. On a 1U with an AMD I've on regular basis had problems. The money customers save on buying an AMD, they'll have to spend on buying a 2U instead of a 1U.
One word: Hammer.
Learn it. Love it.
Dissolve... Resolve... Evolve...
For those of you who didn't bother to RTFA, the P4 Extreme Edition is *NOT* being discontinued, just the 3.2GHz model.
Beyond that though, AMD doesn't really have a problem. They have done a decent job to position the Athlon64 FX chip they are currently selling as simply being the top-of-the-line Athlon64 chip at any given time. The Athlon64 FX-53 is a little bit faster than the current top-end Athlon64 3900+ and sells for only a fairly small ($100-$150) more.
For Intel it is similar except that they have a much larger price gap. Their top-end regular P4 processor (that is actually being sold) is the 3.4GHz model, while the P4EE 3.4GHz is their model up from that. It is a faster chip and therefore can carry a somewhat larger price premium, but Intel is charging a MUCH larger price premium.
As for the 3.2GHz P4EE, it obviously made no sense now that the 3.4GHz P4EE is here, hence the reason why it's being discontinued. Just as AMD only carries one model of Athlon64 FX at any given time, Intel will likely only carry one speed grade of the P4EE at any given time.