AMD's 'Newcastle' Budget Athlon64 Chips Analyzed

Edward Scissorhands writes "CNET News.com reported on Thursday that AMD had released a new "budget" Athlon64 CPU. Appearing on the AMD roadmap under the codename of "Newcastle", these chips are identical to the 754-pin Athlon64 3200+ in every way except for the size of their L2 cache (512KB vs. 1MB). CNET suggests that some of these chips may be 3200's that don't pass QA as having full 1MB caches. Newcastle chips are about half the cost of their 1MB cached counterparts, though preliminary benchmarks from Anand indicate favourable performance/price."

Comments

Many companies do this...

[Anonymous Coward]

This is what many companies do. If certain chips cannot pass Q&A then remark them down and you do not lose your inventory.

cheers
Rick

Re:Many companies do this...

[GrenDel Fuego]

And just in time for Xmas!

Not quite. They're on the roadmap for Q1, which would just miss christmas at the earliest.

Re:Many companies do this...

[billcopc]

If you've been following AMD at all in the last few years, you'll know that roadmaps are a very poor estimate. They will make a press release stating that such-and-such is in production, but you might see it in-stock only 6 months later by the time they get it right.

Pisses me off too, I've been aching for a worthy replacement to this old XP2000, but there hasn't been any meaningful increases in clock speed yet. AMD64 may be the answer once they get to 3.2ghz and up.

Re:Many companies do this...

[Zathrus]

They're on the roadmap for Q1, which would just miss christmas at the earliest.

To be technical, Q1 would just miss this year at the earliest.

That said -- you didn't read the article, did you (feign shocked surprise)? The chips aren't supposed to be out yet according to the roadmap, but they are. You can order them at a reseller near you (they're available cheaper elsewhere, but I like vendors that never, ever give me trouble, ship on time (or ahead of time), and have good return policies) right now and AMD added them to their pricing sheets on Dec 15. So it's an official product that got out ahead of time.

Of course, unless you have someone who stocks them locally you'll be hardpressed to actually get it before Christmas. There's always overnight shipping, but that'll eat a large chunk of the money you're saving over the 3200+.

Re:Many companies do this...

[mocm]

They are already out here in Germany. I still can't decide if i should get one now, or wait for the 939 socket.

Re:Many companies do this...

[Sivar]

Not quite. They're on the roadmap for Q1, which would just miss christmas at the earliest.

Roadmap or not, you can buy 512K cache Athlon64s right now.

Re:Many companies do this...

[Clay Pigeon -TPF-VS-]

I am quite happy with my xp1700+(TBredB). It is currently running fine at 2.2GHz with only a 1.6 Vcore. Get yourself a barton 2500+ and up the bus from 166 to 200 and you have an instant 3200+

Re:Many companies do this...

[Lord Kano]

Not quite. They're on the roadmap for Q1, which would just miss christmas at the earliest.

Not for our Orthodox brothers and sisters, you insensitive clod!

Re:Many companies do this...

[Hoser McMoose]

512K cache Athlon64's exist now, "Newcastle" Athlon64's are scheduled for Q1 of '04. Two different chips here, despite the fact that they'll have the same basic specs. All current Athlon64, Athlon64 FX and Opteron chips use the same die with different features enabled/disabled. The "Newcastle" is a new version of the Athlon64 die with the "extra" features removed rather than simply being disabled.

Re:Many companies do this...

[fastidious edward]

Sir, I found it informative. Would you elaborate on what was not informative, indeed misinformative?

Re:Many companies do this...

[billcopc]

Okay, but even a barton 2500 only runs at like 2ghz or something.. and a 3200 would be something like 2.4ghz. so I would see a 30-40% improvement in speed, but it's not enough. I remember going from a P2-333 to a Celeron OC'ed to 1.0 ghz (back when 1gig cpus didn't exist) and it was absolutely orgasmic. I was suddenly able to encode mp3's in realtime! I just read something about THG running a P4 at 5 ghz (with liquid nitro of course), well I would really be happy if I could get a 4-5ghz cpu today, something that I could really see an improvement.

I work with video quite a bit, and I hate having to wait 2 hours for an encode (often queuing it overnight), I want it to be done in 20-25 minutes so that I can just go read Slashdot or play a round of Bejeweled for a moment and have my results ready before I get sidetracked by the necessities of real life. It's kind of like CD burners, everyone would agree that upgrading from 32x to 48x is kind of pointless in that you save all of 15 seconds out of a 4 minute job. But a few years ago when we went from 2x to 16x almost overnight, that was something!

Re:Many companies do this...

[Clay Pigeon -TPF-VS-]

11x166 == barton 2500+ default 11x200 == barton 3200+ default 11x133 == tbred 1700+ default 11x200 == my overclocked 1700+ I do have to say that I oced my 1700+ without even booting at default settings, so I dont have a frame of reference for the difference. My last processor was a tbird at 1.33GHz and the difference is amazing. I used to do a seti work unit in 7 hours with my tbird, and now I get it done in 3.

Darn...

[Coaster-Sj]

They are out just a bit too late for the christmas rush.

bad bad bad

[Anonymous Coward]

a processor named after a beer?

Re:bad bad bad

[UnknowingFool]

Why not? They already a version of Linux is named after a beer (Redhat 7.0=Guinness). So all they need is motherboard named after a snack, then you've got a whole party:
Yes, I've got Newcastle with Pretzels. They're running with Guinness.

Re:bad bad bad

[the_bahua]

An excellent beer to namesake your chip with, though. Smooth, reliable, and makes you dizzy.

Re:bad bad bad

[NanoGator]

"An excellent beer to namesake your chip with, though. Smooth, reliable, and makes you dizzy."

Crap, mine was served warm.

Re:bad bad bad

[cruachan]

Aye, you'll be able to spot it by the blue star on the packaging.

Re:bad bad bad

[infiniphonic]

a processor named after a beer? Finally !

Re:bad bad bad

[jbrader]

Why would you want pretzels when you already have chips? Sorry it had to come out:-)

[H]ardOCP has had this story for a few days now

[unborracho]

http://www.hardocp.com/article.html?art=NTYw

Kyle of HardOCP makes a bunch of speculations as to AMD's purpose for releasing these chips, and comes to basically the same conclusions that CNet does.

He sugguests that these chips are also just the ones that only had partial working cache (a portion of the cache was working, the other portion was not) and to save money they are selling these as a "budget" chip. Seems like a good idea to me!

Re:[H]ardOCP has had this story for a few days now

[cK-Gunslinger]

That's pretty standard practice in hardware manufacturing. It also explains the reasons why some hardware (Radeon 9500, etc) can be "unlocked" and turned into the real thing. They don't actually test "every" part at first, just samples of a batch. If X% fail the full spec, the entire batch is remarked as reduced-spec parts. They they are individually tested at the lower spec. It stands to reason that a certain number of these part would have passed the more rigorous full-spec tests, thus us "cheap" buyers can sometimes get lucky and get a nice piece of hardware for a small price.

Re:[H]ardOCP has had this story for a few days now

[Rhubarb Crumble]

That's pretty standard practice in hardware manufacturing.

Going back IIRC to the 386SX, which was a 386DX with a nonfunctioning (and hence deactivated) FPU....

Re:[H]ardOCP has had this story for a few days now

[Bun]

Tell me about it. With a resistor mod, and a quick BIOS flash, I turned my 9500 into a FireGL X1. Doesn't OC at all well any more, but it was still worth it. Rock stable with everything I could test it....maybe I should have bought a lottery ticket instead.... ;-)

Re:[H]ardOCP has had this story for a few days now

[Zathrus]

Argh. No.

The 386SX was a completely and utterly different chip from the 386DX. The SX only had a 16-bit data path while the DX had a full 32-bit data path. This is not a minor change in the chip or board layout -- in fact, one of the major reasons that Intel released the 386SX was to reduce transition costs from 286 motherboards -- there's considerably less difference between 286 and 386SX than 286 and 386. The 386SX had no feature differences -- it was just slower.

I believe what you're thinking of is 486DX vs 486SX. The 486DX was the first Intel chip (in the 80x86 line) to integrate the FPU onboard. The 486SX didn't have the FPU, or the FPU was disabled post-manufacture (most likely due to failures in the FPU module, while the rest of the chip was fine). This is something that can be done during product test.

The funny thing about that was the poor schmucks who bought a 486SX and then decided they wanted the FPU after all... there was a second processor socket onboard, and when the "FPU" was plugged in it simply disabled the primary CPU completely -- the "FPU" chip was a full fledged 486DX. IIRC, there was another varient where the second CPU sat on top of the first CPU (and disabled it), but I can't recall for sure.

Re:[H]ardOCP has had this story for a few days now

[Rhubarb Crumble]

I believe what you're thinking of is 486DX vs 486SX.

Damn, you're right.

Re:[H]ardOCP has had this story for a few days now

[ocelotbob]

IIRC, there was another varient where the second CPU sat on top of the first CPU (and disabled it), but I can't recall for sure.

Such a chip existed, but it was for upgrading older systems, like the 286 to a 486. Obviously because other parts of the chipset weren't as fast, the performance wasn't the same as a true mobo swap, but it was good enough for some people. Such upgrades usually used the 486slc2 chip. Information on these chips, much like the chips themselves, is a bit hard to find, but I remember a bit of temptation to get one back in the day.

Re:[H]ardOCP has had this story for a few days now

[afidel]

Evergreen made such an upgrade kit for 486 systems that allowed a low end Pentium CPU to run with the 486's motherboard limitations. They are still around and some info can be foun at their website

Re:[H]ardOCP has had this story for a few days now

[red floyd]

The other people have already got you corrected on the 386/486 thing.

OT, but what the hell... I just wish Intel had released a 286 bus/pincompatible version of the 386SX. At the time, I wouldn't have cared about the performance loss of clocking at 10MHz, I wanted the instruction set!

Re:[H]ardOCP has had this story for a few days now

[Descartes]

Actually, from what I recall the Evergreen chip was really a 486DX running at 100 Mhz or higher. They just billed it as a pentium speed. Really I think it's fair because the very late production 486's (that is after pentiums were out for a while) weren't much slower than the early pentiums.

On a totally pointless side note, I find it annoying that the plural for pentium isn't pentia as it should be.

Re:[H]ardOCP has had this story for a few days now

[zakezuke]

The 386SX had no feature differences -- it was just slower.

If i'm not sadly mistaken, the 386sx could only address up to 16megs of memory (24 address lines), where the 386DX could address a bit more (32 address lines) 4gigs of ram if i'm remembering correctly..

Let's not forget the 486slc and dlc as well. Unlike the 486 sx/dx, this was more like the 286 / 386sx in the fact that it only had a 16-bit data path and could only address 16megs of memory. This was most annoying for largish companies who bought PS/2s who, when they started to upgrade to windows 95, discovered this annoying limitation, and even when upgrading to the full 16megs of memory, it would still not be able to access a small trivial amount of memory, but just enough to apear to some software as not having 16megs of memory. I'm unsure if IBM was unaware of this limitation, but needless to say they got rather shafted on this deal for this chip, which I believe was marketed for laptops.

Re:[H]ardOCP has had this story for a few days now

[gwjgwj]

You mean 486SX. 386SX had narrower busses than 386DX.

Re:[H]ardOCP has had this story for a few days now

[bhtooefr]

Actually, I thought that the 386SX was designed FOR THE 286 BUS! It wouldn't be hard to make an adaptor yourself, considering that there had to be less than 100 pins there. BTW, why won't Intel release a pin-compatible version of the Pentium M for Pentium 4 boards? (Oh, well - PowerLeap is making an adaptor - there'd be a /. or K5 story about it, but both were shot down quickly)

Re:[H]ardOCP has had this story for a few days now

[Zathrus]

If i'm not sadly mistaken, the 386sx could only address up to 16megs of memory

Ah... right you are!

Of course, that was a complete non-issue at the time. I mean, who could possibly need more than 16 MB of memory?!?!? :)

I do recall getting my 486DX up to 20MB of memory in the early 90s... there were very few people with more memory at the time, and I only got that much because my MB had 8 DIMM slots (4x4M, 4x512k).

Re:[H]ardOCP has had this story for a few days now

[zakezuke]

Of course, that was a complete non-issue at the time. I mean, who could possibly need more than 16 MB of memory?!?!? :)

Well, it was sorta a non-issue... when I got my first 386sx-16 somewhere between 1989 or 1990 till web browsers started to hit the scene circa 1994/1995 or so. This was the main reason I upgraded to a true 386dx that took simms, rather then my old sx-16 board that took both chips and sipps.

I also remember trying to argue with the nice folks at the local computer shop. They basicly kept trying to tell me "you need 4 banks full, we don't know why but you do", and I knew perfectly well I only needed two banks full on the sx.

I do recall getting my 486DX up to 20MB of memory in the early 90s... there were very few people with more memory at the time, and I only got that much because my MB had 8 DIMM

Perhaps you mean simm sockets... typicaly 30 72 pin simms, though there was one oddball 64pin simm someone used at the time other then apple. At some point I know I upgraded from a pretty sweet 386dx-33 board with an iit mathco to a true blue intel 486dx-33 board, and found that in bench marks the 386dx was superior under old notrons si which didn't take floating point into account.

But yea, while i'd agree this was a NON issue in 1990, it sorta became an issue in 1995 as OS/2, win95 (win3.1 as well) sorta needed more then the stock 4megs of ram, esp since the price dropped to about $50 per 4megs, then $50 per 16megs, so forth and so on.

But needless to say, I would have kept my 386sx-16 in service longer then 3 years if it could have addressed more memory.

Re:[H]ardOCP has had this story for a few days now

[JK Master-Slave]

I just 'rediscovered' that 386sx memory limit this very afternoon. I reconditioned an ancient NEC 386sx box that I have here and now have Minix 2.0 running on it. It has 6 30 pin SIMM sockets, so I pulled six 4 meg SIMMs out of the hoard to fully populate it (the machine had 4 256K SIMMs in it prior to this upgrade). I quickly discovered that the machine only finds 16 MB of the memory so I yanked the fifth and sixth SIMMs. I suppose back in the days of memory-scarcity someone would have stuck 6 256K SIMMs or 6 1M SIMMs in that box.

The machine is up now, happily running Minix and I have all the (limited) networking services running.

Hooray.

And I'm glad to be able to post this ramble on a thread (supposedly) about the screaming fast Athlon parts.

Re:[H]ardOCP has had this story for a few days now

[steveha]

The 487SX was actually a 486DX in disguise. But Intel made the socket for it just a little different from the socket for the 486; IIRC it had an extra pin.

I was happy to see that the major PC manufacturers (Gateway, for one) simply put a ZIF socket for the CPU, and didn't bother to put in a 487SX socket like Intel wanted them to do. If you bought a 486SX and later wanted an FPU, you could then just unsocket the CPU and swap in a 486DX.

I don't remember any version where one chip was literally on top of another, though.

steveha

Re:[H]ardOCP has had this story for a few days now

[ocelotbob]

Actually, there were Pentium overdrive chips for the 486, however, they were slower than a regular pentium owing to the fact that it only had a 32 bit external data bus. Though Evergreen's solution, was as you said, a very fast modified 486-based processor.

Re:[H]ardOCP has had this story for a few days now

[Zathrus]

Perhaps you mean simm sockets

Most likely... it's been a long, long time...

But needless to say, I would have kept my 386sx-16 in service longer then 3 years if it could have addressed more memory.

I had the "advantage" of bypassing the 386SX... I went from my parent's XT straight to a 486DX/25 I bought myself (only computer I've ever bought that I didn't assemble) for $4k when I was a junior in high school. Some people buy cars... I bought a computer. Lasted me through most of college though. Borland Turbo Pascal 6.0 was too slow to use on the XT, so I needed to upgrade.

Re:[H]ardOCP has had this story for a few days now

[zakezuke]

I had the "advantage" of bypassing the 386SX... I went from my parent's XT straight to a 486DX/25 I bought myself (only computer I've ever bought that I didn't assemble) for $4k when I was a junior in high school. Some people buy cars... I bought a computer. Lasted me through most of college though. Borland Turbo Pascal 6.0 was too slow to use on the XT, so I needed to upgrade.

I personaly never owned an XT, and my parents computer was a ti-99/4a. I did own an atari 8-bit but after I blew my multi io board with serial, printer and scsi as well as 1meg of ram, it was more cost effective to buy a 386 then it was to replace that board. Though I was given a 8086 at one point and time, a lugable, but it was really too slow for my needs.

I guess technicaly it too was the only computer I bought that I didn't assemble (amigas excluded), though I had to remove the drive rails in order to accomidate my 15meg seagate st-419 full height drive. I tried to get an MFM controler from the people who sold me the PC, and I didn't know enough to know that what an IDE was.

I also remember at the time no one sold 1.44 meg 3.5 inch disks, with the exception of a few shops who wanted to chage me $40 for a pack of 10 disks, so I stuck with 720k disks and added a switch to my floppy to allow me to format them to 1.44 meg.

Turbo Pascal 6.0 I also had, but my class required me to use Microsoft Pascal, though techncialy the teacher said I can use Turbo Pascal but then failed me cause it wouldn't compile under microsoft pascal. Most annoying as the text book was geared tward turbo pascal.

But needless to say, I was much happier with a PC, dispite the Amiga being the far more attractive option. Basicly you could upgrade without soldering (for the most part), and faster CPUs were released yearly, rather then the world of atari and commodore.

Re:[H]ardOCP has had this story for a few days now

[Descartes]

Thanks for the link, much better than my hazy recollection of hardware I couldn't afford.

+1 informative

Re:[H]ardOCP has had this story for a few days now

[bhtooefr]

BZZT! You're right about the 386, but the 486SLC and DLC weren't on a 16-bit data path - they were on a full 32-bit data path, and could address 4GB RAM. They were made by Cyrix, and are Socket 0 (132-pin PGA - same as 386DX). I'm fairly sure SLC means SX Low Cost and DLC means DX Low Cost. They're simply Cyrix 486SX2 (if there ever was such a thing) and DX2 CPUs that in some cases are underclocked (available in (not sure on this one) 16x2=32, 20x2=40, 25x2=50, or (not sure here either, but don't know why they wouldn't have one in this speed) 33x2=66 MHz speeds), sitting on a 132 pin socket, and designed to run on a 386 chipset.

Re:I'm sure

[Pingular]

Sorry, that should be Newcastle

Looks like AMD..

[cK-Gunslinger]

.. has another winner on their hands. Excellent performance at a fraction of the price. 2 GHz, 64-bit performance for about $200 is nothing to sneeze at. Bring on the 64-bit apps/drivers! (And, of course, the MS OS.)

Re:Looks like AMD..

[lcde]

(And, of course, the MS OS.)

[flame]insert flame here[/flame]

:D

Re:Looks like AMD..

[cK-Gunslinger]

=P Well, like it or not, everyone knows it's Windows that currently drives the consumer market. The release of a Microsoft 64-bit OS is what will determine if/when the 64-bit desktop market takes off. The release of XP 64, followed by 64-bit aware device drivers will start the snowball. I would love to see some applications written to take advantage of those extra registers! (Linux apps aside.)

Re:Looks like AMD..

[lcde]

I fully agree with you. Just trying to get rid of the bad case of the mondays :).

I've thought for some time now that 64-bit will take off after the gamers latch on. I'm really surpised that MS is taking so long to release a ia32-64 version. I don't remember them taking this long for Intel's ia64.

You would think that with Apple's G5s they would hurry up and get something out to make it seem like Apple's idea isn't that special.

That would just seem like good marketing to me.

Re:Looks like AMD..

[Captain Beefheart]

Not only a fraction of the price, but a fraction of the megahertz as well. I used to disregard the MHz Myth as Apple told it, but a gap like this, with apples-to-apples benchmarks showing them neck-and-neck, is kind of astonishing. I understand now why AMD is trying to push away from MHz speed to determine overall speed, into more of a video card model, where benchmarks determine the top performer, not specs. nVidia cards are always clocked higher than their direct competitors, yet you almost never see this point brought up as an argument in their favor--because the benchmarks show that both GPUs are on the same level.

Also, as Moore's Law seems to be finally bending--"faster" chips aren't coming out with steady increases as before--it becomes more difficult to find an upgrade with a clear performance increase, without spending big bucks for the uber chip. This is a clear performance increase over Intel chips clocked hundreds of MHz faster, and is a true return to accurate PR ratings that we haven't seen, in my opinion, since the 2400+.

And at a little over $200, it's practically a steal, although you do have to buy a new motherboard, which usually means having to format and reinstall your OS partition. I think, though, 64-bit "future proofing," the apparent stability and lack of heat issues, and the reasonable price will convince many people to make the switch. It's also an opportunity to get a motherboard with built-in gigabit LAN, audio, RAID, FireWire and whatever other nice features you've been doing without while saving for that next upgrade.

Re:Looks like AMD..

[drsmithy]

I used to disregard the MHz Myth as Apple told it, but a gap like this, with apples-to-apples benchmarks showing them neck-and-neck, is kind of astonishing.

Except when Apple were telling the "Mhz myth", a) the intel CPU was twice the clock speed, not half again as much and b) the apples-to-apples benchmarks demonstrated the CPUs *weren't* "neck-and-neck" - the x86 CPUs were (generally) faster.

Also, as Moore's Law seems to be finally bending--"faster" chips aren't coming out with steady increases as before--it becomes more difficult to find an upgrade with a clear performance increase, without spending big bucks for the uber chip

That's because so very few things are bottlenecked by the CPU.

I think, though, 64-bit "future proofing," the apparent stability and lack of heat issues, and the reasonable price will convince many people to make the switch. It's also an opportunity to get a motherboard with built-in gigabit LAN, audio, RAID, FireWire and whatever other nice features you've been doing without while saving for that next upgrade.

Personally I'm going to give the motherboards at least a generation - probably two - to work out all the inevitable problems.

Re:Looks like AMD..

[spinlocked]

Not only a fraction of the price, but a fraction of the megahertz as well. I used to disregard the MHz Myth as Apple told it, but a gap like this, with apples-to-apples benchmarks showing them neck-and-neck

There is no such thing as an apples-to-apples benchmark unless you happen to be weighing apples. Every chip architecture is different and will perform in different ways. A 64-bit apple PowerPC chip isn't the same as those that go into an IBM pSeries. A 64-bit Sun workstation UltraSPARCIII chip isn't the same chip that goes in an F15K. A 3GHz Xeon will spank a lot of chips on integer performance, but it's FP performance sucks.

The clock speed is irrelevant. Benchmarks are mostly irrelevant because they can't possibly model your application's workload.

You aren't getting future proofing from 64-bit CPU's (especially since some of these low-end 64-bit boxes can't even take 4GB of memory), you're only getting major performance improvements on applications which are either: a) specifically developed for this architecture or b) exhibit memory address space bottlenecks, i.e. for most people today, relational databases. Memory capacity, memory bandwidth and IO performance then become the dominant factors in scalability.

A fast chip which can't be fed with useful data will just spend it's time running nops. A CPU architecure which is not designed to scale beyond 8 CPUs might also become a constraint.

Re:Looks like AMD..

[drinkypoo]

You don't need to write apps to take advantage of the additional registers. You just need to use a new compiler. You may need to write apps to take advantage of the fact that the registers are 64 bit, however. (For some things, you won't need to do anything but recompile.)

Re:Looks like AMD..

[pinkocommie]

Any info on when they're releasing the OS's (2003/XP). I was reading an article somewhere about the XP-64bit beta release that said the OS would be released to OEM's only and not available as a retail upgrade (sorry can't find the link @ the moment) any info on that? Also here's a blurb about a version of Visual Studio which supports AMD64 also being released someday

Re:Looks like AMD..

[LennyDotCom]

although you do have to buy a new motherboard, which usually means having to format and reinstall your OS partition.
Does anybody remembr when a cpu upgrade didn't require a motherboard upgrade?

I'm sure...

[TwistedSquare]

the people of Newcastle (working link) will. Long has the North been associated with cheap ;-)

Re:I'm sure...

[TheCoop1984]

Its the first time a whole processor has been named after the city where I live... (but you should see the Quayside on a Friday night, most of the people there should technically be dead with the amount of alcohol they've drunk...) Is the 'Newcastle' name only for the 3000+'s, or is it a general group of processors? The 3200+ wasnt referred to as a Newcastle...

AMD changing pin # anyway

[3DKnight]

after 2004, the 754-pin sockets will make way for their new 939-pin sockets. AMD has said that they will continue upgrades for 754-pin 64-bit chips up to i think 3700+ After that you will need to buy a 939 pin motherboards. Though I wonder what the shelf life for the 754 pins are, since not that many programs can even make use of 64bit cpus yet.

Re:AMD changing pin # anyway

[unborracho]

There haven't been any official statements that their new processors are going 939-pin.. that is a speculation by a few highly-voiced individuals and off-the-record statements

Re:AMD changing pin # anyway

[the_2nd_coming]

well since they run 32 bit code natively and FASTER than the top of the line P4, I would say the shelf life is good.

Re:AMD changing pin # anyway

[Jeff DeMaagd]

As it is the 9xx pin FX jobs don't really have that much of a reason to exist yet because the performance increase is marginal, maybe 5%. Sometimes dual channel actually slows things down by a percent or two. I hope that this changes before the 7xx pin version goes away. I imagine that at a higher clock the difference becomes more noticable.

The only reason to get a 9xx pin chip is to get multiprocessing in the form of Opteron.

Re:AMD changing pin # anyway

[ergo98]

Interestingly I've seen this several times now in conversations about the Athlon 64, and it's a reasonable debate that I've tossed around many times in the past myself (i.e. "what is my upgrade path?").

The reality, though, is that I have never, ever upgraded just the chip on any motherboard, no matter how supposedly long its lifespan was -- there was always some new technology on the newer motherboards, be it faster AGP, better interconnect standards (USB2, Firewire, SATA), faster memory standards, etc. In other words who really cares how long the socket lasts? You'll likely replace the motherboard when you replace the chip anyways.

Re:AMD changing pin # anyway

[bhtooefr]

Here's the deal:

Athlon 64 (mid to high-end consumer CPU): Socket 754 now, possibly Socket 939 in the future.
Athlon 64 FX-51 (high-end consumer to mid-end enthusiast/professional CPU): Socket 940 now, next version of the 64 FX will be Socket 939
Opteron 1xx (mid-end professional or server CPU), 2xx (mid to high-end professional or server CPU), 8xx (high-end server CPU): Socket 940

Re:AMD changing pin # anyway

[DrMindWarp]

The reality, though, is that I have never, ever upgraded just the chip on any motherboard,

As a systems administrator, I usually consider this possibility but it doesn't influence the actual buying decision. A drop-in upgrade with a faster CPU is obviously attractive but no vendor will guarantee that over a sufficiently long period (although Intel have come close to long term stability with the Itanium package).

Having said that, the AMD switch to 939 did make me ponder delaying our Opteron roll-out but 1) we need them now and 2) by the time the last 940 comes out it'll still be significantly faster than our existing 1.8GHz 244's so there will be an 'easy' upgrade path.

In the past, I went from two generations of Slot1 CPU through to dual Slot1 (new motherboards, same CPUs) and saved a fortune but that's not likely to be repeated. As for interconnect standards, one can usually rely on the longevity of the PCI(-X) standard to allow PCI cards to take the strain.

Just what I was waiting for

[Anonymous Coward]

This is a very good incentive to go 64 bit. I was thinking of getting a 2500 Barton, since my 1800 finally kicked it last week when the cooling fan gave out (this was right after a re-format, so the temperature monitoring system was not installed yet). However, since this came out, it might be a good time to go 64 bit. The chip still packs punch, so its not really what we would tend to think of when the term "budget" comes up (AMD Duron...Intel Celly). Plus, it won't be that expensive to replace if you take the OC too far.

Re:Just what I was waiting for

[cK-Gunslinger]

Yeah, it makes you reconsider that Athlon XP 2500+ purchase. 64-bit is temping, but you have to keep in mind that the 754-pinout on the chip is doomed. AMD already announced that they will move to a 939-pinout for most future 64s (Opterons are 940, so I assume they are just removing the "multiple-cpu" pin.) If that's the case, you may not have a very long upgrade path (3700?)

Re:Just what I was waiting for

[Coaster-Sj]

I know some people are conecerned about having large upgrade paths but I find that I'm really not one of them any more.

Usually by the time a processor drops in price enough that I think it'd be worth replacing an older CPU there is a new FSB or something that makes me want a new motherboard + ram to go with it.

Lately when I've been buying computers I've came to the conclusion that Motherboard, Processor, and Ram are pretty much a package that will never be upgraded independantly (Short of adding more Ram). Unless I have a processor die I'm really not worried about changing it.

Re:Just what I was waiting for

[maraist]

Anymore, the vid-card, sound-card and high-perf ethernets are the only things I care about. MB + CPU + Mem + Case become an old Linux machine / firewall / vncserver (for aim).

Considering that the average MB+CPU is only $200, this isn't a bad deal.

Re:Just what I was waiting for

[ProtonMotiveForce]

Wait, I'm geniunely curious (I want to see how AMD's marketing machine works)...

Why are you convinced that a 64-bit processor will help you in any way? It won't, of course, unless it also happens to be faster in run-of-the-mill 32-bit code - which I will admin AMD's 64-bit chips sometimes are.

But you specifically mention you want to "go 64 bit". How did they convince you this will help anything?

Re:Just what I was waiting for

[Cajun Hell]

Why are you convinced that a 64-bit processor will help you in any way? It won't, of course, unless it also happens to be faster in run-of-the-mill 32-bit code - which I will admin AMD's 64-bit chips sometimes are.

Speed is not the only advantage the AMD brings with the x86-64 architecture. It also has features. For example, if you run OpenBSD natively on one of these chips, you'll have something to sing about. X86-64 lets to control whether a page of memort is executable or not -- a nice protection against things like buffer-overflow attacks.

People have been saddled with the x86 legacy for so long, that they have forgotten the old days, when processors distinguished themselves from one another, in ways other than price/performance.

Re:Just what I was waiting for

[kernelpanic77]

To answer that, you should know that this was an 1800 with a Thorougbred B core. I had it overclocked to about 2 GHz, and it was running very hot. The fan was a cheap thing I had in place, which was OCed by putting more voltage through it by using a secondary power-supply (mobo intput jack is free, so you can feed all kinds of voltage through different components) so that the fan could keep up with the CPU until I could get a Volcano of some sort for it. However, the engine in the fan and the bearings didn't take too well to having more juice than intended running through it, therefore the fan died.

Re:Just what I was waiting for

[bdraschk]

AFAIK, the 939 Socket will actually be for future FX chips and it won't be "just a pin less", but will allow for 4 layered PCBs, instead of expensive 6 layered ones.

I hope and rely on the 754 not to be doomed, as i usually upgrade my CPUs once (K6-2 to K6-3, Athlon 1100 to Athlon XP 1800+), when the chips become cheaper.

The Athlon64 3000+ was just what i was waiting for:

AMD, not Intel

inexpensive

possible upgrade path (unlike XP 3200+ or P4)

I think I'll order after the Christmas rush

Re:Just what I was waiting for

[WNight]

I'm looking forward to x86-64 because the benchmarks I've seen suggest that sound and video encoding are much faster when compiled for 64 bit.

I run Linux so finding a 64 bit flavor to run shouldn't be hard, and with a short recompile, the few CPU-bound things I do will get faster.

And yes, mp3 (ogg now) encoding is something I wait on. My Barton 2500+ encodes at about 4x realtime for oggs, but I can rip CDs at 12x, so my temp directory fills up. When I run two simultaneous copies of Grip it gets even worse.

I don't do this all the time, but when I do it usually has to be quick. I'll get to borrow my friend's 100-disc collection, but only over a weekend, or something.

Not enough to buy a new machine for, but certainly worth an extra $50 or so to play around with for my next system. Supposedly it'll be easier to make dual-CPU machines too, so maybe I can go to dual Athlon64s for $200 extra, or so. Well within a play budget.

Re:Just what I was waiting for

[owlstead]

Dunno, I always keep my CPU's and motherboards together nowadays. If I upgrade, both my CPU and motherboard go to another machine. Motherboards are not too expensive, and there are always reasons to upgrade your mobo (faster memory, better lan, serial ata, raid etc.).

Re:Just what I was waiting for

[bhtooefr]

Tbe 754's doomed, but it's going to be like Intel was with Socket 2 and 3 and the early Pentium Overdrives - AMD will release up to an Athlon 64 3700+ in Socket 754.

Re:Just what I was waiting for

[DrMindWarp]

AFAIK, the 939 Socket will actually be for future FX chips and it won't be "just a pin less", but will allow for 4 layered PCBs, instead of expensive 6 layered ones.

An AMD engineer assured me that the Opteron will also be going to 939 for the same reason.

Re:Just what I was waiting for

[Bert64]

And Intel chips are consequently more expensive, if you want rock solid reliability you dont buy intel tho, you buy DEC, SUN, IBM or HP.. proper server-class chips that are rock solid. Infact, most of these chips dont have tiny little fans blowing down on the cpu, they have larger fans (bigger fans are more reliable) generating a flow of fresh air into the case, over the cpu and out again

You always disable half the cache!

[Saville]

I remember this back in the PentiumII/Celeron days, but does this happen anymore? Has anybody looked at the chip to see if it is any smaller or the same size as its bigger brother?

Re:You always disable half the cache!

[Saville]

Opps. :)

http://www.hardocp.com/article.html?art=NTYw

"AMD did also verify that the "new" core is 193mm2 in die size. This is the same size as specified on the original 1MB L2 Athlon 64 cores as noted here back in September of this year. "

Re:You always disable half the cache!

[bhtooefr]

How the fuck? BTW, a NetBurst Celery has half the L2 cache, which doesn't sound as bad at first, but when you realize that the cache is what's keeping NetBurst CPUs from going into the gutter performance wise, the Celeries suck - just look at Anand's benchmarks of sub-$100 Intel and AMD CPUs (with a 1.8A P4 as a baseline) - except for the Duron 1.6, AMD CPUs and the P4 MURDERED Celeries. As for the Duron, it usually held it's own against the Celeries and 1.8 P4.

We reviewed this days ago

[ruiner5000]

Didn't we? Athlon 64 3000+ review.

In conclusion the Athlon 64 3000+ is one of the best CPUs AMD has never announced. It makes a sub $1,000 system that is 64 bit capable easy to reach, and is able to perform quite admirably even with half of the cache of the other AMD64 CPUs. Will AMD make more 512kb cache Athlon 64s in the near future? How long will Socket 754 continue? Is this 3000+ an overclocker of merit? Stay tuned. For now if you have been craving for a powerful and cheap system with 64 bit onboard then the Athlon 64 3000+ is your CPU. It has no competition in its class, and likely will not for months to come.

Let's see, 1 year since Slashdot has approved a story I've submitted. Let's keep the streak alive! ;) HP shipping Mandrake biz PCs. Who cares!

Re:We reviewed this days ago

[Saville]

Maybe it's because of the generally grumpy tone found on the main news page of amdzone?

I didn't know HP was shipping Mandrake. I wish that had made it as a story. :(

Re:We reviewed this days ago

[KefabiMe]

Just to let you know, if the /. editors posted a story about HP shipping Mandrake, It would prolly be considered a dupe.

price?

[Nate Fox]

$213USD seems to be the lowest on pricewatch, for those who are wondering
http://www.pricewatch.com/1/3/5867-1.htm

It's all Bush's fault

[Luscious868]

This move by AMD is clearly the result of Intel not contributing enough to Bush's reelection fund. Bush and his evil neo-con allies have it in for Intel and are using AMD as a front to destroy it. No blood for processor preformance!

Re:It's all Bush's fault

[ed__]

uhhh, maybe a little blood. i can always use a faster processor

Re:Does 512k vs 1 meg cache make much diff?

[Anonymous Coward]

If an application is frequently accessing a reasonably small set of code and data, and the total size of the accessed code and data is less than the size of the L2 cache, then the application can run from within the cache, which is much faster than main memory. The size of the L2 cache directly affects the point at which this speed benefit can be realized.

Cheap CPU's get cheaper.

[dilvie]

This is good news. The next month or so will be a great time to buy those boring 32-bit CPU's that nobody cares about anymore. Moore's law rocks.

Anandtech and other review sites

[Anonymous Coward]

Sorry, this is a bit off topic, but are sites like Anandtech trying to be useless? At least they do include comparisons to the P4 in this review, where often they don't give you enough context to really compare across the board, but now they've switched to using Flash for the charts. Why? It's not like they're "live" or anything.

Compared to decent, thoughtful sites like StorageReview, where they at least they let you compare any model to any model, I hope Anand and the other forever-amateur just-enough-info-to-make-you-think-its-worth-buyin g sites soon sink into the mud.

Re:Anandtech and other review sites

[Nasarius]

You're right, it's absolutely insane to use Flash in this case. The charts aren't interactive. Why exactly can't they just store their pretty graphics as JPEGs or whatever and make everyone happier?

Re:Anandtech and other review sites

[cK-Gunslinger]

Obviously, to keep other sites and everyone else from "stealing" their charts. I know it's a pain in the ass, but Anand is a pretty large and profitable review site. If their charts were just JPGs, what's to stop some unscrupulous site from snatching their pics, changing the colors in a batch job, and reposting the results as their own? Granted, they could probably just screen-cap their pics now, but that involves a little more work.

Re:Anandtech and other review sites

[cheesybagel]

They can still copy the Flash files as well. Makes no bloody difference, expect they are using non W3C standards. Bad, very bad.

Re:Anandtech and other review sites

[Coventry]

Bandwith - the major cost of runing any hardware review site. By making a simple flash movie that displays a graph based upon passed in text data, they save bandwith. You download the flash movie once (or once every time they update it), and then everytime they display a graph your browser uses the cached movie, but with the new param set for what data labels and values to display. Sending even 15 lines of data is a lot smaller than sending a 300x120 pixel image for every single graph - especialy when a multi-page review may have 15+ graphs.

Yes, it sucks for the user who has to download whatever version of flash they use - and it sucks even more if the version they require is so new it isn't avail on all platforms, but they Do have a good reason.

Re:Anandtech and other review sites

[jrockway]

Thanks for clarifying that. That's actually a good idea. I thought it was for copy protection or something (and wondered how hard it would be to copy the text into gnuplot or something and make a new graph)

Re:Anandtech and other review sites

[steveoc]

Bandwidth ?

What about this :

A64 3000 __________________ (316fps)
A64 3200 ____________________ (328fps)
P4 3.2EE __________ (142fps)

Works in lynx too.

If it didn't pass QA

[morelife]

Why would anyone want it?

The cache is allocated physically. It isn't a question of it having 1023 Kbytes instead of 1024 so then being sold as a 512K model...

Re:If it didn't pass QA

[Anonymous Coward]

Cache is not all or nothing. Let's say 934k of 1024k work. Turn off 512k including all the non-working memory and voila -- sellable budget processor. AMD chose the 50% line arbitrarily -- they could have chosen 75% and gotten fewer, higher performance processors or 25% and gotten more, lower performance processors.

They can't actually tell you how much of the cache works because OEMs like to sell identical machines, i.e. all with 512k cache, not 512k cache or more.

Re:If it didn't pass QA

[cK-Gunslinger]

I'm not sure, but I would think that as long as they could find 512K of contiguous cache that passes, they could use just that without any major modifications.

Re:If it didn't pass QA

[Rhubarb Crumble]

Why would anyone want it?

Because 99% of the CPU works, and you know which bits don't, and can live without them?

The cache is allocated physically. It isn't a question of it having 1023 Kbytes instead of 1024 so then being sold as a 512K model...

Maybe not 1023K, but something along those lines it exactly what it is. I'm guessing here but presumably the cache part of the die is divided into cells (say, 32 cells of 32K each, or whatever) and if more than half the cells are working, you can use half of them and map them transparently so the processor 'sees' a single 512K cache rather than a 1024K cache with holes in it.

Net result: instead of having a broken CPU which you'd throw away, you now have a perfectly-functional-if-slightly-slower CPU. Makes sense or what?

Re:If it didn't pass QA

[Kazymyr]

The cache is allocated in blocks. There are 2 512k blocks, and if there are bad cells in just one of them, it is disabled and the chip is sold with only 1 512k block enabled. If there are bad cells in both, they throw it away.

Re:If it didn't pass QA

[morelife]

Wrong. It doens't work that way. They don't, on a processor by processor basis, go in and "disable" the random parts of any Cache area that failed testing.

ASS HAT Moderators too.

Re:If it didn't pass QA

[Alien Being]

Actually it makes sense that they would just disable the most significant address line and go straight from a 1024 to a 512.

Re:If it didn't pass QA

[Hoser McMoose]

I don't know about this chip specifically, but for chips in general there would be more than just two blocks. In fact, it would probably be along the lines of 36 cells of 32K a piece. If any of these cells fail validation, you can disable them.

So long as the number of busted cells is 4 or less (just using my example numbers here), you can sell the processor with 1024KB of cache enabled. Obviously, if there are no cells that are bad you still disable 4 of them to keep the parts consistant (OEMs don't want one chip with 1024KB of cache, the next with 1056KB of cache and the next with only 992KB). If more than 4 cells are bad, you disable all but 16 of them and you've got a chip with 512KB of cache.

Intel and AMD both do this (as do others), and at various times both have had chips sold at different performance and price points depending on how much cache (or some other feature, ie hypertransport links, hyperthreading, floating point units, etc.) is enabled.

Re:If it didn't pass QA

[steveoc]

> if there are bad cells in both, they throw it away.

Hmm .. would a chip work at all with Zero cache ?

How about a $2 Athlon64 2999+ ?? - 2.0GHz, 0kb cache. Maybe they could bundle it with rejected motherboards, where the COM port, one of the IDE channels, or the USB ports are fried.

Who has the contract to empty the dumpsters from AMD I wonder ?

Re:If it didn't pass QA

[BlueBiker]

Actually an interesting proposition. IIRC, there were tests on P-II with disabled caches which lost maybe 30% in performance but became more amenable to overclocking. An Athlon64 2200+ for $20 could be an attractive chip for some.

Re:If it didn't pass QA

[xluap]

Yes !

And for the rest of the system:

509Mb of ram, using the bad ram patch so the bad bits aren't used.

And a 79Gb harddisk with bad blocks on it, those will be marked as bad blocks when the disk is formatted.

And a flat lcd panel with bad pixels.

And of course a microsoft operating system !!!!

The new 'dual celeron'?

[Henk Poley]

Could this become the new 'dual celeron' like a couple of years ago?

Re:The new 'dual celeron'?

[cloudmaster]

After you connect a couple o fleads on the top of the chip with heavy pencil lines, of course... :)

BTW, I like my dual Celeron machine. Abit BP6 may be a crappy, unstable board, but I still like it.

Ha ha ...

[Anonymous Coward]

I don't know what's funnier ... your comment or the possibility that, given the political leanings of a lot of the people who post here, some people will think your serious and not realize it's a joke.

Re:Ha ha ...

[Luscious868]

I don't know what's funnier ... your comment or the possibility that, given the political leanings of a lot of the people who post here, some people will think your serious and not realize it's a joke.

:-) ... hehe .... I know ...

Re:Ha ha ...

[nazsco]

/.ers will probably get that it's a joke since 99% of them browse with +5 Funny, +5 Flame Bait and +5 troll... so they don't expect to see anything serious here.

btw, how do you think i found your post :)

Not quite fresh news...

[Kazymyr]

Overclockers.com did a mini-feature on the Newcastles last week, including why you shouldn't buy one too soon.

Re:Not quite fresh news...

[Kazymyr]

Link for the above.

Re:Not quite fresh news...

[Kazymyr]

It's not about getting better results, it's about AMD's pricing scheme. Go read the article.

How many pins is too many???

[manganese4]

Given current mechanical properties of the materials that encase the actual chip, the connections from the chip to the pin and the ability to insert chip into a motherboard, is there any impending barrier to the number of pins for future chips?

Is it not an extra bonus that it's 64 Bit

[XMichael]

Did the article not miss one MAJOR item in the review ---> the Processor is 64 bit capable, taking into account that Windows and nearly all games arn't ready to take advantage of it, this processor will continue to set performance standards as the software becomes 64 bit ready?

Whoop, I hope I understand that correctly (-;

Re:Is it not an extra bonus that it's 64 Bit

[The One KEA]

Yes, but conversely Linux can take advantage of it - SuSE, Mandrake, Red Hat and Gentoo all have functioning offerings available for purchase or download.

The only game I know of off the top of my head is Epic's Unreal Tournament 2003.

Either way, the Athlon 64 3000+, IMO, might just be what AMD is looking for to really break into the market. If the price goes below $200, then things will definitely start to get interesting for Intel.

Re:AMD is overpriced

[voss]

Maybe on taste tests(youre not supposed eat THOSE chips) , but most real world tests show the Athlon 64 3200 going neck and neck with the P4EE (a jury rigged chip with 2mb of cache that sells for $974 on pricewatch. The only way a pentium 4 2.8 could outperform an athlon 64 is if the Athlon 64 was on a PCchips motherboard and I think the Athlon would beat the Pentium 4 2.8 even then! :P

Bad, but actually...

[JawFunk]

I think they dedicated these "flawed" chips to the Newcastle Stock Exchange. Fits the Australian profile of doing business.

The origin of the name Newcastle

[andy666]

In Moby Dick, Newcastle is an assistant to the navigator who does all the computations.

Re:The origin of the name Newcastle

[Anonymous Coward]

Correct, but not the reason why AMD chose the name. Consider the others:

Athens
Troy
San Diego
Winchester
Paris
Odessa
Dublin
Oakville
Egypt...

I would wager that Newcastle was chosen because it was a place, but it's kinda cool that it has another link to computing. Are there any computer links for the others I wonder (apart from Winchester)?

Re:The origin of the name Newcastle

[Anonymous Coward]

Or maybe it's because releasing another 64bit processor is like carrying coals to Newcastle? Or because it's new and your computer is your castle? Or because it can run software that is free as in speech and beer, hence the connection to Newcastle Brown Ale? Which is a.k.a. Nukie Brown, perhaps alluding to Duke Nukem Forever? Or because Newcastle United routinely gets hammered by other teams, hence the connection to AMD's *hammer codenames?

Low-priced alternative

[dorlthed]

I like this idea, and from a product-line standpoint, it's a good one. After the Athlon XP line started, I sort of missed the situation with the Thunderbird/Duron, where there was always a low-priced alternative for budget systems.

Perhaps now they will create a sort of "64-bit Duron," a lower-priced and less-powerful version of the Athlon 64. This way, in the future, if I want to create a bargain version of a AMD64 computer for a family member or friend, or buy one, there is a cheaper processor available for such a system.

I sort of missed having that alternative available; this creates a bit of processor nostalgia for me :p

hmmm....

[theMerovingian]

These 'reject' chips might be the reason Emachines offers such a cheap 64 bit computer.

"32bit computing is dead"

[Space cowboy]

... or so claimed AMD. Maybe this is why - they are releasing 64-bit chips at prices comparable to mid-range 32-bit ones! Way to go AMD :-)

I have no particular beef with Intel, btw, it's just that AMD always seem to aim more at value for money. I like that :-)

Simon.

Re:"32bit computing is dead"

[blunte]

I have no particular beef with Intel, btw, it's just that AMD always seem to aim more at value for money. I like that :-)

More importantly, AMD has provided real competition for Intel, which has helped greatly reduce the price of non-bleeding-edge CPUs.

Cheers for AMD (in general... and cheers for AMD for surviving this long :)

Re:"32bit computing is dead"

[ProtonMotiveForce]

They can release 256 bit processors if they want, it doesn't mean anything to your garden variety desktop user unless it's faster at running 32-bit code (which is sometimes the case, admittedly).

The 64-bit thing is a bragging rights gimmick and doesn't do anything for the vast, vast majority of desktop users out there, who don't even have 1G of memory much less 3-4G. What's funny is people are actually buying into it.

Re:"32bit computing is dead"

[raodin]

Ah yes, good ol' fud. CURRENT top of the line Intel processors run hotter and consume more power than top of the line AMD processors. You're a good 4 major revisions behind on your info. The Thunderbird was the last really hot Athlon (and it really wasn't THAT bad) They cooled off considerably with the release of the Palamino, and again with the Thoroughbred. Barton is basically the same as the Tbred, and the A64 chips don't seem to be running too hot, either.

Who cares how good the retail "hintsinks" are. They're pretty decent, these days, but anyone who cares will replace it with a better performing model. No different than with Intel CPUs.

You're entitled your opinion, but your post is pretty much just 3 year old FUD.

Re:"32bit computing is dead"

[Slack3r78]

Really, I'm more interested in it as a redesign of the x86 architechture. x86-64 uses the older design as a foundation, except it removes much of the cruft and adds quite a bit to the capabillities of the chip.

The most harped upon example is the increase in the number of registers available, which in and of itself should increase processing power - something that's not being taken advantage of yet since there's not really any 64 bit native software available yet.

The fact that the chip is already efficient enough to trounce a top of the line P4 in most tasks in 32bit mode is what piques my interest - I'm interested in seeing how well it'll perform in native x86-64 mode with the extra registers, etc enabled. In short, there's much more to x86-64 that having >4GB of memory.

Re:"32bit computing is dead"

[TheLink]

Just wait till you see the next generation clippy :).

Seriously though, 2GB+ RAM can be very useful if you are using virtual machines. Opening a suspect attachment/file in a quarantine virtual machine (then rollback to pristine condition) etc. Once you start doing this sort of thing the RAM and disk space can start being used up pretty quickly.

Pity AMD doesn't seem to have added better support for virtualization to their AMD64 chips.

Apparently the PowerPC supports full virtualization - you can run a VM in a VM and it won't even know. Whereas with x86 and AMD64 you can't do that without resorting to a lot more emulation.

Re:"32bit computing is dead"

[Rob Simpson]

Here's a cpu wattage chart.

If these numbers represent real life heat production, the athlon 64 is their hottest running processor yet (89 W), as is the P2EE Intel's (93.9 W). While the athlon 64 is the more appealing choice, I'd still expect it to need a pretty noisy heatsink/fan. Personally, I'd like to see a dual-processor Pentium M. ;)

Re:"32bit computing is dead"

[10Ghz]

Your info is WRONG! 89W is the figure AMD told to MoBo-makers as a guideline. As in "design your MoBo's in such way that they can handle CPU's that generate 89W of power". The actual figure for A64/Opteron is way below that number!

89W is the worst-case scenario for the future. From what I have gathered, the real number for A64/FX51 is something like 50-60W. And if the CPU-load is low, the CPU automatically underclocks itself (all the way down to 800Mhz). I have heard of people who run their A64's with the fan on the heatsink being dead-still!

Really?

[Slash Subscriber *]

I hadn't considered upgrade CPU's to be a big contender for stocking stuffers or under the tree, but maybe I'm just out of touch...

Re:Really?

[buck_wild]

Hey, they will be if you let Santa in on your list.

I get a CPU upgrade nearly every year, but I don't like to be more on the value-edge of the proce/technology curve rather than on the bleeding (wallet hemhorage) edge.

Well, my time-travelling webshop...

[Kjella]

Not quite. They're on the roadmap for Q1, which would just miss christmas at the earliest.

...has 5 in stock, and that's here in Norway even. So it might not arrive before Christmas, but if you hurry you'll probably still go 64 bit in 2003 :)

I'm perfectly happy with the PC I have though. Usually, whenever Christmas is nearing I get questions about whether I'd want something for my computer. This year, for the first time in as long as I can remember, the answer is "not really". No big itch to scratch... I have CPU, GPU, RAM, HDD enough, broadband, most everything really.

I must say, I'd still like to improve noise and style though. Performance, well it's not that critical anymore. But the noise is pretty bad, even after I replaced my WD disks with Seagate. And I admit a Shuttle XPC + LCD looks ten times better than my beige box. Maybe next PC, but that one is not now. Not for a while yet...

Kjella

Sure, until you are...

[cnelzie]

...the poor sap that can barely afford to spend the extra money on the 'value' board, you listen to someone's spiel about how you can 'unlock' the 'magic' or something and you end up frying your once perfectly good, yet low-spec'd board and are stuck having to go back to your old parts, if those aren't fried as well.

Re:Sure, until you are...

[Jeff DeMaagd]

On one hand, I think people whould be free to do what they with with hardware they own, on the other, I would expect that they should know what they are doing before voiding their warranties and otherwise ignoring manufacturer warnings and disclaimers.

While I understand that sometimes there really is a marketing reason for makers to down-mark their chips, I pretty much refuse to overclock anything because sometimes the silicon engineering reason to downmark is very real too, I really can't afford to junk perfectly good parts or risk flakiness.

Why Aye Man!

[turgid]

AMD processors, designed by James T. Nail and fabricated on Tyneside.

Crocodile shooooes!

Re:Why Aye Man!

[turgid]

No, I think it's "T" as in James T. Kirk.

Re:Why Aye Man!

[turgid]

...and he's king of Newcastle, and hence the world.

Ed Stroligo is a moron.

[cheesybagel]

If you don't believe me, just read any of his articles.

This CPU is a good deal, Athlon 64 at Athlon XP prices. Some of us actually need to buy machines now, not in the next 6 months. Oh, wait to find a socket 939 processor at $200 like that jackass Ed is saying. With luck, it will only take until Xmas next year.

Re:Ed Stroligo is a moron.

[Kazymyr]

If you *need* it now, then by all means get it. I frankly don't see much of a future in socket 754 anyway (it was created as a transition thingie, like slot A for instance) and since I made the jump to socked 462 in the past six months only, I have no reason to change platforms now. I'll wait and see. So far I've always skipped one generation of CPUs and bought the next one, and it's worked pretty well so far. I can wait for mass socket 939.

Presidents?

[T-Kir]

Be worried if they release CPU's named after presidents... and you have the 'Pretzel' motherboard edition and a certain CPU edition, things are bound to go wrong.

I suppose instead of "hello world", you could have "heimlich maneuver".

;-)

Re: Quieter machines

[WuphonsReach]

Have you considered the Antec Sonata case?

(And of course, the other options like quieter CPU fans, quieter exhaust fans, quieter power supplies. For my home office, 2004 is the year of "quiet", my goal is to make a serious dent in the amount of PC noise going on in here.)

Its safe to innovate again

[steveoc]

Dont you just love the way that AMD dropped this 64bit chip on the market - WITHOUT WAITING FOR MICROSOFT TO CATCH UP - Like, trust that Linux support alone will be enough to push this thing into the low-end 64bit market.

And its selling like hot cakes - so the market is proving them right.

Maybe it is a sign of things to come - hardware vendors pushing forward and bringing real innovation back into PeeCees, knowing that Linux alone will be there to support the innovations, and that Linux support is enough to drive sales.

Remember how back in the good old days, Hardware makers (Commodore, Atari, Apple, etc) were free to introduce radical new hardware every 12 months, with no regard to operating software portability - they knew that the software guys were capable of keeping up back then.

The current situation, with Microsoft being the sole supplier of OS's means that any new hardware has to conform to some horrid, and aging 'standard' invented back in the 80's, simply because Microsoft seems to be incapable of keeping pace with innovations in hardware.

Well done AMD - for daring to break the status Quo, and for sticking one up Microsoft at the same time.

Re:Its safe to innovate again

[Satan's Librarian]

Linux isn't the only thing that supports it right now though - most Windows software companies should be considering purchase of one of these, if they haven't already bought one for each developer. Longhorn 64-bit has been available to anyone with an MSDN subscription for a while now, and the first to market apps with 64-bit versions of software will get a big boost in market share when Longhorn goes live.

Re:Its safe to innovate again

[The Lynxpro]

"Remember how back in the good old days, Hardware makers (Commodore, Atari, Apple, etc) were free to introduce radical new hardware every 12 months, with no regard to operating software portability - they knew that the software guys were capable of keeping up back then."

Examples??? Not on the Atari end. The Atari 400 and 800 were released in 1979. After that, you had the 1200XL, then the 600XL and 800XL, the "unreleased" 1400XL and 1450XLD, and the 65XE, 130XE, and the XE game system. The newer models retained compatibility.

On the Commodore side, you had the PET, the Commodore 16, the Commodore 64 and 128, and the Plus4. The Commodore 128 was backwards compatible with the wildly successful Commodore 64. We won't talk about the others.

Apple had the I, the II, the II+, the IIe, the IIc, the III, and the IIgs. We won't talk about the Apple I or the Apple III, but the later models of the II line maintained compatibility.

Now Apple, Atari, and Commodore broke from backwards compatibility when they jumped to their 16/32 bit computers (I'm not counting the IIgs in this argument). Apple Macs were compatible with Apple Macs, Commodore Amigas with Amigas, and the ST/TT/Falcon line with Atari.

So where's this myth of breaking with backwards compatibility from the other great computer companies that you named?

Re:Its safe to innovate again

[steveoc]

True, but from what I have seen with Windows on AMD64 :

1) The 32 bit mode performance is pretty impressive anyway (so AMD were clever to hedge their bets here)

2) The performance of Win32 code on Win64 base is WORSE than Win32 code on Win32 base using this processor.

I dont trust Microsoft to fully support x86-64 till Intel comes to the party as well.

Keep in mind, that if Linux never existed, then Microsoft would hold all the cards, and would be in a position to sink this chip if they wanted to. Microsoft mouthing support for x86-64 is a reaction to Linux being able to support it. They are playing follow-the-leader now, which is a significant change to the status quo.

If that was not the case, and Linux did not exist, then it would have been way too risky for AMD to try this on. They would have put their resources into a cheaper/better/faster Athlon 32bit core.

My point is that I believe we have reached a significant milestone here, where Microsoft no longer calls the shots to hardware vendors, and that as a result of this change, hardware vendors can feel safe to finally break away from the awful, restricted PeeCee architecture over time, since Linux support will allow such new architectures to be instantly useable.

I also doubt that Windows, being so deeply tied to that old architecture, will find it increasingly difficult to keep pace with such changes .. so, bring em on, I say.

I cant wait to go into a PeeCeeSoftware shop after Longhorn is out and see the titles on the shelves. Are we going to have 8 different shelves, with 'Win98 / WinXP / Win2K / Longhorn32 / Lornhorn64 / .NET CLR / XBox / XBox2' titles. I cant wait for the support calls to start rolling in. What about businesses that 'Need' to standardize on Longhorn64-Office, but also need to keep support up for some legacy WinXP apps which dont run under Longhorn64 for whatever reason. Microsoft's stock answer will be to push complete and uncompromising migration to Longhorn. I cant wait to see the confusion and resentment this is going to cause.

This is also going to add a lot of additional burden onto any proprietry software vendor, making support and porting a lot more complex than it already is. Are we likely to see the average Windows developer in a couple of years time having 4 different PC's on each desk just to keep pace with this mess ? Im buying shares in KVM manufacturers I reckon.

Just watch for China in the coming years to introduce something way better, simpler, and cheaper in the way of architecture and even CPU design. If it only runs Linux, that will suit them well - they can leapfrog the Windows-laden west in 1 easy step. They have the tech and the political will to be able to do this.

Its also quite possible that such a rapid advance in mass-market architecture can happen BEFORE Longhorn hits the market, which will really splash some water into the frying pan. Microsoft's long term outlook is pretty fucked up now.

Re:Its safe to innovate again

[erikdotla]

Enthusiast gamers are what is selling that chip, not Linux. PC Gamers have always been the driving force behind adoption of new technology.

Re:Its safe to innovate again

[Anonymous Coward]

trust that Linux support alone will be enough to push this thing into the low-end 64bit market.

Not just Linux. Solaris, FreeBSD, and NetBSD have working ports already.

On FreeBSD, the AMD64 architecture is a "Tier 1" platform, which means fully supported. See here.

On NetBSD, the port seems to be basically functional and mostly done and is expected to be included as part of the next release. See here

Sun has announced amd64 support along with amd64 hardware products as well (Opteron entry-level servers). See here and here. Solaris 10 (which isn't out yet, but which does have "early access" releases available to the public) supports it too. Sun is rumored to be working hard right now on tuning its own compiler to generate good amd64 code. And, significantly, Sun has been selling 64-bit machines for nearly 10 years, so they have pretty much solved the issues of integration, making it easy to manage the combination of 32-bit and 64-bit applications. (Older Sun machines can run either a 32-bit or 64-bit kernel and will automatically select 32-bit apps on the 32-bit kernel but 64-bit (unless unavailable) on the 64-bit kernel with the same OS install image.)

All of which is good, because the x86 market tends to procastinate, and the 32-bit to 64-bit transition is IMHO a bit overdue. (And because it is possible that it will make Intel pee its pants if this really takes off.)

Re:Its safe to innovate again

[Keeper]

I cant wait for the support calls to start rolling in. What about businesses that 'Need' to standardize on Longhorn64-Office, but also need to keep support up for some legacy WinXP apps which dont run under Longhorn64 for whatever reason.

What do you think VirtualPC is meant to be used for?

As a side note, a 64bit version of WinXP should (in theory) be released in '04 (probably the 1st quarter of it). It's release will probably be fairly close to the release of .Net 2.0, which should have 64bit versions of the runtime (ie: all .Net apps get a nice "boost" from the new architecture).

sample buildouts

[erikdotla]

Here's example systems you could build, with the best possible motherboards. Each assumes you need to buy some DDR400 RAM so that is not included, since it's all the same:

My Athlon XP system:
Athlon XP 2800+: $150
Abit NF7: $100
Total: $250
I'm quite happy with it. Best price/performance choice (last week, anyway.)

Top-End Athlon XP system:
Athlon XP 3200+: $289
Abit NF7: $100
Total: $389
A complete waste of money, especially after today.

P4 3.2 system for comparison:
P4 3.2 CPU: $366
Asus P4C800-E: $164
Total: $530
Better than both of the above, but only by a few percent for most things.

That was the situation last week. Including an Athlon 64, Athlon 64 FX, or P4Extreme in the examples would have been useless since they're insanely expensive.

However:

Athlon 64 system you can build now:
Athlon 64 3000+ 512k cache: $230
Gigabyte GAK8VT800M: $106
Total: $336
Yowza.

So, to jump from the top-end Athlon XP to an entry level Athlon 64 actually saves you $53. I could have spent an extra $86 and got all this. The Athlon 64 system will now save you $194 over the best P4 Intel has to offer, and it will beat it (for virtually all applications.) Of course, if $336 is too much, you can still build a good Athlon XP system and cut costs dramatically, but $336 is very reasonable for building a brand new system. It'll be interesting when Intel gets it's P4Extreme down to a reasonable price, and AMD starts ratcheting up the Athlon 64 speeds.

Proves it's always better to wait just one more week. I should have known that there would be major cuts in the 64-bit world soon after the processor debut.

Hope all this is useful to anyone considering building a system. Keep in mind that 1gb of dual-channel DDR400 RAM is gonna run at least $150.

All prices are PriceWatch.com and the Athlon 64 CPU price is from a link on AnandTech. I know PriceWatch prices are hard to get and you have to deal with shipping and all that.

Re:sample buildouts

[bhtooefr]

Athlon XP 2500+: $83
Abit NF7: $100 (I'd use an A7N8X Deluxe here, but...)
A multiplier adjustment to up it to 2800+ speeds (I've heard the 2500+ is an underclocked 2800+): $0
Total: $183

Re:sample buildouts

[Thing 1]

I know PriceWatch prices are hard to get and you have to deal with shipping and all that.

For the last year or so, Pricewatch prices have included shipping.

Re:sample buildouts

[icebones]

ZipZoomFlyhas these for$236 w/ free 2nd day air shipping.

Re:sample buildouts

[erikdotla]

Obviously when comparing system buildouts, overclocking is not factored in since it throws a wrench into everything. But your comments are useful. I chose the 2800 not knowing it was just an "idiot tax". :) For the price of these class of Athlons, you can't really make any mistakes. I wasn't even aware of the overclocking capabilities of the 2500. I'm usually on a budget so I prefer to never overclock, the risk is too great. Too much hassle to monitor the temp, and one slip while mucking with the cooling and you break your board/cpu, etc. I am an enthusaist, but I like to use my computer, not tweak it all the time.

I chose the Abit NF7 after reading a Tom's review of all the latest NForce3/Ultra400 boards. They're all very similar in performance (almost identical) and I was lazy when I bought it - they had it at my store, so I bought it. Would have got the Asus too if they had it. I also would have bought the Shuttle board because it's only about $60 ($110 for the Abit), but the Shuttle didn't have SATA.

I own dozens of Asus boards and do prefer them in general. One problem I've had with Asus, which I could demonstrate on at least four or five boards, is that it doesn't detect the keyboard reliably on boot. Seen this problem across many generations of Asus. Never had this problem with an Abit. It's been annoying me a lot lately so I've decided to give other manufacturers a shot at my wallet.

Re:sample buildouts

[erikdotla]

I meant you have to deal with the shipping process itself, instead of having the convenience of picking it up at your local store.

We all know how much it sucks to order something and then realize you won't get to play with it for a week. :)

Also, if there's a problem with whatever you buy, sometimes it's a lot less hassle to just take it back to the store, yell at a clerk and get a replacement (sometimes even a free upgrade.)

I actually convinced a local store to give me a free Athlon 2000+ when a $25 cpu cooler siezed up and fried my existing chip. A friend bet me and swore up and down that they wouldn't do it, heh heh.

Additionally, I recently ordered from tufshop.com via PriceWatch. After 4 days they still hadn't shipped my order and I got an email saying it would be shipped "soon". I called, talked to "this one guy", and was told I would I would get a call back in 10 minutes. An hour later nobody called, so I called again. Got the same guy, got told the same thing. I got my call 10 minutes later - SAME GUY. I couldn't understand him due to his thick accent. He appeared to spend 15 minutes trying to upsell me some sort of warranty. I screamed and yelled (nicely) that I just wanted my frickin order shipped as-is. I'm still not sure if he understood. He said something about how the warranty option I chose wasn't available for this item and that I had to choose something else. He claimed they tried to call me, but I had no messages and no emails. I said screw the warranty, just ship it. I was assured it would be shipped soon.

I called back three days later (this is 7 days total, 5 working days) and it still hadn't shipped. Utterly disgusted at this point, I cancelled my order. I bought it at my local retailer for $154 instead of $134, and wish I would have done that in the first place. A lousy $20 isn't worth the hassle.

Re:sample buildouts

[Thing 1]

Yes, we are in complete agreement.

Actually I have found cheaper prices than Pricewatch for 250 GB 7200 RPM 8 MB cache drives (Western Digital). They're $150-160, after taking the rebate into account (which means you're out an extra $90 for a short period of time).

Cheapest on Pricewatch is $203.

I've recently lost an 80 GB drive, so my Christmas present to myself is a RAID-5 array of 4 x 250 GB drives, so I'll have 750 GB of redundant storage. Should be enough for a couple years, at least. ;-)

And although I thought I'd get a better deal ordering them, I saw the rebates and decided to support my local merchants.

Now, on things like media (CD-R, DVD+/-R, etc.), you can't beat the Internet. Plus they don't have 90 min CD-Rs anywhere locally.

Re:But.....

[jonbryce]

Of course. I saw an AMD-64 version of SuSE in PC World today.

Haven't seen a 64 bit version of Windows yet.

I just hope...

[twoslice]

That they are free (as in beer).

Review with Win XP, they must be kidding

[Schugy]

There should be additional Athlon-only tests with SuSE SLES 8 SP3 or/and SuSE Linux 9.0.

Your testing programs with intel optimized drivers

[voss]

even then the athlon 64 beats the 3.2 p4 80% of the time...and the fx-51 runs pretty close to the P4EE, remember the P4EE is an optimized THIRTY TWO bit chip... Athlon 64 is not even running in 64 bit mode on those games.

Re:AMD is overpriced

[raodin]

I've found that Tom's benchmarks almost never line up with what the rest of the hardware sites are reporting. Now... Which is more likely fudged a bit. A single, large hardware site, or just about every other hardware site on the net?

Nope

[ZxCv]

These 'reject' chips are being sold as the Athlon 64 3000+. The chips in the Emachines box you linked to are the Athlon 64 3200+. Same clock speed, the 3200+ just has twice the cache over the 3000+.

Performance / Price

[rffmna]

Whatever AMD does, they always make better performance for price.

Budget Athlon 64

[FrostedWheat]

Commodore 64!

(spooky to think that this processor alone has more memory than that entire computer!)

Yes

[DAldredge]

123,345,334.4 pins.

Any more and the world will turn into one like the one Homer was on that rained donuts!

MMMmmmm Newcastle

[Holi]

Oh wait were not talking about beer are we?

Re:When I was your age...

[bhtooefr]

Umm... no. The 68000 was a 16-bit bus, 32-bit core CPU, just like the good ol' 386SX. The 68008 (I think) is an 8-bit bus version. I don't know if there was a 32-bit bus version of the original 68000.

Re:I'm sure

[ChaoticLimbs]

Damn, too cheap for a 'k' huh?

I agree...

[cnelzie]

...people should be completely free to do with as they please. On the other hand, I have seen to many people that weren't tech savvy fall for the ole 'unlock the magic' spiel that someone sold them on regarding some 'el-cheapo' board that they picked up.

There are some people that you can just tell needed that computer to bring a little bit of joy into their life and barely had the money to do much else then get that computer and when you have to tell them that they are out of luck, it's not a great feeling.

Still farting sunshine?

[cwolfsheep]

I found this little gem while going over AMD64 reviews...

#8, I'm typing this on an XP2100+ and we can both sleep soundly tonight knowing this has nothing to do with my ego or with blinders. I am farting sunshine over the fact the A64 performs so well clock-for-clock and I am looking forward to building my next workstation as a 64-bit platform.

Granted it was a few months back, but I think we're all farting sunshine now.

That's right you commie pinko bitches!

[Lord Kano]

Pro Bush, Pro AMD, and Pro DiVX encoding. Gimme some of that 64 bit love!

Re: Quieter machines

[greygent]

I have an Antec Sonata and fucking love it. I don't ever hear it running, I don't hear the fans, etc. Also the fans and hard drive cages use rubber grommets to reduce vibration. Great case.

Multiprocessing Possible?

[clubin]

Hmm... I already decided that my next system would be dual-CPU. Can that be done with this processor? I've so far remained pretty ignorant as to the technicalities behind dual-CPU systems. Does one need an Athlon MP processor to go dual in an AMD world?

Re:Multiprocessing Possible?

[ItMustBeEsoteric]

AMD's Athlon MP is their 32 bit dual processor capable offering, Opteron is their 64 bit dual capable. Athlon64/64-FX are NOT dual capable.

Re:Multiprocessing Possible?

[clubin]

Ahh... well, then, to exploit this learning experience I must also ask what motherboard chipsets are capable of using the mentioned processors in a dual configuration? Are there any obvious motherboard choices?

Re:Multiprocessing Possible?

[BlueBiker]

I've been running a pair of unmodified Palomino XP 2000+ in a Tiger MPX for two years. Other non-MP Athlons can be run SMP using pencil bridge tricks. Try 2cpu.com and other sources.

Of course, this AMD motherboard chipset is ancient now -- you might be better off with a low-end Opteron.

Re:xp2500 all the way...

[aardwolf204]

I got my xp 2500 about 6 months ago for about $100 (retail box too ;), choose it over a 2600 because it was barton (333) and at the time the 2600 was not. anyway, with my MSI K7N2 nForce2 board I noticed that I was able to adjust the multiplier and fsb without having to do any pin shorting tricks. My room mate got the same chip from a different vendor and found that in his gigabyte nforce2 he cannot change the multiplier and if he ups the fsb by 1mhz the system will hang. mine on the other hand is stable up to 2.16ghz.

What gives? were the early xp2500's really unlocked? are they really underclocked 2800's?

someone help me clear this up. any links aprechiated. thanks.

--
1001011001101001 0110100110010110 0110100110010110 1001011001101001

Re:xp2500 all the way...

[TheLink]

"Yeah, but since AMD recently started locking the 2500 chips, it's not such a great deal any longer. The 2800 is the next fastest unlocked chip."

They started locking them? Darn. I have an unlocked one, and so far I've underclocked it a few times and I find that more useful - I run it 24/7. I haven't bothered overclocking it at all.

In my environment, running burnK7 when normally clocked takes it to 69 degrees C, which doesn't give a great safety margin. My O/S is set to powerdown if the temp goes over that, but I'd rather it not happen at all. So overclocking seems out of the question unless I want to change heatsinks and lose the warranty (I'm using a boxed CPU+fan which comes with a warranty etc).

With the 32 bit Athlons, it seems like they come "pre-overclocked". Most are already running near enough to their limits to make overclocking not as useful. Whereas the P4s appear to have a lot more headroom for the low end chips.

The Athlon64 budget CPU looks very interesting. I hope AMD hang around for a lot longer. 64 bit is the way to go but I'd rather not sail on the Itanic :).

Re:When I was your age...

[chrisbtoo]

The 68000 was fully 32-bit. It was common to implement it as 16 bit at the time, but was fully 32 bit.

There was nothing less than 32 bit in the 680x0 line. The full line was fully 32 bit from day one. There definately were NOT any 8 bit versions.

No, the parent was correct. The original 68000 had 32-bit registers, a 16-bit external data bus and a 24-bit address bus. The Atari ST ("Sixteen/ Thirty-two") was so named because of this architecture.

The 68020 was the first in the series to have a 32-bit external data bus, and Atari used it in their TT ("Thirty-two/ Thirty-two").

There was indeed a 68008, which had an 8-bit external bus.

And it's "definitely" ;-)

Or possibly a city

[eggmaster]

I work in the city of Newcastle-Upon-Tyne, where Newcastle Brown Ale is brewed. I find it offensive that you think it is just a name for beer, rather than the name of the town where the beer is brewed.

Re:When I was your age...

[bhtooefr]

If there weren't a 68008, there might not be Linux. The computer that Linus used before his 386 PC that he first developed Linux on was a Sinclair (forget the model) with a 68008. He wouldn't have devoted almost a whole chapter to it in his book if the 68008 didn't exist.

Also, why aren't early Macs "32-bit clean"? It's because of several factors, including that early 68k CPUs didn't have 32-bit bussez. The other factor was that Macs were still optimized to work with a 16-bit bus for the longest time...

Re:When I was your age...

[AndroidCat]

It would have been the Sinclair QL that had the 68008.

With early Macs, one nasty reason that they weren't 32-bit clean was that the 68000 only supplies 24 bits addressing to the outside world, but has those 32 bit address registers .. so they used the upper 8 bits for storing other stuff like memory handles or something.

Re:When I was your age...

[AndroidCat]

When I say "they used the upper 8 bits", I mean Apple of course.

Re:When I was your age...

[Urkki]

• With early Macs, one nasty reason that they weren't 32-bit clean was that the 68000 only supplies 24 bits addressing to the outside world, but has those 32 bit address registers .. so they used the upper 8 bits for storing other stuff like memory handles or something.

Wow. That sounds at the same time both rather clever and incredibly stupid :-). I wonder if they just didn't think about it when they decided to use those upper 8 bits fot that, or if they figured nobody will ever need more than 16 megs of memory...

Re:When I was your age...

[Bert64]

Indeed, Amiga had that problem too, a lot of games designed for the 68000 based A500 used the upper 8 bits of the address registers to store other data, and this resulted in a crash on a 68020 cpu such as that found in the A1200, even tho it was the stripped down 'EC' version which could only address 16mb of ram, it still had 32bit addressing.
However, you were violating Commodore and Motorola's programming guidelines if you did this, properly written software worked all the way up to the 68060

Re:When I was your age...

[AndroidCat]

Since the first Mac shipped with a whole 128K (unexpandable without a solding iron), the 16M limit and the 68020 were way way in the future.

Re:When I was your age...

[Urkki]

Well, considering where the primary(?) competitor MS-DOS had set the limit, 16megs does sound a lot.

Re:xp2500 all the way...

[bhtooefr]

Swap CPUs and try again - see what happens. Also, look at the date code - if it's 0339 or greater, it's locked (has to do with the release of the Athlon 64 - AMD wants you to buy an A64, not overclock an AXP 2500+... and the A64s are toplocked - only the FX series is not toplocked, and this is because the FX is designed to allow overclocking (Opterons are fully locked)).

Re:xp2500 all the way...

[bhtooefr]

Yes, a 2.4C will take you to about 3GHz, but an Athlon XP 2500+ can go from 1.83 to 2.2GHz (sometimes 2.4 if you've got a good heatsink) for a LOT less money). Also, it puts less strain on the system bus when it's a multiplier tweak instead of an FSB tweak. Now, for some math:

The 2.4C gains 600MHz, or it runs at 1.25 times it's normal speed, in an overclock to 3GHz.
The 2500+ gains 370MHz, or it runs at approximately 1.20 times it's normal speed, in an overclock to 2.2GHz. However, basing this on performance ratings means that the overclock is going from 2500 to 3200, running at 1.28 times it's normal speed on performance ratings. This means one of two things: AMD was getting more aggressive with their performance ratings, or they were following the performance curve of the Pentium 4 line (unlikely). Look at the Athlon 64: it runs at 2.2GHz to get a PR of either 3000 or 3200. Someone needs to bench a P4 3.2 800 HT, a P4EE, an Athlon XP 3200+, and an Athlon 64 3200+ to see how AMD holds up, and how aggressive their PRs are. BTW, test your mobos first, and use the best performing mobo money can buy for each CPU. Obviously, you only need one Socket 478 mobo... BTW, why not add in a Pentium M 1.7 once PowerLeap releases their adaptors in 1Q 2004 (or so they said in their e-mail to me) or Intel releases a Socket 478 Pentium M (hint: 1000 signatures on my petition before the end of January 1, 2004 would help GREATLY).

Re:When I was your age...

[BlueBiker]

The 68000 may've been fully 32-bit, but it lacked 32-bit multiply finally added with the 68020. Not a trivial omission for a CISC machine.

MS-DOS !!

[xluap]

Of course you need the chip with 1024KB and not the one with 512KB !

Then all your 640KB dos memory and even your high memory will run from the cache!