AMD Releases Barton: Athlon 3000+
Harle writes "Today AMD has introduced a new version of the Athlon, codenamed "Barton," that features twice as much L2 cache as previous chips. Along with the increase in L2 cache comes an increase in the Athlon's performance rating -- specifically the new 2.17 GHz chip is rated at 3000+.
The clockrate is actually slighly lower than the Athlon XP 2800+'s 2.25 GHz speed, so the question becomes "Does the cache improve performance enough to counter the loss in clockspeed?" For the most part, the answer seems to be "yes," however, it doesn't unilaterally stand up to the 3.06 GHz Pentium 4.
With the recent delay of the Athlon 64 to September, this is AMD's top desktop chip for some time to come. The reviews are starting to pop up at Ace's Hardware and Extremetech." There's also reviews on The Tech Report, SimHQ, HotHardware, EarthV, in Norwegian on Hardware.no, and last but not least AMD's press release. I'm sure there's many many more links, but I'm tired of pasting them all in here, so post 'em below. *grin*
AMD better get there act together and get the Athlon 64 out in september or sooner. Intel is just kicking there ars right now and AMD has nothing to compete with that 3.06 ghz with multi-threading and whatnot. Dont get me wrong i dislike intel but unless AMD's next big thing is BIG, then they could be in for some trouble
Good point, except this will never happen. Intel have based their marketing around the clockspeed, and to go back on that would be distasterous for them.
AMD have to pander to the 'OMG 3.2 Gigawats is better than 2.3Googawits!' idiocy.
The people who care about the difference between AMDese and real megahertz, already know. Joe Public doesn't honestly care; an Athlon 2000 is a match for a Pentium IV 2000, and that's all that really matters. AMD aren't on the fiddle; they've been entirely fair with the ratings at which they market their chips (and the temptation to inflate a little would be considerable...)
As for asking Intel nicely to help out AMD's marketing department, what colour is the sky where you live? The Pentium IV is designed to get big megahertz at the expense of actual performance; why would Intel throw away their chip's advantage like that?
Real Daleks don't climb stairs - they level the building.
I'm still using a pentium 3 on my workstations, but I recently upgraded a server to XP 2ghz and it does help for CPU intensive stuff like parsing text or XML. Other than doing benchmarks or running heavy weight apps, anything over 1ghz is pointless for average user. You're better off getting a 1ghz with 1gig of ram.
The first Athlon's came out in the summer of 1999 at 500-650 speeds. By my count, that was about 3.5 years ago. Athlon is obviously prepping the Athlon 64 for launch within the year, so they are moving on to a new design, but the Windows world doesn't seem ready to make that big jump quite yet, so they've done a good job of extending the line of the 32 bit CPUs. They've redesigned the core a couple times, and increased the bus speed from 100/200 to 166/333, and it's quite possible they'll make it to 200/400 by the time they're done.
There could be significant compiler differences. For instance, how well does gcc use SSE2/HT? 3DNow!? etc. If you don't use good compilers, then you will get criticism like "You are favoring AMD because you aren't using the advanced features of the P4 - SSE2/HT" (or vice versa).
Benchmarking is tough to do "fairly".
Wow, a great idea, I can see it now...
Intel has a faster processor than us, and we can catch up or even beat us, but that would only be an small percentage increase.... I say we just sit on our asses until we can double our speed. We'll be bankrupt by then, but users don't care about the releases.
If you are saying they should be doubling the speed at the same intervals as current incremental changes, you are being ridiculous. They are moving as fast as they can. This is what a competitive market does. They try to move any slower than they possibly can and competition leaves them behind. This is why the x86 platform is becoming much faster much more quickly than other platforms, the fierce competition.
If you don't like the small speed increase releases, just ignore them and pretend they never happen. For example, if you have an Athlon XP1500+, pretend that every successive release until now never happend. Then you'll be happy.
These are not meaningless speed increases. If you have the 2700+ processor, the 3000+ is faster, but not worth an upgrade (not yet anyway). If you have a 1500+ processor, this release is bound to make the 2700+ more reasonable, or even 3000+. Manufacturers do not expect a consumer to buy into every release cycle, but they expect different consumers to be ready to buy into different cycles at different intervals.
XML is like violence. If it doesn't solve the problem, use more.
Processors that can move at such speeds - and, incidentally, keep requiring larger and larger heatsinks - only have practical applications on high-performance server (clusters). The average desktop user simply doesn't need all that, even for gaming; if you go to the store it's rare that you'll find a game that also requires you to have a computer with over a Ghz of computing power. Unless you're toying with some seriously heavy graphic or music manipulation software, it's next to useless.
So why is the hype aimed at so many desktop users? Simple: it's the largest market. Do we, the endusers, need it? No. Are we going to buy it, with the economy in the tight spot it's in? Nope. We're going to upgrade their memory sticks and leave it at that. I've got a trusty P3 600 which works fine with my GeForce 4 to run NWN at awe-inspiring resolutions and graphics, it's got 512mb so it's smoother than a narwhal, and I for one see no need to buy a new one anytime soon. The net result is that the intel/amd power struggle has been so intense that there's no point to it anymore. My system is still configured for gaming, but a lot of people - in companies as well as at home - only use their computers to email and write letters and maybe listen to some music. Like as not they'd much rather save for a 19" TFT than another tower. I own my own little IT company and generally advise my clients to stick to their 450mhz machines and upgrade a few choice parts.
The only thing I'm wondering is how big is the group that seriously uses such powerful machines? I can understand major websites or software companies will have clusters, but that can't be much more than a few percent can it? Anyone have an idea?
- Jynx
A positive attitude may not solve all your problems, but it will annoy enough people to make it well worth the effort.
As the AC who replied to you said, on Athlons, the caches are exclusive. The data will be in only one at a given time, so it is really effectively 640k. On the Intel chips, the data in L1 is also in L2, so the total is exactly the same as just the L2 size.
My server
I've assembled many systems for value-conscious people around, most being Duron 700-950. Nowadays the Duron 1200-1600 are nice bets since any higher clock would push the price by a bigger margin.
The vast majority of customers, both OEM and custom-assembled, really couldnt care about the psychological 1GHz or 2GHz bump, or getting the very latest processor. The real competition is the number of processors sold, and everyone can see the Athlon has always outperformed the Pentium4 in price/performance competitions. Give the new power ratings, with the Athlon chugging lesser watts while pushing the cache for a more reasonable performance figure, I'd say AMD will come out the winner.
Most customers quite simply dont want a $3000 system with the latest and greatest parts. Sure such a market exists, but theyre a loud minority. There are parents buying systems for their kids, their offices and college guys for their collegework. There are ordiniary people who want ordiniary computers that just do the job reliably, you know, use MS Word, browse the net, maybe watch a DVD. You dont even need a processor clocked over 1GHz for this, a Duron 800 with 256MB RAM and a Geforce2MX card can hold its own even in todays market. Remember very few are really buying Windows XP right now.
In planning for the future, AMD should not ignore the FSB for the Hammer, nor should they ignore the power ratings. The price has always been their edge, but having the only 32/64-bit processor, they could even afford to jack it up a little assuming Intels 64-bit doesnt do too well with 32-bit code. What bothers Intel and Microsoft right now is that people are perfectly happy with a low-end machine, and will continue to be for a while. The whole North Amerian market is coming closer to saturation, and poorer countries would have a bell curve centered much closer to the very low end of America's computer buying bell curve. All this points to the next boom in extreme value systems, where AMD again has kept their edge over Intel, only to compete with transmeta and the C3.
"Give orange me give eat orange me eat orange give me eat orange give me you." -Nim Chimpsky
A more important question: would you be able to even notice a performance gain?