Slashdot Mirror
"Budget" Chips go Head-to-Head
StewedSquirrel writes "Anandtech has published an article taking a look at the low-end of the CPU market today. It takes Intel's newest Celeron processors against the AthlonXP and Duron with a Pentium 4 1.8GHz thrown in for comparison. All of these processors will cost you under $120, but the article shows that the old Duron (at barely $40) can out-perform Intel chips costing nearly 3x as much. In addition, it shows that the performance of the Athlon XP is head and shoulders above the Celeron processors, while costing roughly the same."
I notice that Anandtech describe an 800MHz machine as "chugging along". Hardly. One of my older machines is an 800MHz Athlon Thunderbird machine with 256MB RAM and a 40GB disk. It runs GNOME and WinXP without any problems and continues to be extremely responsive and perfectly adequate for the vast majority of tasks I throw at it (except Games).
The same is true for budget chips - if you want a machine to go online, to do Word Processing, play a few older games or whatever, these chips are perfect. Putting together a full-blown capable system for $400, or buying secondhand, is a great way for people to get in to computing who couldn't otherwise afford it. Getting them on the bandwagon is the important thing, and whilst the hardware is so far ahead of the majority of software (at least until Longhorn comes out...) getting more people using computers in their homes is a really good idea.
I see AMD advertisements on the web all the time, but they don't seem to have much of the "big name maker" market. Why not? Is Intel so intreched that their value doesn't even matter any more?
AMD seems to have been kicking Intel's butt for a little while now technically.
I'd love to see some brand name servers start using AMD chips, look at what AMD's doing on the low end!
-Pete
Soccer Goal Plans
I read the article with two Intel banners. There goes your conspiracy theory.
And a "not trying to troll" back at you, but I've used both AMD and Intel at home (seems every office PC I've used has been Intel) and never had a problem. I run a webserver, both SQL and Oracle DB's on them, do all of my side gig development on them. I have to admit I did once have a mobo problem, but that was an aBit KT-7 RAID board, which turns up in google all the time with problems. I turned it into a pretty cool looking wall clock.
I actually have a K6-2 (400 MHz) still running at home, as a matter of fact. My "fastest" is an Athlon XP 1900 (time to upgrade again)... never a problem with any of them.
I wonder, then, what the difference between your experience and mine is? Do you typically buy the top of the line or one-offs? I usually stick to one-off's regarding performance, and I wonder if you've been experiencing newest run problems.
I dunno, it just makes me curious.
Saying Android is a family of phones is akin to saying Linux is a family of PCs.
... that whenever I bought a new motherboard + CPU, and then after 6 months decided to upgrade I would ALWAYS have to by a new motherboard + CPU.
They changed their CPU specs faster than I change between my two pair of socks. (almost..)
It was like whenever they released a faster Celeron or P3 you would have to buy a new motherboard because the number of pins were (your current pins) + 1, and then we had the Slot-1 to socket 370,371,372,373.... Dunno where we are now.
I'll be modded down as flamebait or offtopic, when this is stuff that matters (...) Still stick there at 1.25GHz on your G4s apple? Tsk.
May not be that much off-topic, actually. It's blatantly obvious nowadays that clock frequency isn't closely linked to performance, especially when comparing different architectures like PPC and 386.
I don't think it would be nonsensical to run a benchmark comparing PPC, Intel x86 and AMD x86: if you read a few of the articles at Ars Technica, you will see how incredibly complex the P4 is. In these conditions, it wouldn't be surprising that Intel's chip is as different from PPC as from AMD's chips.
OK, I don't see how AMD AthlonXPs and Durons outperforming Intel Pentium4s and Durons is news, but does it really matter much anymore? Those are desktop CPUs and that's a declining market. High-density servers and portable devices need low power-consumption CPUs and that's where Intel is way ahead of the competition. I'd like to see a performance/watt comparison of AMD's notebook CPUs and Intel's Pentium M line.
1. RISC vs. CISC.
Dude, go do some research on the latest Pentiums. They may still be saddled with having to support a CISC instruction set, but they are primarily RISC processors "under the covers". Plus one of the primary selling points of RISC way back when was that since it used a "simpler" instruction set, they could clock much faster. Well uh, that blows your comment out of the water. Fact is, a 2 Ghz PPC can outperform a 3Ghz P4 because of design decisions made by Intel on how to achieve performance.
2. Bus speeds.
How of the things that has hamstrung Apple for a long time WAS their lame bus speeds and general lack of performance in the systems surrounding the cpu (slow memory, slow support buses, etc).
3. instruction speed. this is where AMD shines!
Once again you blow your first point right out of the water with this statement, and reaffirms what I said earlier. Intel chose to go the route that says "we'll achieve performance improvements by architecting a system that allows us to easily increase clock speed in leiu of doing more per clock". Others, like AMD, choose to achieve more per clock at the expense of making it harder to scale their clock rates up. Go read about pipelines to get a better feel for where these tradeoffs are made.
Oh a couple more things. 1) Comparing Apples to Sun (by which I assume you mean to compare PPC to SPARC) is just as meaningless in the context in which we're speaking. 2) you should be modded down, but not because of your opinion, just the general lack of accuracy in your post.
Is that AMD chips burn out 3x faster than Intel chips.
AMD procs run hot but that is why there is such a thing as a heat sink. Many people do not know how to properly use a fan with their heat sink, and many have never heard of silver compound. If you dont cool the AMD proc well you will have problems. So the word is heatsink... say it with me now h-e-a-t-s-i-n-k.
If I wanted easy I wouldnt be an engineer or a patriot.
That has top do with compilers.
The tests that showed the 2.6's having an advantage were optimized with VC++ for the p4. Especially the fortran tests.
The macs software was optimized for the G4. Newer optimized compilers should come out soon. It explains also why Jaguar was 32 bit and not 64.
http://saveie6.com/
Ahh, yes the "proof" of some ranting idiot on Slashdot. Excuse me while I change my company's server standards from HP Proliant/Intel to err, now hang on a sec, there is something, err, nope I still don't see any server with anything like a pedegree that runs AMD chips.
This is nothing to do with any conspiricy, it is merely because Intel has been the best manufacturer for many years and most of us who spec out servers for companies have been bitten badly in the past by AMD chips and are not likely to choose a chip from a manufacturer which has only been targeted at the server market for a few months.
You don't get overclocking advertised for servers.
I'm running an AMD K6 1.1 GhZ chip on the machine I'm using at home, and I feel no need to upgrade right now. Nothing I'm running demands the extra performance... even though I know the cheapest eMachines on the market today is running at double the clock speed.
It seems to me that most consumer users don't need 2+ GhZ chips, but marketing over the years have told consumers that higher clock speeds always equate to better chips, even though that's a myth that Apple has worked hard to counteract.
Maybe it's time for a new generation of "budget" chips that don't headline their clock speed but instead compete on cost, power consumption, or other factors....
Has anybody noticed that Anandtech is testing this budget, $60 to $90 processors using a radeon 9800Pro w/256mb of video ram? That's a $400+ video card!
Is people really buying this kind of video card on a budget PC? I'd rather test the processors using a budget video card instead. It might become the bottleneck in some games, but I think that's what the consumer wants after all... an idea on how much faster their game will run on a realistic machine, not in this monstrosity.
What kind of a sad joke is that?
That's okay, I won't really miss any source of info that stretches the graph to make a 1% difference look like a 200% difference.
It seems like AMD has fixed their heat problem with the XP line. My old Thunderbird 1.2 GHz ran pretty hot (60+ c), but this shiny new 2600+ actually runs much cooler (40-50 c), not to mention more powerful.
Also when AMD came out with their new "product rating", like most prople I was skeptical. However, the ratings do seem to be accurate. In every one of these benchmarks, the XP 1700+ smokes the P4 1.8 GHz.
*cough cough*
If that damn capitalism didn't wipe out DEC, and the next gen alpha was developed, it would have blasted all those other CPUs. Alphas always were the best. I shed a tear every day for the murder and slaughter of Alpha.
"If anyone needs me, I'm in the angry dome."
Heh heh - - I pushed it to around 160 on a generic VLB mainboard and thought I was in heaven. I've built more systems for myself, friends, family, and work than I can remember, and every one has been built on AMD. CPU related stability issues have never - - and I do mean NEVER - - been a problem. My years of system building have convinced me that, when stability is a problem, you should eliminate drivers, physical connections, adapter cards, and mainboard components in that order. I know bad CPUs do surface occasionally, but I think that most people get themselves in trouble through pushing voltages/clock cycles and not compensating with good cooling.
I hate seeing money wasted, and the Intel name to me has the same connotations as "BMW" - - it's more about hype than bang-for-buck.
It's only funny until someone gets hurt. Then, it's hilarious.
I have many old(er) computers around here. My fastest is a Pentium III laptop, and my slowest is an old Pentium 133 box. Basically, whenever I got to the point where I needed an upgrade, I just got more RAM. Cutting down on swapping has brought me significantly greater performance improvements than having a "faster" processor.
Further, I am sick and tired of the market hype that surrounds clock speeds. It's not the processor, but the software that needs to be made more efficient. And because many programs spend a lot of time processing graphics and GUI stuff, I think that making video boards "smarter" by adding GUI-specific processing features would bring a significantly greater performance improvement, by offloading crap from the main processor, than speeding up the main processor.
All of that said, it doesn't surprise me that an old processor is "faster" than a newer one. There are hundreds of variables affecting their respective performances, the biggest one being the software used to test them. In most cases, I think it's like comparing apples to oranges.
Personal computers have gotten so fast and powerful in the past couple of years that I think what's under the hood is totally irrelevant to 90% of the users. The other 10% have specific needs because of high-end applications or something.
I've got a desktop system with an AMD K6-2 500 Mhz processor and 512 MB of RAM. The hard drive is a Western Digital 7200 rpm with 8 MB of cache.
And Dell still ships new machines with 4200 rpm hard drives.
Sure, I could buy a new 3.6 GHz system, but it would be slower than the one I've already got.
I've been building fast machines on a budget for the last 7 years. What most people fail to realize is that the average desktop user never uses more than about 300Mhz of processing speed. The rest of the clock cycles are spent waiting on the hard drive, memory bus, ethernet card, or the modem. My system building strategy is this:
- I buy the fastest hard drive I can afford. I get one with the largest cache offered.
- I use motherboards with the fastest system bus offered.
- I buy as much memory as I can afford.
- I spend the rest on the processor.
Anything above 1 GHz is simply irrelevant; I'll never use the processing speed. However, adding RAM and a faster hard disk does noticeably improve performance.And I always smile when people compliment me on the speed of my Macintosh (I've got a blue case) and I tell them it's a 500MHz PC. They can't believe that a processor "that slow" could be so fast. As if the processor speed made any difference.
It's not the hardware, it's how you configure it...
The society for a thought-free internet welcomes you.
Great Googly Moogly! When did not wanting to hassle with building your own box start equating to "major brand"?
Look, I'm well past the stage in my life where I can afford to piss away a weekend putting a box of parts together (though fun it may be). However, I haven't bought brand-name in years. My last 3 PC purchases (and those of a couple of clients) were spec'ed out by myself and built/certified by a local PC shop that I've come to trust over the years.
These guys are great, can acquire any parts I request, have a generous support/replacement policy (I know I can walk in with a flakey piece of hardware and they'll give me a new one on the spot), and they sell at a fair price. Sure, they can't compete price-point wise with Dell or hand-building parts from Pricewatch, but I try to support the local guys, and these guys reward my loyalty with solid machines and great service.
Method of processing duck feet
For the next few years/decades, there is a definately a market for desktop CPUs. So, yes, this article definately matters.
Biodiesel : domestic, renewable, clean, and in the fuel tank of my bone stock 2002 New Beetle TDI
Hmm, you're experience with Intel chipsets (or, more to the point, the drivers for said chipsets) must have been better then mine. Sure, Intel's drivers have almost always been better than those for VIA, ALi or SiS, but I'd hardly call them top-notch. Especially early revisions of certain chips were particulaly bad. For example, the first drivers for the PIIX4 southbridge (used with the 430TX and 440LX northbridges) were a disaster. Intel ended up getting them right, and by the time the 440BX came around (still using the PIIX4 southbridge) they were great. Similarly the early i8xx series of drivers were rather troublesome, but by the time the i845 chipset came out they were rock-solid.
:> ).
When it comes to the competition though, nVidia has Intel beat solid. Their drivers simply install and work on any one of their chipsets. Intel has been working to improve their drivers, but they still have a little ways to go to match nVidia if you ask me. And if you want an nVidia chipset, you need an AMD processor (or an XBox
My Athlon box sounds like a buzzsaw Mine does too. Athlon 1.4. The generic fan that I had would work for the desktop, but as soon as I tried to convert some movies or anything using 100% cpu it would overheat and (thankfully) shut down ( vs. melting down). Got a ThermalTake Volcano on there, and it is extremely loud, but there have been no overheating problems. I am tempted to try that Arctic Silent thing.. really getting tired of the noise. calamari
On a related note, ExtremeTech has an article detailing how to build a fast PC for $800. The final recommendation uses an Athlon XP 2500+ CPU with 512 MB RAM, 120 GB hard drive, and a GeForce 5600XT video card.