posted
If you want absolute best performance, you're getting AMD. If you want really high performance, but want to save a good chunk of money, you're getting Intel. If you want best performance/price ratio, you're getting AMD. If you want best cheap yet solid and versatile motherboard, you're getting Intel. If you want a motherboard that'll continue to be upgradable for a while you're (likely) getting AMD, due to Intel's lack of a clear development course. If you want a power efficient mobile computer with wireless capabilities, you're getting Intel.
There are all sorts of other criteria they can be evaluated upon, but those're some of the big ones.
Posts: 15770 | Registered: Dec 2001
| IP: Logged |
posted
Okay, here's how all this works. There isn't a huge difference between performance of AMD vs. Intel right now. Intel's last few chipsets/chips have been utter garbage (High heat, excessive power loss, etc. ), but they're starting to get back on the train with their stuff. Anything between 3ghz and 3.5ghz for Intel is terrible, aside from that they're great for processor intensive applications (Music and video editing among these) and for a number of business applications.
The big difference between the two is this, AMD does just as much as Intel with 900,000 fewer clock cycles per second, which means they have much more room to work with than Intel does. That's the problem with Intel now. They're reaching the physical limits of current CPU manufacturing. AMD has been working with the Architecture of their CPU's, allowing them to get the same ammount of performance without getting as close to that limit, so they have more experience with the problem that Intel is having right now. As for cost comparisons, Intel has always been more expensive for the performance level of their chips. AMD is starting to price their chips at the same level as Intel, so that's probably not going to be a difference in the future. A couple years down the road, we're probably going to see AMD pulling closer to Intel in the market share area, what with the experience they're getting with 64bit processing now. When Longhorn comes out in 2006, Intel's going to have to play catch up. As for upgradablitiy, Intel, for at least another month, has the upper hand, since they were the first Motherboard manufacturer to produce PCI-Express motherboards. AMD doesn't make Motherboards, so they have had to let other manufacturers do this for them. NVidia has released their latest chipset which has PCI Express capability. The big thing that has determined what market share each company has had is a result mainly of the number of major manufacturers that sell Intel only computers. However, if you use any small computer company, they will invariably use AMD in their lower-end systems and Intel in their business and multimedia computers(That is, if they're any good at what they do).
If you want a Low end computer, Intel is terrible. The Celeron is the most worthless chip that has ever been made. AMD's Sempron blows it out of the water without even trying, and the 2400+ is about 10 bucks cheaper than the Celeron D 2.4ghz. Above High-end, it all depends on what you do with your computer. If you do mostly visual/audio work, go with Intel. The chip has a long Data Pipeline that allows it to handle constant processing better than AMD. However, that same pipeline slows it down for non-CPU intense software, like games (Modern games depend more on the Video card for processing than the CPU, a 1.5ghz processor with a high end video card will only drop performance by about 25-35% for most games as compared with a 3ghz CPU. The newer games take a little more, though). As a result, the AMD tends to outperform Intel in games, especially on a pure frequency level.
So, there's my lengthy and jargon-filled response to that question...I bet I confused more than I helped.
Posts: 3003 | Registered: Oct 2004
| IP: Logged |
posted
Out of my third CPU report. Feedback welcome.
<Opening of report>
Intel, with all due respect, has been meddling with itself for the last year and its advancement is appalling. AMD, since Q4 2003, did the following: Created a major Athlon 64 line, with 2 cores ,an Athlon 64 FX line, manufactured about 12 different kinds of Opterons (models), released several Semprons, made 2 new sockets (I don't count the 940 here,) and replaced the European Support Centre secretary (Just kidding).
Intel, made a Prescott core (bad performance, the cache does not cover up), Celeron Ds; and had two new CPU clock speeds, only 1 for the P4EE. As for the server market, the less said, the better. Just to make Intel feel bad, since early 2003 they didn't upgrade the Itanium 2; the Itanium is STILL not obsolete (even though the Itanium 2 was made because the Itanium's performance was appalling), and the Xeon wasn't significantly changed and now Pentium 4s are at the virtual same capabilities. Intel is lagging behind.
You referred on Messenger to the mobile market. The Centrino idea (of combining Pentium-Ms and wireless chips, need I say with Intel Motherboards and chipsets,) was good, very good. If I had money and a real need of a wireless notebook (and if I lived in hot spots) I would get the Centrino. The Dothan core was a very good one and I would get a Dothan Centrino, if possible. AMD, even though the Athlon XPs (and more relevant now, cheaper Semprons) can have a very good battery life time (assuming they didn't change the Thoroughbred-B and Barton cores and they work the same way; but I was assured that the CPUs are just relabelled), I wouldn't say they are very much MEANT to be mobile chips. I would say, though, that the Athlon XP-M (still marketed?) is a good competition, primarily in price (it lacks automatic Centrino-style wireless networking in hotspots); this, though, is not my area of expertise .
I do not know what's Intel planning to do with the Itanium CPU not being ‘re-architectured’ until 2007, but it MUST be bad if- "For a corporate desktop, the Pentium M is a wonderful chip". How far down did we go? In addition, the "Potomac"'s two benefits are a 90nm process and a 4 GHz speed by its release in (Q1?) 2005. But for Christ's sake, if the 4 GHz line isn't broken, the Athlon 64 4000+ will be a vast improvement over Intel (remember that lowers 3400+, 3500+ and 3700+ prices). As far as I can see, the next "Nehalem" CPU core (according to Q2 Intel roadmaps in PC Magazine ) is to be released in H2 2005, which is a LOOONG time for the predecessor Prescott's decline. By then it will be a "Pentium 5" (is it not ironic that the Pentium was released in the first place to 'change' the numbers' marketing CPUs were being marketed at? Isn't it now a Pentium II, III, 4, 5? Never a "Hexium" or any other thing).
AMD is already going through their "Winchester" plans, and their Paris and Victoria (possibly Dublin too) could be out by December. Since the plans are [virtually] ALWAYS postponed in technology, it is quite amazing that AMD might still keep up almost entirely (in time) with their roadmap.
AMD more or less spread out their plans evenly so that in one year they do not have products lagging behind without any real use. Intel, which had the Celeron like that, suddenly made a big (yes, it was big) leap with the Celeron D (that is finally a Value CPU and not a cheap old lump of useless crap); who guarantees the Itanium will not be there too in 6 months?
What is even more shocking is Intel's press release. Here are some quotes:
"Over the past few years, Intel has been making a conscious shift towards delivering broader enhancements to the PC platform beyond the traditional "Clock Speed" metric."
Kindly excuse my term, but it is absolute Bullshit!
As The INQUIRER says:
"Years ago, when the architecture of the P4 was being developed, someone, somewhere, made the decision to prioritise clock speeds over everything else. The design goal was to deliver MHz numbers that no one could possibly match even if they wanted to."
Alternatively, as my “AMD and Intel” CPU report says: "Yeah, right. Intel was the only company that decided that in order to make the Celeron a good 'value' CPU, the Cache (On-Die) could be removed. The decision was so pathetic and the CPU was so poor-performing, that they hurried to replace it by the Celeron® A."
The stupidity does not end there:
"The most recent example of this shift occurred in 2003, when Intel launched a new computing platform for notebook PCs: Intel® Centrino™ mobile technology." What do you say…! Intel did that for a very specific reason, the lower power consumption that those CPUs will have, allowing laptops to last longer without the need or recharging. Apparently, until 2003 it never quite sank into Intel's mind that mobile computers would benefit by having more cache, and all sorts of weird technologies (that are developed non-stop) to increase productivity at lower power costs?
Best of all:
"Market research had shown that notebook users valued other features beyond clock speed."
I feel like dropping on my head. Of course it is that users prefer notebooks that have lower clock speeds, higher productivity and long-lasting batteries! Now, hasn't Intel ever "learned from their enemies (in this case rivals)"? Haven't they ever heard that the Athlon XP's basically equal performance to the Willamette Pentium 4, and even the early Northwood CPUs (no point arguing, they got fairly similar benchmarks for a whole GHz difference in speed and double the cache in Intel's CPUs) was not despite QuantiSpeed but rather because of it?
So, whatever Intel is planning, going the AMD way full-scale won't increase their reputation. What The INQUIRER said was – "So, if you can't increase the performance through clock or bus speeds, how do you compete? Cache. Intel has no peer in putting out innovative silicon in volume. If anyone can add cache on a whim, Intel can. It has already increased the cache from 512KB to 1MB from Northwood to Prescott, and now it is going to double it again. It has the capacity, and will be using it to the fullest extent possible". But moving the market towards the Pentium 4 Extreme Edition will be another mistake. When people said that Intel should start concentrating on things other than clock speed, they meant that they should not be obsessed with those other features. Now, AMD's clock increments seem immensely faster than Intel's, and they are as usual as always. Furthermore, the Pentium 4 Extreme Edition with 3MB of cache does not cost around US$1000 for nothing! It is the cache! What Intel are doing now, is try and move the Pentium 4 to the EE's direction. What will be left of the market?
I'll tell you what. A failed Itanium and Itanium 2 series, a Pentium 4EE that costs more than an Athlon 64 FX and whose performance isn't THAT higher to justify that. The Pentium 4 (5)'s price will sky-rocket and the Celeron will be (as usual) a handicap CPU whose performance won't even be close to proportions between the Pentium and the Pentium EE. The mobile CPUs will be used for servers thus obsolescing the Xeon.
Now, does that sound like a company that will easily push AMD back to the 2002-early 2003 status it had (or better, 1996-7)? Will AMD be unable to start taking shares out of Intel?
I quote Paul Hudson quoting Jerry Sanders (the Chairman of AMD): "I'd like to say they (Intel) are big, strong and wrong. They can't stop us!" That statement appears to be something that is quite true, all-round; Intel is big, strong, and lately they appear to be miserably wrong. They are unable to stop AMD until 2006. As for the statement, that even of AMD does everything possible to catch-up and it spends these 5 Quarters perfectly, it gets stuck with the problem of the actual “chunk” that AMD “chunks” out of the market: it isn’t too big. Even so, assuming all goes right, Intel’s current 78% (or so) of the CPU market could easily drop to 60%, and AMD’s could rise from 21% to 37%. (The numbers are rough assumptions, the actual figures could vary by ± 4%.)
<Finalisation of report>
What I'd say is that AMD's now offering a bit more here and a little more there and just beating Intel step-by-step.
Printer-friendly view of this topic
![[Razz]](tongue.gif)
