|
No doubt that if you are an avid reader of tech sites, you're well aware of the latest and greatest from AMD and Intel. Last week, AMD released their Athlon 2600+ to reviewers, and this week, Intel did likewise with their Pentium 4 2.8GHz. You've seen the benchmarks, and likely you were impressed with some of the numbers, as well as the overclocking abilities of both CPUs. More likely however, you probably balked at the pricing of the chips. Well, probably not so much for the Athlon, but it's probably a good idea to bring back some cans for recycling.
I'll admit, when I saw the benchmarks, I wanted the chips, but in truth, economics are a factor. Contrary to popular belief, not all webmasters get everything for free. Yessir, I'm a commoner, just like most people, and if I want something, it's going to come out of my pocket. Don't get me wrong, if I could, I'd buy every new thing that gets released everyday, but when you got a family to care for and bills to pay, I could probably do without the $500 "Butt Massager Chair".
I would imagine that many people live through the same scenario. Granted, if you're an enthusiast, chances are, you already have a half decent CPU. Maybe if you're still saddled with an Athlon Classic (pre-Thunderbird) or a Pentium 3, you might want to start saving up sooner, but judging from the system specs posted by our readers in the forums, I'd think most of you have decent setup, so you need to ask yourself, "Do I really need more?". I asked myself this very question, as I gazed at my Pentium 4 2.4GHz (533 FSB) and my Athlon 2100+ systems. I dropped a lot of cash picking these up, so am I ready to spend more?
Before I go any further, let me just say that if you don't do anything more intense than web surfing and email, the answer is a resounding no. You will not see any benefit from a 2.5GHz+ CPU. For the rest of you, it's not such an easy decision. Nobody plays just games exclusively, nor do I think people exclusively work all day on their home PCs. I actually push my computers pretty hard, but about 80% of the time is spent doing "real-world" chores, such as work and gaming, whereas he other 20% is spent doing synthetic benchmarks for various articles.
As you probably know, reviews use a great deal of benchmarks to demonstrate the power of a particular item. Although benchmarks tell a great deal about the product, in the cases of motherboards and video cards, benchmarks don't always tell the whole story. Extra hardware and software features should be weighed equally in these cases, and a reader, along with reading several reviews to get a feel for the product, should make a decision based on the overall package. CPUs are a little different, where the benchmarks tell a greater tale, but in the end, it should be the price/performance ratio that determines if it's a worthy purchase. Any legitimate review will always have a comparison product, and seeing the benchmark differences between processors is one part of it. After that, most reviewers will give general impressions, and any special notes about the CPU. Things such as overclocking and CPU steppings are usually discussed, which may or may not be important to you. Although not a mandatory requirement in reviews, most reviewers will discuss the pricing, which for most readers, does matter.
Going back to whether or not you need to upgrade, you have to factor the following: benchmarks, analysis, and pricing. This will also have to set in the grand scheme of things, that is; what are you going to be doing on your computer. Although we're not a gaming site, most of the writers here are gamers, and I'd imagine many of you do play some of the more modern games (and no, Minesweeper XP does not qualify). Let's face it, despite the mainstream press only , I'd think most of us already know that if it weren't for the 3D games on the market, there would not be any reason for the latest high-end computers.
What's important to note, especially for gamers, is that the majority of the 3D action games are video card limited. What does that mean? Well, at low resolutions, yes the CPU you have will matter, but at higher resolutions, it's the video card in most cases. Now, for arguments sake, let's just assume the majority of you have at least a 17" monitor. I'm going to go out on a limb and also assume most of you have at least a GeForce 3 class video card. In such a situation, most of your gaming is likely at 1024x768 and up. With the higher speed processors, the performance delta will actually taper off at these resolutions as you increase resolution. Long story short, you can play at 1600x1200 with an Athlon XP1600+, and still get the same performance with an Athlon XP2100+. You can read some of the scaling articles at , but I whipped up a few examples here&
Test Setup:
Athlon XP1600+, MSI KT3 Ultra-ARU, 512MB PC2700 DDR, Visiontek GeForce 3 Ti500.
Athlon XP2100+, MSI KT3 Ultra-ARU, 512MB PC2700 DDR, Visiontek GeForce 3 Ti500.
Intel Pentium 4 2.4GHz, MSI Max2-BLR, 512MB PC2700 DDR, Visiontek GeForce 3 Ti500.
All tests are run at 1600x1200 with everything on, although sound is off.
The differences between the slowest clocked processor and the highest clocked is about 2 frames per second.
At 5478 for the Athlon XP1600+, and 5487 for the XP2100 (the Pentium 4 finished in between), there isn't a whole lot of difference between the 3 CPUs.
This isn't a video card review, I know, so for arguments sake, let's lower the resolution a bit where the CPU should play a role in performance. We'll be knocking ourselves down to 640x480. Up until this point, we're comparing two different CPU architectures to one another, so the differences in clock speed don't really illustrate identical architectures to one another. With some overclocking, let's see the performance differences between the CPUs&
There's a 6fps difference between the stock speed XP2100+ and the overclocked one. I omitted the Pentium 4 results because overclocking the P4 requires increasing the FSB. The reason this is bad (for these tests for this article) is increasing the FSB will further increase performance than changing the multiplier alone.
Now, I know not everyone plays games, so let's take a quick look at SiSoft Sandra&
Scores are very close again, within 100 points for all tests. Although the faster speed CPU does win out, you have to keep in mind that these tests are synthetic.
Final Words
I think it's pretty clear that the faster CPUs get, the better benchmark scores you get. Of course, everyone knew this before I wrote this article. Unfortunately, I don't have any of the newest CPUs to test, as I'm sure the SiSoft scores would have been more interesting with an Athlon XP2600+. But in the end, these are just benchmarks. The truth is, between my P4 and AMD systems, I'd be hard pressed to tell you which one is flat out faster. Both pretty much "feel" the same to me in all my PC chores. If you're read our P4 2.4GHz review, you'd know we got a nice 2.754GHz overclock. Although the benchmarks were nice, there wasn't a whole lot of real world performance gained in our day-to-day activities. I don't run any database servers at home, nor do I do a lot of video editing, so the extra speed didn't do much for me other than bragging rights on our block.
I'm not trying to say people shouldn't upgrade. Heck, if you got money to burn, go for it. If you got a slow CPU, you probably should consider it. I have a Pentium 3 500, and I'll flat out say that this machine is incredibly slow in all areas when compared to my other systems.
For the rest of us though, unless the CPU offers a completely new architecture, you may be better off holding out, especially if your current CPU is less than 6 months old. These CPUs should provide enough horsepower until AMD and Intel release their next big guns. Until then, save your money, and wait a bit until something really new and worthwhile comes along. I know I am.
Update: Leave it to our informed readers to pose some interesting points. One reader pointed out that a lot of gamers are interested in a minimum frame rate. As a gamer, I'd agree, but at the same time, with the power of todays processors, I really don't think anything over 2GHz is going to make much of a difference. Sure, having more power will always be handy, but let's use Quake 3 as an example. It's been long discussed in various Quake forums that the minimum FPS needed for all the fancy trick jumps is about 125fps. Assuming you have a half decent video card, I was doing this with my 1.2GHz Thunderbird. Now, I understand that it's best to not have a lot of framerate drops in excess of 60fps, but most games will allow you to cap the fps at a set number.
Another reader pointed out minimum system specs for a lot of todays games. Although I understand his (or hers) point, I'm going to have to disagree there. Just because a game requires a 500MHz, and recommends an 800MHz CPU, doesn't mean it's going to run the way designers meant it to run. Shoddy game coding aside, I personally try to double whatever the recommended settings are. In my opinion, that'll provide the best gaming experience.
I'm not opposed to buying the fastest hardware you can afford. It's true, as demonstrated in our benchmarks, that a 100PR (Athlon 2100+ to Athlon 2200+) jump isn't going to do a whole lot. A jump to the Athlon 2600+ will make a difference in benchmarks, and may make a difference in day-to-day computing (I can't elaborate since I don't own one) so whether or not you should drop some ducats on one is your choice. I think if you're hurting for power, you can't really lose, but for the sake of bragging rights, well, that's your choice.
When is the best time to buy? Even if you buy the 2600+ or 2.8GHz P4 today, and sell your 2GHz+ CPU, you'll be crying the same tune come Christmas time. There isn't really such a thing as "the best time". I would suggest waiting another couple months for some new technologies to come out, and maybe perform a full system overhaul, but then we'll end up spending more money. Sucks to be an enthusiast, eh?
Agree? Disagree? Discuss it in our forums
Home
|
