Original Link: http://www.anandtech.com/show/1117
Intel Pentium 4 3.2GHz - The Real Slim Shadyby Anand Lal Shimpi on June 23, 2003 11:23 PM EST
- Posted in
Around a month ago AMD released their Athlon XP 3200+ based on the Barton core. While only outpacing the 3000+ by a meager 34MHz, the new processor carried a much higher model number courtesy of AMD's move to a 400MHz FSB. The review community unanimously agreed that the processor was not deserving of its 3200+ rating, but none were as infuriated by AMD's model number than the folks at Intel.
Ever since the introduction of AMD's model numbering system, Intel has been pulling their hair out - trying to get the rest of the market to see things as they do. In the early days of the Athlon XP, the model numbers were quite conservative as you may recall. Although AMD always insisted that the model numbers were used to compare the Athlon XP to the old Thunderbird core, no one really bought that - the model numbers were architected to draw parallels between the Athlon XPs and higher clocked Pentium 4s. Back in the early days, the Pentium 4 wasn't nearly as competitive as it is today and AMD's model numbers seemed to be a bargain - an Athlon XP 1800+ would outperform a 2.0GHz Pentium 4 most of the time - thus questioning AMD's ratings seemed silly.
Fast forward to today and the picture has changed considerably; Intel's Northwood core has been quite successful and is now paired with features such as an 800MHz FSB and Hyper Threading. The end result is that while AMD was once very conservative with their model numbers, these days they seem much more liberal. So what's the big deal? The benchmarks speak for themselves and everyone will easily find out if a 3200+ can compete with a 3.2GHz Pentium 4 soon enough, so why all the fuss?
Although folks in this community tend to glaze over marketing silliness, the millions that we cater to are only a percentage of the total number of computer users out there. What Intel fears the most is that your average Joe is walking into a computer store and now sees a cheaper, "equal performing" Athlon XP 3200+ next to a Pentium 4 3.2 and makes the obvious choice. Thus Intel's crusade against AMD's model numbers has continued on, to some avail - after countless attempts, reviewers are starting to evaluate the Athlon XP not only on its relative performance to the Pentium 4 but also on the merits of AMD's model numbers.
We chastised AMD when they first introduced the marketing plan for the Athlon XP, but since then have remained relatively quiet on the topic. As far as we're concerned, you all are smart enough to determine what to purchase and what to recommend to the less inclined. If you spend your time reading through, understanding and learning from our reviews then you can come to your own conclusions just fine. Yes, AMD needs to update their benchmarking suite, Yes the Pentium 4 has continued to dominate in performance and as you will see by the end of this review, Yes the 3.2GHz Pentium 4 is noticeably faster than the Athlon XP 3200+.
With that out of the way, let's talk tech, shall we?
FSB Scaling - The Athlon XP
The Athlon XP 3200+ is based on AMD's 0.13-micron Barton core (see our review for a full description of the Barton core) and runs at a clock speed of 2.20GHz. The first thing to keep in mind that at 2.20GHz the Athlon XP 3200+ still carries a lower clock speed than the short lived Thoroughbred-B based Athlon XP 2800+ (2.25GHz). Although the 3200+ only holds a 34MHz clock speed advantage over the 3000+, AMD has moved to a 200MHz DDR FSB for the chip (effectively 400MHz).
Alongside the chip, NVIDIA has revealed that they've been certifying the nForce2 chipsets for use with the faster FSB for quite some time and thus re-badged them as nForce2 Ultra 400 in order to indicate support for the FSB.
So how much of a boost do we see from increasing the FSB on the Barton core to 400MHz? First be sure to read our explanation of why a faster FSB results in better performance here, then take a look at the following tests:
The 400MHz FSB provides a 0 - 7% improvement in performance for the Athlon XP at 2.20GHz, with the average being a 3.2% increase. While we'd normally say that the benefits of a higher FSB frequency will best be realized as the processor reaches higher clock speeds, but with AMD keen on replacing the Athlon XP with the Athlon 64, we'd say the move to a 400MHz FSB wasn't absolutely necessary.
FSB Scaling - The Pentium 4
AMD wasn't the only one to introduce a higher speed FSB, prior to the XP 3200+ launch Intel introduced the first 800MHz FSB Pentium 4s at speeds ranging from 2.4 - 3.00GHz. The fastest FSB Pentium 4s were introduced alongside the 875P and 865PE series of chipsets; we've already covered these chipsets in great detail as well as already published a roundup of the first 20 875P/865PE motherboards to hit our labs.
We've briefly taken a look at what the 800MHz FSB can offer for the Pentium 4, but with this review we're able to see exactly how much the Pentium 4 benefited from increasing its FSB past the introductory 400MHz frequency.
In order to be as thorough as possible, we even threw in a 600MHz FSB which will give you an idea of which cases the 800MHz FSB is overkill and where it isn't. Also remember that Intel will be using a 667MHz FSB in future Xeon processors so these numbers are also more useful than just an analytical tool:
Here the Pentium 4 shows exactly how much of its performance today is courtesy of improvements that Intel has made to the platform overall and not just the CPU. Although we didn't explicitly compare the 800MHz FSB to the 533MHz FSB here, you can find those numbers in our Pentium 4 3.0C Review.
Windows XP Professional Test Bed
AMD Athlon XP 3200+ (2.20GHz) Barton
AMD Athlon XP 2800+ (2.25GHz)
AMD Athlon XP 2600+ (2.083GHz)
AMD Athlon XP 2500+ (1.83GHz) Barton
AMD Athlon XP 2400+ (2.00GHz)
AMD Athlon XP 2200+ (1.80GHz)
AMD Athlon XP 2100+ (1.73GHz)
AMD Athlon XP 1600+ (1.40GHz)
Intel Pentium 4 3.20
Intel Pentium 4 3.06
Intel Pentium 4 3.0C
Intel Pentium 4 2.8C
Intel Pentium 4 2.80
Intel Pentium 4 2.6C
Intel Pentium 4 2.53
Intel Pentium 4 2.4C
Intel Pentium 4 2.40
Intel Pentium 4 2.26
Intel Pentium 4 1.6A
ASUS A7N8X - NVIDIA nForce2 Chipset
Intel D875PBZ - Intel 875P Chipset
2 x 256MB DDR400 CAS2 Corsair XMS3200 DIMM
80GB Western Digital Special Edition 8MB Cache ATA/100 HDD
ATI Radeon 9700 Pro
Content Creation Performance
General Usage Performance
Although not as performance-critical as content creation applications, it is the set of every day applications like Office and other general usage programs that the majority of users find themselves interacting with the most, thus performance here is also very important.
Gaming Performance - Unreal Tournament 2003 (Flyby)
With this review we continue to use the final retail version of Unreal Tournament 2003 as a benchmark tool. The benchmark works similarly to the demo, except there are higher detail settings that can be chosen. As we've mentioned before, in order to make sure that all numbers are comparable you need to be sure to do the following:
By default the game will detect your video card and assign its internal defaults based on the capabilities of your video card to optimize the game for performance. In order to fairly compare different video cards you have to tell the engine to always use the same set of defaults which is accomplished by editing the .bat files in the X:\UT2003\Benchmark\ directory.
Add the following parameters to the statements in every one of the .bat files located in that directory:
For example, in botmatch-antalus.bat will look like this after the additions:
..\System\ut2003 dm-antalus?spectatoronly=true?numbots=12?quickstart=true -benchmark -seconds=77 -exec=..\Benchmark\Stuff\botmatchexec.txt -ini=..\\Benchmark\\Stuff\\MaxDetail.ini -userini=..\\Benchmark\\Stuff\\MaxDetailUser.ini -nosound
Remember to do this to all of the .bat files in that directory before running Benchmark.exe.
Gaming Performance - Unreal Tournament 2003 (Botmatch)
With previous versions of UT2003, Botmatch couldn't be used to compare different systems as there was a bug in the benchmark that could cause inflated numbers on AMD systems vs. Intel systems. We went to Epic with the problem and they provided us with a beta patch in time for this review, the fix will make it into the next publicly available patch release in several weeks.
For those of you that aren't familiar, the Botmatch test focuses mostly on physics and artificial intelligence performance in UT2003, the two areas that are the most CPU dependent in the game.
Splinter Cell Performance (Average Frame Rates)
A new addition to our benchmarking suite is Splinter Cell from Ubisoft. Taken from a 3rd person perspective, the game focuses much more on sneaking around and hiding in the shadows rather than high frame rates and raking up frags as quickly as possible. The Splinter Cell benchmark is a part of the game's 1.2 patch and is publicly available to all owners of the full version of the game.
For reference, we used the first demo and forced the use of "class 1" rendering and left all detail settings at their maximum levels.
Gaming Performance - Quake III Arena
An extremely dated benchmark, Quake III Arena has become much more of a CPU and platform test than anything because of the fact that current generation graphics cards are no where near stressed by it. We used our old 1.29f build of the game with the classic demo "four" at High Quality defaults, with everything maxed out at 1024x768.
Gaming Performance - Jedi Knight 2
Gaming Performance - Serious Sam 2
Video Encoding Performance - DiVX/XMpeg 4.5
What was once reserved for "professional" use only has now become a task for many home PCs - media encoding. Today's media encoding requirements are more demanding than ever and are still some of the most intensive procedures you can run on your PC.
We'll start off with a "quick" conversion of a DVD rip (more specifically, Chapter 40 from the Star Wars Episode I DVD) to a DiVX MPEG-4 file. We used the latest DiVX codec (5.03) in conjunction with Xmpeg 4.5 to perform the encoding at 720 x 480.
We set the encoding speed to Fastest, disabled audio processing and left all of the remaining settings on their defaults. We recorded the last frame rate given during the encoding process as the progress bar hit 100%
3D Rendering Performance - 3dsmax R5
When the Athlon was first released over 3 years ago, 3D Studio MAX was a strong point of its performance. The Athlon's raw FPU performance was right up 3dsmax's ally and thus it put Intel's competing solutions (at the time, the Pentium III) to shame. Things have changed a bit, the latest version of 3ds max (R5) does have some Pentium 4 optimizations that keep things quite competitive between the Athlon XP and the Pentium 4.
3D Rendering Performance - Lightwave 3D 7.5
While 3dsmax 5 is SSE2 optimized, the level of optimization is nowhere near what NewTek reported with Lightwave upon releasing version 7.0b. The performance improvements offered by the new SSE2 optimized version were all above 20% using NewTek's supplied benchmarking scenes.
With the introduction of the 800MHz, Intel has put the nail in the Athlon XP's coffin - whatever chances AMD had at regaining the performance crown with the Athlon XP were lost when Intel introduced the 865PE and 875P platforms. Luckily for AMD, the Athlon 64 is just around the corner but it's clear who the winner of the Northwood vs. Barton battle is.
At high clock speeds, there's no denying that the Pentium 4 is the processor to have - especially the new 'C' models with the 800MHz FSB. We included results from a slower Pentium 4 1.6A and an AMD Athlon XP 1600+ to prove another point, at lower clock speeds the Athlon XP is still a better option although as higher clocked P4s drop in price AMD's advantage in this area will begin to fade as well.
So if you're looking to buy the best on the block today, look no further than a Pentium 4 but we'd caution against purchasing the 3.2GHz Pentium 4. The price premium you're going to be paying doesn't justify the performance advantage you get over a 3.0C, not to mention that the 3.2GHz Pentium 4 is the hottest running CPU on the block. We'd recommend going with one of the slightly lower clocked 800MHz FSB models (the 2.6C or 2.8C come to mind) and holding off on upgrading again until the Socket-775 Prescott CPUs hit next year.
For more information on Intel's future CPU plans be sure to read through our analysis of their roadmap here.