Today, AMD is taking things up another notch with the introduction of the Athlon XP 1900+. At 1.6GHz, this chip runs only 66MHz faster than the 1800+ model. Otherwise, the two chips are basically identical. However, we’ve decided to make the Pentium 4 vs. Athlon XP race a little more interesting by testing the fastest Athlon XP on a different platform this time around: VIA’s new KT266A chipset, which offers significantly better performance than the AMD 760 chipset we’ve used in our Athlon test systems for ages now. The word on the street is that AMD is canceling production of the 760 now that superior alternatives are finally on the market, so most new Athlon XP systems aren’t very likely to be built with the 760 chipset.
Can the combination of the new Athlon XP 1900+ and the KT266A chipset remove all doubts about who owns the x86 processor performance crown? Read on to find out.
Like all new Athlon XP and MP processors, the 1900+ model comes in AMD’s new organic packaging.
And for those of you really into the markings and stepping numbers on a chip, here’s a close-up look at the core of our Athlon XP 1900+:
The “AGKGA” markings on this chip are the same as those on our Athlon XP 1800+ chip.
Enough ogling. Let’s see how this puppy performs.
Our testing methods
I told you last time around that we would be moving right away to Windows XP for most of our benchmarking, but, uhm, not quite yet. Because of time constraints and a few new wrinkles thrown at us by Windows XP’s self-tuning services, we stuck with Windows 2000 once more.
As ever, we did our best to deliver clean benchmark numbers. Tests were run at least twice, and the results were averaged.
Our test systems were configured like so:
|Processor||AMD Athlon XP 1900+ 1.6GHz||AMD Athlon 1.2GHz
AMD Athlon MP 1.2GHz
AMD Athlon 1.4GHz
AMD Athlon XP 1800+ 1.53GHz
|AMD Athlon 1GHz
AMD Duron 1GHz
AMD Duron 1.1GHz
|Intel Pentium 4 1.4GHz
Intel Pentium 4 1.6GHz
Intel Pentium 4 1.8GHz
|Intel Pentium 4 2GHz||Intel Celeron 900MHz
Intel Celeron 1.1GHz
Intel Pentium III 1.2GHz
|Front-side bus||133MHz (266MHz DDR)||133MHz (266MHz DDR)||100MHz (200MHz DDR)||100MHz (400MHz quad-pumped)||100MHz (400MHz quad-pumped)||100MHz (Celeron)
|Motherboard||Epox EP-8KHA+||Gigabyte GA-7DX rev. 4.0||Asus A7VI-VM||Intel D850GB||Intel D850MD||Intel D815EEA2|
|Chipset||VIA KT266A||AMD 760/VIA hybrid||VIA KM133||Intel 850||Intel 850||Intel 815EP|
|North bridge||VT8366A||AMD 761||VT8365||82850 MCH||82850 MCH||82815 MCH|
|South bridge||VT8233||VIA VT82C686B||VT8231||82801BA ICH2||82801BA ICH2||82801BA ICH2|
|Memory size||256MB (1 DIMM)||256MB (1 DIMM)||256MB (1 DIMM)||256MB (2 RIMMs)||256MB (2 RIMMs)||256MB (1 DIMM)|
|Memory type||Micron PC2100 DDR SDRAM CAS 2||Micron PC2100 DDR SDRAM CAS 2||Infineon PC133 SDRAM CAS 2||Samsung PC800 Rambus DRAM||Samsung PC800 Rambus DRAM||Infineon PC133 SDRAM CAS 2|
|Graphics||NVIDIA GeForce3 64MB (12.41 video drivers)|
|Sound||Creative SoundBlaster Live!|
|Storage||IBM 75GXP 30.5GB 7200RPM ATA/100 hard drive|
|OS||Microsoft Windows 2000 Professional|
|OS updates||Windows 2000 Service Pack 2, Direct X 8.0a|
You’ve probably noticed that we’ve only tested the Athlon XP 1900+ on the KT266A motherboard; the other Athlons were tested on older platforms, so the 1900+ should have a bit of an advantage over the other Athlons. Keep that fact in mind as you take in the test results.
The test systems’ Windows desktops were set at 1024×768 in 32-bit color at a 75Hz screen refresh rate. Vertical refresh sync (vsync) was disabled for all tests.
We used the following versions of our test applications:
- SiSoft Sandra Standard 2001.3.7.50
- Compiled binary of C Linpack port from Ace’s Hardware
- ZD Media Business Winstone 2001 1.0.1
- ZD Media Content Creation Winstone 2001 1.0.1
- POV-Ray for Windows version 3.1g (multiple compiles)
- Sphinx 3.3
- ScienceMark 1.0
- SPECviewperf 6.1.2
- MadOnion 3DMark 2001 Build 200
- Quake III Arena 1.17
- Serious Sam v1.02
All the tests and methods we employed are publicly available and reproducible. If you have questions about our methods, hit our forums to talk with us about them.
These memory performance tests are a good way to start, because they’ll show us whether the KT266A chipset gives our Athlon XP 1900+ test system any advantage. However, memory performance alone does not make a fast computer, so keep in mind that these scores are primarily just for show. The results are useful mainly because they help explain what comes on the following pages.
First up is the memory test from SiSoft’s Sandra, which is a modified version of the Stream benchmark. Here’s how things stack up:
As it has been for the past year, the Pentium 4 is fastest here. However, Athlon XP 1900+ gets a nice boost from the KT266A chipset; its FPU and integer scores jump up by a fair amountmore than one would expect with a 66MHz clock speed jump alone.
Now let’s look at it another way with Linpack.
To understand what this wild-looking graph is telling us, consider the results at a couple of different places. First, compare the megaflops numbers at a matrix size of about 64K. Here, the CPUs are able to do the math entirely inside their L1 caches, and the Athlon XP is far and away the fastest.
Next, compare the numbers at about 192K, where the data to be processed fits entirely in the processors’ L2 caches (with the exception of the Celeron and Duron). The Pentium 4’s L2 cache is fastest here, though the Athlon XP 1900+ isn’t too far behind.
Finally, compare the numbers out at 1800K, where the processors must all access main memory constantly in order to retrieve data. Here, the processors compare about like they do in the Sandra memory test; the P4 is fastest, followed by the Athlon XP/KT266A combo, then by the other DDR systems. The Pentium III and the value processors use slower, plain ol’ SDRAM, so they’re quite a bit slower.
Until the Athlon XP’s front side bus gets faster than 266MHz, I wouldn’t expect to see it challenging the Pentium 4 in raw memory bandwidth. The Pentium 4 simply excels here, and its ample memory bandwidth is a big contributor to the P4’s solid gaming and multimedia performance. However, these numbers don’t mean nearly as much as the scores in the performance benchmarks that follow…
Business Winstone 2001
Business Winstone tests performance in general office applications, like word processors, spreadsheets, and web browsers. This test is also usually a good indicator of overall system performance, especially where “light use” patterns are concerned (in other words, not gaming, heavy computation, or multimedia).
For general business use, the AMD processors are easily fastest. The Athlon XP 1900+ only builds on this lead.
Content Creation Winstone 2001
Content Creation Winstone is arguably more important than Business Winstone, since it tests more performance-sensitive apps, like image and audio processing suites, desktop publishing, web layout programs, and the like. As with Business Winstone, the test runs scripts using code from real applications, not just generic simulations.
The Athlon XP opens up a formidable ten-point gap between itself and the 2GHz Pentium 4. It’s just no contest here.
POV-Ray 3D rendering
POV-Ray is a freeware software ray-tracing program that creates high-quality 3D scenes. It’s also a very useful measure of a processor’s performance, particularly on floating-point math. Our POV-Ray tests use the original release of POV-Ray 3.1, plus Steve Schmitt’s recompiled versions, just to see what difference the various compilers and compiler settings can make.
The recompiled POV-Ray comes in two flavors: “PIII” and “P4”. Both were produced with Intel C v. 5.0. The “PIII” version doesn’t use any instructions proprietary to Intel processors or to the PIII; it runs just fine on the Athlon and the P4. The “P4” version uses a small bit of SSE2 code, but it doesn’t take advantage of the P4’s SIMD capabilities. I’ve indicated which version of POV-Ray was used in the graphs below next to the processor/speed labels, so it should be easy to track.
I’ve omitted results for the value processors here.
The Athlon XP is hard to beat in POV-Ray, even with newly compiled binaries, because the Athlon’s floating-point unit is a monster. The next scene, chess2.pov, is even more complex than the first.
Now we’re talking real differences. The Athlon XP 1900+ finishes rendering this scene over 100 seconds before the Pentium 4 2GHz, even when the P4 is running a specially optimized executable.
LAME MP3 encoding
LAME is the encoder of choice around Damage Labs for high-quality output, so this test holds some interest for me. More speed for MP3 encoding is always good.
Once again, the Athlon XP 1900+ is just a little bit faster than the 1800+ model.
Quake III Arena
Now comes the moment of truth. The Pentium 4 has reigned in Quake III Arena performance since its introduction. Say what you will about the Pentium 4’s performance, but in Q3A, it’s been king of the hill since day one. Can the Athlon XP 1900+ together with the KT266A’s additional memory bandwidth finally catch the Pentium 4 in Quake III?
Indeed it can; the Athlon XP wins in style, breaching the 200fps mark for the first time. (For the record, we test with visual quality and geometry detail settings turned to the max in Q3, and we leave the sound on. These are real-world gameplay performance numbers.) Even on the Pentium 4’s home turf, the Athlon XP 1900+ runs fastest.
Now we’ll try another OpenGL-based first-person shooter, Serious Sam. This game enginewhich produces the best looking visuals of anything on the market, as far as I’m concernedoffers us a little bit different perspective than Quake III.
Ow. That’s just brutal. Serious Sam likes the Athlon XP almost as much as Quake III has favored the Pentium 4.
There’s been a see-saw battle in the past six months over the performance lead in 3DMark 2001, but recently, the Pentium 4 has been leading. Can the Athlon XP 1900+ push the Pentium 4 2GHz off the top of the heap?
Not quite. The P4 still holds the lead in this BapCo/MadOnion benchmark.
SPECviewperf workstation graphics
Viewperf measures a different brand of graphics performance: professional OpenGL applications like CAD and 3D modeling. These workstation-class apps often stress a system in different ways than the gaming tests.
Normally, we’d throw all the viewperf results into one large graph, but this time around, we had too many different processors and clock speeds to do so comfortably. Individual graphs are the order of the day. I’ve omitted results for the ProCDRS test, because every single CPU scored between 16.03 and 16.07 on ProCDRS. I’ve also left out the Awadvs for similar reasons; Awadvs is obviously limited by the performance of the video card, not the processor.
The Athlon XP 1900+ with the KT266A chipset sets a new standard for workstation-class 3D performance.
The Sphinx speech recognition tests came to us via Ricky Houghton, who works in a speech recognition effort at Carnegie Mellon University. They’re based on Sphinx 3.3, which is an advanced system that promises greater accuracy in speech recognition. However, Sphinx 3.3 still can’t quite run fast enough on a standard PC to handle tasks in real time; it’s limited primarily by memory bandwidth.
The Pentium 4 2GHz retains the top spot by a narrow margin. The Athlon XP 1900+ turns in the second fastest score overall, even though our graph is sorted in such a way that the P4 1.8GHz looks like it took the second spot. The P4 is faster with the Microsoft compiler, while the Athlon is fasteroddly enoughwith the Intel compiler.
On to Tim Wilkens’ computational benchmark, ScienceMark. This suite of tests measures number-crunching ability by running some computationally intensive scientific equations. Like 3DMark, ScienceMark then spits out a composite number denoting a system’s overall score in the suite.
Here’s how our contenders fared:
It’s another convincing victory for the Athlon XP 1900+. Let’s break it down into some of ScienceMark’s more interesting individual tests.
The Pentium 4 is fastest in Primordia, but the Athlon XP comes out way ahead on the QMC and Liquid Argon calculations.
For the record, our Athlon XP 1900+ was able to overclock to 1.74GHz on a 290MHz bus without any extreme measures. However, the Athlon XP’s new organic package makes overclocking via the CPU multiplier a little more difficult. The cuts that sever the connections on the Athlon XP’s L1 bridges are deeper and wider than on a T-bird. The ol’ pencil trick didn’t work for us. That’s not the end of the world, since it’s probably still possible to connect the L1 bridges and unlock the CPU multiplier. It’s just going to take more work with the Athlon XP.
Also, in light of what we’ve seen, I wouldn’t be shocked if the current 0.18-micron Athlon XP were selling at 1.7 or even 1.8GHz before AMD makes the transition to its 0.13-micron fab process.
The Athlon XP 1900+ utterly dominated our suite of performance tests, winning every test save two: 3DMark 2001 and Sphinx speech recognition. In both of those tests, the Athlon XP 1900+ came in a close second place behind the 2GHz Pentium 4. The P4 is a formidable competitor, particularly in the most performance-sensitive applications, where bus and memory bandwidth are at a premium. But the Athlon XP delivers strong performance across the board, with no apparent weakness. When coupled with VIA’s KT266A chipset, the Athlon XP removes all doubt about which x86 processor is fastest.
Makes you wonder if the “1900+” rating isn’t a little counterproductive, actually.
Regardless, AMD enjoys a sizeable performance lead right now. That’s a good thing for AMD, too, because Intel’s upcoming “Northwood” variant of the Pentium 4 looms large on the horizon. I can’t wait to see how the Athlon XP and the new Pentium 4 will match up.
Until then, the Athlon XP is the undisputed king of the hill.