Original Link: https://www.anandtech.com/show/1004
AMD Athlon XP 2800+ - Introducing the first 333MHz FSB Athlon XP
by Anand Lal Shimpi on October 1, 2002 8:13 AM EST- Posted in
- CPUs
Things haven't been so peachy for AMD as of late; Hammer was officially moved back to next year, the 0.13-micron Thoroughbred-A core didn't scale as well initially as AMD had hoped and facing increasing pressure from Intel, AMD was forced to "paper launch" (launch without beginning volume shipments) the Athlon XP 2600+ and 2400+ in August.
Compare this to the launch of the Athlon XP and its clock bumped successors and you'll see that things aren't as bright for AMD as they once were. This is partially because Intel has finally made the Pentium 4 a worthy competitor, with the Northwood core scaling extremely well. Another contributor to AMD's recent tough times is the strong internal focus on Hammer, but with the CPU around 4 - 6 months away it is impossible to ignore the Athlon XP.
Even though the Athlon XP 2400+ and 2600+ (launched on August 21st) are just now appearing in retail channels, AMD is going ahead with the announcement of their Athlon XP 2700+ and 2800+ processors. The processors won't be widely available for another couple of months, clearly not something to get excited about.
As you'll probably hear all over the web, there's nothing but displeasure from the community about AMD's strategy behind paper launching the Athlon XP. Coming from the company that stuck it to Intel when they were paper launching Pentium IIIs, it's pretty disappointing to see the exact opposite happening here today.
With that said, today we're here to bring you a performance evaluation of the paper-launched Athlon XP 2800+ running on a chipset that was launched 3 months ago and still isn't available in retail. Let's just say this trend isn't something we're too happy about, and neither should you.
Some Good News: 333MHz FSB
As we just mentioned, the two new processors being announced today are the Athlon XP 2700+ and 2800+. The 2700+ runs at 2.17GHz while the 2800+ runs at 2.25GHz, both using a 166MHz DDR FSB (effectively 333MHz). The 25% increase in FSB frequency will apply to all CPUs after the 2700/2800+, meaning that although the 2600+ won't get a 333MHz FSB, the 3000+ will. Both of these CPUs are based off of the new Thoroughbred-B core.
In terms of chipset support, both NVIDIA's nForce2 and VIA's KT400 will officially support the 333MHz FSB. If you've been following any of our Socket-A motherboard reviews over the past year you'll know that quite a few motherboards have no problem running at a 166MHz FSB frequency, so owners of older motherboards should have no problem upgrading even if the faster FSB isn't "officially" supported.
There's not much of a tangible performance improvement as a result of the faster FSB, but as clock speeds increase the added bandwidth will come in handy. So while the faster FSB doesn't make too much of an impact today, it will mean much more 6 - 8 months from now. Intel employs a similar strategy when it comes to introducing new FSB frequencies; the goal is to make the transition as early as possible so when the added bandwidth is necessary, the majority of your user base already has it available to them.
Quite possibly the most important aspect of the 333MHz FSB isn't the added bandwidth, rather the fact that now the FSB runs synchronously with the memory bus if you're running DDR333 SDRAM. The benefit of a balanced architecture where your FSB and memory buses are running in sync is reduced memory latency.
The reduction in memory latency ends up being significant enough that it outweighs the benefits of a slightly higher bandwidth memory solution. This is why the nForce2 platform runs faster with DDR333 than with DDR400 when using the new 333MHz FSB.
The vast majority of the performance increase that accompanies the 333MHz FSB is because of this reduction in latency and not because of an increase in FSB bandwidth. As applications become more bandwidth intensive and as CPU clock speeds increase, this paradigm will shift towards favoring more bandwidth over slightly lower latency; but the idea is that by then we'll be able to move to higher bandwidth FSBs and memory buses while still keeping the two synchronous.
So although the Athlon XP 2800+ is only clocked 5% higher than the 2600+, you'll end up seeing anywhere between a 3 - 6% performance improvement over the 2600+ thanks to the new synchronous platform.
Thermal Comparison
One of the commonly overlooked but most important purchasing decisions for many users is heat dissipation and power consumption of CPUs. Whether you're building a server farm and are looking to stay within a certain power envelope or if you're building a quiet home theater PC, knowing what processors in your performance range run the coolest is quite useful.
As a new addition to our CPU reviews we're providing tables of heat dissipation of every major desktop CPU since the Athlon 500 & Pentium III 500. At the same time, we're also providing a graph comparing the heat dissipation of all of the latest CPUs (all Athlon XPs vs. Pentium 4s). To access this data just make your selection according to what CPU (core) you're interested in:
AMD Athlon | AMD Athlon (Thunderbird) | AMD Athlon XP (Palomino) | AMD Athlon XP (Thoroughbred)
Intel Pentium III | Intel Pentium 4
We hope you find this feature useful, please provide us with any feedback you may have on how to improve and expand our reviews in the future.
|
Maximum
Heat Dissipation
|
Manufacturing
Process
|
||
| AMD Athlon 500 (Model 1) |
42.0W
|
0.25-micron
|
|
| AMD Athlon 550 (Model 1) |
46.0W
|
0.25-micron
|
|
| AMD Athlon 600 (Model 1) |
50.0W
|
0.25-micron
|
|
| AMD Athlon 650 (Model 1) |
54.0W
|
0.25-micron
|
|
| AMD Athlon 700 (Model 1) |
50.0W
|
0.25-micron
|
|
| AMD Athlon 550 (Model 2) |
31.0W
|
0.18-micron
|
|
| AMD Athlon 600 (Model 2) |
34.0W
|
0.18-micron
|
|
| AMD Athlon 650 (Model 2) |
36.0W
|
0.18-micron
|
|
| AMD Athlon 700 (Model 2) |
39.0W
|
0.18-micron
|
|
| AMD Athlon 750 (Model 2) |
40.0W
|
0.18-micron
|
|
| AMD Athlon 800 (Model 2) |
48.0W
|
0.18-micron
|
|
| AMD Athlon 850 (Model 2) |
50.0W
|
0.18-micron
|
|
| AMD Athlon 900 (Model 2) |
60.0W
|
0.18-micron
|
|
| AMD Athlon 950 (Model 2) |
62.0W
|
0.18-micron
|
|
| AMD Athlon 1000 (Model 2) |
65.0W
|
0.18-micron
|
|
|
Maximum
Heat Dissipation
|
Manufacturing
Process
|
||
| AMD Athlon 900 (Thunderbird) |
51.0W
|
0.18-micron
|
|
| AMD Athlon 950 (Thunderbird) |
53.1W
|
0.18-micron
|
|
| AMD Athlon 1000 (Thunderbird) |
55.1W
|
0.18-micron
|
|
| AMD Athlon 1100 (Thunderbird) |
60.3W
|
0.18-micron
|
|
| AMD Athlon 1133 (Thunderbird) |
62.1W
|
0.18-micron
|
|
| AMD Athlon 1200 (Thunderbird) |
65.7W
|
0.18-micron
|
|
| AMD Athlon 1266 (Thunderbird) |
66.9W
|
0.18-micron
|
|
| AMD Athlon 1300 (Thunderbird) |
68.3W
|
0.18-micron
|
|
| AMD Athlon 1333 (Thunderbird) |
69.8W
|
0.18-micron
|
|
| AMD Athlon 1400 (Thunderbird) |
72.1W
|
0.18-micron
|
|
AMD Athlon XP vs. Intel Pentium 4 - Thermal Comparison
|
49.4
51
54.3The Test
|
Windows
XP Professional Test Bed
|
|
|
Hardware
Configuration
|
|
| CPU |
AMD
Athlon XP 2800+ (2.25GHz)
AMD Athlon XP 2600+ (2.13GHz) AMD Athlon XP 2400+ (2.00GHz) AMD Athlon XP 2200+ (1.80GHz) AMD Athlon XP 2100+ (1.73GHz) AMD Athlon XP 2000+ (1.67GHz) AMD Athlon XP 1800+ (1.53GHz) Intel Pentium 4 2.80GHz Intel Pentium 4 2.66GHz Intel Pentium 4 2.60GHz Intel Pentium 4 2.53GHz Intel Pentium 4 2.50GHz Intel Pentium 4 2.40B GHz Intel Pentium 4 2.40GHz Intel Pentium 4 2.26GHz Intel Pentium 4 2.20GHz Intel Pentium 4 2.0A GHz |
| Motherboard |
ASUS
A7N8X - NVIDIA nForce2 Chipset
Intel D850EMV2 - Intel 850E Chipset |
| RAM |
1
x 256MB DDR400 CAS2 Corsair XMS3200 DIMM
2 x 128MB PC800 Samsung RIMMs |
| Sound |
None
|
| Hard Drive |
80GB
Western Digital Special Edition 8MB Cache ATA/100 HDD
|
| Video Cards (Drivers) |
NVIDIA GeForce4 Ti 4600 (30.82) |
Content Creation Performance
We're still delving deeper into the performance issues behind SYSMark 2002 but in the meantime we will continue to provide performance numbers from the benchmark as well as other similar benchmarks to compare/contrast with. Keep in mind that the real-world tasks SYSMark 2002 runs are known to be noticeably faster on the Pentium 4 platform which is why even the fastest Athlon XP cannot reach the performance of a 2GHz Pentium 4.
|
370Content Creation Winstone 2002 is another Content Creation benchmark that focuses on performance under applications like Adobe Photoshop, Premier and Macromedia Dreamweaver. This benchmark is much more I/O bound than its SYSMark 2002 counterpart, although the results are very similar:
|
General Usage & Business Application Performance
For general/business usage performance we turn to Office Productivity SYSMark 2002 and Business Winstone 2001:
|
The Pentium 4 2.80GHz ends up being around 10% faster than the Athlon XP 2800+ here, which is a noticeable performance difference.
|
Business Winstone paints a similar picture, except here the next 2800+ delivers about the same level of performance as the 2.66 and 2.80GHz Pentium 4s. Your specific experience will vary depending on the type of workload you put on your CPU but generally speaking, the Athlon XP 2800+ proves its competitiveness in the business world with these benchmarks; it's not the fastest, but it can hold its own.
3D Rendering Performance - 3ds max 5
When the Athlon was first released 3 years ago, 3D Studio MAX was a strongpoint 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.
For our 3ds max 5 benchmarks we chose two benchmark scenes that ship with the product - SinglePipe2.max and Underwater_Environment_Finished.max
|
Once again we see that the XP 2800+ is on the heels of Intel's 2.66 and 2.80GHz Pentium 4s.
|
The Athlon XP 2800+ actually takes the lead over the 2.80GHz Pentium 4 here, showing off its x87 floating point muscle.
|
The Athlon XP continues to have a strong showing under Maya, with the XP 2800+ extending a clear lead over the competition.
3D Rendering Performance using SSE2
While 3ds max 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.
We chose two benchmarks to use, one of the least SSE2 optimized scenes and another that is more optimized just to get an idea of the potential that lies for Pentium 4 users running heavily optimized applications
|
|
Intel would definitely like to see more of these sort of optimizations in other applications, unfortunately for them it's rare that the Pentium 4 sees this sort of a lead across the board.
Media Encoding Performance
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 (5.02) 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%.
|
MPEG-4 encoding is a very bandwidth intensive procedure, which is why we see a healthy boost when moving to the 333MHz FSB Athlon XP 2800+; this is also part of the reason why the Pentium 4 does so very well in the test.
MP3 audio encoding is another great CPU test, although this sort of an application is much less platform intensive than the MPEG-4 test from above. Raw computational power is mostly stressed in the following MP3 encoding test; we encoded a 170MB wav file into a VBR (Variable Bit Rate) MP3 using the highest quality settings (-V 0) allowed by the LAME 3.91 MP3 encoder.
|
Gaming Performance - Unreal Tournament 2003
We introduced the latest Unreal Tournament 2003 benchmark in our GPU Shootout article a while back and we're continuing to use it as an example of a next-generation game test. This review doesn't use the final demo that was released publicly but we will be transitioning to that demo shortly.
|
The faster FSB helps the XP 2800+ considerably here as it rises to the top of the charts.
Gaming Performance (continued)
Jedi Knight 2 is another great gaming benchmark for CPUs as it is based on the popular Quake III engine but is demanding enough that it doesn't break the 300fps barrier at every chance it gets.
We used the 1.03 patch which still retained the game's built in demo (JK2FFA) and ran the test at high detail settings with anisotropic filtering disabled.
|
We continue to see a competitive display by the XP 2800+ but it fails to claim the absolute victory in this test, conceding to both the 2.66GHz and 2.80GHz Pentium 4s.
|
|
Final Words
The numbers speak for themselves, the Athlon XP 2800+ is yet another competitive part from AMD. While it fails to regain the absolute performance crown for AMD, it keeps them in the running with Intel.
As we mentioned earlier, there's only a 5% increase in clock speed between the XP 2800+ and the 2600+. Much of the performance gain we saw here today wasn't due to the increase in clock speed, but a result of the FSB and memory buses now running synchronously at a 166MHz frequency. The resulting reduction in latency provides an increase in performance across the board that helps make the jump to the XP 2800+ more than just a measly 5% bump.
The only real problem (and it's a big one at that) with this processor is that you can't get your hands on one, and you won't be able to for quite some time. Remember that at the time of publication the Athlon XP 2400+ and 2600+ parts just started popping up in the channel, it's going to be a matter of months before you can easily pick up a 2800+. By then Intel will have launched the 3.06GHz Pentium 4 with Hyper-Threading support, thus extending their performance lead even further while maintaining a steady grip on the performance crown.
It will take more than a series of paper launched processors to keep AMD truly competitive in the market; what worries us the most is that if AMD is having difficulty getting these CPUs out and into mass production in a timely manner, then what's to say that Hammer will be any different? With the release of Hammer now several months away, we can only hope that AMD can sort things out before then.
Facebook
Twitter
RSS
