
Original Link: https://www.anandtech.com/show/1215
AMD's Athlon 64 3400+: Death of the FX-51
by Anand Lal Shimpi on January 6, 2004 4:35 AM EST- Posted in
- CPUs
What a weird way to end the year; at the beginning of 2003 we expected AMD to fall short of clock expectations and for Intel to trample all over the Athlon 64 with Prescott. With 2004, still in its infancy, being a meager 6 days old we know that the outcome of the processor wars of last year was not as expected. AMD surprised us all with a far more competitive Athlon 64 launch than we had originally expected, and Prescott didn’t exactly make it out the gates.
Instead we were left with a new class of processors with the Athlon 64 FX and the Pentium 4 Extreme Edition; cash cow CPUs marketed for our community but priced well above our comfort levels. Granted if you look back at the $1000+ price tag the Pentium II had upon its release a $700 CPU today isn’t asking too much, but we’ve grown far too accustomed to sub-$200 CPU prices for that to work.
With just under three-and-a-half months under AMD's 64-bit belt, we're ready for the first speed bump for the Athlon 64 line.
You'll remember from our initial coverage that the major difference between the Athlon 64 and the Athlon 64 FX that the latter boasts a 128-bit memory controller as opposed to the 64-bit interface of the regular 64. The only other differences (other than price) were that the Athlon 64 FX was available at 2.2GHz (compared to the fastest 2.0GHz 64 offering) and the FX ships without a multiplier lock. With today's launch, the focus is further shifted away from the pricey FX and onto the latest reason not to buy AMD's most expensive CPU – the Athlon 64 3400+.
Now boasting a 2.2GHz clock, equaling that of the flagship FX51, the Athlon 64 has become an even more powerful force to reckon with. With a 10% increase in clock speed, can AMD begin to eat into Intel's lead in encoding/content creation applications? Let's find out…
A Diamond in the Rough
When we first looked at the Athlon 64 and FX we realized that the performance difference between the two was negligible at best, but what truly sealed the fate of the Athlon 64 FX in our eyes was the quiet release of the Athlon 64 3000+ based on AMD's Newcastle core.
Newcastle is the mainstream successor to Claw Hammer, what all current Athlon 64s are based on right now. The only difference between Newcastle and Claw Hammer is that Newcastle has half the L2 cache, totaling 512KB instead of the original 1MB L2 that AMD launched. Why AMD would introduce the Athlon 64 with a 1MB L2 only to scale it back a couple of months later is anyone's guess. Perhaps AMD felt that it would be necessary to compete with Prescott or perhaps there were design issues with getting it to market in time, needless to say that slowly but surely all Athlon 64's will be Newcastle derived.
You caught a glimpse of the performance of the Athlon 64 3000+ in our earlier preview, but you will get a full taste of the price-effective performance that Newcastle offers in this review. Performance close to the Athlon 64 3200+ (which was close to the Athlon 64 FX51) at about half the price can't really be beat, and you'll surely see that here.
The Test
We have already gone into great depth about the Athlon 64 and the underlying architecture so be sure to check out our previous articles for all you could possibly want to know about the Athlon 64.
Performance Test Configuration | |
Processor(s): | AMD Athlon 64 3000+ AMD Athlon 64 3200+ AMD Athlon 64 3400+ AMD Athlon 64 FX51 Intel Pentium 4 3.2GHz EE Intel Pentium 4 3.2GHz Intel Pentium 4 3.0GHz Intel Pentium 4 2.8GHz |
RAM: | 2 x 512Mb OCZ 3500 Platinum Ltd 2 x 512Mb Mushkin ECC Registered High Performance 2:3:2 |
Hard Drive(s): | Seagate 120GB 7200 RPM (8MB Buffer) |
Video AGP & IDE Bus Master Drivers: | VIA Hyperion 4.51 (12/02/03) Intel Chipset Drivers |
Video Card(s): | Sapphire ATI Radeon 9800 PRO 128MB (AGP 8X) |
Video Drivers: | ATI Catalyst 3.10 |
Operating System(s): | Windows XP Professional SP1 |
Motherboards: | Intel D875PBZ (Intel 875P Chipset) FIC K8-800T (VIA K8T800 Chipset) ASUS SK8N (NVIDIA nForce3 Pro Chipset) |
All performance tests were run with ATI 9800 PRO 128MB video card with AGP Aperture set to 128MB with Fast Write enabled. Resolution in all benchmarks is 1024x768x32.
Business Application Performance
The Athlon 64 3400+ is hot on the trails of the FX51, and the Newcastle is hot on the trails of the 3200+. What's most interesting is that even the mighty Extreme Edition isn't able to dethrone the mainstream Athlon 64.
DirectX 9 Performance – AquaMark 3
For Aquamark3 (and all our subsequent game tests), we followed the same procedure that we used in our recent video card articles. For this test, we only looked at the free version of Aquamark3, as that is what our readers use. Aquamark uses DX8 and DX9 pixel and vertex shaders.
If we isolate the CPU's contribution to the overall performance we can see that the Pentium 4 takes the lead, with the Extreme Edition sitting pretty just beyond the 10,000 mark. Although CPU-only scores like these are not representative of real world performance they are interesting to look at to get an idea for what performance differences do exist among the CPUs, in this case we see that there isn't much at all.
DirectX 9 Performance – Halo
For the Halo benchmark, we used the PS/VS 2.0 rendering mode (DX9), and ran the timedemo with vsync off at 60Hz.
The only somewhat appreciable difference in performance is seen by the Extreme Edition, but in terms of a tangible performance difference between the CPUs there's not much to look for.
DirectX 9 Performance - GunMetal Benchmark 2 This game is a DX8 game with DX9 vertex shaders. It isn't possible to disable Antialiasing or Anisotropic filtering in this benchmark, so we ran at the same 4 x AA settings we used in our graphics card benchmarks. The situation in GunMetal mimics the previous three game benchmarks we've seen, there's virtually no appreciable difference between the CPUs in any of the GPU-intensive gaming environments.
DirectX 7/8 Performance – UT2003
If we look at the botmatch numbers we see even more of an advantage for the Athlon 64.
OpenGL Performance – Quake III Arena
DivX Encoding Performance
We have been using a DivX encoding test as a part of our CPU benchmarking suite for quite some time now, however the performance test has never been truly realistic as it wasn't geared towards producing a high quality DivX rip - rather it was designed to stress CPU performance.
We have since revised our benchmark and now follow the DivX 5 encoding guide published at Doom9.net . For our test title we use Chapter 9 from The Sum of All Fears DVD. We conduct a 2-pass encoding process and report the encoded FPS from both passes averaged together. The results are lower than our previous Xmpeg tests, however they are much more applicable to real-world usage.
Hyper-Threading along with other features of the NetBurst architecture give Intel the performance advantage in video encoding as you can see by our DivX results above. The Athlon 64 3400+ performs respectably but it isn't the CPU that's best for these sorts of tasks.
3D Rendering Performance
Development Workstation Performance
One test we've always been asked to run is a Visual Studio/Visual C compile test, however we never had a project large enough to compile. Sitting around talking about Athlon 64 testing one day we came up with the idea of using the publicly available Quake 3 source code as a compile test for CPUs, and thus for this next test we timed how long it took for the Quake 3 source code to compile.
This compile test should be a relatively good indicator of overall compile performance, which will be very useful for those of you that have very large projects that can take countless minutes to well over an hour to compile.
The Athlon 64/FX continue to put the Pentium 4 to shame in our VC++ compile test; serious developers will want to save their compiling for Athlon 64 machines.
Price/Performance Ratio
We've been asked to compare the price/performance of the CPUs we review so what we've done is taken the average prices from the top vendors off of DealTime and divided each benchmark result by that average price and multiplied by 100. In the case of some of the tests where we could group multiple results into one category we averaged the price/performance index values together.
Final Words
With the Athlon 64 3400+ AMD has effectively killed any reason to shell out the extra bucks for an Athlon 64 FX; the 3400+ is basically as fast as the Athlon 64 FX 51 at a lower price point. Soon AMD will launch the FX53 which will restore some balance to their 64-bit universe but until then you can expect the FX line to take an even more passive role in AMD's marketing and in our lives.
AMD's Athlon 64 3000+ has provided us with an impressive show of force, especially considering its very low price point. If you look at the price/performance index values you'll see that the Newcastle based 3000+ is the only CPU able to offer a better value than Intel's Pentium 4 2.8C - the previous best-bang-for-your-buck title holder.
The overall performance picture hasn't changed much since we first looked at the Athlon 64; AMD is still the best bet when it comes to business/gaming/2D workstation either from a performance perspective or because of a superior price/performance ratio, while Intel offers the best in encoding and 3D rendering performance.
The one worry that is worth taking into account is the fact that AMD will significantly revamp their Athlon 64 line over the course of 2004 with the introduction of Socket-939 CPUs. Although AMD has committed to supplying Socket-754 and 940 CPUs throughout 2004, those looking to hold onto their motherboards and upgrade their CPUs beyond the next 12 months will want to wait until the Socket-939 platforms hit in the next few months.