The FX-64 Lives! (AMD Athlon 64 X2 6000+)

The Cat is Finally Out of the Bag - X2 6000+

AMD's Athlon 64 X2 series has seen a proliferation of many new models in the last two months, bringing customers many more options to build their new system upon. Expanding both on the low and high-end scales of their successful dual-core processor series, AMD now has a vast (and confusing) lineup to fulfill almost every price point to configure a computer system. Focusing on the high-end, there have been some silent additions to the lineup such as the Athlon 64 X2 5200+ late last year, and just earlier this month, we saw speedier models such as the Athlon 64 X2 5400+ and X2 5600+. Today, AMD hits another milestone with the Athlon 64 X2 6000+ model - their first 3GHz dual-core processor (discounting the earlier launched FX-74 for the pseudo quad-core platform).

6000+ is an impressive sounding model number, but for all intents, the processor model numbers carry little meaning these days, but it still does a dandy job to bamboozle the less-informed. The Athlon 64 X2 6000+ is equipped with 2MB of full speed L2 cache and based on the Windsor core, this 3GHz processor is essentially the rumored . So yes, the FX-64 is alive and kicking finally, but with the start of the Quad FX platform, only processors for this platform bear the FX nomenclature and the stand-alone dual-core processors fall under the Athlon 64 X2 lineup - which is a good thing in our opinion to do away with the differentiated naming scheme for their fastest model.

So what crazy wattage can we expect out of the new Athlon 64 X2 6000+ processor? Fear not as AMD has been fine-tuning the manufacturing process and of course tweaking the processors with various steppings along the way. The result? The Athlon 64 X2 6000+ shares the same 125W TDP as the Athlon 64 FX-62 processor. In fact, the FX-62 is now available in an 89W TDP variant which is known as the Athlon 64 X2 5600+. While we don't have the X2 5600+ at hand to try our hand to verify AMD's claims, we do have the X2 6000+, which is the highlight today and will be under close scrutiny with the FX-62 and Intel's stable of Core 2 Duo processors.

DDR2-750 and Notes on Memory Clocks

While the Athlon 64 X2 series can handle up to DDR2-800 memory, the key words are "up to". Since the memory controller is built into the processor, it can only handle what it has been programmed to do so. Remember, AMD only supports integer divisor numbers and the final memory clock speeds are derived by dividing the CPU clock speed with a suitable memory divisor. For example on the Athlon 64 FX-62 (and the X2 5600+), clock speeds are 2.8GHz and a memory divisor value of 7 is used to derive the base clock speed of DDR2-800 memory (which is 400MHz).

Unfortunately on the A64 X2 6000+, the 3GHz clock speed poses an issue to obtain a true 400MHz base clock for DDR2-800 memory operation. If a divisor of 8 is used, the base memory clock is 375MHz and this is what the A64 X2 6000+ uses. It can't use a divisor of 7 because it would then be overclocking the memory at 425MHz base clock; thus the processor automatically falls back to a divisor where the next highest speed is supported.

Enthusiasts should note that this less than efficient memory clock speed could affect performance in certain scenarios, but more often than not, the impact is negligible in real world use. The benchmarks in the following pages would further attest to this, but even then we did find a few occasions when the dent in memory bandwidth reduced the lead of the Athlon 64 X2 6000+ over the Athlon 64 FX-62 processor.

Overclocking

The Athlon 64 FX-62 was found to be a poor overclocker in our past encounters as it couldn't nudge past 3.1GHz - just 300MHz more than its recommended clocks speeds. Since the Athlon 64 X2 6000+ is based on the same core, we expected no better and our opinions were spot on. Pushing the processor to 3.15GHz is about the most stable overclock we can crank out with the stock cooler. At 3.2GHz however, we weren't even able to get Windows to boot, and that's how impractical it gets to overclock this 'new' processor. Our advice is to skip overclocking should you feel like splurging on the Athlon 64 X2 6000+. Spending more money on aftermarket coolers and other equipment wouldn't be as efficient as switching to an Intel platform for even more performance, which also gives you much more mileage as seen and .

Test Setup

With the Athlon 64 FX-62 being its direct predecessor, the new X2 6000+ will be closely compared against it. In fact, the performance is identical to our previously previewed FX-64, which we derived by overclocking the FX-62. The X2 6000+ is essentially this and nothing more. As depicted on the comparison table earlier, the speedier Conroe processors will be the main competitive environment which the X2 6000+ has to stand up against if it is to be a hit with AMD enthusiasts.

The following testbed configurations will be used throughout our comprehensive benchmarking segment:-

Intel dual-core Core 2 Configuration

• Intel Desktop Board D975XBX (Intel 975X Express chipset)
• Intel Core 2 Extreme X6800, Core 2 Duo E6700, E6600 and E6400
• 2 x 512MB Corsair XMS DDR2-800 memory modules (CAS 4. 4-4-12)
• Seagate Barracuda 7200.7 80GB SATA hard disk drive (one single NTFS partition)
• MSI GeForce 7900 GT 256MB - with NVIDIA Detonator XP 84.21
• Microsoft Windows XP Professional with Service Pack 2
• AcBel 400W Power Supply Unit
• Default Intel CPU cooler

AMD dual-core Athlon 64 Configuration

• ASUS M2N32-SLI Deluxe (nForce 590 SLI chipset)
• AMD Athlon 64 FX-62 and Athlon 64 X2 6000+
• 2 x 512MB Corsair DDR2-800 memory modules (CAS 4, 4-4-12)
• Seagate Barracuda 7200.7 80GB SATA hard disk drive (one single NTFS partition)
• MSI GeForce 7900 GT 256MB - with NVIDIA Detonator XP 84.21
• Microsoft Windows XP Professional with Service Pack 2
• AcBel 400W Power Supply Unit
• Default AMD CPU cooler

Benchmarks

The following benchmarks were used in this review:-

• BAPCo SYSmark 2004
• SPEC CPU2000 v1.3
• Lightwave 3D 7.5
• Futuremark PCMark 2005
• SPECviewperf 8.01
• Cinebench 2003
• XMpeg 5.0.3 (DivX 6.2.5 encoding)
• Futuremark 3DMark06
• Unreal Tournament 2004
• AquaMark3
• Quake 4 ver.1.20
• F.E.A.R.

Results - SYSmark 2004

Firing up our real-world user-simulation benchmarking program, SYSmark 2004's overall scores are vital to understand the performance standings when dabbling in heavy office applications as well as multimedia creation activities. Right off from the start, the newcomer leverages on its higher clock speeds to post results that are 4.5% faster than the FX-62. However, this is still a far cry from the performance figures of the faster Core 2 Duo series. Even the Core 2 Duo E6600 has no problems copping against the Athlon 64 X2 6000+. Worst of all is that the E6600 is half the cost of an Athlon 64 X2 6000+. The nearest competitor in terms of price would be the Core 2 Duo E6700 (which is still a lot cheaper) and it too has at least a 10% lead over the newcomer.

Results - Lightwave 3D 7.5

In this 3D modeling program's rendering simulation tests, the Athlon 64 X2 6000+ made better headway. In the more taxing Tracer-Radiosity render set, AMD's new star almost matched the performance of a Core 2 Duo E6700, while in the easer Sunset render set, it went up against even an a Core 2 Extreme X6800. Finally, the Athlon 64 X2 6000+ is showing some sparks of competition, but can it keep in more tests to come? Read on.

Results - Futuremark PCMark05

Here's another synthetic test suite in PCMark05, but this time, we measure subsystem performance based on a few quick test sets built into the benchmark. Unfortunately, the outcome is again somewhat similar to SYSmark 2004 where even the Core 2 Duo E6600 stands in better light. Surprisingly, the memory tests didn't really pickup any performance deficit from an under efficient memory subsystem.

Results - SPECViewperf 8.01

SPECViewperf's OpenGL nature has a mix of workloads that stresses the memory subsystem and scales well with CPU and GPU performance. With the GPU being capped in every test scenario, we've highlighted a couple of viewsets that react the most to processor scaling. In these, we saw the Athlon 64 FX2 6000+ chalking up to 6% performance gains against the FX-62. In spite of that, this OpenGL benchmark is unfortunately too much for the AMD processors to even dent a lower class US$220 Core 2 Duo E6400 processor!

There are however, more memory dependent viewsets (which we've not shown here), and these actually showed that the Athlon 64 X2 6000+ fared only just as well as the FX-62 or a tad worse because of its inefficient utilization of DDR2-800 memory.

Results - Cinebench 2003 and XMpeg 5.03

In Cinebench 2003's highly-threaded rendering benchmark, we once again saw similar performance standings where the Athlon 64 X2 6000+ leads its predecessor by around 6%, but still ends up slower than a Core 2 Duo E6700 processor.

In our Xmpeg DivX encoding tests, we found the Athlon 64 X2 6000+ to be 7% speedier than the FX-62 part by shaving 42 seconds off the 1GB test file encoded. Unfortunately, this is still over 10% slower than even the Core 2 Duo E6600 which manages to complete nearly a minute faster than the newcomer from AMD.

Results - Futuremark 3DMark06

In 3DMark06's tests, the CPU is also actively involved in the normal graphics benchmarks for some physics and flight path calculation of the jet-packed airship rendered on screen, but when considering its influence in the big picture of a gaming performance indicator, graphics performance is still key. Thus you can see why the Athlon 64 X2 6000+ fared well in the overall 3DMark scores, but not so when considering the CPU tests only. While the GPU may have the last say today in pure gaming performance, this might not be the case considering games like Alan Wake in the near future; so don't write-off the influence of the CPU just yet.

Results - Unreal Tournament 2004 and AquaMark3

Gaming used to AMD's forte and we can still see some remnants of that aspect peering through with the A64 X2 6000+ as it beat the Core 2 Duo E6700 in Unreal Tournament 2004 and managed a decent game performance standing in AquaMark 3 as well. However the CPU test segment of AquaMark 3 once again gave way to its weakness just like in 3DMark06.

Results - Quake 4

In the popular Quake 4 benchmark, we see the Athlon 64 X2 6000+ on equal footing with the Core 2 Duo E6600 processor, but still falls short of the E6700's standard by a notable margin. While the Athon 64 X2 6000+ is plenty fast enough with a GeForce 7900 GT in this game (at this particular test settings), but so are all other processors in this comparison. The real question is which processor offers more performance to better outlast the rest of them when faced with the real crunch. Obviously, the higher the performance figures, the better, so the Core 2 Duo series still gets the nod here.

Power Consumption

It's been a while since we poked into the power consumption department but we hope to do that more often since we just acquired a new power meter. Take note that our previously measured power consumption numbers were based on the AcBel power supply's own meter and thus readings from it cannot be cross compared. Thus we have a fresh set of power readings in this article based on our newly acquired power meter. The system configuration is as mentioned in the test setup page and only consists of the bare necessities like one hard drive, one optical drive, a single graphics card and the platform itself.

Take note that we've enabled all platform power saving options like Enhanced Intel Speedstep Technology (EIST) as well as Cool'n'Quiet (CnQ) Technology on AMD processors to simulate real world setup experience in this segment. On a side note, with or without EIST enabled, we saw only small differences to the net power consumption on Intel processors, but this was the exact opposite observation for AMD processors. Without CnQ, these high-speed AMD processors drew up to 30% more power (depending on the situation). So if you are an Athlon 64 user, make sure you have installed the CnQ driver (from AMD's website) for Windows XP operating system. Windows Vista users don't require any drivers, but for both operating systems, you should load the 'Minimal Power Management' profile in the control panel to enable this power savings technique to kick into effect. With that said, the following are our observations:-

Idling on Windows Desktop

With AMD's CnQ enabled, the power consumption of the Athlon 64 X2 6000+ is just 8% higher than the Core 2 Duo processors. Without CnQ, we measured up to 125 watts power consumption - just at idle. It certainly pays to configure your power savings options as detailed above.

3DMark CPU Test 2

We spent much time trying out various programs from Prime95 to the highly acclaimed SPECCPU industry benchmark, but after a lengthy trial and error affair, we found that Futuremark's 3DMark06 CPU test happens to tax the CPU more than any other benchmarks with the highest temperatures recorded as well as the highest power draw among CPU related tests. In this max CPU loading scenario, the Athlon 64 X2 6000+ requires 55% more power than the Core 2 Duo E6600 and Core 2 Duo E6700 processors which it competes against in performance. So if you dabble in a lot of lengthy processing tasks, it's quite apparent where to channel your resources when planning a new system.

3DMark06 HDR/SM3.0 Deep Freeze Test

What about power consumption during gaming? Well, we have that covered with 3DMark06 too. This was again our choice after trying several games like F.E.A.R., Quake 4 and others which seemed to tax the system on a similar level as the 3DMark06 game tests. The Deep Freeze test in 3DMark06 seemed to have more impact and thus we stuck with it. Again the Core 2 processors had more than 40% power savings advantage and another clean sweep for the blue team.

Conclusion

There's actually nothing really new for us to convey since we've already previewed the Athlon 64 X2 6000+ in the guise of the rumored Athlon 64 FX-64 more than six months ago. Our findings today only further reinforced our earlier conclusion that the Athlon 64 X2 6000+, on average, performs slightly better than the Core 2 Duo E6600, but less than that of the Core 2 Duo E6700 model. However performance can dip below the E6600 in some scenarios. Judging by its estimated street price of just under US$500, this is still overpriced for what it offers. The Core 2 Duo E6600 retails for around US$315, while the E6700 model goes for about US$515. Now that's quite a discrepancy; from the savings obtained going the E6600 route, you can bag much better components like a speedier graphics card or a bigger monitor which directly lends its hand to better enjoyment. You can even overclock the living daylights out of the Core 2 Duo just using the stock cooler - an excellent performance to price ratio for those looking to maximize their hardware. The new AMD processor gives you no leeway for safe overclocking in our opinion when using the stock cooler (or even general air cooling for that matter).

Even besides the usual price and performance comparisons that are convincing enough to steer one away from the Athlon 64 X2 6000+, the power consumption figures are also not in favor of the new linchpin from AMD. At idle, the differences are small, but if you often load your system for intense gaming sessions or lengthy processing tasks, watch out for your monthly power bill. And the reason is because the Athlon 64 X2 6000+ consumes 50% more power than Core 2 Duo processors when busy crunching away.

No doubt that the Athlon 64 X2 6000+ is AMD's personal milestone in offering a 3GHz processor for the consumers, offering up to 6% performance improvement over its FX-62 predecessor, it still falls short of being in our recommendation list. On all grounds, Intel's Core 2 Duo is still the favorable choice. One step down, the Athlon 64 X2 5600+ (which is the former FX-62) at US$325 squares off with the Core 2 Duo E6600, for which it does hold its own in some of the tests. The same however can't be said of the Athlon 64 X2 6000+ unless it retails for well below the US$450 mark.

Our articles may contain affiliate links. If you buy through these links, we may earn a small commission.