Logo   Banner   TopRight
TopUnder
Transparent
AMD Athlon X2 BE-2350 and BE-2300 "Brisbane" Processors
Transparent
Date: Jun 05, 2007
Section:Processors
Author: Jeff Bouton
Transparent
Introduction and Product Specifications

 

Although they're hard at work on Barcelona and plan to release a new high-end, native quad-core CPU in the coming months, AMD has turned a good part of their attention to a different segment of the market, the low power segment.  Back in August of 2006, AMD released several low power CPUs that aimed to balance performance and power consumption.  One was the 'Windsor' based Athlon 64 X2 4600+ with a TDP of 65w and the other was the Athlon 64 X2 3800+ which had a low TDP of 35w.  While the processors were impressive from an efficiency standpoint, AMD struggled to keep up with demand and also offered the models at premium price points.  In February of 2007, AMD changed their approach somewhat, introducing its "Brisbane" core.  The new 65nm Athlon 64 X2 processor came in a smaller package and dropped maximum power requirements from 89w with the "Windsor" cores to 65w with "Brisbane".  Furthermore, the more power efficient processors were released in mass quantities, with no additions to the price structure.

Today brings us to the next step of AMD's low power strategy, with the release of AMD's 65nm Athlon X2 BE-2350 and BE-2300 processors.  These 'Brisbane' based models sport a maximum power rating of 45w and weigh in below $100, making for a cost effective processor that is cheap to buy and cheap to run.  The cores of these new processors are still 'Brisbane' based, so there isn't a lot new here to go over as far as their architectural characteristics go.  However, as you can see, AMD did find it the right time to introduce a new numbering scheme in an effort to more accurately represent a processor's features and performance.  

Specifications of the 65nm AMD Athlon X2 BE-2350 and BE-2300
"Brisbane"

Processor:
AMD Athlon™ X2 dual core processor BE-2350
AMD Athlon™ X2 dual core processor BE-2300
OPN:
AH2350IAA5DD

AH2300IAA5DD

CPU Core Count:
2

Operating Frequency:
BE-2350:  2.1GHz

BE-2300:  1.9GHz

L1 Cache Size:
64K - L1 instruction + 64K - L1 data cache per-core (256KB total L1)

L2 Cache Size:
512KB L2 data cache per-core (1MB total dedicated L2 cache)

Manufactured:
Fab 30 and 36 / Dresden, Germany

Process Technology:
65-nanometer DSL SOI (silicon-on-insulator) technology

Packaging:
Socket AM2 (940-pin organic micro PGA)

HyperTransport Spec:
One 16-bit/16-bit link @ 2.0GHz (1GHz DDR) full duplex (up to 8.0 GB/sec bandwidth)

Memory Controller:
One integrated 128-bit dual-channel memory controller (up to 12.8GB/sec bandwidth)

Supported Memory Speeds:
DDR2 memory up to and including PC2 6400 (DDR2-800) unbuffered

Total Processor bandwidth:
Up to 20.8 GB/sec

Approximate Transistor count:
221 million

Approximate Die Size:
118 mm2

Nominal Voltage:
1.15- 1.20 V

Max Thermal Power:
45W

Max Ambient Case Temp:
61o Celsius to 78o Celsius

Max Processor Current:
36.5 A

Min P-State  (power management):
-1.0 GHz
-Nominal Voltage @ min P-state:  1.10 V
-Max Thermal Power @ min P-state:  27.7 W
-Max Current @ min P-state:  22.5A


Processor Number Methodology

Along with the new low power processors comes a newly revamped numbering methodology.  As technology has advanced and the market has changed over the years, AMD felt now was the time to introduce a numbering system that is more reflective of their processors' characteristics.

The new numbering methodology is still in its infancy, with AMD cutting their teeth with the new BE-2350 and BE-2300.  At this stage of the game, "BE" is the only classification, with the first letter signifying the processor class and the second identifying the TDP, which in this case is 45w.  So with this new line up, the "BE" is an indicator that the processor is a member of the sub-65 Watt processor class.  The four digit numbering thereafter represent the processor's relative position within its family.  So, in this case the "2" indicates the chip is a member of the Athlon X2 family of processors and the "350" and "300" represent performance positions relative to others in its family, with the higher number representing higher performance.  The numbers are arbitrary, but AMD hopes they'll provide a clear picture of where a processor's performance should fall relevant to other chips in its family.

It may take a while to get a handle on the new numbering scheme, but as new chips become available, it should help consumers to differentiate the characteristics more clearly as new products are produced.  One last change AMD has thrown in as well is the dropping of "64" from the naming altogether.  This makes perfect sense since all of AMD's processors are 64-bit capable, and labeling it a 64-bit chip is a bit redundant. 

Transparent
AMD Athlon 64 X2 BE-2350 and BE-2300 "Brisbane"

AMD Athlon X2 BE-2350 and BE-2300
More from Brisbane

Aside from a lower Maximum Thermal Power rating of 45w, the AMD Athlon X2 BE-23xx series isn't all that different from other "Brisbane" based CPUs.  The major difference is the newer processors sport a lower operating voltage range of 1.15-1.20v where the 65w "Brisbane" ranges from 1.25-1.35v and the "Windsor" cores are rated at 1.3-1.35v.  Other than that, there’s no new features to report as CPU-Z shows below.  (Note: The latest release BIOS reports our review sample as a BE-235, which is an error in the BIOS code.  The newer Beta BIOS has this corrected, however, we didn't feel it appropriate to do our testing with a Beta BIOS).

small_011.JPG  small_014.JPG

small_cpuz1.JPG  small_cpuz2.JPG

BE-23xx series processors are manufactured at 65nm and use the 940 Pin AM2 specifications.  The "Brisbane" still follows the standard model of a 512KB L2 cache which is implemented across all "Brisbane" processors.  The core sports the same 64 L1 instruction and 64K L1 data cache while the integrated 128-Bit dual channel memory controller supports up to 800MHz DDR2.  When looking at the picture of the CPU, don't be confused by the "64" stamped on it, retail models will not carry the "64" designation any longer.  While the Athlon X2 BE-2350 is clocked at 2.1GHz, the Athlon X2 BE-2300 comes clocked 200MHz slower at 1.9GHz and costs a few dollars less.

Transparent
Market Position and DTX Form Factor

AMD Athlon X2 BE-2350 and BE-2300
Market Positioning and DTX Form Factor

While not targeted at enthusiast, AMD is taking aim at those interested in a low wattage system, whether we're talking about simple workstations or personal users looking for a small form factor low-power PC for desktop usage or HTPC applications.  There are a lot of areas where the biggest, fastest, more power hungry CPUs are not needed and that's the market AMD is turning their attention to here.  By making improvements to their fabrication process, AMD is able to churn out the new BE-23xx series processors at an attractive price point while anticipating no issues keeping up with demand.  This seems like a solid strategy as Intel has been sitting pretty atop the high end CPU sector, and AMD continues their attempts to solidify their position in the mainstream markets, where the bulk of the volume resides.

With so much attention being focused on the environment and becoming more efficient in every way possible, the low power BE-23xx series processors aim to fill a specific need.  Whether your motives are environmentally ethical or you simply want to save on electric costs, the BE-23xx series appears to be an attractive option.  This is even more so with businesses of all sizes who stand to save a fair amount of money in operating costs, minimizing power consumption whether the system is idle, with Cool'n'Quiet, or under load, where AMD claims the BE-23xx will consume a minimum of 20 watts less than their non-low power counterparts.

AMD is also readying these chips to aid in the adoption of the DTX form factor. DTX aims to facilitate a broad range of small, low-power systems through the cooperation of OEMs from all sectors of manufacturing.  DTX is focused around reducing the overall footprint of the PC while accommodating existing ATX and mini-ITX form factors.  The new form factor will come in both DTX and mini-DTX flavors which aim to improve compatibility with existing hardware while reducing production costs for OEMs.  Another benefit of DTX is its potential to help OEMs flex their creativity with SFF PCs by alleviating the need for a proprietary solution, which isn't uncommon.  This may sway some potential consumers worried about the ability to upgrade in the future, which is valid.  With DTX, a standardized format can help potential consumer gain confidence that they will be able to update their hardware as technology improves. For a more in-depth explanation of DTX, we suggest checking out www.dtxpc.org

Transparent
HH Test Bed, Power & Thermal Characteristics (IDLE)

The HotHardware Test Systems
AMD Athlon X2 BE-2350 Powered
Power Supply: 
Thermaltake Toughpower 750 W Cable Management

Processor
AMD Athlon X2 BE-2350
2.1GHz w/512KB L2 Cache
AMD Athlon 64 X2 4000+ Brisbane
2.1GHz w/512KB L2 Cache

AMD Athlon 64 X2 3800+ Windsor
2GHz w/512KB of L2 Cache

Motherboard
Asus M2A-VM HDMI
AMD 690G


Video
Radeon X1250 Integrated

Memory
2048MB PQI24200 Turbo RAM

Audio
Integrated on board


Hard Drive
1
- 74GB Western Digital "Raptor"

CD/DVD-ROM

1 - Artec
16X DVD +/-R/RW

Watt Meter
P4400 KILL A WATT Meter

Software

Prime95 (Torture Test)
SANDRA XI - CPU,
Memory , Latency and Cache Testing
3DMark06 - CPU Test
PCMark05 - CPU and Memory Test
Cinebench R9.5
Quake 4 - Low Resolution Testing
WorldBench 5 - Office XP & Photoshop 7

When the "Brisbane" core made its first appearance in February of 2007, we maintained an apples-to-apples setup, comparing various thermal, power and performance data between the AMD Athlon 64 X2 5000+ 90nm "Windsor" and the AMD Athlon 64 X2 5000+ 65nm "Brisbane".  In this review, we are going to stick to that approach, focusing on a three-way comparison between the BE-2350 and a Athlon 64 X2 4000+ Brisbane core, which is virtually identical except the 4000+ carries a 65w power rating vs the BE-2350's 45w spec.  Lastly, we threw in an Athlon 64 X2 3800+ Windsor core which runs 100MHz slower and carries a 65w power rating as well.

AMD Athlon X2 BE-2350
Power & Thermal Characteristics - IDLE

Starting out with thermal and power testings, we recorded the idle wattage of our test system with Cool'n'Quiet enabled.  All measurement were recorded at 30 minute intervals using a Kill-A-Watt meter at the outlet.  In all testing, a stock cooler was used along with Arctic Silver 5 thermal grease.

wattidle.PNG

With Cool'n'Quiet enabled, each processor dropped down to an x5 multiplier, running each CPU at 1000MHz.  As expected, all three processors returned similar results, which is expected.  All three test platforms hovered in the 67-68 watt range while idle, yielding no surprises.

While monitoring the system wattage at the outlet, we also had a copy of ASUS Probe II running in the background to monitor the CPU temperature.

tempidle.PNG

As with the wattage, all three processors ran at nearly the same range, each with 1°C of each other.  The AMD Athlon X2 BE-2350 was the coolest of the three chips, idling at 15°C.

Transparent
Power & Thermal Characteristics (LOAD) & Overclocking

AMD Athlon X2 BE-2350
Vitals:  Power & Thermal Characteristics - 100% LOAD

When running the CPUs at idle, the power and thermal difference are not going to be all that evident.  Only once we stress each CPU can we truly see the relationship between each core's wattage rating as well as the relationship between wattage draw and CPU temperature.  To stress each CPU, we ran two iterations of Prime95 with the Torture test set for maximum heat load.  In each test, the system was left to run for 30 minutes before recording the results.

wattload.PNG

With each CPU running at its maximum, the benefit of the lower power BE-2350 becomes much more evident.  Both the Athlon 64 X2 4000+ and Athlon 64 X2 3800+ are 65w CPUs, so their performance deltas were expectedly similar, however, the Athlon X2 BE-2350 demonstrated a 16w drop, which was close to the 20w difference that was anticipated.

tempload.PNG

When measuring the CPU temperatures during each test using ASUS Probe II, we recorded a drop of 10°C between the two "Brisbane" cores.  The Athlon X2 BE-2350 and the Athlon 64 X2 3800+ "Windsor" had a 7°C difference. 

Overclocking the Athlon X2 BE-2350
Kicking Things Up a Notch

Before launching into the benchmarking segment, we took a moment to see what kind of clock speed headroom the Athlon X2 BE-2350 had to offer.  The first thing we did was drop our memory speed to 400MHz to ensure the memory didn't interfere with the maximum overclock and set the CPU voltage to 1.5v.  Next, we pushed the CPU clock up step by step to see how high we could go.   

AMD Athlon X2 BE-2350

small_oc.JPG

With a stock cooler, we managed a stable CPU reference clock of 240MHz which pushed the CPU itself from 2.1GHz to 2.52GHz, an even 20% increase.  Next, we put stability to the test and ran two iterations of the Prime95 Torture test for 15 minutes with no operational issues noted.  Obviously we were pushing the CPU temperature a bit here, but with an aftermarket cooler we suspect the CPU has a bit more muscle to share.  We should also note that the added clock speed did tack on close to 80 watts to the system draw, averaging a steady 184 Watts during load testing.  With a near 20% improvement in raw performance, this nearly doubled the overall wattage draw under load.

Transparent
Synthetic Testing Analysis with SiSoft Sandra - CPU & Memory

Next we are going to shift our focus to performance to see what impact, if any, the reduced power design of the Athlon X2 BE-2350 has on the CPU's actual performance.

Synthetic Performance Metrics - CPU & Memory
SANDRA XI

In our first round of tests, we focused on synthetic testing with SANDRA XI.  First we started with the basics, running the Arithmetic, Multi-Media, and Memory Bandwidth modules.

sancpu.PNG

sanmm.PNG

Neither CPU based test returned any results that were out of the ordinary.  With the Athlon X2 BE-2350 essentially the same as the reference 4000+, we expect these results to track closely.  In both the Arithmetic and Multi-Media modules, the BE-2350 lagged slightly behind the 4000+, however, the differences were minimal. 

sanmem.PNG

Looking at memory performance, we saw the Athlon X2 BE-2350 lag slightly behind the Athlon 64 X2 4000+ by a small margin, whereas the Windsor based 3800+ was more noticeable with its lower clock speed.

Transparent
Synthetic Testing Analysis with SiSoft Sandra - Latency

Synthetic Performance Metrics - Latency
SANDRA XI

When the Brisbane core was first introduced in February, there was a fair amount of talk about latency issues.  This probably would not have gotten as much attention as it did except for the fact that the processor was specifically billed as having a smaller die size and no loss in performance.  It turns out, the Brisbane core had higher latency when accessing cache, but the performance impact showed up more so in synthetic testing rather than real world tests.  Experience told us we should take a look at several latency tests and overall cache performance to see if anything unusual shows up with the new low power Athlon X2 BE-2350.  To start, we ran SANDRA XI's Memory Latency test. 

sanmemlat.PNG

According to SANDRA XI's Memory Latency component, there was a nominal increase in latency recorded between the Brisbane based Athlon 64 X2 4000+ and the Athlon X2 BE-2350, with 6ns leaning in favor of the 4000+

sancachemem.PNG

Contrary to the Memory Latency test, the Cache and Memory module reported improved performance compared to the Athlon 64 X2 4000+, with the Athlon X2 BE-2350 posting an additional 234MB/s over all.  Ultimately, these results are all very close in the end and we'd rather rely on real world application testing before drawing any major conclusions.

Transparent
Futuremark 3DMark06 - CPU Test and PCMark05

 

Futuremark 3DMark06 - CPU Test
CPU Performance

3DMark06's test is a multi-threaded "gaming related" DirectX metric that's useful for comparing relative performance between similarly equipped systems.  This test consists of different 3D scenes that are generated with a software and hardware GPU renderers, which is also dependant on the host CPU's performance. In its CPU tests, the calculations normally reserved for your 3D accelerator are instead sent to the host processor. 

3dmk.PNG 

The performance between the Athlon 64 X2 4000+ and Athlon X2 BE-2350 was nearly identical, with the low power Athlon X2 BE-2350 recording a minor lead over the 4000+.

Futuremark PCMark05
Synthetic CPU and Memory Assessment

Next, we turned our attention to PCMark05's CPU and Memory modules.  Before each test we inserted a snip of FutureMark's explanation for each segment.

"The CPU test suite is a collection of tests that are run to isolate the performance of the CPU. The CPU Test Suite also includes multithreading: two of the test scenarios are run multithreaded; the other including two simultaneous tests and the other running four tests simultaneously. The remaining six tests are run single threaded. Operations include, File Compression/Decompression, Encryption/Decryption, Image Decompression, and Audio Compression" - Courtesy FutureMark Corp.

pcmkcpu.PNG

FutureMark’s PCMark05's CPU module posted similar performance deltas to what we recorded with the 3DMark06 CPU test.  Here the Athlon X2 BE-2350 posted an even smaller lead over the Athlon 64 X2 4000+ while the 3800+ posted a predictable third position.

" The Memory test suite is a collection of tests that isolate the performance of the memory subsystem. The memory subsystem consists of various devices on the PC. This includes the main memory, the CPU internal cache (known as the L1 cache) and the external cache (known as the L2 cache). As it is difficult to find applications that only stress the memory, we explicitly developed a set of tests geared for this purpose. The tests are written in C++ and assembly. They include: Reading data blocks from memory, Writing data blocks to memory performing copy operations on data blocks, random access to data items and latency testing."  - Courtesy FutureMark Corp.

pcmkmem.PNG

With memory performance, the "Brisbane" based processors were on the same level, with the elder Athlon 64 X2 4000+ posting slightly higher performance.  Ultimately, it's the "Windsor" based Athlon 64 X2 3800+ that posted the highest memory scores, adding a 100 point lead over the "Brisbane" cores.

Transparent
World Bench 5.0: Office XP SP2 and Photoshop 7

To see how the synthetic test results we've seen thus far translate into real world performance, we ran several common tests using PC World Magazine's WorldBench 5.0. This Business and Professional application benchmark suite consists of a number of performance modules that each utilize one, or a group of, popular applications to gauge performance.  In this segment, we ran Office XP SP2 and Photoshop 7  Modules.

PC World's World Bench 5.0: Office XP SP2 and Photoshop 7
Business And Content Creation application performance

wboff.PNG

The WorldBench 5 Office XP test recorded the Athlon X2 BE-2350 as the slowest of the three chips by 19 seconds compared to the Athlon 64 X2 4000+.  The best performer was the Windsor based 3800+ which completed the test 23 seconds faster than the low power BE-2350.

wbps.PNG

With the Photoshop 7 test, the results swung in favor of both "Brisbane" cores, with the two tied at 349 seconds in the end.  The Athlon 64 X2 3800+ fell to 361 seconds, reflecting the 100MHz lower clockspeed.

Transparent
Cinebench R9.5 & Quake 4 Low Res.

Cinebench R9.5
Single and Multi Threaded

The Cinebench R9.5 benchmark is an OpenGL 3D rendering performance test, based on the commercially available Cinema 4D application.   This is a multi-threaded, multi-processor aware benchmark that renders a single 3D scene and tracks the length of the entire process. The time it took each test system to render the entire scene is represented in the graph below (listed in seconds).

cine.PNG 

Cinebench yielded little surprises.  The two "Brisbane" based processors returned identical results while the 2GHz Athlon 64 X2 3800+ was a tad slower in each test.

Performance Comparisons with Quake 4
OpenGL

Quake 4
id Software, in conjunction with developer Raven, recently released the latest addition to the wildly popular Quake franchise, Quake 4. Quake 4 is based upon an updated and slightly modified version of the Doom 3 engine, and as such performance characteristics between the two titles are very similar.  Like Doom 3, Quake 4 is also an OpenGL game that uses extremely high-detailed textures and a ton of dynamic lighting and shadows, but unlike Doom3, Quake 4 features some outdoor environments as well. We ran this Quake 4 benchmark using a custom timedemo with the game set to its "Low-Quality" mode at a resolution of 640 x 480 with AA and aniso disabled.

q4low.PNG 

We continued to see a similar trend with Quake 4 low resolution testing where the two "Brisbane" based processors returned the same results while topping the Windsor based 3800+ by half a frame per second.

Transparent
Performance Summary and Final Thoughts

Performance Summary:  When comparing the new Athlon X2 BE-2350 to the similar Athlon X2 64 4000+ and Athlon X2 64 3800+, virtually all of the benchmark results fall within expected ranges.  Power consumption and thermal testing, however, did show a significant decrease compared to both 65w processors.  In actual performance, the Athlon X2 BE-2350 and Athlon 64 X2 4000+ were very close with only WorldBench 5's Office XP SP-2 test recording any major difference.  We also recorded a slight latency increase with respect to Memory, which may have influenced the results in the Office XP SP-2 test.  As whole, however, the performance was on par with similarly equipped "Brisbane" processors while reducing power consumption and temperature under load nearly 20%

In AMD's quest to reduce power consumption, they seem to have hit a good balance with the Athlon X2 BE-23xx series processor line.  It surely would have been more impressive if they were able to maintain the 35w TDP of the older "Windsor" Athlon 64 X2 3800+, but as whole, the "BE" models do a nice job of improving power consumption efficiency while maintaining the same performance, clock for clock, as comparable Athlon X2 models.  With the Athlon X2 BE-2350 tagged with an MSRP of $91 and the BE-2300 weighing in at $86, these chips certainly offer decent bang for the buck.

At the moment, Intel doesn't appear to have a direct competitor to the BE-Series in regard to power consumption, although the new 65nm Pentium E2140 Dual-Core chip does consume much less power than higher-clock Conroe-based processors.  What the future holds is unclear, but at the moment, the BE series Athlon X2s do look to be some of the most energy efficient processors available.  While they are targeted at the mainstream market, we think these chips also offer the enthusiast a bit as well.  With the BE-2350 having the strong potential to overclock to 2.5GHz, this brings the performance level up to an Athlon 64 X2 4800+ which sells for $38 more and may even reach 5200+ speeds, which sells for nearly $170.  So while the processor does appeal to more to basic workstation and small form factor uses, enthusiasts may find themselves a low cost gem in the BE series.

As we all become more aware of power consumption and efficiency, it's good to see that a number of major manufacturers are making efforts to do their part.  In actuality, AMD seemed to sign on a bit faster than others and appear to be making a concerted effort to reduce power consumption on a broad scale with their processors and core logic chipsets.  We look forward to see how these processor evolve. 

  • Low Power Requirements
  • Low Purchase Price
  • Low Operating Cost 
  • Excellent Overclocker
  • No New Performance Features 

Discuss This Article and Share Your Comments



Content Property of HotHardware.com