Most of you are aware by now of the problems with AMD’s quad-core Phenom processors. As we have chronicled closely, the chips were late to market and debuted with unexpectedly low clock frequencies. As a result, their performance was underwhelming compared to Intel’s Core 2 offerings. Worse yet, shortly before its release, AMD discovered a bug in the Phenom that could cause a system hang in certain, very specific circumstances. This so called “TLB erratum” caused AMD to cease shipments of the Phenom’s server-oriented counterpart, the Opteron 2300 series, but the firm went ahead with its plans to sell Phenoms in consumer PCs. For those systems, AMD offered a workaround in the form of a BIOS update, but the cure was arguably worse than the affliction, causing substantial performance hit in many applications.
During our coverage of the unfolding TLB erratum story, AMD told us it planned to deliver a new revision of its quad-core silicon in “mid-to-late Q1” of this year. This revision, dubbed B3, would include a proper chip-level fix for the TLB erratum. Since then, we’ve been waiting impatiently for these new chips to arrive while agitating for AMD to provide consumers with more information about whether and how they might disable the TLB workaround on existing Phenoms.
Fortunately, today, the wait appears to be over. Like a tired cliché rising from the keyboard of a website author, the Phenom has been resurrected in the form of silicon revision B3. Accordingly, AMD is announcing a whole new lineup of Phenom processors that should be available for purchase almost immediately. Even better, AMD seems to have found some additional clock frequency headroom in the B3 chips, so that lineup extends to new territory in the form of the Phenom X4 9750 and 9850 Black Edition processors.
New Phenoms: It’s all about the Xs
As expected, AMD’s new Phenom lineup includes a number of “xx50” model numbers intended to denote the presence of B3 silicon and a proper hardware fix for the TLB erratum. For instance, the Phenom 9550 replaces the Phenom 9500. Aside from the silicon rev and the TLB fix, the two products are essentially the same: both have four cores, a 2.2GHz core clock, 2MB of L3 cache, and all the rest. Of course, the Phenom 9550 should be faster in a default configuration because it doesn’t suffer the performance penalty caused by the TLB workaround.
The new Phenom line includes a number of surprises, though. Here’s a look at the whole list, with vitals for each model.
X4 9850 Black Edition
Perhaps the biggest surprise for the pedantic little man inside of 98.2% of geeks is the addition of “X3” and “X4” indicators in the new Phenom model names, whose numbers correspond to the number of active processing cores on each chip. This naming scheme harkens back to the Athlon 64 X2, of course. We saw the “Phenom X4” name bandied about prior to the Phenom’s initial release, but then AMD canned it and went with straight model numbers. Apparently, that didn’t take. AMD says its customers liked the Xs, and so they’re backthis time, presumably, for good.
Once you’re over the shock of that nomenclature recalibration, you’ll probably notice the “X3” Phenoms in the list. Yep, these are the vaunted triple-core variants of the Phenom we’ve expected for some time; they’re official now, but only for large PC makers known as original equipment manufacturers (OEMs). Over half the new Phenom lineup is intended only for OEMs, in fact, and that’s why we have only approximate prices listed for those processors, if we have them at all. The big dawgs get all sorts of interesting stuff, including the tri-core chips, a low-power variant of the Phenom with a 65W thermal/power envelope (or TDP), and a Phenom X4 9750 with a 95W TDP. We can expect to see consumer versions of these products eventually; those should all be B3 silicon when they arrive.
AMD saved some of the fun for the rest of us, though. It claims the Phenom X4 9550 is the lowest-priced quad-core CPU on the market at under $200, and the 9750 at 2.4GHz doesn’t look like a bad deal, either. The most intriguing product of the lot, however, is the Phenom X4 9850 Black Edition. This puppy comes with a 2.5GHz core clock and a 2GHz north bridge clockimportant because the north bridge clock governs the speed of the L3 cache. That should make the 9850 a little bit quicker than it might otherwise be. Like AMD’s other “Black Edition” processors, the 9850 has an unlocked upper clock multiplier that makes overclocking ridiculously, guilt-inducingly easy. And although it’s AMD’s flagship model, the 9850 lists for only $235, well below the list prices for ostensible competitors like the Core 2 Quad Q6600. This combination of attributes should make the 9850 the one to have, in my view.
Keeping tabs on the competition
The new Phenoms have the fortune of making it to market just as Intel is struggling to meet demand for its new 45nm chips, which means the Phenoms face a slightly less lethal mix of competitors.
For instance, this review is our first look at Intel’s new 45nm Core 2 Duo E8400 and E8500 processors. Although they have only two cores to the Phenom’s four, those two cores run at 3GHz and 3.16GHz frequencies, with (roughly) up to 20% higher clock-for-clock performance than the 65nm version of the Core 2 Duo. As a result, they’re able to give the lower frequency Phenoms a run for their money, even in widely multithreaded applications. But, as we’ve noted, the availability of these processors at online vendors is rather spotty, and prices have risen where they are available. The E8500’s ostensible $266 list price is already higher than the Phenom X4 9850’s, too.
The Phenom will face even deadlier competition in the form of Intel’s new 45nm Core 2 Quad processors, but those are even harder to find right now. We pinged Intel about those CPUs in preparation for this review, and the company says all of the new 45nm Core 2 Duo and Quad processors are indeed shipping nowincluding the Core 2 Quad Q9450, Q9550, and the Core 2 Extreme QX9770, interestingly enough. We weren’t able to get any specific ETA for when the availability picture should improve, but Intel says it expects supplies to rise to meet demand as 45nm production ramps up.
One of the more notable developments in Intel’s 45nm lineup, incidentally, is the recent and very quiet introduction of the Core 2 Quad Q9300. The Q9300 runs at 2.5GHz and, like the other 45nm Core 2 Quads, has a 1333MHz front-side bus. However, the Q9300 has only half the L2 cache6MB totalof its siblings and looks tailor-made to take on the new Phenoms. We’ll try to get our hands on a Q9300 to test soon, along with other speed grades of the 45nm Core 2 Quads.
Please note that we’ve included a “Phenom ES 2.6GHz” processor in the results on the following pages. This is an engineering sample chip clocked at 2.6GHz with a 2GHz north bridge that AMD supplied to us back when the Phenom first launched. No real Phenom product is yet shipping at this speed, but we’ve included it for comparison’s sake, nonetheless.
Also, we have included performance results for a couple of very high-end systems, including a dual Xeon X5365 system and a Skulltrail dual Core 2 Extreme QX9775 rig. With eight cores and price tags in the many thousands of dollars, such systems aren’t direct competitors for any Phenom AMD currently offers. If their presence annoys you, hold up two fingers to the screen to block out the bars representing them while reading the performance graphs.
Our testing methods
As ever, we did our best to deliver clean benchmark numbers. Tests were run at least three times, and the results were averaged.
Our test systems were configured like so:
|Processor|| Core 2 Quad Q6600 2.4GHz
Core 2 Extreme QX6800 2.93GHz
|Core 2 Duo E6750 2.66GHz
Core 2 Extreme QX6850 3.00GHz
2 Extreme QX9770 3.2GHz
2 Extreme QX9775 3.2GHz
| Athlon 64 X2 5600+ 2.8GHz
Athlon 64 X2 6000+ 3.0GHz
Athlon 64 X2 6400+ 3.2GHz
|Dual Athlon 64 FX-74 3.0GHz||Phenom
Phenom engineering sample (ES) 2.6GHz
|Core 2 Extreme QX9650 3.00GHz||Phenom
X4 9750 2.4GHz
Phenom X4 9850 Black Edition 2.5GHz
2 Duo E8400 3.0GHz
Core 2 Duo E8500 3.16GHz
|System bus||1066MHz (266MHz quad-pumped)||1333MHz (333MHz quad-pumped)||1600MHz
|1GHz HyperTransport||1GHz HyperTransport||1GHz HyperTransport||1GHz HyperTransport|
|Motherboard||Gigabyte GA-P35T-DQ6||Gigabyte GA-P35T-DQ6||Gigabyte
|Asus M2N32-SLI Deluxe||Asus L1N64-SLI WS||MSI
V1.2B1 (TLB patch)
|North bridge||P35 Express MCH||P35 Express MCH||X38
|nForce 590 SLI SPP||nForce 680a SLI||790FX||790FX|
|South bridge||ICH9R||ICH9R||ICH9R||6231ESB ICH||6321ESB ICH||nForce 590 SLI MCP||nForce 680a SLI||SB600||SB600|
|Chipset drivers||INF Update 126.96.36.1993
Intel Matrix Storage Manager 7.5
|INF Update 188.8.131.523
Intel Matrix Storage Manager 7.5
|INF Update 184.108.40.2063
Intel Matrix Storage Manager 7.5
Intel Matrix Storage Manager 7.5
Intel Matrix Storage Manager 7.8
|ForceWare 15.01||ForceWare 15.01||–||–|
|Memory size||4GB (4 DIMMs)||4GB (4 DIMMs)||4GB (4 DIMMs)||4GB
|4GB (4 DIMMs)||4GB (4 DIMMs)||4GB (4 DIMMs)||4GB (4 DIMMs)|
|Memory type||Corsair TWIN3X2048-1333C9DHX
DDR3 SDRAMat 1066MHz
DDR3 SDRAMat 1333MHz
DDR2 SDRAMat 800MHz
|Samsung ECC DDR2-667
FB-DIMM at 667MHz
ECC DDR2-800 FB-DIMM at 800MHz
DDR2 SDRAMat ~800MHz
DDR2 SDRAMat ~ 800MHz
DDR2 SDRAMat 800MHz
DDR2 SDRAMat 800MHz
|CAS latency (CL)||8||8||4||5||5||4||4||4||4|
|RAS to CAS delay (tRCD)||8||9||4||5||5||4||4||4||4|
|RAS precharge (tRP)||8||9||4||5||5||4||4||4||4|
|Cycle time (tRAS)||20||24||18||15||18||18||18||18||18|
with Realtek 220.127.116.1149 drivers
with Realtek 18.104.22.16849 drivers
with Realtek 22.214.171.12449 drivers
with Realtek 126.96.36.19949 drivers
with SigmaTel 6.10.5511.0 drivers
|Integrated nForce 590 MCP/AD1988B
with Soundmax 188.8.131.5200 drivers
|Integrated nForce 680a SLI/AD1988B
with Soundmax 184.108.40.20600 drivers
with Realtek 220.127.116.1132 drivers
with Soundmax 18.104.22.16880 drivers
|Hard drive||WD Caviar SE16 320GB SATA|
|Graphics||GeForce 8800 GTX 768MB PCIe with ForceWare 163.11 and 163.71 drivers|
|OS||Windows Vista Ultimate x64 Edition|
|OS updates||KB940105, KB929777 (nForce/790FX systems only), KB938194, KB938979|
Please note that testing was conducted in two stages. Non-gaming apps and Supreme Commander were tested with Vista patches KB940105 and KB929777 (nForce systems only) and ForceWare 163.11 drivers. The other games were tested with the additional Vista patches KB938194 and KB938979 and ForceWare 163.71 drivers.
Thanks to Corsair for providing us with memory for our testing. Their products and support are far and away superior to generic, no-name memory.
Our single-socket test systems were powered by OCZ GameXStream 700W power supply units. The dual-socket systems were powered by PC Power & Cooling Turbo-Cool 1KW-SR power supplies. Thanks to OCZ for providing these units for our use in testing.
Also, the folks at NCIXUS.com hooked us up with a nice deal on the WD Caviar SE16 drives used in our test rigs. NCIX now sells to U.S. customers, so check them out.
The test systems’ Windows desktops were set at 1280×1024 in 32-bit color at an 85Hz screen refresh rate. Vertical refresh sync (vsync) was disabled.
We used the following versions of our test applications:
- SiSoft Sandra XI.SP4a 64-bit
- CPU-Z 1.40
- WorldBench 6 beta 2
- Team Fortress 2
- Lost Planet: Extreme Condition with DirectX 10
- BioShock 1.0 with DirectX 10
- Supreme Commander 1.1.3260
- Valve VRAD map build benchmark
- Valve Source Engine particle simulation benchmark
- Cinebench R10 64-bit Edition
- POV-Ray for Windows 3.7 beta 21a 64-bit
- CASE Lab Euler3d CFD benchmark multithreaded edition
- MyriMatch proteomics benchmark
- notfred’s Folding benchmark CD 8/8/07 revision
- picCOLOR 4.0 build 598 64-bit
- The Panorama Factory 4.5 x64 Edition
- Windows Media Encoder 9 x64 Edition
- LAME MT 3.97a 64-bit
- VirtualDub 1.7.6 with DivX 6.7
The tests and methods we employ are usually publicly available and reproducible. If you have questions about our methods, hit our forums to talk with us about them.
Memory subsystem performance
We’ll kick things off with our usual set of synthetic memory performance tests. As always, these tests tell us more about the particular characteristics of each CPU and system architecture than they do about real-world performance prospects. They let us look at how fast (in terms of bandwidth) and quick (in terms of latency) each CPU’s memory subsystem can be.
No big surprises here. The B3-revision Phenoms slide right in where expected given their clock frequencies. As we’ve noted before, the Phenom achieves a nice combination of memory bandwidth and access latencies, due in part to its integrated memory controller. However, its access latencies to main memory are inflated somewhat due to the presence of its L3 cache. The Athlon 64 X2 lacks an L3 cache and is about 10ns quicker to main memory as a result.
Team Fortress 2
We’ll kick off our gaming tests with some Team Fortress 2, Valve’s class-driven multiplayer shooter based on the Source game engine. In order to produce easily repeatable results, we’ve tested TF2 by recording a demo during gameplay and playing it back using the game’s timedemo function. In this demo, I’m playing as the Heavy Weapons Guy, with a medic in tow, dealing some serious pain to the blue team.
We tested at 1024×768 resolution with the game’s detail levels set to their highest settings. HDR lighting and motion blur were enabled. Antialiasing was disabled, and texture filtering was set to trilinear filtering only. We used this relatively low display resolution with low levels of filtering and AA in order to prevent the graphics card from becoming a primary performance bottleneck, so we could show you the performance differences between the CPUs.
Notice the little green plot with four lines above the benchmark results. That’s a snapshot of the CPU utilization indicator in Windows Task Manager, which helps illustrate how much the application takes advantage of up to four CPU cores, when they’re available. I’ve included these Task Manager graphics whenever possible throughout our results. In this case, Team Fortress 2 looks like it probably only takes full advantage of a single CPU core, although Nvidia’s graphics drivers use multithreading to offload some vertex processing chores.
Since TF2 doesn’t make use of any more than two CPU cores, the Phenoms have no advantage over dual-core chips. Clock for clock, Intel’s Core 2 chips are faster here; at 2.4GHz, the Core 2 Quad Q6600 outperforms the Phenom X4 9750. And the Core 2 Duo E8400 and E8500 are both well ahead of the Phenom X4 9850.
Lost Planet: Extreme Condition
Lost Planet puts the latest hardware to good use via DirectX 10 and multiple threadsas many as eight, in fact. Lost Planet‘s developers have built a benchmarking tool into the game, and it tests two different levels: a snow-covered outdoor area with small numbers of large villains to fight, and another level set inside of a cave with large numbers of small, flying creatures filling the air. We’ll look at performance in each.
We tested this game at 1152×864 resolution, largely with its default quality settings. The exceptions: texture filtering was set to trilinear, edge antialiasing was disabled, and “Concurrent operations” was set to match the number of CPU cores available.
I’m not sure what’s happening in the Snow level, but oddly, a couple of the lower-clocked Phenoms do unusually well there. The primary bottleneck in that test is probably the GPU, since scores are bunched tightly together for the various CPUs. Things change rather dramatically in the Cave level, where we get a rare taste of a game that uses more than two CPU cores to good effect. Here, the new Phenoms shine, outperforming the Core 2 Quad Q6600 and nearly matching the Core 2 Extreme QX6800.
We tested BioShock by manually playing through a specific point in the game five times while recording frame rates using the FRAPS utility. The sequence? Me trying to fight a Big Daddy, or more properly, me trying not to die for 60 seconds at a pop.
This method has the advantage of simulating real gameplay quite closely, but it comes at the expense of precise repeatability. We believe five sample sessions are sufficient to get reasonably consistent results. In addition to average frame rates, we’ve included the low frame rates, because those tend to reflect the user experience in performance-critical situations. In order to diminish the effect of outliers, we’ve reported the median of the five low frame rates we encountered.
For this test, we largely used BioShock‘s default image quality settings for DirectX 10 graphics cards, but again, we tested at a relatively low resolution of 1024×768 in order to prevent the GPU from becoming the main limiter of performance.
Our Bioshock results are an object lesson in CPU performance in today’s games: most of the time, you don’t need an especially fast CPU in order to get acceptable performance. Even at this modest display resolution, our graphics card (the very fast GeForce 8800 GTX) or some other constraint looks to be limiting frame rates. That said, the new Phenoms rank in the upper echelon of all of the processors we tested.
We tested performance using Supreme Commander‘s built-in benchmark, which plays back a test game and reports detailed performance results afterward. We launched the benchmark by running the game with the “/map perftest” option. We tested at 1024×768 resolution with the game’s fidelity presets set to “High.”
Supreme Commander’s built-in benchmark breaks down its results into several major categories: running the game’s simulation, rendering the game’s graphics, and a composite score that’s simply comprised of the other two. The performance test also reports good ol’ frame rates, so we’ve included those, as well.
The new Phenoms handle Supreme Commander easily, with the 9850 finishing just behind the Core 2 Quad Q6600. Once more, the Core 2 Duo E8400 and E8500 place higher, but the margins of difference here are very smalljust a few frames per second, when it comes down to it.
Valve Source engine particle simulation
Next up are a couple of tests we picked up during a visit to Valve Software, the developers of the Half-Life games. They had been working to incorporate support for multi-core processors into their Source game engine, and they cooked up a couple of benchmarks to demonstrate the benefits of multithreading.
The first of those tests runs a particle simulation inside of the Source engine. Most games today use particle systems to create effects like smoke, steam, and fire, but the realism and interactivity of those effects are limited by the available computing horsepower. Valve’s particle system distributes the load across multiple CPU cores.
The Phenom’s four cores give it a clear advantage over the 45nm Core 2 Duo chips here, but the Core 2 Quad Q6600 proves to be even faster.
Valve VRAD map compilation
This next test processes a map from Half-Life 2 using Valve’s VRAD lighting tool. Valve uses VRAD to precompute lighting that goes into games like Half-Life 2. This isn’t a real-time process, and it doesn’t reflect the performance one would experience while playing a game. Instead, it shows how multiple CPU cores can speed up game development.
Much like in the last test, the new Phenoms perform well here, but not quite well enough to catch the Q6600.
WorldBench’s overall score is a pretty decent indication of general-use performance for desktop computers. This benchmark uses scripting to step through a series of tasks in common Windows applications and then produces an overall score for comparison. WorldBench also records individual results for its component application tests, allowing us to compare performance in each. We’ll look at the overall score, and then we’ll show individual application results alongside the results from some of our own application tests. Because WorldBench’s tests are entirely scripted, we weren’t able to capture Task Manager plots for them, as you’ll notice.
Like most of the desktop applications out there today, including its component apps, WorldBench doesn’t gain much from more than two CPU cores. In fact, the Core 2 Duo E8500 nestles into fourth place just behind the quad-core QX6850. Clock for clock, Intel’s Core 2 architecture again looks to be faster, with the Core 2 Quad Q6600 scoring quite a bit higher than the Phenom X4 9750.
Productivity and general use software
MS Office productivity
Firefox web browsing
Multitasking – Firefox and Windows Media Encoder
WinZip file compression
Nero CD authoring
The new Phenoms put in a series of respectable showings in these productivity apps, with the 9850 matching up particularly well against the Core 2 Quad Q6600. The two exceptions are WinZip, where I’d wager the Core 2 chips’ larger caches may be helping them, and Nero, which depends greatly on disk controller performance. Here, the Phenoms’ scores suffer due to the 790FX chipset’s problems with ACHI and NCQ.
There’s just no denying the Intel processors’ dominance in this Photoshop test. The 45nm ones are especially potent, as the Core 2 Duo E8500’s third-place finish indicates.
The Panorama Factory photo stitching
The Panorama Factory handles an increasingly popular image processing task: joining together multiple images to create a wide-aspect panorama. This task can require lots of memory and can be computationally intensive, so The Panorama Factory comes in a 64-bit version that’s multithreaded. I asked it to join four pictures, each eight megapixels, into a glorious panorama of the interior of Damage Labs. The program’s timer function captures the amount of time needed to perform each stage of the panorama creation process. I’ve also added up the total operation time to give us an overall measure of performance.
This application uses four and even eight cores quite well. Nonetheless, the Core 2 Duo E8500 is nipping at the heels of the Phenom X4 9750, amazingly enough. Still, the Phenoms perform well for their price range.
picCOLOR image analysis
picCOLOR was created by Dr. Reinert H. G. Müller of the FIBUS Institute. This isn’t Photoshop; picCOLOR’s image analysis capabilities can be used for scientific applications like particle flow analysis. Dr. Müller has supplied us with new revisions of his program for some time now, all the while optimizing picCOLOR for new advances in CPU technology, including MMX, SSE2, and Hyper-Threading. Naturally, he’s ported picCOLOR to 64 bits, so we can test performance with the x86-64 ISA. Eight of the 12 functions in the test are multithreaded, and in this latest revision, five of those eight functions use four threads.
Scores in picCOLOR, by the way, are indexed against a single-processor Pentium III 1 GHz system, so that a score of 4.14 works out to 4.14 times the performance of the reference machine.
Some of picCOLOR’s functions use four threads, yet the E8400 and E8500 outperform the Q6600, as well as both new Phenoms.
Video encoding and editing
VirtualDub and DivX encoding with SSE4
Here’s a brand-new addition to our test suite that should allow us to get a first look at the benefits of SSE4’s instructions for video acceleration. In this test, we used VirtualDub as a front-end for the DivX codec, asking it to compress a 66MB MPEG2 source file into the higher compression DivX format. We used version 6.7 of the DivX codec, which has an experimental full-search function for motion estimation that uses SSE4 when available and falls back to SSE2 when needed. We tested with most of the DivX codec’s defaults, including its Home Theater base profile, but we enabled enhanced multithreading and, of course, the experimental full search option.
Obviously, the 45nm Intel CPUs with SSE4 are fastest here, as expected. The Phenoms, however, finish before the Core 2 Quad Q6600.
Windows Media Encoder x64 Edition video encoding
Windows Media Encoder is one of the few popular video encoding tools that uses four threads to take advantage of quad-core systems, and it comes in a 64-bit version. Unfortunately, it doesn’t appear to use more than four threads, even on an eight-core system. For this test, I asked Windows Media Encoder to transcode a 153MB 1080-line widescreen video into a 720-line WMV using its built-in DVD/Hardware profile. Because the default “High definition quality audio” codec threw some errors in Windows Vista, I instead used the “Multichannel audio” codec. Both audio codecs have a variable bitrate peak of 192Kbps.
Wow, this one’s tight. The Phenom X4 9850 is just one second behind the Q6600a virtual tie. With only two cores, the E8500 trails.
Windows Media Encoder video encoding
Roxio VideoWave Movie Creator
The virtual tie continues with Worldbench’s Windows Media Encoder test, but the stalemate breaks in VideoWave, where the Core 2 processors take a decisive lead.
LAME MT audio encoding
LAME MT is a multithreaded version of the LAME MP3 encoder. LAME MT was created as a demonstration of the benefits of multithreading specifically on a Hyper-Threaded CPU like the Pentium 4. Of course, multithreading works even better on multi-core processors. You can download a paper (in Word format) describing the programming effort.
Rather than run multiple parallel threads, LAME MT runs the MP3 encoder’s psycho-acoustic analysis function on a separate thread from the rest of the encoder using simple linear pipelining. That is, the psycho-acoustic analysis happens one frame ahead of everything else, and its results are buffered for later use by the second thread. That means this test won’t really use more than two CPU cores.
We have results for two different 64-bit versions of LAME MT from different compilers, one from Microsoft and one from Intel, doing two different types of encoding, variable bit rate and constant bit rate. We are encoding a massive 10-minute, 6-second 101MB WAV file here.
Regardless of which compiler we use, the Phenoms just aren’t that strong, relatively speaking, with LAME audio encoding.
Graphics is a classic example of a computing problem that’s easily parallelizable, so it’s no surprise that we can exploit a multi-core processor with a 3D rendering app. Cinebench is the first of those we’ll try, a benchmark based on Maxon’s Cinema 4D rendering engine. It’s multithreaded and comes with a 64-bit executable. This test runs with just a single thread and then with as many threads as CPU cores are available.
The X4 9850 outduels the Core 2 Quad Q6600 here, and notice how it does so. Although the Q6600 is faster with one thread, the Phenom scales better to four.
We caved in and moved to the beta version of POV-Ray 3.7 that includes native multithreading. The latest beta 64-bit executable is still quite a bit slower than the 3.6 release, but it should give us a decent look at comparative performance, regardless.
Clock for clock, the Phenom proves faster in POV-Ray than the Core 2 Quad, with one to four threads. However, the tables turn when we use the built-in POV-Ray benchmark scene, which largely relies on a single execution thread.
3ds max modeling and rendering
Once again, the Phenoms can’t quite keep pace with the Q6600. It’s close, though.
Next, we have a slick little Folding@Home benchmark CD created by notfred, one of the members of Team TR, our excellent Folding team. For the unfamiliar, Folding@Home is a distributed computing project created by folks at Stanford University that investigates how proteins work in the human body, in an attempt to better understand diseases like Parkinson’s, Alzheimer’s, and cystic fibrosis. It’s a great way to use your PC’s spare CPU cycles to help advance medical research. I’d encourage you to visit our distributed computing forum and consider joining our team if you haven’t already joined one.
The Folding@Home project uses a number of highly optimized routines to process different types of work units from Stanford’s research projects. The Gromacs core, for instance, uses SSE on Intel processors, 3DNow! on AMD processors, and Altivec on PowerPCs. Overall, Folding@Home should be a great example of real-world scientific computing.
notfred’s Folding Benchmark CD tests the most common work unit types and estimates performance in terms of the points per day that a CPU could earn for a Folding team member. The CD itself is a bootable ISO. The CD boots into Linux, detects the system’s processors and Ethernet adapters, picks up an IP address, and downloads the latest versions of the Folding execution cores from Stanford. It then processes a sample work unit of each type.
On a system with two CPU cores, for instance, the CD spins off a Tinker WU on core 1 and an Amber WU on core 2. When either of those WUs are finished, the benchmark moves on to additional WU types, always keeping both cores occupied with some sort of calculation. Should the benchmark run out of new WUs to test, it simply processes another WU in order to prevent any of the cores from going idle as the others finish. Once all four of the WU types have been tested, the benchmark averages the points per day among them. That points-per-day average is then multiplied by the number of cores on the CPU in order to estimate the total number of points per day that CPU might achieve.
This may be a somewhat quirky method of estimating overall performance, but my sense is that it generally ought to work. We’ve discussed some potential reservations about how it works here, for those who are interested. I have included results for each of the individual WU types below, so you can see how the different CPUs perform on each.
The Phenoms place well here on the strength of decent all-around performance, especially with the Tinker and Amber WU types.
Our benchmarks sometimes come from unexpected places, and such is the case with this one. David Tabb is a friend of mine from high school and a long-time TR reader. He recently offered to provide us with an intriguing new benchmark based on an application he’s developed for use in his research work. The application is called MyriMatch, and it’s intended for use in proteomics, or the large-scale study of protein. I’ll stop right here and let him explain what MyriMatch does:
In shotgun proteomics, researchers digest complex mixtures of proteins into peptides, separate them by liquid chromatography, and analyze them by tandem mass spectrometers. This creates data sets containing tens of thousands of spectra that can be identified to peptide sequences drawn from the known genomes for most lab organisms. The first software for this purpose was Sequest, created by John Yates and Jimmy Eng at the University of Washington. Recently, David Tabb and Matthew Chambers at Vanderbilt University developed MyriMatch, an algorithm that can exploit multiple cores and multiple computers for this matching. Source code and binaries of MyriMatch are publicly available.
In this test, 5555 tandem mass spectra from a Thermo LTQ mass spectrometer are identified to peptides generated from the 6714 proteins of S. cerevisiae (baker’s yeast). The data set was provided by Andy Link at Vanderbilt University. The FASTA protein sequence database was provided by the Saccharomyces Genome Database.
MyriMatch uses threading to accelerate the handling of protein sequences. The database (read into memory) is separated into a number of jobs, typically the number of threads multiplied by 10. If four threads are used in the above database, for example, each job consists of 168 protein sequences (1/40th of the database). When a thread finishes handling all proteins in the current job, it accepts another job from the queue. This technique is intended to minimize synchronization overhead between threads and minimize CPU idle time.
The most important news for us is that MyriMatch is a widely multithreaded real-world application that we can use with a relevant data set. MyriMatch also offers control over the number of threads used, so we’ve tested with one to eight threads.
I should mention that performance scaling in MyriMatch tends to be limited by several factors, including memory bandwidth, as David explains:
Inefficiencies in scaling occur from a variety of sources. First, each thread is comparing to a common collection of tandem mass spectra in memory. Although most peptides will be compared to different spectra within the collection, sometimes multiple threads attempt to compare to the same spectra simultaneously, necessitating a mutex mechanism for each spectrum. Second, the number of spectra in memory far exceeds the capacity of processor caches, and so the memory controller gets a fair workout during execution.
Here’s how the processors performed.
When sorted by their best times, the Phenoms bracket the Q6600, and look at how they manage it. Although the Q6600 is quicker with one and two threads, the Phenoms scale better to four threads.
STARS Euler3d computational fluid dynamics
Charles O’Neill works in the Computational Aeroservoelasticity Laboratory at Oklahoma State University, and he contacted us to suggest we try the computational fluid dynamics (CFD) benchmark based on the STARS Euler3D structural analysis routines developed at CASELab. This benchmark has been available to the public for some time in single-threaded form, but Charles was kind enough to put together a multithreaded version of the benchmark for us with a larger data set. He has also put a web page online with a downloadable version of the multithreaded benchmark, a description, and some results here.
In this test, the application is basically doing analysis of airflow over an aircraft wing. I will step out of the way and let Charles explain the rest:
The benchmark testcase is the AGARD 445.6 aeroelastic test wing. The wing uses a NACA 65A004 airfoil section and has a panel aspect ratio of 1.65, taper ratio of 0.66, and a quarter-chord sweep angle of 45º. This AGARD wing was tested at the NASA Langley Research Center in the 16-foot Transonic Dynamics Tunnel and is a standard aeroelastic test case used for validation of unsteady, compressible CFD codes.
The CFD grid contains 1.23 million tetrahedral elements and 223 thousand nodes . . . . The benchmark executable advances the Mach 0.50 AGARD flow solution. A benchmark score is reported as a CFD cycle frequency in Hertz.
So the higher the score, the faster the computer. Charles tells me these CFD solvers are very floating-point intensive, but oftentimes limited primarily by memory bandwidth. He has modified the benchmark for us in order to enable control over the number of threads used. Here’s how our contenders handled the test with different thread counts.
Sometimes, having a better building block trumps having more blocks. Here, for example, the Core 2 Duo E8500 exactly ties with the Phenom X4 9850, despite the fact that this test can benefit from even eight threads on the right system. The Phenom’s showing is respectable, but the E8500’s is remarkable.
SiSoft Sandra Mandelbrot
Next up is SiSoft’s Sandra system diagnosis program, which includes a number of different benchmarks. The one of interest to us is the “multimedia” benchmark, intended to show off the benefits of “multimedia” extensions like MMX, SSE, and SSE2. According to SiSoft’s FAQ, the benchmark actually does a fractal computation:
This benchmark generates a picture (640×480) of the well-known Mandelbrot fractal, using 255 iterations for each data pixel, in 32 colours. It is a real-life benchmark rather than a synthetic benchmark, designed to show the improvements MMX/Enhanced, 3DNow!/Enhanced, SSE(2) bring to such an algorithm.
The benchmark is multi-threaded for up to 64 CPUs maximum on SMP systems. This works by interlacing, i.e. each thread computes the next column not being worked on by other threads. Sandra creates as many threads as there are CPUs in the system and assignes [sic] each thread to a different CPU.
We’re using the 64-bit version of Sandra. The “Integer x16” version of this test uses integer numbers to simulate floating-point math. The floating-point version of the benchmark takes advantage of SSE2 to process up to eight Mandelbrot iterations in parallel.
Yep, uh, there you have it.
Power consumption and efficiency
Now that we’ve had a look at performance in various applications, let’s bring power efficiency into the picture. Our Extech 380803 power meter has the ability to log data, so we can capture power use over a span of time. The meter reads power use at the wall socket, so it incorporates power use from the entire systemthe CPU, motherboard, memory, graphics solution, hard drives, and anything else plugged into the power supply unit. (We plugged the computer monitor into a separate outlet, though.) We measured how each of our test systems used power across a set time period, during which time we ran Cinebench’s multithreaded rendering test.
Almost all of the systems had their power management features (such as SpeedStep and Cool’n’Quiet) enabled during these tests via Windows Vista’s “Balanced” power options profile. The exception here was the Skulltrail system, since its BIOS didn’t support SpeedStep.
Anyhow, here are the results:
Let’s slice up the data in various ways in order to better understand them. We’ll start with a look at idle power, taken from the trailing edge of our test period, after all CPUs have completed the render.
The new Phenoms’ idle power use is in line with their predecessors’, but not quite a nice as the quad-core Intel systems’, overall.
Next, we can look at peak power draw by taking an average from the ten-second span from 30 to 40 seconds into our test period, during which the processors were rendering.
Under load, our test systems based on the new Phenoms consume just about as much power as those based on Intel’s fastest 65nm quad-core CPUs. That’s not too bad, but Intel’s 45nm processors draw quite a bit less power than their 65nm counterparts.
Another way to gauge power efficiency is to look at total energy use over our time span. This method takes into account power use both during the render and during the idle time. We can express the result in terms of watt-seconds, also known as joules.
The Phenom systems’ combination of idle and peak power draw ends up requiring more total power than those based on the Core 2 Quad Q6600 and Core 2 Duo E8500.
We can quantify efficiency even better by considering the amount of energy used to render the scene. Since the different systems completed the render at different speeds, we’ve isolated the render period for each system. We’ve then computed the amount of energy used by each system to render the scene. This method should account for both power use and, to some degree, performance, because shorter render times may lead to less energy consumption.
AMD hasn’t quite caught up with Intel’s 65nm quad-core processors in terms of power-efficient performance. Thus, they have quite a bit of work to do in order to match Intel’s 45nm parts.
We’ve haven’t had much luck with Phenom overclocking in the past, but our experiences with the 9850 Black Edition were more pleasant. I was able to get it running stable at 3GHz using some extra voltage and the same Cooler Master Hyper 212 cooler we used with the Phenom 9600 Black Edition. Rather than tell you a long tale of what I ran into during my overclocking attempts, I’ll give you a look at my notes, which look like so:
2.7GHz/stock – pass
2.8GHz/stock – pass
2.9GHz/stock – no POST
2.9GHz/1.36V – pass
3.0GHz/1.36V – hang on Windows boot
3.0GHz/1.403V – BSOD on boot
3.0GHz/1.442V – BSOD on boot
3.0GHz/1.481V – pass
3.1GHz/1.481V – hang on boot
3.1GHz/1.519V – hang on boot
A “pass” means the CPU made it through 4-5 minutes of the stability test in the AMD Overdrive utility. Getting to 2.9GHz was very easy; the 9850 needed only a minor bump in voltage to make it there. Hitting the 3GHz mark took more effort and a lot more voltage. In fact, I finally settled on 1.519V at 3GHz for my testing, and even then, the system wasn’t perfectly stable. I’d say this is a 2.9GHz chip for most intents and purposes. Still, I was able to take a screenshot and do some testing at 3GHz.
The CPU-Z voltage readout lies! It lies!
At 3GHz, the Phenom looks primed to take on the QX6800, at least in these two apps. I think I might have been able to extract some additional performance out of the Phenom by overclocking its north bridge (and thus its L3 cache), but sadly, our MSI K9A2 Platinum motherboard’s BIOS doesn’t allow for that. We may have to try it on another board soon.
I’m more than curious to find out whether my experiences here were in any way representative of what most folks will experience with the 9850. Who knows?
The best thing I can say about the Phenom X4 9750 and 9850 is this: AMD is back in the game. The banishment of the TLB erratum to the history books is, of course, a welcome development, and the higher clock frequencies now available are a small but important step in the right direction. The Phenom X4 9850 Black Edition can’t always keep pace with the Core 2 Quad Q6600 or the Core 2 Duo E8500, but it’s close. AMD still hasn’t caught up to Intel’s 65nm “Kentsfield” processors in terms of overall performance or power efficiency, yet it has produced a credible alternative to those products. That fact, combined with aggressive pricing and the bold move of offering an unlocked upper multiplier on a $235 quad-core processor, has enabled the Phenom to grab our attention. We can finally say with confidence that if you have an existing Socket AM2 system and want to upgrade, buying a Phenom looks like a more attractive upgrade path than making the switch to Intel.
Around here, we tend to think in terms of guide-worthiness, in reference to our helpful system guides with specs for building a new PC. I think we can now say that the Phenom is also a guide-worthy proposition, if not as a primary recommendation, then at least as an alternate. The MSI K9A2 Platinum motherboard we used in our Phenom test rig sells for as little as $159 online and would make a nice foundation for a Phenom systemwe’re talking about a board with four second-gen PCIe x16 slots, a board capable of three- and four-way CrossFire X configs. AMD’s so-called “Spider” platform has had an incredibly rough childhood, but it’s beginning to mature into a reasonable proposition, believe it or notnot the fastest or the best in any sense, but a potentially solid value.
That, for now, is enough to get AMD back onto the PC enthusiast’s radar and into contention in the middle of the market, thanks in part to the window opened up by the slow ramp of Intel’s 45nm chips. The next steps for AMD may prove to be considerably harder, but they’ve at least established a toehold again.