Socket-AM2 Performance Preview
Without major architectural changes to the new AM2 CPUs, we wanted a quick and easy way to showcase the performance differences between AM2 and Socket-939. What we've got is a massive table below with all of our usual CPU benchmarks and their results for the same CPU in both Socket-939 and AM2 varieties, and the performance benefit offered by AM2:
| Benchmark | Socket-939 (DDR-400) | Socket-AM2 (DDR2-800) | % Advantage (Socket-AM2) |
| PC WorldBench 5 | 115 | 115 | 0% |
| Business Winstone 2004 | 23.3 | 23.2 | -0.4% |
| Multimedia Winstone 2004 | 38.4 | 38.9 | 1.3% |
| SYSMark 2004 | 220 | 224 | 1.8% |
| ICC SYSMark 2004 | 282 | 286 | 1.4% |
| OP SYSMark 2004 | 171 | 175 | 2.3% |
| 3dsmax 7 | 2.38 | 2.38 | 0% |
| Adobe Premier Pro 1.5 (Export w/ Adobe Media Encoder) | 130 s | 128 s | 1.5% |
| Adobe Photoshop CS2 | 210.6 s | 210.3 s | 0.1% |
| DivX 6.1 | 11.6 fps | 12.0 fps | 3.4% |
| WME9 | 35.2 fps | 35.6 fps | 1.1% |
| Quicktime 7.0.4 (H.264) | 3.63 min | 3.63 min | 0% |
| iTunes 6.0.1.4 (MP3) | 43 s | 43 s | 0% |
| Quake 4 - 10x7 (SMP) | 111.3 fps | 117.4 fps | 5.5% |
| Call of Duty 2 - 10x7 | 59.3 fps | 60.1 fps | 1.3% |
| F.E.A.R. - 10x7 | 92 fps | 94 fps | 2.1% |
| Multitasking Test (LAME + WME + Anti Virus + Zip) | 216.3 s | 213.4 s | 1.4% |
| ScienceMark 2.0 (Bandwidth) | 5007 MB/s | 6805 MB/s | 36% |
| ScienceMark 2.0 (Latency 512-byte stride) | 53.83 ns | 49.77 ns | 7.5% |
We'll start at the bottom of the table and go up from there. Rev F processors feature a 128-bit DDR2-800 memory controller, which works out to offer a peak theoretical bandwidth to/from memory of 12.8GB/s. As you can expect, that's twice the bandwidth of Rev E CPUs' 128-bit DDR-400 controller at 6.4GB/s. Thus to see a 36% increase in memory bandwidth according to ScienceMark is to be expected, albeit a bit on the low side. The old DDR-400 memory controller is able to deliver 5GB/s out of a maximum of 6.4GB/s, but now we're only seeing 6.8GB/s out of a maximum of 12.8GB/s with AM2. This however is a huge step for AMD, as it is the first spin of the Rev F silicon that we've been able to see such a significant advantage in theoretical memory bandwidth over previous DDR-400 cores.
What's even more important than the increase in memory bandwidth is that access latency has been reduced by 7.5% over the DDR-400 memory controller in the Rev E cores. Lower latency and more bandwidth means that, at bare minimum, performance won't go down. At least, not perceptibly: .4% slower in one test that has a 1-2% variability is nothing to worry about.
It also doesn't guaranee that performance will go up, as you can see from the results above. If we only count the overall SYSMark score and leave out the synthetic tests, the real world performance advantage averages out to a little under 1.3%. There are some special cases such as Quake 4 and DivX were performance goes up fairly reasonably, which can be expected since both of those tasks are fairly bandwidth intensive and make good use of both cores. However similar benchmarks, such as F.E.A.R. and Windows Media Encoder 9 show lower improvements, so it is very dependent on the specific application and workload.
It's important to note that until recently, AM2 samples were not able to produce scores even on par with Socket-939, so the fact that we're seeing a performance increase at all is a major step from where we were just a couple of months ago. The real question is, is this all we get?
Facebook
Twitter
RSS
107 Comments
View All Comments
savantu - Monday, April 10, 2006 - link
Anand , please correct the values in the table at the Adobe and multitasking tests , they should be negative.savantu - Monday, April 10, 2006 - link
..It's time so less is better.Anemone - Monday, April 10, 2006 - link
Don't forget capacity. If you loaded up 4 memory slots to bring 939 machines to 4gb, you harmed timings. DDR2 can bring 2gb on a module at good timings and thus make a higher performance 4gb machine. Why is this important? In Vista (yep I know it's delayed) you will find that gamers and power users of a wide variety will want 4gb machines, much the same as a good number of folks are starting to use 2gb machines now. The larger footprint of Vista ups the ante. It's not that 2gb won't be enough, it will do. Just as 1gb today "will do". But come the end of this year 4gb will start to be that cutting edge amount and that suits the use if DDR2 very well.You already mentioned the power issue, with 1.8V for DDR2. That is something AMD needed in the mobile arena to stay competitive, so if they were already designing a top notch DDR2 controller, might as well do the entire cpu line. Since much of the market is actually starting to shift to mobile solutions, from a growth standpoint, being competitive here is going to be telling in each company's numbers. Since Merom is having heat issues (hence why Conroe is coming out so far ahead of Merom, meaning less of a heat budget constraint) you can expect that Turion X2's with DDR2 will put some pain on Yonah machines. I suspect that AMD knows their bandwidth is superb with DDR2 and are designing Turion X2's with 512 cache because it doesn't hurt them much. Watch this area in 2006, because the world won't be painted blue if AMD does well here. Intel is likely well aware and will push as hard as possible to bring Merom out to keep AMD from making ground. Mobile designs under S1 socket are coming aplenty.
$.02
coldpower27 - Monday, April 10, 2006 - link
I must have missed it by what clock rate and what amount of cach were the models used???
Anand Lal Shimpi - Monday, April 10, 2006 - link
We purposefully didn't publish this information to protect our source of the CPU.Take care,
Anand
flyck - Monday, April 10, 2006 - link
well ofcourse we don't know the clock. But it is an important factor. @ 1.8GHz A64 didn't bennifit from dual channel. but it think @ 2.8GHz that story was totally different. So if you are testing @ 2.4GHz. there could be a larger gap @ 3GHz for example.Anand Lal Shimpi - Monday, April 10, 2006 - link
Very true, but from what I've seen the picture doesn't change much from the 3800+ up to the FX-62, at least with current CPUs/platforms. The main thing that shows a performance difference is when going to even lower latency memory. You are right though, the hungrier the CPU gets (faster clock, wider core, etc...) the more it depends on a high bandwidth memory bus. However, I do believe that AMD's own model numbers tell the tale of what to expect.I think that at the end of the day the 2 - 7% increase range is what will hold, with the vast majority of applications falling at the 2% end of the spectrum.
Take care,
Anand
ozzimark - Monday, April 10, 2006 - link
though one thing to keep in mind: memory efficiency is directly related to the cpu mhz. i've found that going from 2ghz to 3ghz while keeping the ram at 250mhz increases bandwidth by around 1800mb/s on my s939 rig :eek:Sunrise089 - Monday, April 10, 2006 - link
I think there will be a lot of current users taking the upgrade path in my subject line, or something close to it. I may very well buy a new AMD chip in 2007, I'd sure like to be able to, but for the next 12 months I only see myself buying a used, overclock proven 2x1meg cache s939 X2 at the end of summer, and then trading up to Conroe during the holidays. I can't really see anyone but the strongest fanboys (and I'm pretty strong) buying AM2.AnnonymousCoward - Monday, April 10, 2006 - link
Why in the world do you upgrade so often? My path might be AthlonXP 1800+ --> Conroe.