AMD's K10: a "dead" product or not?
by Johan De Gelas on May 12, 2008 12:00 AM EST- Posted in
- IT Computing general
A few years ago it was fashionable to bash Intel's Pentium 4 as a braindead architecture. The fact that the Pentium 4 Northwood (533 MHz FSB) was the best performing processor from mid 2002 until late 2003 in many applications, and that the Pentium 4 Northwood remained competitive until early 2004 was conveniently forgotten: nuances do not make good headlines.
It is now trendy to bash AMD. One" PC doctor" at ZDNet goes as far to say that:
"When I look at AMD’s current product line, all I see is a forest of
deadness. Intel has products trump every category of products
going. Server, desktop, mobile, low-end, high-end, dual-core,
quad-core. Intel has all these markets stitched up."
Nuances, who needs them when you can make a sensational headline? And indeed, the lastest desktop CPU articles here at Anandtech show that Intel's midrange CPU have a significant lead over the fastest Phenom processors.
Like any design, the K10 is a trade-off. And most trade-offs were made in favor of the applications in the server and HPC market, at the expense of games and other desktop applications.
First take a look at this page which compares a Core 2 Duo 4400 (2 GHz, 2 MB L2 and 800 MHz FSB) with a slower 1.86 GHz Core 2 Duo E6320 (4 MB of L2 and a 1066 MHz FSB). One thing is for sure: games prefer the larger L2 cache. Some of the games were up to 10% faster on the CPU which was clocked 7% lower but with twice the L2-cache. The fact that games prefer a 4 MB L2 is not going to change when you run it on a AMD CPU with integrated memory controller. A L2 can deliver the necessary data in 12-20 cycles, an IMC needs about 100 cycles.
Now, take a look at the Cache architecture of AMD's K10/Barcelona. If your run a single threaded game on it, it gets a fast 512 KB L2-cache and after that a relatively slow (44-48 cycles!) 2MB L3. If you know that the same game can benefit from more than 2 MB cache, it is pretty clear that the 512 KB L2 is not going to cope, you'll end up using the L3 a lot. A dual threaded game might need a little less per thread, but the same problem will happen again: it needs to go to that slow L3 cache all too often. Run that same game on Intel Core CPU and each thread of your dual threaded game gets a low latency 4 MB (or 6 MB) L2.
Now let us now imagine that we run 4 threads of an HPC workload on it. Each thread has a very limited number of instructions, which perfectly fit in each of the L2 caches. You get 4 threads which gets a total of 4x the bandwidth of L2. In case of Intel, each two threads have to share the available bandwidth of the L2. The amount of data is huge, so caching the data is hardly possible. The fast IMC does wonders for the K10 chip.Data that is shared between the 4 cores remains in the L3-cache and all L2 caches are kept coherent at a incredibly fast SRI. So your cache coherency overhead does not increase with the number of caches, it increases per socket. Going from 2 to 4 sockets means that you double the amount of cache coherency traffic. Compare that to the Intel platform where all L2 caches need to be kept coherent.
It is just one example why we could never expect the K10 chip to be a super desktop chip. But how is Barcelona doing in the server world? Is it limited to an HPC niche market? Well, let us see what Intel thinks. First of all, where do most of the 45 nm chips go? Just a few weeks ago, Anand reported that Intel had no intention of flooding the desktop with 45 nm Core 2 chips quickly.
Those 45 nm chips are going to the server market. Why? Several reasons.
First of all, the server market might be only 20% of Intel's revenue. But look at this:
CPU
ASP
Profit margin (estimate)
Percentage of revenue Intel Server CPU >$400 >$300
+/- 20%
AMD Server CPU
$300-$400
$220-$330
+/- 16%
Intel Mobile/Desktop CPU
$100
$40-$50
+/- 80%
AMD Mobile/Desktop CPU
$50-65
$5-$30
>80%
Secondly, Intel needs those 45 nm to be competitive in the HPC market. A 2 GHz Barcelona is capable of keeping up with the best 65 nm Xeons in those applications.
It is pretty clear why AMD focused on the server market. Without a complete redesign it is not possible to beat Intel's integer crunching power and the fast and big L2-cache and that is exactly what a modern game needs. Barcelona built further on the K8 architecture and inherited the relatively inflexible integer pipeline. While Core 2 has sophisticated reordering of loads and stores, Barcelona does a limited reordering of loads. While Core 2 offers a 32 entry queue to the integer units, Barcelona has 3 rather inflexible separated 8 entry queues.
So the right way forward for AMD was to focus on HPC and server applications where it could leverage it's strong points. We can bash AMD for being so late, and coming up with relatively low clocked CPUs, but even a 2.8 GHz Phenom would not have raise AMD's ASP significantly in the desktop market.
We are almost done with our first round of quad socket benchmarking and we can tell you that we are having a lot more fun than Anand: it is a good old exciting fight between AMD and Intel. Don't believe us? Let Intel do the talking again:
Yes, projecting the bad performance of the desktop chip to say that "AMD's products are a dead forest" is ... just silly. If you have missed the previous entries of our IT blog, just go to it.anandtech.com
Facebook
Twitter
RSS
74 Comments
View All Comments
Itany - Tuesday, May 13, 2008 - link
What innovations of Intel in the past two years?What about Penryn? A 3GHz CPU may fit into a notebook?
What about quad core processor?
What about ATOM?
It's true that Intel did not refresh the uArch for two years, hower, at least AMD just can't outperform the "old" Q6600 on desktop, with so many "innovations" of K10. All that AMD is good at inherites from the IMC and HT bus of K8. K10? Innovations? hehe...
JPForums - Tuesday, May 13, 2008 - link
Intel's Core2 architecture is based on the Core and thus P6 architecture, so what wrong with AMD leveraging the strengths of their previous architectures. If it isn't broke, don't fix it. (Not saying they can't fix other areas of the processor)Atom is somewhat innovative, but is mostly based on older designs using newer processes with an extremely low power overall design goal. It's a very good design, but the innovation is more in creating a market where one didn't really exist before.
Intel didn't innovate quad core, they innovated double dual core. This was really an innovation that they made with the double core Pentium Ds. While not a long term performance enhancing innovation, it should not be shrugged off either. It allowed Intel to offer two cores when they couldn't have otherwise. They also maintained higher margins this way as the yields were much higher than they would have been if they tried to get two cores on a single die. And guess what, AMD is following suit.
Penryn wasn't all that innovative as an architecture. However, the new process is a huge innovation. It's not a simple refinement in lithography technology. They had to change the way transistors were made. You can only deny that it is innovative if you call it inventive.
To be fair, AMD has been innovative as well. AMD's HT interconnect technology is routed like a crossbar switch. Effectively, this means that core 0 and core 2 can talk while core 1 and core 3 talk (or any other mutually exclusive combo) with no penalties. I don't know what Intels variation on HT will do, but if they design it like time division multiplexing (TDM) switches, they will have additional latency in some cases. More importantly, the processors will get extremely hot if they run the link fast enough to allow all processor simultaneous communications using TDM.
The integrated memory controller is innovative in the sense that it is a progression towards a system on a chip. Bulldozer is another innovative step in this direction.
As for K8 vs K10, there is a nice excerpt on the K10s shared cache in the article. They wouldn't be matched with Intel if they hadn't improved in some way.
Both companies have been quite innovative. You just have to keep in mind that some innovations are for reasons other that higher frame rates in the latest shooters (I've mentioned nothing of the energy saving innovations deployed by both companies). Also, some innovations don't pan out the way they were intended. Netburst is a good example of that. It was a radical redesign and extremely innovative. It also allowed Intel to hit higher clock speeds that they ever had before nor have since. Unfortunately, like the shared L3 cache on Phenom, this doesn't always translate into the highest performance.
7Enigma - Monday, May 12, 2008 - link
Intel /= crap, fanboy = crap.Sorry had to fix the formula for you. I'm sorry but your reply is currently the worst of the bunch. It's sad to see someone so hypocritical in their own post. Do us all a favor fanboys, just put the keyboard down, you only hurt your own company. So sad, so sad....
If you buy anything other than the best performance for the price you are a monkey and a slave to a company. Use your noodle...
Ensoph42 - Tuesday, May 13, 2008 - link
To be honest I don't think you're post is any better than the one above yours.Keep in mind it's absolutely necessary for AMD to stay afloat, unless you want to be be paying $500 for Intels low end chips and have performance improvements in a 5 year time frame rather than 2. So if a few fanboys want to throw their money at AMD, more power to them.
And, although being an AMD fanboy myself, I'm not unrealistic about the situation. But as long as they are able to supply me with the performance I want at a price I'm willing to pay I'll toss them my money. Since I don't O/C, do play my games at resolutions significantly highter than 1024x768, don't sit around converting DVDs to Divx and compressing 1GB into Rar files I find Phenom's performance to be more than adequate. I've never had a top of the line system so the company that has the performance crown means little to me.
I'd also like to say that I've seen many reviews with questionable test setups (DDR2-800 on a phenom? why?) or data represented in such a way to be misleading (bar graphs scaled to show intel having some massive lead when really it's more like 1 or 2 percent.)
Every system I've had for about the last 12 years has been AMD. They've been sturdy, stable, and solid and I've not regretting owning any of them. AMD has supplied me with positive product experiences and I would like that to continue.
varneraa - Tuesday, May 13, 2008 - link
I'm not sure AMD sinking would creste an Intel monopoly, they could easily fumble the next processor or two and be bought out by someone big(Samsung for example) who would take over the competition with Intel. I'm sure AMD knows that they won't survive simply on fanboys and their CURRENT server performance, and Intel's not going to give them a break. If they don't execute better in the future then someone else will step up to take their place.Locutus465 - Tuesday, May 13, 2008 - link
Agreed, I've been using AMD for 10 years and have never really been disappointed. I did consider a core 2 system this time around, but at the end of the day AMD's platform inititives got me... Spider is exciting and AMD Live! seems to be doing better than Vive.My Phenom quad core seems to have more than enough horse power to handle anything I could possibly throw at it, including compressing/decompressing 3GB text files, breaking down said files and converting formats from pipe delimted to valid CSV, any kind of game I could want to play and having media center burn my HD recordings to DVD.
I'm very happy with my upgrade, I went from an Athlon 64x2 with 2GB DDR400 memory and Gefore 7800GT to a Phenom x4, AMD/ATI Radeon 3870 and 4GB of DDR2-800. To be perfectly honest while I was expecting game performace to increase a (good) bit I wasn't expecting over all system performace to increase as dramatically as it did. AMD still has good stuff, yes intel currently has better CPU's but amazingly enough they're lacking in the platform arena.... Ironically exactly where the P4 was better than the Athlons.
antifanboys - Sunday, May 18, 2008 - link
Innovations aside, Intel right now still offer the best bang for the buck. Performance is what matters and they are delivering it at the moment. Don't get me wrong, I love AMD and ATI, I would love nothing more than to see them bring something better to the table. Better products from all sides will help make things cheaper for everyone.jamori - Monday, May 12, 2008 - link
AMD does not consider the Phenom to be their "K10" core. Phenom is merely a revision of their previous K9 core. Bulldozer is the successor architecture to the K9; Phenom is simply a tide-over (granted, they hoped it would have good performance and clockspeeds, but it turned out not to)It would be more appropriate (but still wrong) to refer to Bulldozer as K10 -- AFAIK, AMD insists that there's no such thing as K10.
Visual - Tuesday, May 13, 2008 - link
There is no K9. Phenom is K10.Remember, this is not as simple as counting. It is Marketeering ;)
RamarC - Monday, May 12, 2008 - link
"Just a few weeks ago, Anand reported that Intel had no intention of flooding the desktop with 45 nm Core 2 chips very soon."according to the consumer desktop chart, 45nm desktop chips should make up half of the total chip volume in Q3 '07. Q3 starts in less than 3 weeks, so the chart seems to contradict the statement.
could it be that intel's production capacity is ramping up allowing them to provide as many chips as either the desktop or server supply chain needs?