One issue I'm having is the weird color scheme in your graphs. Just make all AMDs green and all Intel's blue. Stop mixing and matching randomly. It makes the colors useless and misleading.
I only highlighted the new chips we were focusing on. I debated doing it the other way (AMD green, Intel blue) but figured the focus should be on the new chips. I can understand the confusion though. I've updated the graphs to reflect green for AMD and blue for Intel, if everyone is ok with it I'll keep it :)
This sounds good to me - just don't make the contrast too different that we have to look at 4 different colors. The greens should be close enough in spectrum that they can be distinguished but close to the same hue (same for the blue).
That way looking at a glance your brain can quickly compare overall AMD vs Intel, but then giving it more consideration you can tell what's new/old.
I've updated the graphs to use the dark/light colors, though I'm not sure how "new" some of the Intel parts are. Anyway, at least there's a bit of separation to make things "visible".
Have you guys ever thought about placing the prices of the chips in the benchmark tables for easy price reference? Perhaps just to the right of the benchmarks?
Just a suggestion!
Thanks for this update as well, AMD is starting to regain my interest.
A few price to performance graphs would be nice. You know, "sysmarks per dollar", or that sort of thing. It would help identify the sweet spot in processor reviews.
I've been thinking about doing it for a while, it looks like there's overwhelming desire for it so I'll begin working on the best way to put it together :)
I'm not sure "Sysmarks/$" is all that more useful than a general recommendation that the results reveal quite clearly. In this article for example, the Athlon X3 is a stellar value while the Athlon X4 and i5 quad cores are also very good.
However, this has been common knowledge for over a year now, so are we really getting anything we didn't already know from a "Sysmarks/$"-type of graph?
I haven't been in the market for a new processor/chipset for over a year now, so price/performance is something that would be very helpful for me right now with all these newer chips since Core2 and Phenom I. So I concur with the OP, I would love to see some price references in or near some graphs.
If you do that, please integrate the differences in power consumption somehow. It would be silly to save $20 by buying an AMD processor if it costs $50 more to run over its lifetime.
How long would that lifetime be? 1 year? 4 years? Is the machine on all the time but idling 22 hrs a day? Is it gamed on 10 hrs a day but off the rest?
I'm sure it's fairly easy to come up with a couple of realistic scenarios to use. There are only really three variables - # of hours idle over lifetime, # of hours load over lifetime, and average cost of power over lifetime - where it should be fairly easy to come up with some numbers to give a good idea. A useful metric might be to say "given a cost of electricty of X, the breakeven point between processors A and B is after Y hours of use".
The whole point of benchmarks isn't to say "here's how it will perform for everyone!", but to give readers an idea of how the product performs in some specific (and hopefully realistic) situations, thereby allowing them to make informed decisions based on their own needs.
I ditto that! Price is such an important factor it should be part of the graphs or do up a separate set of graphs with performance per $ per application, that'd be awesome. I know that ill be quickly outdated but just put date next to the graph and that'll make it very obvious to take that into consideration when looking at the graph.
The sweet spot is the AMD lineup seems to be the Phenom II X4 955BE. For $ 145, you get a fully enabled 3.2 GHz quad-core with L3 cache, that you can easily set yourself @ 3.6 GHz with little or no voltage increase.
Yet it will still lose to the i5 750 in virtually every benchmark. The real "sweet" spot is being able to get an i5 750 for a good price. I've seen them on sale at some places for as low as $170-175. At that price you'd be crazy to choose a Phenom II X4 955BE over it.
Depends if all you have to do is just replace whatever processor you have in your socket, atm.
In AMD's case, you can just swap out the processor on most AM2+ motherboards and it'll just work. On Intel's case, chances are you'll have to get a new motherboard (and quite possibly RAM, if your system is old enough).
Been thinking of getting one of those 955BEs for myself. Not much of an upgrade to my main machine that's running a 940BE, but then I can pass on my older processor to my second box which is running an X2-6000 of ancient times.
That's the big selling point for me. I'm willing to spend an extra $50 to save even 10 watts because over the life of the computer I will get that money back in electricity savings, maybe even more so.
Part of it is probably because the older processor were tested with older mobos/BIOSes/chipsets. We can ask Anand to confirm my speculations. I just do not think that they have rerun old CPUs for this review and have used numbers from the original reviews.
Anand, I do not think you gave this CPU justice. It performs similarly to the Phenom x2 555 and it is at least $20 cheaper. It uses a lot less energy than the Phenom and in a office environment would be more responsive than the Athlon II X4 and X3 due to the higher clocks. I hate looking at the Task Manager, because most of the time one of my cores is loaded almost 100% and the other ones are near idle. I just think that the extra L2 and higher clocks make the Athlon II X2 265 better CPU for most of the people most of the time and the benches would never be able to display that.
This highlights what the real problem is for the PC industry today that they don't want you to know.
Do you know why Anand doesn't benchmark much more common tasks like Windows 7 boot time or MS Word launch time? It's practically the same across the board on from $65 CPUs to $999 CPUs! Unless you play new games or use professional artist software, you have almost no reason to upgrade a PC that is a few years old. Power savings yes, but not enough over the life of a computer to justify spending hundreds of dollars.
Application launch time is actually reasonably CPU dependent if you've got a SSD. While I agree that common tasks don't vary in performance much between similarly clocked processors, I'd say the gap between generations is very visible. Many have said that Core 2 was fast enough but I'd argue there's a noticeable difference in just general PC usage between Lynnfield and Penryn simply because of the benefits you get from high turbo frequencies (windows pop up faster, there's a tangible reduction in response time). That gap will grow with each subsequent generation.
The problem with these subjective performance tests is they are difficult to quantify.
I agree that unless you have a specific application that is heavily influenced by CPU performance you don't need to upgrade every generation. But that's also why the PC upgrade cycle is typically 3 - 5 years. Original Core 2 owners will probably want to upgrade next year, while Lynnfield or AM3 owners can definitely wait.
My upgrade will probably come sometime between 2011Q2 and 2012Q3.
Note: I don't typically upgrade parts unless they fail, so the mobo/cpu/gpu combo are generally the same as the initial build. Holding onto them for 5-6 years is terrific.
I've got the same CPU and was thinking of upgrading to SB or BD next year. My problem is that my old machine is still perfectly functional, I just wish it'd load SC2 or run through x264 a little quicker.
What do you with this older perfectly functional hardware? Does anyone have any good ideas for how to re-task it? I've already got a similar spec'd HTPC along with a pre-ordered Boxee Box and an old XP2600 running the file server.
1) donate it to friend or family - I still have parents and none are tech savy or willing to put money into it, so that's the most likely option
2) keep it offline for spare parts or backup in case of disaster (laptop is probably good enough though)
3) really look into setting up a home server, possibly get a dedicated IP and host some files for family/friends. It's not the most powerful hardware, but it's good enough for a small number of people.
4) try to throttle down and turn it into an HTPC
5) take all laptops and old cpus and play with Hadoop/MapReduce at home
I would really like to see a "Bang for your Buck" graph. In many test, AMD comes close to the i7, but the i7 cost more. So when you do a performance per dollar ratio, AMD may come out on par, or maybe even ahead in some cases.
I also find it interesting that some of the Phenom II x4's have a lower power consumption at idle than the X2's. I realize the X2's use the same silicon, but with the cores turned off, shouldn't they use less?
You write "Bulldozer is out of the question however, AM3+ chips aren’t backwards compatible with AM3 motherboards (although the opposite is true, you will be able to use your 970 in an AM3+ motherboard)."
Is this the definitive truth? :-)
I have read a lot of people guessing whether AMD will have AM3 compatibile Bulldozer modes available, or only AM3+ (AM3Rev...).
This information came directly from AMD :) Not saying it's 100%, just saying it's the source.
You'll see AM3+ motherboards in advance of the Bulldozer client launch so you'll be able to buy a new board for your CPU and then later drop in a Bulldozer.
"AMD also told us that it will introduce a new AM3+ socket for consumer versions of Bulldozer CPUs. AM2 and AM3 processors will work in the AM3+ socket, but Bulldozer chips will not work in non-AM3+ motherboards."
"For the desktop, the Zambezi processor is good news and bad news. The good news is it's an 8 core product, the bad news is it needs a new socket - AM3r, or AM3+. This is an electrical upgrade of the AM3 platform, to provide the power phases and planes/states required by the power gating features of Zambezi. As you might have guessed from the name, this socket is backwards compatible with existing AM3 processors, so you'll be able to piecemeal your upgrade - motherboard for your birthday, CPU for Christmas, or however your upgrade cycle works."
"When we initially set out on the path to Bulldozer we were hoping for AM3 compatibility, but further along the process we realized that we had a choice to make based on some of the features that we wanted to bring with Bulldozer. We could either provide AM3 support and lose some of the capabilities of the new Bulldozer architecture or, we could choose the AM3+ socket which would allow the Bulldozer-base Zambezi to have greater performance and capability.
The majority of the computer buying public will not upgrade their processors, but enthusiasts do. When we did the analysis it was clear that the customers who were most likely to upgrade an AM3 motherboard to a Bulldozer would want the features and capability that would only be delivered in the new AM3+ sockets. A classic Catch-22.
Why not do both you ask? Just make a second model that only works in AM3? First, because that would greatly increase the cost and infrastructure of bringing the product to market, which would drive up the cost of the product (for both AMD and its partners). Secondly, adding an additional product would double the time involved in many of the development steps.
So in the end, delivering an AM3 capability would bring you a less featured product that was more expensive and later to market. Instead we chose the path of the AM3+ socket, which is a path that we hope will bring you a better priced product, with greater performance and more features - on time.
When we looked at the market for AM3 upgrades, it was clear that the folks most interested in an AM3-based product were the enthusiasts. This is one set of customers that we know are not willing to settle for second best when it comes to performance, so we definitely needed to ensure that our new architecture would meet their demanding needs, for both high performance and overclockability. We believe they will see that in AM3+."
<i>But Socket AM3+ motherboards will be backward compatible with existing AM3 processors.</i>
This was not true from AM2+ to AM3, and I don't believe for a second this is true across the board from AM3 to AM3+. Many motherboard manufacturers (even top tier ones) will not bother with BIOS updates to older boards that went EOL some time ago.
If your board has a 7xx or Nvidia chipset and has been EOL for a year, I wouldn't count on anything. Better safe than sorry.
"This was not true from AM2+ to AM3, and I don't believe for a second this is true across the board from AM3 to AM3+. Many motherboard manufacturers (even top tier ones) will not bother with BIOS updates to older boards that went EOL some time ago."
From what I read, it is not the same situation before: Most AM2+ boards can run AM3 cpus, pending on BIOS support, AM2+ cpus will not work on AM3 boards, because they do not have a DDR3 memory controller. AM3+ (AM3r2) cpus will not work on AM3 boards, because the power gating logic on Bulldozer based cpus is not compatible with the power plane setup on AM3.
Basically, AM3 cpus are both backwards and forwards compatible with previous and upcoming sockets in their line.
Benchmarks that cover something more than common tasks (that low and mid range processors do adequately every day) would be helpful.
With regard to the competition issue...It's blatantly obvious where Intel sees AMD as competition. All the way up to the 4 core I7 series. Beyond that, no competition and ridiculous prices.
If not for AMD's efforts, you wouldn't have the I5's and I7s in the 200-300+ range right now. So, yes, they are providing strong competition in certain market segments.
Would you consider doing something of a Folding@Home benchmark, possibly running the SMP client for a day or something and seeing what kind of PPD figures you would get with these processors? Seems like F@H is really what I do most with my computers.
All these Athlon II and Phenom, are expected to complete against Sandy Bridge next year? ( Since Bulldozer are not coming to Fusion type and low end CPU )
SandyBridge is coming to kick some ass then by the looks of it. Higher IPC, Higher Clock speed, lower Idle power, better Hyper Threading......
Apart from Fusion i am not excited by AMD's roadmap at all.......
I believe Llano is supposed to compete with Sandy Bridge, but the problem is Llano is coming to market later than Sandy Bridge. Until Llano comes, yes these Athlon IIs and Phenom IIs will be forced to compete with Sandy Bridge, and they are going to lose badly.
Even when Llano comes, it's not supposed to be a big performance increase, so I think Sandy Bridge will rule the mainstream market.
As for Bulldozer, it will have to go up against Intel's socket 2011 Sandy Bridge models that are likely to be without integrated graphics. These models are rumored to have up to 8 physical cores, 40 lanes of PCI-E 3.0, and quad-channel DDR3.
You write: "The Phenom II X6 1075T is an interesting chip as you get a lot of compute but it's only useful in heavily threaded apps."
I strongly disagree, unlike you are calling virtualization "apps".
These chips are a godsend to developers, testers and sales who are running entire tiles of software. Need to test your client on XP, Vista and Windows 7? Need to run a web application server on Windows and an Oracle database backend on Linux? Need to do all of these at the same time? This is what a Hex Core chip is made for, not running a single heavily threaded app on a single O/S.
I'd like to see an AT Bench benchmark that covers this sort of application. I'd like to see how comparably priced Intel CPU's do versus AMD chips. I'm thinking this would be a huge differentiator between HyperThreading and actual cores, but I'd like to see the numbers before making any claims.
this is not quite comparable. on moderately IO dependant workloads (your web server/oracle db example), HT will do just fine. on compute dependant applications with little memory access, HT will bottleneck. However you will hardly run 4-6 compute dependant VMs at once every time.
BTW I mean "unless" not "unlike". Where's the edit button?
Can I get a reply from AT on VM? Can we stop treating every desktop chip as a gaming candidate-only chip and get some kind of desktop VM benchmark going here?
I'm thinking running one per VM: compile/link/install (push application from dev VM to web server VM over virtual network), web server with integrated JAVA application, three web clients, and a database backend. Run a mix of Linux & Windows, 32 and 64 bit O/S's. The app being compiled doesn't even have to be the app being run.
Yes I know they can do it, but we have no data on how similarly priced desktop chips compete. HTT versus Cores is only one aspect. I want to see the AMD hex chips stretched against Intel quad cores in VM's and shake out the bang for buck. AT's attitude about hex cores only being useful for single heavily threaded apps just doesn't match how this chip is being marketed or how it is being used in the field.
AT should realize that a lot of their PC gaming readers have daytime jobs not just as server administrators, but as software developers and testers. Management comes to us for advice and no one is providing us with the data to make informed decisions. 7Zip benchmarks are fine but give us a software development lifecycle benchmark using VM's, give us data on what we are already working with for several years in real life (work). Give us a benchtop VM benchmark! Please!
Annoying to see that we still have no competition in the high end segment. It's still being dominated by intel which makes for higher prices for us, the consumers.
Just think about it for a minute - had not AMD acquired an almost complete DEC team around 1997, there would have been NO real competition and the prices would WAY above what Intel is charging now. Think VIA Nano on steroids as "competition".
Let's be realistic, shall we? It is simply not feasible for 2 competitors in such a complex field to be BOTH on top concurrently for extended periods of time (as has happened during the Slot A heydays). These situations are bound to be more of an abnormality than a standard.
Considering the HUGE gap in resources, we shall be very happy that AMD is still within 30% of Intel's TOP despite using an 12yrs old K7 architecture. A true testament to Alpha actually.
And, for killing IA64 on desktop alone, they deserve a Nobel prize! No contest.
Speaking of Alpha, I wonder where many former Alpha employees have gone? Oh right, they're now working at Intel on future Itanium processors.
Sure we were spoiled, if you will, years ago when AMD introduced the K7, Intel missed with the P4, and AMD introduced the K8.
Thing is, historically Intel was almost always on top in the consumer CPU market.
AMD did provide some real competition back then, but not so much anymore. Intel can price AMD out of existence right now if they wanted to, but Intel doesn't want to.
Unless Bulldozer is a gigantic performance leap forward, AMD will continue to struggle to provide competition and will only be "good enough" or barely catching up to Intel's latest.
Intel has real competition in non-consumer markets, and that is challenging Intel.
I love such thorough and insightful answers such as these.
People talk all the time about how competition is good. Real competition is good yes, mediocre or weak competition is almost pointless. Right now AMD is mediocre/weak competition from Intel.
You can call people names, point fingers, and complain about it all day long, but it is what it is.
I have nothing against seeing some STRONG competition for Intel, but AMD currently isn't providing strong competition.
Really? Then why did I decide to buy a 1075T instead of an Intel chip in the same price range? For my needs (all of my single threaded apps are quite fast enough with any modern processor, but as a software developer I need as many cores as I can get for parallel compiles) the 1075T is a great deal. Throw in the fact that AMD motherboards are better than Intel at the same price point and it was an easy decision. I am slightly miffed that Bulldozer won't be AM3 compatible becase I was banking on that a little but when deciding to go with an AMD platform, but whatever. The rest of the value proposition is still in AMD's favor.
Now of course my experience is just anecdotal; I can say with certainty that in some segments, AMD is very competitive with Intel. I think the question of whether or not AMD provides "strong competition" with Intel is really just about how many segments AMD provides strong competition with Intel and how strong the competition is in those segments, and from that, you can decide whether or not the overall competition is "strong" or "weak".
Given the success AMD is having in netbooks/notebooks lately (at least all of the Fry's/Best Buy/Office Depot/whatever ads I have seen in the past few months have been heavily featuring AMD, I assume because the platform is selling well), I think we can say that AMD is providing strong competition in that space.
In the 'midrange' CPU market - which is the majority of the PC processor market - AMD clearly provides strong competition. See this Anandtech article, my purchasing decision above, countless similar articles and testimonies all over the internet.
On the high end PC CPU market, AMD is not competitive. We all know that.
On the server CPU front, AMD has not been competitive lately but is starting to gain a little ground back. At least, if my impressions given what I've read on Anandtech and other sites is anything to go by.
So in the end, I'd say that in 3 out of 4 x86 CPU markets, AMD is providing strong competition with Intel. For this reason I think it's erroneous to call AMD's current competitiveness 'weak'. I think it's overall 'fairly strong', and I only hope it gets stronger, for everyone's benefit.
Oh and I forgot to mention integrated graphics. Of course AMD is extremely competitive with Intel in that segment, and in fact is the dominating player (at least, performance-wise).
Yes they have problems when Conroe was launched in the high end. Most of this is due to the problems they encountered with Barcelona. Without those issues they would have been very competive with Core2. i7 lineup is now the strongest technology but AMD can compete ok using more cores. The biggest issue AMd is having is that most of their products had to be postponed. But with Ontario platform they will have their first real low power chip which is very promesing. llano will offer higher speed than x4 with very good GPU, which again is a pretty solid effort to gain the low and mainstream market due to embedded graphics that do not suck for desktop useage. Bulldozer, although to late, seems to be a chip on which AMD can build further upon. Performance is one area, the ability to price competitive is also an area and that is an area AMD is really competing with. Their future shows alot of promise (bit like in the K8 days where the XP couldn't cut it anymore). AMd is the only reason why Intel (lead x86 manufacturer) is actually making processor worth the money. This alone implies that AMD is competitive with Intel. Without the clausules and market abuse iItel did in teh years before, K7 and K8 could have damaged them alot more, they can't let that happen again.
I love competition. I want choices, not just at the lower end price points.
The real issue is that AMD has not improved their instructions per clock cycle ratio substantially since the earliest K8's. AMD caches have gotten larger, HT has gotten faster, power saving features have gotten more sophisticated, clock rates have gone up, but we've yet to see any real jump in core processing efficiency.
Until AMD addresses that they will stay on the low end, with the low end margins that come with that market.
$179.99 for the Phenom II 1075T seemed like a good deal to me so I picked one up. I still don't even have any of the other system components I need to run this thing; but the deal was so good that I just had to jump. I'll buy motherboard, memory, etc later ...
The weird thing is the AMD seal on the top of the box lists it as a 1075T and says its clock rate is 2.8 Ghz. Didn't notice this until I got home. I wrote to AMD to ask about it and they replied that they had some printing problems with the labels and some boxes went to vendors with that printing error. But they assured me that it's really a 3.0 Ghz part.
While I like Intel CPUs, it seems that good 1156/1366 mainboards are so expensive compared to Socket AM3.
It'll be awhile before I replace my Q9650, but when it happens, if Intel hasn't worked with vendors to make mainboards more reasonably-priced, I see AMD as a real possibility.
It's not board OEMs that are causing motherboards to be unreasonably-priced, it is Intel with their chipset pricing. Despite moving the IGP off of the chipset (a huge cost elimination), the wholesale price stayed the same.
3rd party chipset competition is needed desperately but Intel will have none of that! AMD boards definitely give you more for less.
Actually this sad fact is affecting the AMD market too.
It seems Nvidia has completely occupied the low-end market while AMD IGP boards have moved up on the price scale around $10.
Also check CPU prices - there is no AMD dual core below $60 with artificially sold single cores for $40.
Basically while the performance you can buy at $60 has gone up 2x over last 3 yrs on ADM side, their $40 offing is now useless while previously it was relatively reasonable. AMD sees no competition from Intel there => no low end dual cores.
For the majority of the market out there, you may be getting to a level where the performance of the CPU is good for what, embedded applications? Linux servers that can be run on used hardware for $20? Niche applications.
At least there is speed at the low price points to fulfill a multitude of uses.
I would think AMD's can't sell any CPU for less than about $65 without going deep into the red on every unit. Think about raw material costs, electricity, clean room maintenance, shipping, packaging, ... This is before accounting for any R&D or payroll for a single employee!
Similarly, I think we're at an absolute floor at about $40 right now for hard drives.
Once again, I believe the biggest release AMD could make is the platform. I ended up going with a Q6600 a few years ago when I would have loved to choose AM2+. I have been happy with the Q6600, but I spent a lot more on it than I really wanted to at the time. I could have put together a system using a CPU that cost $200 less, then upgraded one or two years later with a significantly faster CPU.
Fast forward to 2010 and the situation is quite similar. I would love to pick up an AM3+ based system with a Phenom II X4 and be ready for whatever AMD has coming. I would build that system right now, and AMD would be getting 2 CPU sales (unless things went horribly, HORRIBLY wrong with BD).
Is there any news about the 95W x6 cpu's, particularly the 1055? Apparently its on sale outside US, but I can't seem to find it from US based retailers. Thanks.
I read alot of benchmark reviews. "This benchmark favors Intel" is a line I read quite frequently, and I always assume it's because compiler writers and/or software developers consciously choose to optimize for Intel chips (which makes sense since it's the larger part of the market). "This benchmark favors AMD" is not something I think I've ever read, at least not in a context that led me to believe that it's due to specific optimizations targeted at the processor. Why would 3DMax 2010 perform relatively better on Intel than 3DMax 2005 did? Is it because the newer version has been better optimized for Intel?
In that case, is it more valid to use a newer version or an older version? I guess we want our benchmarks to reflect the non-benchmark software that will run on the platform, so I suppose that if most software is Intel-optimized, Intel-optimized benchmarks make sense. If not, then not.
I personally use Linux almost exclusively and I feel pretty confident that the GNU compiler toochain that is used for this operating system is not more optimized for one processor vendor than another - at least not intentionally, anyway.
In the for-money world of Microsoft and Intel with their backroom deals and shady business practices, however, I can't say for sure.
I think it would be more helpful if you gave percentages when overclocking in addition to ran clock numbers. It managed 4.0GHz does not mean as much as it achieved a stable overclock of 25% vs the other processor which managed a tiny 10%. Makes comparing things a lot easier.
I just bought the 95W version of the 1055T and it took a few months for it to be available. I put it in my ECS A780GM-A Black Edition Motherboard and it works for awhile then the screen goes blank.
Going to have to get a new motherboard. So much for backward compatibility with AM2+ socket.
Looking at these charts. The Core i5 750 consistently beats the X4 970 and they're priced to compete against each other. Not only does Intel win more, but they win in all the areas that really matter to me; not to mention the power consumption advantage. Secondly, where AMD actually does very well, encoding video, that advantage is negated by the fact that regardless of how good the CPU is a encoding and rendering video doing those things on the CPU is almost entirely obsolete. I know I do all that stuff using my Nvidia GPU and it's exponentially faster. In other words, AMD can't win anywhere at a similar price compared to intel. The ONLY reason I can see buying an AMD CPU and that price point right now is if the CPU is all your changing and you already have an AMD Motherboard. Unfortunetly for AMD if you're building a whole new system Intel is the way to go and every price point with decent profit margins.
Ivy Bridge is meant to be still on track for 2H 2011. This just seems crazy. Intel are not going to release a whole new generation in Sandy Bridge and then replace it within 6 months!??
I guess they might just release a few 22nm parts like they did with 32nm but even that seems so unlikely right now. I mean we won't even have high-end Sand Bridge parts out yet and we will have the replacement for Sand Bridge at our door step!
I am quite confused as to WHY Athlon II X3-450 with higher speed is actually cooler and more efficient than X3-440? Also, the same between X4-645 and X4-635?!
I thought they're all the same core, no?
If these are more efficient, are they ideal for HTPC stored in a cabinet? using Antec Fusion Case?
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
98 Comments
Back to Article
KikassAssassin - Tuesday, September 21, 2010 - link
On page 1: "The Core i3 540 is priced similarly but you only get two cores, and no Hyper Threading to bridge the gap."The Core i3 does have HyperThreading. The only Clarkdale CPU without HyperThreading is the Pentium G6950.
quiksilvr - Tuesday, September 21, 2010 - link
One issue I'm having is the weird color scheme in your graphs. Just make all AMDs green and all Intel's blue. Stop mixing and matching randomly. It makes the colors useless and misleading.Anand Lal Shimpi - Tuesday, September 21, 2010 - link
I only highlighted the new chips we were focusing on. I debated doing it the other way (AMD green, Intel blue) but figured the focus should be on the new chips. I can understand the confusion though. I've updated the graphs to reflect green for AMD and blue for Intel, if everyone is ok with it I'll keep it :)Take care,
Anand
quiksilvr - Tuesday, September 21, 2010 - link
Much appreciated! A suggestion could be this to differentiate between old and new:Light Blue for New Intels
Dark and less saturated Blue for Old Intels
Light Green for New AMDs
Dark and less saturated Green for Old AMDs
vol7ron - Tuesday, September 21, 2010 - link
This sounds good to me - just don't make the contrast too different that we have to look at 4 different colors. The greens should be close enough in spectrum that they can be distinguished but close to the same hue (same for the blue).That way looking at a glance your brain can quickly compare overall AMD vs Intel, but then giving it more consideration you can tell what's new/old.
KikassAssassin - Tuesday, September 21, 2010 - link
This sounds like the best option to me.JarredWalton - Tuesday, September 21, 2010 - link
I've updated the graphs to use the dark/light colors, though I'm not sure how "new" some of the Intel parts are. Anyway, at least there's a bit of separation to make things "visible".quiksilvr - Tuesday, September 21, 2010 - link
You guys are awesome! You're like, one of the busiest gadget sites on the web yet you took the time to read my suggestion and actually implement it!High five!
(turns off Adblock for Anandtech.com)
vol7ron - Wednesday, September 22, 2010 - link
i think the light green is a little too light, but much betterfoundchild1 - Tuesday, September 21, 2010 - link
Have you guys ever thought about placing the prices of the chips in the benchmark tables for easy price reference? Perhaps just to the right of the benchmarks?Just a suggestion!
Thanks for this update as well, AMD is starting to regain my interest.
Guspaz - Tuesday, September 21, 2010 - link
A few price to performance graphs would be nice. You know, "sysmarks per dollar", or that sort of thing. It would help identify the sweet spot in processor reviews.Anand Lal Shimpi - Tuesday, September 21, 2010 - link
I've been thinking about doing it for a while, it looks like there's overwhelming desire for it so I'll begin working on the best way to put it together :)Take care,
Anand
Taft12 - Tuesday, September 21, 2010 - link
I'm not sure "Sysmarks/$" is all that more useful than a general recommendation that the results reveal quite clearly. In this article for example, the Athlon X3 is a stellar value while the Athlon X4 and i5 quad cores are also very good.However, this has been common knowledge for over a year now, so are we really getting anything we didn't already know from a "Sysmarks/$"-type of graph?
RyuDeshi - Monday, September 27, 2010 - link
I haven't been in the market for a new processor/chipset for over a year now, so price/performance is something that would be very helpful for me right now with all these newer chips since Core2 and Phenom I. So I concur with the OP, I would love to see some price references in or near some graphs.marraco - Tuesday, September 21, 2010 - link
No, please. Do not do bar charts. Do X-Y price-performance charts. They are far more useful.vol7ron - Tuesday, September 21, 2010 - link
I'm kind of partial to the smallnetbuilder's price-performance chart: http://www.smallnetbuilder.com/index.php?option=co...You can hover to see the item.
evilspoons - Tuesday, September 21, 2010 - link
Seconded. X-Y performance charts are the way to go!!evilspoons - Tuesday, September 21, 2010 - link
Err, price-performance.Brucmack - Wednesday, September 22, 2010 - link
If you do that, please integrate the differences in power consumption somehow. It would be silly to save $20 by buying an AMD processor if it costs $50 more to run over its lifetime.strikeback03 - Wednesday, September 22, 2010 - link
How long would that lifetime be? 1 year? 4 years? Is the machine on all the time but idling 22 hrs a day? Is it gamed on 10 hrs a day but off the rest?Brucmack - Thursday, September 23, 2010 - link
I'm sure it's fairly easy to come up with a couple of realistic scenarios to use. There are only really three variables - # of hours idle over lifetime, # of hours load over lifetime, and average cost of power over lifetime - where it should be fairly easy to come up with some numbers to give a good idea. A useful metric might be to say "given a cost of electricty of X, the breakeven point between processors A and B is after Y hours of use".The whole point of benchmarks isn't to say "here's how it will perform for everyone!", but to give readers an idea of how the product performs in some specific (and hopefully realistic) situations, thereby allowing them to make informed decisions based on their own needs.
AtaStrumf - Wednesday, September 22, 2010 - link
I ditto that! Price is such an important factor it should be part of the graphs or do up a separate set of graphs with performance per $ per application, that'd be awesome. I know that ill be quickly outdated but just put date next to the graph and that'll make it very obvious to take that into consideration when looking at the graph.BernardP - Tuesday, September 21, 2010 - link
The sweet spot is the AMD lineup seems to be the Phenom II X4 955BE. For $ 145, you get a fully enabled 3.2 GHz quad-core with L3 cache, that you can easily set yourself @ 3.6 GHz with little or no voltage increase.Dark_Archonis - Tuesday, September 21, 2010 - link
Yet it will still lose to the i5 750 in virtually every benchmark. The real "sweet" spot is being able to get an i5 750 for a good price. I've seen them on sale at some places for as low as $170-175. At that price you'd be crazy to choose a Phenom II X4 955BE over it.LordanSS - Tuesday, September 21, 2010 - link
Depends if all you have to do is just replace whatever processor you have in your socket, atm.In AMD's case, you can just swap out the processor on most AM2+ motherboards and it'll just work. On Intel's case, chances are you'll have to get a new motherboard (and quite possibly RAM, if your system is old enough).
Been thinking of getting one of those 955BEs for myself. Not much of an upgrade to my main machine that's running a 940BE, but then I can pass on my older processor to my second box which is running an X2-6000 of ancient times.
Lolimaster - Tuesday, September 21, 2010 - link
Why an obsolete CPU when for the price of an i5 quad you get a 1055T wich a better chip overall with a lot of future proof.Besides 1055T, X4 955 is only one option below it's price (then the value segment with AII X4/X3).
Dark_Archonis - Wednesday, September 22, 2010 - link
Future-proof? You DO know that Bulldozer won't work in current AMD motherboards, right? What would you be upgrading to exactly from a 1055T? A Llano?Madmanden - Wednesday, September 22, 2010 - link
Perhaps he meant 6 cores vs. 4 cores, so more future proof in regards to newer, more threaded software and games.KaarlisK - Tuesday, September 21, 2010 - link
Just look at the power consumption (even idle) of the newer chips compared to the older ones.Or maybe it's just binning.
quiksilvr - Tuesday, September 21, 2010 - link
That's the big selling point for me. I'm willing to spend an extra $50 to save even 10 watts because over the life of the computer I will get that money back in electricity savings, maybe even more so.jonup - Tuesday, September 21, 2010 - link
Part of it is probably because the older processor were tested with older mobos/BIOSes/chipsets. We can ask Anand to confirm my speculations. I just do not think that they have rerun old CPUs for this review and have used numbers from the original reviews.Anand Lal Shimpi - Tuesday, September 21, 2010 - link
Correct, that is part of it. The earlier 8-series boards/BIOSes appear to have higher power consumption for example.There have also been general improvements in the manufacturing process.
Take care,
Anand
jonup - Tuesday, September 21, 2010 - link
Anand, I do not think you gave this CPU justice. It performs similarly to the Phenom x2 555 and it is at least $20 cheaper. It uses a lot less energy than the Phenom and in a office environment would be more responsive than the Athlon II X4 and X3 due to the higher clocks. I hate looking at the Task Manager, because most of the time one of my cores is loaded almost 100% and the other ones are near idle. I just think that the extra L2 and higher clocks make the Athlon II X2 265 better CPU for most of the people most of the time and the benches would never be able to display that.Taft12 - Tuesday, September 21, 2010 - link
This highlights what the real problem is for the PC industry today that they don't want you to know.Do you know why Anand doesn't benchmark much more common tasks like Windows 7 boot time or MS Word launch time? It's practically the same across the board on from $65 CPUs to $999 CPUs! Unless you play new games or use professional artist software, you have almost no reason to upgrade a PC that is a few years old. Power savings yes, but not enough over the life of a computer to justify spending hundreds of dollars.
Anand Lal Shimpi - Tuesday, September 21, 2010 - link
Application launch time is actually reasonably CPU dependent if you've got a SSD. While I agree that common tasks don't vary in performance much between similarly clocked processors, I'd say the gap between generations is very visible. Many have said that Core 2 was fast enough but I'd argue there's a noticeable difference in just general PC usage between Lynnfield and Penryn simply because of the benefits you get from high turbo frequencies (windows pop up faster, there's a tangible reduction in response time). That gap will grow with each subsequent generation.The problem with these subjective performance tests is they are difficult to quantify.
I agree that unless you have a specific application that is heavily influenced by CPU performance you don't need to upgrade every generation. But that's also why the PC upgrade cycle is typically 3 - 5 years. Original Core 2 owners will probably want to upgrade next year, while Lynnfield or AM3 owners can definitely wait.
Take care,
Anand
vol7ron - Tuesday, September 21, 2010 - link
I'm a C2D E6600 and I agree about the 3-5 years.My upgrade will probably come sometime between 2011Q2 and 2012Q3.
Note: I don't typically upgrade parts unless they fail, so the mobo/cpu/gpu combo are generally the same as the initial build. Holding onto them for 5-6 years is terrific.
icrf - Tuesday, September 21, 2010 - link
I've got the same CPU and was thinking of upgrading to SB or BD next year. My problem is that my old machine is still perfectly functional, I just wish it'd load SC2 or run through x264 a little quicker.What do you with this older perfectly functional hardware? Does anyone have any good ideas for how to re-task it? I've already got a similar spec'd HTPC along with a pre-ordered Boxee Box and an old XP2600 running the file server.
Pirks - Tuesday, September 21, 2010 - link
ebayvol7ron - Wednesday, September 22, 2010 - link
I am looking at a couple options:1) donate it to friend or family - I still have parents and none are tech savy or willing to put money into it, so that's the most likely option
2) keep it offline for spare parts or backup in case of disaster (laptop is probably good enough though)
3) really look into setting up a home server, possibly get a dedicated IP and host some files for family/friends. It's not the most powerful hardware, but it's good enough for a small number of people.
4) try to throttle down and turn it into an HTPC
5) take all laptops and old cpus and play with Hadoop/MapReduce at home
6) sell it
Any other suggestions?
vol7ron - Tuesday, September 21, 2010 - link
But that's also why they're beefing up the ondie gpuStuka87 - Tuesday, September 21, 2010 - link
I would really like to see a "Bang for your Buck" graph. In many test, AMD comes close to the i7, but the i7 cost more. So when you do a performance per dollar ratio, AMD may come out on par, or maybe even ahead in some cases.I also find it interesting that some of the Phenom II x4's have a lower power consumption at idle than the X2's. I realize the X2's use the same silicon, but with the cores turned off, shouldn't they use less?
Great over all article as always though :)
Zanfib - Tuesday, September 21, 2010 - link
Anand, as always great work,You write "Bulldozer is out of the question however, AM3+ chips aren’t backwards compatible with AM3 motherboards (although the opposite is true, you will be able to use your 970 in an AM3+ motherboard)."
Is this the definitive truth? :-)
I have read a lot of people guessing whether AMD will have AM3 compatibile Bulldozer modes available, or only AM3+ (AM3Rev...).
Thanks!
Anand Lal Shimpi - Tuesday, September 21, 2010 - link
This information came directly from AMD :) Not saying it's 100%, just saying it's the source.You'll see AM3+ motherboards in advance of the Bulldozer client launch so you'll be able to buy a new board for your CPU and then later drop in a Bulldozer.
Take care,
Anand
stmok - Tuesday, September 21, 2010 - link
Yeah, its true.Bulldozer isn't backward compatible with current Socket AM3 motherboards.
But Socket AM3+ motherboards will be backward compatible with existing AM3 processors.
Mentioned on a number of sites...
=> http://www.extremetech.com/article2/0,2845,2368186...
"AMD also told us that it will introduce a new AM3+ socket for consumer versions of Bulldozer CPUs. AM2 and AM3 processors will work in the AM3+ socket, but Bulldozer chips will not work in non-AM3+ motherboards."
=> http://www.rage3d.com/articles/amd_heat_vision_hot...
"For the desktop, the Zambezi processor is good news and bad news. The good news is it's an 8 core product, the bad news is it needs a new socket - AM3r, or AM3+. This is an electrical upgrade of the AM3 platform, to provide the power phases and planes/states required by the power gating features of Zambezi. As you might have guessed from the name, this socket is backwards compatible with existing AM3 processors, so you'll be able to piecemeal your upgrade - motherboard for your birthday, CPU for Christmas, or however your upgrade cycle works."
=> http://www.xtremesystems.org/forums/showpost.php?p...
"When we initially set out on the path to Bulldozer we were hoping for AM3 compatibility, but further along the process we realized that we had a choice to make based on some of the features that we wanted to bring with Bulldozer. We could either provide AM3 support and lose some of the capabilities of the new Bulldozer architecture or, we could choose the AM3+ socket which would allow the Bulldozer-base Zambezi to have greater performance and capability.
The majority of the computer buying public will not upgrade their processors, but enthusiasts do. When we did the analysis it was clear that the customers who were most likely to upgrade an AM3 motherboard to a Bulldozer would want the features and capability that would only be delivered in the new AM3+ sockets. A classic Catch-22.
Why not do both you ask? Just make a second model that only works in AM3? First, because that would greatly increase the cost and infrastructure of bringing the product to market, which would drive up the cost of the product (for both AMD and its partners). Secondly, adding an additional product would double the time involved in many of the development steps.
So in the end, delivering an AM3 capability would bring you a less featured product that was more expensive and later to market. Instead we chose the path of the AM3+ socket, which is a path that we hope will bring you a better priced product, with greater performance and more features - on time.
When we looked at the market for AM3 upgrades, it was clear that the folks most interested in an AM3-based product were the enthusiasts. This is one set of customers that we know are not willing to settle for second best when it comes to performance, so we definitely needed to ensure that our new architecture would meet their demanding needs, for both high performance and overclockability. We believe they will see that in AM3+."
Taft12 - Tuesday, September 21, 2010 - link
<i>But Socket AM3+ motherboards will be backward compatible with existing AM3 processors.</i>This was not true from AM2+ to AM3, and I don't believe for a second this is true across the board from AM3 to AM3+. Many motherboard manufacturers (even top tier ones) will not bother with BIOS updates to older boards that went EOL some time ago.
If your board has a 7xx or Nvidia chipset and has been EOL for a year, I wouldn't count on anything. Better safe than sorry.
JMC2000 - Tuesday, September 21, 2010 - link
"This was not true from AM2+ to AM3, and I don't believe for a second this is true across the board from AM3 to AM3+. Many motherboard manufacturers (even top tier ones) will not bother with BIOS updates to older boards that went EOL some time ago."From what I read, it is not the same situation before: Most AM2+ boards can run AM3 cpus, pending on BIOS support, AM2+ cpus will not work on AM3 boards, because they do not have a DDR3 memory controller. AM3+ (AM3r2) cpus will not work on AM3 boards, because the power gating logic on Bulldozer based cpus is not compatible with the power plane setup on AM3.
Basically, AM3 cpus are both backwards and forwards compatible with previous and upcoming sockets in their line.
Dark_Archonis - Tuesday, September 21, 2010 - link
Yes, for AMD owners, the party is over. Bulldozer is not compatible with AM3 motherboards, so anyone who wants Bulldozer will need a new motherboard.Madmanden - Wednesday, September 22, 2010 - link
The party's over for both AMD and Intel users. They both have to get a new board for upcoming CPUs.mike23 - Thursday, September 23, 2010 - link
this is a very good point.Benchmarks that cover something more than common tasks (that low and mid range processors do adequately every day) would be helpful.
With regard to the competition issue...It's blatantly obvious where Intel sees AMD as competition. All the way up to the 4 core I7 series. Beyond that, no competition and ridiculous prices.
If not for AMD's efforts, you wouldn't have the I5's and I7s in the 200-300+ range right now. So, yes, they are providing strong competition in certain market segments.
DMisner - Tuesday, September 21, 2010 - link
Would you consider doing something of a Folding@Home benchmark, possibly running the SMP client for a day or something and seeing what kind of PPD figures you would get with these processors?Seems like F@H is really what I do most with my computers.
iwodo - Tuesday, September 21, 2010 - link
All these Athlon II and Phenom, are expected to complete against Sandy Bridge next year?( Since Bulldozer are not coming to Fusion type and low end CPU )
SandyBridge is coming to kick some ass then by the looks of it. Higher IPC, Higher Clock speed, lower Idle power, better Hyper Threading......
Apart from Fusion i am not excited by AMD's roadmap at all.......
Dark_Archonis - Tuesday, September 21, 2010 - link
I believe Llano is supposed to compete with Sandy Bridge, but the problem is Llano is coming to market later than Sandy Bridge. Until Llano comes, yes these Athlon IIs and Phenom IIs will be forced to compete with Sandy Bridge, and they are going to lose badly.Even when Llano comes, it's not supposed to be a big performance increase, so I think Sandy Bridge will rule the mainstream market.
As for Bulldozer, it will have to go up against Intel's socket 2011 Sandy Bridge models that are likely to be without integrated graphics. These models are rumored to have up to 8 physical cores, 40 lanes of PCI-E 3.0, and quad-channel DDR3.
Eeqmcsq - Tuesday, September 21, 2010 - link
On the Video Encoding & Data Archive page, the graph for Par2 is in reverse. It has the Athlon II X2 255 at the top and the Core i7 860 at the bottom.Makaveli - Tuesday, September 21, 2010 - link
Eeqmcsq the graph is correct.Look under the heading of the graph....
(Lower is better)
Eeqmcsq - Tuesday, September 21, 2010 - link
Yes, I saw that. But Anand usually puts the best CPU at the top, not at the bottom. See the previous review for the Core i7 970.http://www.anandtech.com/show/3833/intels-core-i7-...
hangfirew8 - Tuesday, September 21, 2010 - link
You write: "The Phenom II X6 1075T is an interesting chip as you get a lot of compute but it's only useful in heavily threaded apps."I strongly disagree, unlike you are calling virtualization "apps".
These chips are a godsend to developers, testers and sales who are running entire tiles of software. Need to test your client on XP, Vista and Windows 7? Need to run a web application server on Windows and an Oracle database backend on Linux? Need to do all of these at the same time? This is what a Hex Core chip is made for, not running a single heavily threaded app on a single O/S.
I'd like to see an AT Bench benchmark that covers this sort of application. I'd like to see how comparably priced Intel CPU's do versus AMD chips. I'm thinking this would be a huge differentiator between HyperThreading and actual cores, but I'd like to see the numbers before making any claims.
haplo602 - Tuesday, September 21, 2010 - link
this is not quite comparable. on moderately IO dependant workloads (your web server/oracle db example), HT will do just fine. on compute dependant applications with little memory access, HT will bottleneck. However you will hardly run 4-6 compute dependant VMs at once every time.hangfirew8 - Wednesday, September 22, 2010 - link
I agree on your HTT assessment.BTW I mean "unless" not "unlike". Where's the edit button?
Can I get a reply from AT on VM? Can we stop treating every desktop chip as a gaming candidate-only chip and get some kind of desktop VM benchmark going here?
I'm thinking running one per VM: compile/link/install (push application from dev VM to web server VM over virtual network), web server with integrated JAVA application, three web clients, and a database backend. Run a mix of Linux & Windows, 32 and 64 bit O/S's. The app being compiled doesn't even have to be the app being run.
Accord99 - Tuesday, September 21, 2010 - link
There's some tests done by the IT side of Anandtech:http://www.anandtech.com/show/2774/10
Relatively speaking, more cores offers a bigger boost than Hyperthreading, but a single Nehalems core is much more powerful than a AMD K10 core.
hangfirew8 - Wednesday, September 22, 2010 - link
Yes I know they can do it, but we have no data on how similarly priced desktop chips compete. HTT versus Cores is only one aspect. I want to see the AMD hex chips stretched against Intel quad cores in VM's and shake out the bang for buck. AT's attitude about hex cores only being useful for single heavily threaded apps just doesn't match how this chip is being marketed or how it is being used in the field.AT should realize that a lot of their PC gaming readers have daytime jobs not just as server administrators, but as software developers and testers. Management comes to us for advice and no one is providing us with the data to make informed decisions. 7Zip benchmarks are fine but give us a software development lifecycle benchmark using VM's, give us data on what we are already working with for several years in real life (work). Give us a benchtop VM benchmark! Please!
IceDread - Tuesday, September 21, 2010 - link
Annoying to see that we still have no competition in the high end segment. It's still being dominated by intel which makes for higher prices for us, the consumers.Dark_Archonis - Tuesday, September 21, 2010 - link
Annoying? Well it is what it is, AMD just can't compete at the high end and I don't think Bulldozer will help much going up against Sandy Bridge E.Madmanden - Wednesday, September 22, 2010 - link
Well that's a surprise...mino - Tuesday, September 21, 2010 - link
Well, I think we are a bit spoiled actually.Just think about it for a minute - had not AMD acquired an almost complete DEC team around 1997, there would have been NO real competition and the prices would WAY above what Intel is charging now.
Think VIA Nano on steroids as "competition".
Let's be realistic, shall we?
It is simply not feasible for 2 competitors in such a complex field to be BOTH on top concurrently for extended periods of time (as has happened during the Slot A heydays).
These situations are bound to be more of an abnormality than a standard.
Considering the HUGE gap in resources, we shall be very happy that AMD is still within 30% of Intel's TOP despite using an 12yrs old K7 architecture.
A true testament to Alpha actually.
And, for killing IA64 on desktop alone, they deserve a Nobel prize! No contest.
Dark_Archonis - Wednesday, September 22, 2010 - link
Speaking of Alpha, I wonder where many former Alpha employees have gone? Oh right, they're now working at Intel on future Itanium processors.Sure we were spoiled, if you will, years ago when AMD introduced the K7, Intel missed with the P4, and AMD introduced the K8.
Thing is, historically Intel was almost always on top in the consumer CPU market.
AMD did provide some real competition back then, but not so much anymore. Intel can price AMD out of existence right now if they wanted to, but Intel doesn't want to.
Unless Bulldozer is a gigantic performance leap forward, AMD will continue to struggle to provide competition and will only be "good enough" or barely catching up to Intel's latest.
Intel has real competition in non-consumer markets, and that is challenging Intel.
crucibelle - Wednesday, September 22, 2010 - link
fanboi.Dark_Archonis - Wednesday, September 22, 2010 - link
I love such thorough and insightful answers such as these.People talk all the time about how competition is good. Real competition is good yes, mediocre or weak competition is almost pointless. Right now AMD is mediocre/weak competition from Intel.
You can call people names, point fingers, and complain about it all day long, but it is what it is.
I have nothing against seeing some STRONG competition for Intel, but AMD currently isn't providing strong competition.
bji - Wednesday, September 22, 2010 - link
Really? Then why did I decide to buy a 1075T instead of an Intel chip in the same price range? For my needs (all of my single threaded apps are quite fast enough with any modern processor, but as a software developer I need as many cores as I can get for parallel compiles) the 1075T is a great deal. Throw in the fact that AMD motherboards are better than Intel at the same price point and it was an easy decision. I am slightly miffed that Bulldozer won't be AM3 compatible becase I was banking on that a little but when deciding to go with an AMD platform, but whatever. The rest of the value proposition is still in AMD's favor.Now of course my experience is just anecdotal; I can say with certainty that in some segments, AMD is very competitive with Intel. I think the question of whether or not AMD provides "strong competition" with Intel is really just about how many segments AMD provides strong competition with Intel and how strong the competition is in those segments, and from that, you can decide whether or not the overall competition is "strong" or "weak".
Given the success AMD is having in netbooks/notebooks lately (at least all of the Fry's/Best Buy/Office Depot/whatever ads I have seen in the past few months have been heavily featuring AMD, I assume because the platform is selling well), I think we can say that AMD is providing strong competition in that space.
In the 'midrange' CPU market - which is the majority of the PC processor market - AMD clearly provides strong competition. See this Anandtech article, my purchasing decision above, countless similar articles and testimonies all over the internet.
On the high end PC CPU market, AMD is not competitive. We all know that.
On the server CPU front, AMD has not been competitive lately but is starting to gain a little ground back. At least, if my impressions given what I've read on Anandtech and other sites is anything to go by.
So in the end, I'd say that in 3 out of 4 x86 CPU markets, AMD is providing strong competition with Intel. For this reason I think it's erroneous to call AMD's current competitiveness 'weak'. I think it's overall 'fairly strong', and I only hope it gets stronger, for everyone's benefit.
bji - Wednesday, September 22, 2010 - link
Oh and I forgot to mention integrated graphics. Of course AMD is extremely competitive with Intel in that segment, and in fact is the dominating player (at least, performance-wise).flyck - Wednesday, September 22, 2010 - link
AMD is competing with Intel just fine.Yes they have problems when Conroe was launched in the high end. Most of this is due to the problems they encountered with Barcelona. Without those issues they would have been very competive with Core2. i7 lineup is now the strongest technology but AMD can compete ok using more cores. The biggest issue AMd is having is that most of their products had to be postponed. But with Ontario platform they will have their first real low power chip which is very promesing. llano will offer higher speed than x4 with very good GPU, which again is a pretty solid effort to gain the low and mainstream market due to embedded graphics that do not suck for desktop useage. Bulldozer, although to late, seems to be a chip on which AMD can build further upon. Performance is one area, the ability to price competitive is also an area and that is an area AMD is really competing with. Their future shows alot of promise (bit like in the K8 days where the XP couldn't cut it anymore). AMd is the only reason why Intel (lead x86 manufacturer) is actually making processor worth the money. This alone implies that AMD is competitive with Intel. Without the clausules and market abuse iItel did in teh years before, K7 and K8 could have damaged them alot more, they can't let that happen again.
hangfirew8 - Wednesday, September 22, 2010 - link
I love competition. I want choices, not just at the lower end price points.The real issue is that AMD has not improved their instructions per clock cycle ratio substantially since the earliest K8's. AMD caches have gotten larger, HT has gotten faster, power saving features have gotten more sophisticated, clock rates have gone up, but we've yet to see any real jump in core processing efficiency.
Until AMD addresses that they will stay on the low end, with the low end margins that come with that market.
bji - Tuesday, September 21, 2010 - link
Fry's had a sale on these last week; you can see the advertisement for this sale at:http://newspaperads.mercurynews.com/FSI/Page.aspx?...
$179.99 for the Phenom II 1075T seemed like a good deal to me so I picked one up. I still don't even have any of the other system components I need to run this thing; but the deal was so good that I just had to jump. I'll buy motherboard, memory, etc later ...
The weird thing is the AMD seal on the top of the box lists it as a 1075T and says its clock rate is 2.8 Ghz. Didn't notice this until I got home. I wrote to AMD to ask about it and they replied that they had some printing problems with the labels and some boxes went to vendors with that printing error. But they assured me that it's really a 3.0 Ghz part.
Gilbert Osmond - Tuesday, September 21, 2010 - link
I notice on the photos of the chip packages it says:DIFFUSED IN GERMANY
MADE IN MALAYSIA
I've not seen the "Diffused in..." marker ever before on a chip package. What does it refer to / mean?
Gilbert Osmond - Tuesday, September 21, 2010 - link
I've found a complete (exhaustive) answer to my own question, here:http://www.faqs.org/rulings/rulings2007HQW968421.h...
The long and short is that the separate markings help to resolve an ambiguity about the country-of-origin.
LoneWolf15 - Tuesday, September 21, 2010 - link
While I like Intel CPUs, it seems that good 1156/1366 mainboards are so expensive compared to Socket AM3.It'll be awhile before I replace my Q9650, but when it happens, if Intel hasn't worked with vendors to make mainboards more reasonably-priced, I see AMD as a real possibility.
Taft12 - Tuesday, September 21, 2010 - link
It's not board OEMs that are causing motherboards to be unreasonably-priced, it is Intel with their chipset pricing. Despite moving the IGP off of the chipset (a huge cost elimination), the wholesale price stayed the same.3rd party chipset competition is needed desperately but Intel will have none of that! AMD boards definitely give you more for less.
mino - Tuesday, September 21, 2010 - link
Actually this sad fact is affecting the AMD market too.It seems Nvidia has completely occupied the low-end market while AMD IGP boards have moved up on the price scale around $10.
Also check CPU prices - there is no AMD dual core below $60 with artificially sold single cores for $40.
Basically while the performance you can buy at $60 has gone up 2x over last 3 yrs on ADM side, their $40 offing is now useless while previously it was relatively reasonable.
AMD sees no competition from Intel there => no low end dual cores.
iamkyle - Wednesday, September 22, 2010 - link
For the majority of the market out there, you may be getting to a level where the performance of the CPU is good for what, embedded applications? Linux servers that can be run on used hardware for $20? Niche applications.At least there is speed at the low price points to fulfill a multitude of uses.
Taft12 - Friday, September 24, 2010 - link
I would think AMD's can't sell any CPU for less than about $65 without going deep into the red on every unit. Think about raw material costs, electricity, clean room maintenance, shipping, packaging, ... This is before accounting for any R&D or payroll for a single employee!Similarly, I think we're at an absolute floor at about $40 right now for hard drives.
SonicIce - Tuesday, September 21, 2010 - link
good article and good cpu's. you should highlight which chips are the new ones in the beginning chartTanclearas - Tuesday, September 21, 2010 - link
Once again, I believe the biggest release AMD could make is the platform. I ended up going with a Q6600 a few years ago when I would have loved to choose AM2+. I have been happy with the Q6600, but I spent a lot more on it than I really wanted to at the time. I could have put together a system using a CPU that cost $200 less, then upgraded one or two years later with a significantly faster CPU.Fast forward to 2010 and the situation is quite similar. I would love to pick up an AM3+ based system with a Phenom II X4 and be ready for whatever AMD has coming. I would build that system right now, and AMD would be getting 2 CPU sales (unless things went horribly, HORRIBLY wrong with BD).
iuqidids_sm - Tuesday, September 21, 2010 - link
Is there any news about the 95W x6 cpu's, particularly the 1055? Apparently its on sale outside US, but I can't seem to find it from US based retailers. Thanks.Lolimaster - Tuesday, September 21, 2010 - link
Why don't you use 3DMax 2010 instead of an ancient version from 2005. Enough with the bias.Maybe because in this updated version Phenom II X6 perform better than any i7 quad?
bji - Wednesday, September 22, 2010 - link
I read alot of benchmark reviews. "This benchmark favors Intel" is a line I read quite frequently, and I always assume it's because compiler writers and/or software developers consciously choose to optimize for Intel chips (which makes sense since it's the larger part of the market). "This benchmark favors AMD" is not something I think I've ever read, at least not in a context that led me to believe that it's due to specific optimizations targeted at the processor. Why would 3DMax 2010 perform relatively better on Intel than 3DMax 2005 did? Is it because the newer version has been better optimized for Intel?In that case, is it more valid to use a newer version or an older version? I guess we want our benchmarks to reflect the non-benchmark software that will run on the platform, so I suppose that if most software is Intel-optimized, Intel-optimized benchmarks make sense. If not, then not.
I personally use Linux almost exclusively and I feel pretty confident that the GNU compiler toochain that is used for this operating system is not more optimized for one processor vendor than another - at least not intentionally, anyway.
In the for-money world of Microsoft and Intel with their backroom deals and shady business practices, however, I can't say for sure.
Lolimaster - Tuesday, September 21, 2010 - link
http://www.tomshardware.com/charts/desktop-cpu-cha...flyck - Wednesday, September 22, 2010 - link
I would expect Anand to have some form of moderation on this forum?Taft12 - Friday, September 24, 2010 - link
You would think they would with this new forum software, but alas....lwatcdr - Friday, September 24, 2010 - link
Slashcode does.Why do we have to waste time reading dumb stuff like this?
I agree.
but alas....
hacksquad - Wednesday, September 22, 2010 - link
I hope you stop using intel's new cpu's cause it contain AMD technology which is x86-64/AMD64 :PStaples - Wednesday, September 22, 2010 - link
I think it would be more helpful if you gave percentages when overclocking in addition to ran clock numbers. It managed 4.0GHz does not mean as much as it achieved a stable overclock of 25% vs the other processor which managed a tiny 10%. Makes comparing things a lot easier.Madmanden - Wednesday, September 22, 2010 - link
OMG YOURRRR SO INSITEFULLL!!!11111Learn to spell, moron.
MrPerlishells - Thursday, September 23, 2010 - link
I just bought the 95W version of the 1055T and it took a few months for it to be available. I put it in my ECS A780GM-A Black Edition Motherboard and it works for awhile then the screen goes blank.Going to have to get a new motherboard. So much for backward compatibility with AM2+ socket.
Bennyto - Thursday, September 23, 2010 - link
I have seen a few desktops form OEMs featuring an AMD Phenom II X6 1035T.Is this an OEM only CPU ? Could you include it in your performance charts? Thanks
biggbigg - Wednesday, September 29, 2010 - link
The 970BE looks interesting especially if it has headroom for more OC.coronaJet - Monday, October 4, 2010 - link
I think its astonishing, that the new Athlons X2 and X3 have lower power consumption than the old ones, even if they work at a higher frequency.How is that possible? Is that a new stepping or better voltage magagement ?
Hrel - Wednesday, October 6, 2010 - link
Looking at these charts. The Core i5 750 consistently beats the X4 970 and they're priced to compete against each other. Not only does Intel win more, but they win in all the areas that really matter to me; not to mention the power consumption advantage.Secondly, where AMD actually does very well, encoding video, that advantage is negated by the fact that regardless of how good the CPU is a encoding and rendering video doing those things on the CPU is almost entirely obsolete. I know I do all that stuff using my Nvidia GPU and it's exponentially faster.
In other words, AMD can't win anywhere at a similar price compared to intel. The ONLY reason I can see buying an AMD CPU and that price point right now is if the CPU is all your changing and you already have an AMD Motherboard. Unfortunetly for AMD if you're building a whole new system Intel is the way to go and every price point with decent profit margins.
loadwick - Tuesday, October 12, 2010 - link
Ivy Bridge is meant to be still on track for 2H 2011. This just seems crazy. Intel are not going to release a whole new generation in Sandy Bridge and then replace it within 6 months!??I guess they might just release a few 22nm parts like they did with 32nm but even that seems so unlikely right now. I mean we won't even have high-end Sand Bridge parts out yet and we will have the replacement for Sand Bridge at our door step!
Anyone know what is really going on?
alexb1 - Wednesday, October 13, 2010 - link
I am quite confused as to WHY Athlon II X3-450 with higher speed is actually cooler and more efficient than X3-440? Also, the same between X4-645 and X4-635?!I thought they're all the same core, no?
If these are more efficient, are they ideal for HTPC stored in a cabinet? using Antec Fusion Case?