AMD Undercuts Intel With Six-Core Phenom IIs
EconolineCrush writes "As Slashdot readers are no doubt aware, Intel's latest 'Gulftown' Core i7-980X is an absolute beast of a CPU. But its six cores don't come cheap; the 980X sells for over a grand, which is more than it would cost to build an entire system based on one of AMD's new six-core CPUs. The Phenom II X6 line starts at just $200 and includes a new Turbo capability that can opportunistically raise the clock speed of up to three cores when the others are idle. Although not as fast as the 980X, the new X6s are quick enough to offer compelling value versus even like-priced Intel CPUs. And the kicker: the X6s will work in a good number of older Socket AM2+ and AM3 motherboards with only a BIOS update."
We all should hope AMD does well. I use AMD chips in about 90% of my systems. Value is the main reason. Intel makes excellent products however you invariably have to upgrade the motherboard to use a new chip. AMD has been kinder in this regard recently. I go with a middle of the pack system anyhow and I really appreciate the value AMD provides.
On a price performance basis AMDs Phenom IIs have consistenly been a better buy for some time now. To the point it's hard to suggest anyone buying intel at all, unless money is no object. (I don't know why I bought Intel anyway :S).
Honest hardware review sites (that aren't far up the ass of vendors) are at the point of recommend AMD CPUs on a price/performance basis.
http://www.tomshardware.co.uk/best-gaming-cpu,review-31857.html
It seems Intel doesn't get even a "honorable mention" until page 3. At $120 price point, Core i3 gets a look in. Oh, they also don't recommend anything above about $160 to quote Tom's: "Best gaming CPU for $190: None".
To add further insult, money saved from AMD motherboards being cheaper (in particular SLI/xfire AMD boards are a good whack cheaper) will let you put money towards more storage, a SSD or a step up in CPU speed.
After logging in slashdot still does not take you back to the page you were on. It's been that way for 20 years.
It depends on what you consider "general use". Some cases demand more cores. GTA 4 is more or less unplayable on dual core systems, so an AMD is the cheapest option. As always, look at what you want to do, and then buy a computer, and don't be a fucking idiot.
Although not as fast as the 980X, the the new X6s are quick enough to offer compelling value versus even like-priced Intel CPUs. And the kicker: the X6s will work in a good number of older Socket AM2+ and AM3 motherboards with only a BIOS update.
So doubly pointless
Indeed as this is the "the" new X6s.
I still like the underdog and hope they do well. The latest and greatest is often overkill.
Views expressed do not necessarily reflect those of the author.
For less than the price of Intel's top desktop chip, you can get an uber-1337 AMD Opteron with 12-cores. Beat that, Intel...
Prices start at $750.
I do live music production, and I can use all the horsepower I can get.
My software is quad-core optimized, at present. On my dual-core machine, I can definitely do more than I could before that optimization. So I'm wondering, when I'm doing a crapload of DSP computations and demanding consistently ultra-low latency performance, where does the processors/threads/cpu speed/ memory speed equation balance out?? It's hard to believe my HD speed matters much, when I've got 8 gb of RAM. I do stream samples from disk, but that stopped being a bottleneck 5 years ago.
No one does these kinds of real world tests in reviews, even if they offer frame-rates for 17 games!!
"Regular Joe" users play with home videos and photo editing - two embarrassingly parallel situations right there where you can always find something to do with as many CPU cores as you can get.
Four is a relative large number though.
So where's that 8P Intel PC? Granted, it's less of an issue now that there are 6 core CPUs out there, but let's not forget that Intel still can't effectively put one together.
Let's look at that Top 500 list. The top 2 are AMD systems, and that the 3 Intel systems are in the bottom 5.
If I want to browse the web and not heat my house at the same time, AMD really does offer the better chip, and cheaper by far too.
If I want to play that FPS and have an extra frame or two, then the Intel chip is a winner. If I want to participate in something like Folding@Home, then Intel is a winner if power and heat aren't considered. If I'm rendering video, Intel might be a winner, depends on whether heat matters.
Very few things even most on /. would utilize a computer for will only see an Intel advantage maybe 1% of the time.
After all - does using an AMD or Intel chip make any difference rendering /.?
The cesspool just got a check and balance.
What part of that 14 minutes improvement -- is actually the result of a speedy RAM, improvements to the processor-memory bus interface? If we strip it down to its bare, I have a high confidence that the cores added only a fraction of improvement.
Totally off base. First thing I tried when I got a new quad core dev box at work was to try the standard build with 1, 2, 3, and 4 cores. Imagine my total lack of surprise when the 2 core build ran in about half the time, the 3 core build in about a third the time, and the 4 core build in about a quarter of the time. How shocking.
No, it's not perfect, as in T(N) = T(1)/N, but it's certainly much stronger a correlation than "only a fraction."
Hell, I'd love to see it turned on for all the official build machines, but it's just a little bit scary, since I have seen it fail (as in the build driver crashing) about 1 in a 100 times. But for daily dev work, it's awesome.
To me the new AMD six core is a little bit of "me too!" from AMD. Not that there won't be people who won't find practical uses, and no buying one to OC it so that you can get a higher folding score does not count imo, for them but it's still as many have pointed out hard to find real world scenarios where people need that type of a CPU on their desktop.
AMD not only has to compete with Intel on the technology front but marketing as well. And again I don't want to take anything away from AMD and the idea behind pushing the envelope on new tech. But when it comes to end users they really don't know and or care what is driving what they do with their computer. I see people's eyes glaze over when I even start to talk about what type of hardware I'm going to set them up with. They simply do not care.
However I have seen where people have been brainwashed by the marketing. People have asked me if their system is Intel Inside. I try to explain to them that at most price points AMD is a better buy and the more brainwashed come back to me with some very clueless lines like, "But if I don't have and Intel I won't be able to run what I need to." I even remember back in the early 2000's walking into a local computer shop, I needed a mobo asap, and one of the sales reps told me that AMD CPU's were, "Garbage. We don't even stock any AMD parts."
I asked if he knew about the, at the time very high end, computing array that was I think setup at GT that was using AMDs and he started to sputter. "Well, I don't know about that." Of course you don't you idiot I felt like saying, but I just left and have since made it a point to make sure that people that I know and do work for look out to be wary of that place.
My main point is that AMD serves many purposes in what our modern computing landscape is. I personally do like them a lot but as someone who deals with many systems I deal with Intel plenty too. And hell I like a lot of Intel's products. They have top notch R&D and blah blah blah. But we would be a poorer group of computer users without AMD even without all of the other reasons to like them.
Really, I know what I'm doing...Ohhhh, look at the shiny buttons!
ECC support is disabled on all non-Xeon chips. There isn't a technical reason, it was purposeful market segmentation.
Per Intel FAQ
Knowing this before hand, that's why opted to build a new AMD based computer that has ECC enabled. The parts I used includes current prices below from Newegg.com and Crucial.com
Motherboard: Asus Crosshair III Formula = $199
CPU: AMD Phenom II X4 955 Black Edition = $159
RAM: Crucial 4GB kit (2GBx2) ECC DDR3-1333 (P/N: CT2KIT25672BA1339) = $149
Total (minus shipping) = $507
It's damn cheap for a fast performing ECC workstation.
Life is not for the lazy.
I am amazed at the shifts that happen sometimes here on slashdot. A few years ago, multiple cores were the cat's meow and were going to change the world, and all the programmers who didn't learn Haskell or other non-parallel languages would be left behind. If you tried to post a comment saying that a parallel algorithm doesn't always make things better, you would get modded down or ignored, or laughed at.
Now it's gone to the other extreme, you've said multiple cores are essentially useless for the average person, and got modded up. Can there be no middle ground? Something like, multiple cores are better than not having them, but they aren't going to change the world?
Qxe4
Lots of cores are good. A good amount of RAM helps too. And when you put a bunch of servers that do real services in one box, storage and network bandwidth and latency are also important. At the moment memory seems to be the sticking point rather than cores.
For many virtualization scenarios right now with VMWare there is the VMWare licensing to consider. The 8GB DIMMS are still spendy but so is the VMWare licensing so there's tension between server density with expensive DIMMs to minimize rackspace and licensing cost, or twice as many servers with the cheaper DIMMS and paying the increased licensing.
HP and Cisco both have interesting propositions in this area, with blades that do 10GbE and FC for good bandwidth, have dual 6-core processors and support 192GB of RAM or more. When better processors and cheaper memory come out the Cisco UCS solution may have challenges because the architecture may become I/O bound with only 2 10Gbps links for both network and storage per half-width blade, and sharing at least half of that in the chassis uplink. The HP blade solution supports full line rate between servers and an insane amount of uplink - and it's denser than the UCS so it takes up less rack space. The UCS solution uses an ASIC to more than double the number of memory sockets, so for example a full-width server supports up to 192GB using the cheaper 4GB DIMMS and they claim they all work at 1333MHz. I don't know what IBM and Dell are doing here but I know they have products too.
All of the basic virtual environments are basically free (except Microsoft's Hyper-V, of course). Microsoft software is practically free in education environments so they're making inroads there. But in the enterprise the high availability and reliability features of the advanced commercial packages are compelling. There's something awesome about asking an admin management type to evacuate a server so you can work on it (because the local IT support is out today), and watching her migrate the VMs off in a few seconds so you can take it down.
One of the really neat things about VM consolidation is that 20 physical servers with 4x 1Gbps Ethernet don't actually use it so by consolidating them you eliminate waste. Not only that, but by moving to the VM host with 10GbE you get virtual servers that have multiple 1Gbps connections but each has a 0.2ms ping to the external gateway and each other. This makes many things work faster like databases, websites and such. The downside is the downside of sharing: if you don't plan carefully and get a storm load, the servers will contend for bandwidth and knock each other offline.
Now that DDR3 is becoming cheaper than DDR2, I'm glad to see AMD adopt it. I like their 8-core server chips for workstation stuff - a coworker and I are building out dual-8 core boxes for virtual machines and such.
Help stamp out iliturcy.
Anyone who knows a bit about cgi (I mean ray tracing, not the the real time rasterization you see in games) knows how much of a pain in the ass it is to have to wait 48 hours to render a scene just to be able to see if the light scatters correctly on that orc's forehead. Ray tracing is one of the few practical applications that parallelizes really well, and you will actually use close to 100% of your six core grunt. Since I got an AMD Phenom II 810 quad core, the difference in rendering time for blender has been, well, phenomenal.
Lets not be too quick to jump on the "omg moar cores is teh suxorz" bandwagon, some people actually need it.