• What
    is this?
    You've landed on the AMD Portal on AnandTech. This section is sponsored by AMD. It features a collection of all of our independent AMD content, as well as Tweets & News from AMD directly. AMD will also be running a couple of huge giveaways here so check back for those.
    PRESENTED BY
POST A COMMENT

174 Comments

Back to Article

  • tipoo - Thursday, January 27, 2011 - link

    Goodbye, crappy Atom. Reply
  • ssj4Gogeta - Friday, February 04, 2011 - link

    @Anand:

    "Remember that the Atom 330 is a dual-core CPU with SMT (4-threads total) so you’re actually getting 17.2% of four hardware threads used, but 34.4% of two cores."

    Why would you say that, when those 4 threads are actually running on the 2 CPU cores? Or is it that the average for the 4 threads can never exceed 50%? Only in that case your doubling of the number makes sense.
    Reply
  • ClagMaster - Thursday, January 27, 2011 - link

    This is amazing performance for an 18W processor if you really think hard about it.

    This rig will beat my Athlon 1700+/KT266A/DDR-266/GeForce 4200 rig I had six years ago.

    AMD will sell a lot of these Bobcats after the final bugs are worked out.
    Reply
  • Dark_Archonis - Friday, January 28, 2011 - link

    No, no it's not. Intel will very soon release 17W ULV Sandy Bridge CPUs.

    Even a current ULV Arrandale CPU will take out a Brazos in terms of CPU performance.
    Reply
  • Kiijibari - Friday, January 28, 2011 - link

    Hahaha ... attacking tiny Zacate with a Sandy Bridge ... yes it may have triple or quadruple performance, but so will be it's price ^^ Reply
  • Shadowmaster625 - Friday, January 28, 2011 - link

    yeah and the tray price will be $188 per cpu, in qty's of 1000. Haha good luck with that. Reply
  • djgandy - Friday, January 28, 2011 - link

    Yes I am looking forward to ULV Sandybridge. I still have a Core Duo CULV which is starting to show its age. I'm sure it is still way faster than Atom and Zacate.

    It may cost more but you'll have 2-3 times the CPU performance with SB. Plus it should have dedicated video decode.
    Reply
  • coconutboy - Friday, January 28, 2011 - link

    You're only talking about the speed and if that's all you need, of course focus on a much pricier alternative. Now add in power envelope and price which is the entire point of this thing; moderate performance, very low power draw, dirt cheap. AMD nailed all of that.

    Brazos e-350 can be built as a tiny, passively cooled unit that costs a mere $100. If my neighbor asks me to build her a new system (again), this is EXACTLY what she'd want. It'll sit quiet 'n' cool in her living room right next to the TV and cost peanuts.

    Toughest part of building a ~$200-250 Brazos rig will be getting an inexpensive (but not cheaply made), passively cooled PSU and case. RAM prices are barely a concern.
    Reply
  • sebanab - Friday, January 28, 2011 - link

    "you'll have 2-3 times the CPU performance with SB" -> the things i use notebooks for don't need allot of CPU.
    If you do allot of PI calculations / video encoding / SETI , well , you should get a job. And when you do get a job, you will also understand why a 500$ notebook is better than a 1K$.

    I hope i don't annoy anyone , but Intel really has to pay for the four years of selling Atom saying "it is good enough".
    Reply
  • knedle - Saturday, January 29, 2011 - link

    I have a job and I'm running SETI (now it's called BOINC).
    When I have to make decision - get computer that will give me more SETI points, or computer that will give me more "money saved" points, I go for the money. :D
    Reply
  • Scootiep7 - Friday, January 28, 2011 - link

    The major problem is that Intel's drivers are flat out junk for anything Gaming/HTPC related. With ULV Sandy Bridge, you'll be paying three times as much for a complete system that performs less than half as well on most of the stuff you want to do with it in this market segment. Reply
  • Marlin1975 - Thursday, January 27, 2011 - link

    Talk about good for a first try.
    Think about it; AMDs first shot into this area is as good or way better than the much updated Atom from intel.

    Anand can you ask when amd thinks they will be able to move this to 32nm? Seems the design is good just needs updated shrinkage to increase the Mhz.

    Also since this will not be in a netbook/laptop can you overlclock???
    Reply
  • nitrousoxide - Thursday, January 27, 2011 - link

    No 32nm Brazos parts, they will come directly with 28nm next year. Yes, 4 bobcat cores, VLIW 4D architecture GPUs, 2x performance at minimum or even higher if AMD gives it Turbo Boost feature.

    Overclocking doesn't make any sense on this chip, because no matter how you overclock it, the chip cannot do the things it can't do at default clock. it simply brings you more power consumption.
    Reply
  • Marlin1975 - Thursday, January 27, 2011 - link

    Overclock for the desktop not the netbook.

    Could make some items go from unusable to usable as a low power desktop replacement.
    I like to see what would happen and what would help the most or not.
    Reply
  • nitrousoxide - Thursday, January 27, 2011 - link

    Try AMD Overdrive Reply
  • knedle - Saturday, January 29, 2011 - link

    You can overclock everything (even a notebook), but since the board didn't support it, it's much harder. ;)
    I'm sure someone will release overclockable boards, it's just the matter of time.
    Reply
  • Gigantopithecus - Thursday, January 27, 2011 - link

    ...ETA for Brazos mini-ITX availability in retail channels?

    I can't think of a reason to buy anything, be it a netbook or nettop or htpc, with an Atom in it given Brazos's performance and power consumption. Can't wait to start building Brazos systems!
    Reply
  • codedivine - Thursday, January 27, 2011 - link

    I am a little confused. If I put a graphics card in the x16 slot, will it run at x4 PCIe 2.0 speeds or x16 PCIe 2,0 speeds? Reply
  • rpsgc - Thursday, January 27, 2011 - link

    It will run at x4 (comparable to PCI-E 1.1 x8) Reply
  • rpsgc - Thursday, January 27, 2011 - link

    Since I can't edit my post...

    ... look here for the performance penalty of running at x4 versus x16
    http://www.techpowerup.com/reviews/AMD/HD_5870_PCI...
    Reply
  • codedivine - Thursday, January 27, 2011 - link

    Thanks! Reply
  • codedivine - Thursday, January 27, 2011 - link

    I was wondering if it is possible to overclock this mobo+APU combo? Reply
  • allzhat - Thursday, March 15, 2012 - link

    I use K10stat to increase the multiplier for my notebook (e350) from x16 to x18 (1800Mhz)or even x22 (2200Mhz), but I can't feel more performance for this APU but it draw more power consumption and heat of course

    I suggest to use gamebooster for heavy application

    PS: I'm sorry for my bad english
    Reply
  • nitrousoxide - Thursday, January 27, 2011 - link

    So an SSD really makes difference on Brazos platform? Now I can place my order :)

    My choice is Corsair Nova 64GB since it is the cheapest drive with the minimum accpetable space I need. Any drive cheaper than this one is not big enough and bigger ones are more expensive... Hope its low Random Read performance doesn't impact much, at least faster than an HDD!
    Reply
  • Dark_Archonis - Friday, January 28, 2011 - link

    Being cheap with SSDs? Good luck when you suffer reliability problems with your SSD. Reply
  • nitrousoxide - Friday, January 28, 2011 - link

    Guess what? Somebody ran an durability test for an Intel X-25V, a non-stop read--write-erase cycle for 300 days and still working. Performance began to drop after 6 months, but keep in mind that this stress test is over 20 times beyond daily use. So it actually takes years to destroy an SSD, without SF controller, of course :) Reply
  • mariush - Sunday, January 30, 2011 - link

    A sample of ONE is by no means enough. The same person could have tested a second SSD and have it die on him after a week.

    After all that's what MTBF is all about - the average time for a failure in a population of identical devices.
    Reply
  • Ethaniel - Thursday, January 27, 2011 - link

    Is it me or the Nano X2 just beats the crap out of both Brazos and Atom? Can we get a review of that chip, Anand? Reply
  • tno - Thursday, January 27, 2011 - link

    http://www.anandtech.com/show/4017/vias-dual-core-...

    Not a full review but darn close.
    Reply
  • e36Jeff - Thursday, January 27, 2011 - link

    they did, its here:

    http://www.anandtech.com/show/4017/vias-dual-core-...

    Problem is that what they tested was basically an engineering sample built on the wrong node, they havent gotten anything to market yet, so actual numbers from real products are unknown. having said that, yeah it does look like it might be better, but until someone makes a product based on it, we'll never know.
    Reply
  • nitrousoxide - Thursday, January 27, 2011 - link

    The price itself is stopping Nano X2 from dominating APU or Atom in the compete. The Nano platform consists of 3 different chips (aka CPU, NB, SB as the traditional layout) but the latter two managed to do that with only one chip, especially APU with very small die size. While Nano X2 is an impressive part compared to APU and Atom in absolute performance, it's not competitive in other features such as die size, price and power consumption. Reply
  • Tralalak - Friday, January 28, 2011 - link

    AMD Zacate TDP Configs@18W + Hudson M1 = Fusion Controller Hub = ("South Bridge") TDP Configs@2,7W to 4,7 W for typical configurations === AMD Brazos 20,7W to 22,7W TDP. (2 chip solution)

    VIA say: VIA Nano X2 have some TDP than VIA Nano Single-Core.

    VIA Nano X2 1.4GHz (40nm TSMC) have some TDP than VIA Nano U3200 1.4GHz (65nm Fujitsu) TDP = 6.5W

    VIA Nano X2 1.4GHz max. TDP@6.5W + all-in-one chipset VIA VX900 MSP (Media System Processor = North Bridge (IGP) + South Bridge) max. TDP@4,5W === max.TDP@11W (2 chip solution)

    I mean that In "minibotebook market" is very competitive.

    VIA's 40nm next all-in-one chipset VIA VX MSP with DirectX 11 IGP refresh will appear in Q4 2011.
    Reply
  • mczak - Monday, January 31, 2011 - link

    VX900 has Chrome9 HC3 graphics core at 250Mhz - the same as VX800. Its 3d performance can barely keep up with the atoms anemic IGP (I could not find ANY review of VX900, just VX800), so totally no match for Brazos (though, in contrast to atom, it should support video decode much better). So your TDP comparison basically ignores the 3d part of it (surely the graphics core won't consume that much given its performance).
    Alright, even if I were to believe the DC Nano has same TDP as the current single core one, traditionally perf/power has never been that good with those VIA chips (not terrible, just not really good). Maybe their TDP definition is different (btw I've never seen a published TDP figure for any of the nano u3xxx series, which you seem to use as reference), also keep in mind runtime of notebooks is barely affected by TDP, much more important if you can get low idle power figures. I have no idea how the VIA platform would compete there (granted the publish idle power of the nano u3xxx cpus is only 100mW), but based on past designs I have to assume not very well (fwiw, this article here doesn't help for that neither, since the atoms don't have all of their mobile siblings power management features enabled).
    So, unless VIA delivers, I remain sceptical if they can be competitive. Yes, a 1.4Ghz Nano DC should be quite competitive with 1.6Ghz Zacate performance wise - maybe also power wise, but 3d graphics will be very very sub par. To catch up in that area the new IGP is needed which as you mentioned is q4 (if the graphic core is like VN1000, it should do quite well, though I'll note that VN1000 plus the required southbridge has a 12W TDP).
    Reply
  • bjacobson - Saturday, January 29, 2011 - link

    I for one will be clicking any links with more info on these nanos! That sorely beat the Brazos out of nowhere hah!

    I doubt the it'll be much good at games though, and the drivers will be rough.,,
    Reply
  • silverblue - Monday, January 31, 2011 - link

    I've seen the Nano in comparison with Brazos before so I knew it was capable of being faster, however it'll take a load more power.

    The last time I saw a proper Nano rundown, we were talking a 65nm chip...
    Reply
  • Iketh - Thursday, January 27, 2011 - link

    power consumption is considerably higher on Nano Reply
  • AmdInside - Thursday, January 27, 2011 - link

    Just curious if it was a fad or are people still buying Atom systems? I bought an Atom netbook but sold it within 30 days cause I couldn't stand it (Dell Mini 10v). Then I bought an ION Zotac Atom 330 system to use as a video streaming device for my bedroom and while I do like it for what it does, I just can't see myself buying another cheap lower powered Atom/AMD E-350 like device. I bought mine because the price made it seem like a great deal but once I got past that, I just lost interest in netbook/low powered mini-ITX platforms. Tablets on the other hand I am still hooked on. Love my iPad and may pick up a second this year and pass mine onto my wife. Reply
  • nitrousoxide - Thursday, January 27, 2011 - link

    What APU/Atom can do is still far beyond ARM-based tablet's reach. Just look at how much superiority x86 have in absolute performance. E-350 has roughly 5 times the performance of Tegra 2, and even an Atom is significantly faster. So tablet is just an alternative of nettops, not a replacement. If you are fine with you iPad that's cool, but saying that nettop is dead is still far too early.
    The superiority of x86 is just unmatched by ARM, and that's why Intel claims that ARM is not a big deal for it. Want to be as fast? Then add more instructions sets, design more complicated architectures, and what you get is no longer an ARM.
    Reply
  • ocre - Friday, January 28, 2011 - link

    hahaha, You need to educate yourself a little more on the advantages/disadvantages of both the x86 and ARM architectures before you go around posting ignorant statements like these. If we assumed your statement: is true: "E-350 has roughly 5 times the performance of Tegra 2" it would still be nothing to brag on becaue the same tegra 2 uses 7-9 times less energy than the E-350. This makes a mere 5 time performance a joke.

    But on top of that, there is no good reasonable way to conclude where these architectures rate towards one another. The software is radically different. There is not enough to go on for any useful conclusion and what little we do have is subject to a very special cases where there is SW that in the end results to similar obtained data. But the entire process is different and while we know the x86 is pretty well optimized, any ARM based counter SW is in its beginnings. X86 has the luxury of optimizations, ARM will only get better when SW engineers find better ways to utilize the system. But this is considering an actual case where there is SW similar enough to even reasonably attempt to measure x86 vs ARM performance. For the most part there the SW of each architecture is so radically different. ARM is extremely good at some things and not so good at others and this doesnt mean it cant be good, in some cases the SW just hasnt matured yet. All this matters little cause at the end of the day, everyone can see that the exact opposite from your statement; The superiority of ARM is just unmatched by x86 when it comes to performance per watt. This is undisputed and x86 has a long long way to go to catch up with arm (and many think it never will). Alls Arm has to do is actually build 18w CPUs, they will be 3 to 4 times more powerful than the e-350 based on the current ARM architecture.
    Reply
  • jollyjugg - Friday, January 28, 2011 - link

    Well your last statement is simply laughable. Because all we have from ARM right now is chips which run in smartphones and tablets. ARM based processors are not exactly known for their performance as much as their power. While it is true that performance/watt is great in ARM today, it also true as any modern microprocessor designer would say that, as you go up the performance chain (by throwing more hardware and getting more giga Hz and other tweaks like wider issue and out of order etc), the mileage you get out of the machine improves so does the drop in performance/watt. There are no free lunches here. I doubt a good performance ARM architecture will be a whole lot different that an x86 architecure. While the lower power is an achilles heal for x86 as is the performance an achilles heal for ARM. You mentioned software maturity. It is laughable to even mention this in this cut throat industry. Intel showed its superiority over AMD by tweaking the free x86 compiler it gave away to developers to suit its x86 architecture compared to its rival and the users got cheated until European commision exposed Intel. So dont even talk about software maturity. The incumbant always has the advantage. ARM first has to kill Intel's OEM muscle and marketing muscle before it can start dominating. Even if it did the former two, there is something it can never do, which is matching intel's manufacturing muscle. Intel by far is way better than even the largest contract manufacturer TSMC whose only task is to manufacture. Reply
  • Shadowmaster625 - Monday, January 31, 2011 - link

    Nonsense. ARM is more optimized that x86. x86 code is always sloppy, because it has always been designed without having to deal with RAM, ROM, and clock constraints. When you code for an ARM device, you are presented with limits that most software engineers never even faced when writing x86 code. When writing software for Windows, 99.9% of developers will tell you they never even think about the amount of RAM they are using. For ARM it was probably 80% 10 years ago. Today it is probably less than 20% of ARM software engineers who would tell you they run into RAM and ROM limitations. With all this smartphone development going on today, ARM devices are getting more sloppy, but still nowhere near as bad as x86. Reply
  • Shadowmaster625 - Monday, January 31, 2011 - link

    Best buy is still littered with them. Literally. Littered. Reply
  • e36Jeff - Thursday, January 27, 2011 - link

    what review were you reading? The only bug that is actually mentioned is the issue with flash, which AMD and Adobe are both aware of and should be fixed in the next iteration of flash. Stop seeing anything from AMD as bad and Intel as good. For where AMD wants this product to compete, this is a fantastic product that Intel has very little to compete with now that they locked out Nvidia from another Ion platform. Reply
  • codedivine - Thursday, January 27, 2011 - link

    Ok one last question. Is it possible to run your VS2008 benchmark on it? Will be appreciated, thanks. Reply
  • Anand Lal Shimpi - Thursday, January 27, 2011 - link

    Running it now, will update with the results :)

    Take care,
    Anand
    Reply
  • Malih - Sunday, January 30, 2011 - link

    I'm with you on this.

    I'm thinking about buying a netbook and may be a couple net tops with E-350, which will mostly be used to code websites, may be some other dev that require IDE (Eclipse, Visual Studio and so on).
    Reply
  • micksh - Thursday, January 27, 2011 - link

    how can it be that "1080i60 works just fine" when it failed all deinterlacing tests? Reply
  • Anand Lal Shimpi - Thursday, January 27, 2011 - link

    It failed the quality tests but it can physically decode the video at full frame rate :)

    Take care,
    Anand
    Reply
  • nitrousoxide - Thursday, January 27, 2011 - link

    That's cool. When Intel owned everything, try buy a Core i5 with $200 as today. Actually, I'm not quite sure if an i3 will cost less than $150.

    And that's what happened during G80 era. nVidia has the best GPU, much much faster while the R600s are craps, ATI's on the edge of extinction. See what you got? $1000 for the flag-ship 8800 Ultra, $600 for a high-end 8800GTX. A decent card may cost you $400 (8800GTS) and even a crappy (though not as slow as Radeon 2600) 8600GT costs you $250.
    Reply
  • Enlightenment777 - Thursday, January 27, 2011 - link

    Rumors, Rumors, Rumors, Reviews, Reviews, Reviews, when can we buy E-350 motherboards from NewEgg? Reply
  • GTaudiophile - Thursday, January 27, 2011 - link

    What are the differences between high- and low-end HTPCs?

    I am thinking of making a combo HTPC/NAS box of sorts...playing the role of both media streamer (online content, mkvs, etc.) as well as a storage box for machines on the network.

    Will this do it? With what OS?
    Reply
  • djfourmoney - Thursday, January 27, 2011 - link

    Go to the AVS Forums and look at Rene's HTPC Guide, the latest version is $10 but well-worth it.

    Low End HTPC especially AMD in recent years rely on IGP for all video decoding. But this equals the performance of a 785/880 based CPU/Chipset combo and cost about 25% less, less than half the power needed and tiny footprint.

    Unless you need expansion slots, as I said this thing can be built for such a low price you should replace your standard def Cable Box with this. Even with a PCIe Tuner Card that read Clear QAM as most do, it will pay for itself (no box rental fee) in about a year if not less.

    You can not buy a i3 for less than the whole board. In fact for the price of a i3, you can get the Asus with Wifi, Bluetooth and Heat Pipe cooling which is going for 141 Euro listed on European sites. It should cost roughly the same in the States some sources have it coming in around $150.

    Even if that's the case if you don't need USB 3.0 or SATA 3.0, the ASRock and Jetway boards will be the cheapest and think should be under $100 street price.

    Too bad AMD's current drivers don't support HDMI Audio for Linux installs. If they ever solve that, then you really can build a sub-$325 with all brand new parts.

    Just to fully answer your question, a high end HTPC depending on your desires, will have a Core i5 or AMD Phenom II CPU and Mid Range Video Card (HD5770) for some post-processing and 7.1 over HDMI. You can do that or a $200 Asus Xonar card, the GPU is cheaper and will do the same thing...

    Cost would be roughly twice that of a Atom/ION or AMD E350 system ($500-$700)

    Would be Windows 7 with Media Center if you want DVR capabilities or you can run XMBC strictly for media playback or Mediaportal.

    As I said Linux drivers for AMD hardware does not support audio over HDMI, so you would have to go with Toslink.
    Reply
  • GTaudiophile - Thursday, January 27, 2011 - link

    Thanks for the reply.

    I meant high- vs. low-end more in terms of capability? What more can I do with the high-end vs. low? Can a Zacate-based HTPC, armed with 2x 2TB HDDs in RAID 1 stream 1080p from say Netflix and play the 720p mkv episodes of TOP GEAR, smoothly? And when not streaming it will be a locally networked storage array.

    You think Win7 is the best? I currently run FreeNAS now for my NAS...why do I feel like Win7 would be some sort of downgrade?
    Reply
  • Anand Lal Shimpi - Thursday, January 27, 2011 - link

    High vs. Low mostly has to do with the responsiveness of the machine more than anything else (for a purely dedicated HTPC). Remember that only video decode is fixed function hardware assisted, everything else (launching apps, navigating around menus, general use stuff) is still run on the CPU core - which in this case may be around the performance of a 2.66GHz Pentium 4 depending on the workload.

    A higher end HTPC can also be multifunction (e.g. do your ripping, transcoding and watching on the same machine).

    Take care,
    Anand
    Reply
  • tecknurd - Friday, January 28, 2011 - link

    I disagree. A low end processor can still be as responsive as a high end processor. The difference between a low end and a high end processor depends on how much work you want to do with the processor. Also it depends on how long it takes a re-encoding or any task to be done. A low processor can do two things if they both consume 50% of CPU usage. One of the tasks could be video playback and other is re-encoding a video. The re-encoding will take longer than a high-end processor.

    Responsiveness is how the user interface software is written. Using a high end processor with a user interface that is slow will not help the responsiveness of the whole setup.

    A example of a low end and a high processor is a computer nerd and muscle builder pounding as many nails in a wood. The muscle builder is able to pound more nails than the computer nerd. Another test is pulling an airliner and again the muscle builder is able to move it. In this example shows that bigger muscles helps to do the heavy lifting, but the efficiency of the smaller muscles of the computer nerd uses less energy although it takes more time.
    Reply
  • zodiacfml - Friday, January 28, 2011 - link

    he refers the system as DVR or HTPC, not just a file server which most will do with Brazos. Reply
  • tecknurd - Friday, January 28, 2011 - link

    Digital audio over HDMI does work in Linux including for AMD hardware, so I do not where you are getting your information. Getting digital audio to work in Linux is a whole different matter all to its self. This goes for any hardware. It is best to use analog output for ease of setup.

    Do not need a $200 Xonar card. A $90 Xonar is all you need although the only times you need this card is its analog audio capabilities.

    An Athlon II is a better buy for a HTPC than a Phenom II.

    The problem using AMD graphic cards in Linux is no video codec hardware acceleration support. Sure there is some, but the short list and its problems makes it has no support. AMD graphic users have to wait until the Xorg team provides VA API or UVD support.

    FYI, ASRock does not have a warranty for end-users, so you have to rely on the store wherever you bought it from. Jetway may seem cheap, but their support is not good.
    Reply
  • djfourmoney - Thursday, January 27, 2011 - link

    I was thinking the same thing -

    You can build a Server with MSI or Gigabyte because those will be the cheapest SATA 3.0 boards. Maybe 12TB (6 SATA ports x 2TB@90 each) plus case and E350, about $800? I found a mITX Server Case that holds

    Get a Sapphire Pure White Fusion which has Bluetooth and 2nd PCIe x1 slot.

    Go Powerline Ethernet, 60GB Kingston SSD, Windows 7 and Windows Home Server with Recorded TV Manager an uTorrent Plug-in.

    Turn your old PC into a Workstation for Ripping Blu-Ray and Encoding.

    Reply
  • djfourmoney - Thursday, January 27, 2011 - link

    You can build a sub-$350 HTPC with this! If you can recycle some parts from any of your other builds you might be able to get it under $300. I built a ASRock based HTPC based on the price expected for that board ($110) and it comes in at $319 before taxes and shipping. Careful shopping might avoid that.

    Run Mediabrowser with TV and GameTime! Plug-ins.

    You can now throw away your Cable Box SD or HD. If you have standard cable, turning in your box and building a HTPC around one of these boards will pay for itself in about a year.

    If you get HD and Premium Channels, hopefully SiliconDust's 3 Tuner CableCard adapter will be out before NFL Training Camp.

    For Direct TV/Dish Network/AT&T U-verse, you'll be able to use Hauppauge Colossus with Component Input, eliminate issues with the HD-DVR USB version. As long as they don't cripple the component output, there's no PQ difference.

    Reply
  • Khato - Thursday, January 27, 2011 - link

    I've been somewhat disappointed with the lack of actual investigation into how changes in memory bandwidth affect this new generation of integrated GPUs - both on Brazos here as well as Sandybridge earlier. The direct comparison to a 5450 here is interesting, but since it wasn't stated I'm guessing those were stock 5450 numbers, not a 5450 underclocked to be the same frequency core/memory?

    The primary reason for it being a point of interest is that the current rumor has Llano at anywhere from 4x to 6x the shader resources, but only 2.4x the potential memory bandwidth. More likely 2x in any actual systems though given that anything above DDR3 1333 carries a decent premium. So if Brazos is already seeing hints of memory bandwidth limitations...
    Reply
  • Anand Lal Shimpi - Thursday, January 27, 2011 - link

    Those were stock 5450 numbers, and you are correct - memory bandwidth is an issue (one of Sandy Bridge's "tricks" is the shared L3 cache, it helps mask memory bandwidth limitations quite well as it is currently used for Z operations among other things). I expect that Llano will be much quicker than the E-350, remember that in many cases we're not necessarily GPU bound but rather CPU bound in these game tests.

    I will continue to play with performance on Brazos but I expect that once I've got Llano in house I'll be able to get a better idea of how bad the memory bandwidth limitations actually are.

    Take care,
    Anand
    Reply
  • Khato - Friday, January 28, 2011 - link

    Thanks for the reply. The CPU vs GPU bound comment actually sparked another point of curiosity - how does the performance picture change as resolution increases?

    There are quite a few games where the performance increase going from integrated to either the 5450 or 5570 is basically the same, implying that it's CPU limited and something about the integrated graphics decreases the performance. The unknown being whether that something is a constant, a slight latency hit due to whatever arbitration scheme is used between CPU and graphics for example, or if it will scale with the load placed upon the GPU, as would be the case for memory bandwidth.

    This certainly has me looking forward to at last getting to see how Llano graphics performance is in a few months. I'd find it all too amusing if the better integration in Sandybridge resulted in graphics performance on par with a memory constrained Llano.
    Reply
  • bjacobson - Saturday, January 29, 2011 - link

    I was thoroughly surprised the e350 coped as well as the dedicated cards using shared RAM.

    Was not expecting that.
    Reply
  • Speed3mon - Thursday, January 27, 2011 - link

    this comment... kinda gay.. sry but true Reply
  • Aloonatic - Friday, January 28, 2011 - link

    I don't think that even the staunchest homophobe would even go as far as to associate being gay with the nonsense that the OP wrote.

    It'[ a shame that there isn't a store that fanboys have to shop in, where their beloved companies can rip them off royally, as that seems to be their want.

    His comment wasn't gay, it was just sad, and ill-conceived, which might be how their parents probably view them too, in hindsight :o)
    Reply
  • etudiant - Thursday, January 27, 2011 - link

    The Zacate die size is comparatively minute at 75mm2, about a quarter the size of the Thuban.
    That translates to perhaps one twentieth of the manufacturing cost, given yields are much better for smaller die. Should be very helpful for AMD if they can deliver in quantity.
    Reply
  • GeorgeH - Thursday, January 27, 2011 - link

    I'm very impressed. Does anyone have an idea when Intel and VIA might respond with updated products? Reply
  • Anand Lal Shimpi - Thursday, January 27, 2011 - link

    Intel's 32nm Atom refresh will appear in Q4 2011.

    Take care,
    Anand
    Reply
  • Silver47 - Friday, January 28, 2011 - link

    Surely by then Anand AMD would be starting to show off the next gen Bobcat?

    Are Intel factoring this in and going to try and compete with the next gen, because if not they would look rather silly having a better performing part for a few months and AMD come along and crash the party?
    Reply
  • Anand Lal Shimpi - Friday, January 28, 2011 - link

    AMD indicated we should see it follow a ~12 month cadence with Brazos and its successors. Assuming perfect execution that would mean we'd see the followup in Q1 2012.

    The next-gen Atom part is just going to run at a faster frequency and have better media functionality (e.g. H.264 decode). I believe Intel is one more generation away from a significant performance boost with Atom.

    Take care,
    Anand
    Reply
  • Tralalak - Friday, January 28, 2011 - link

    VIAl's 40nm Nano X2 (Eden X2) with all-in-one chipset VIA VX900 MSP = VIA EPIA M900 Mini-ITX Q2 2011 (04/2011).

    VIA's 40nm next all-in-one chipset VIA VX MSP with DirectX 11 IGP refresh will appear in Q4 2011.
    Reply
  • SilentSin - Thursday, January 27, 2011 - link

    Don't mean to nitpick as I think you meant 5570 due to the market this board is aiming, but on page 5 when you added a discrete GPU did you use a 5770 or a 5570? The paragraph text switches back and forth, graph shows 5570. Reply
  • Anand Lal Shimpi - Thursday, January 27, 2011 - link

    Thanks for the correction - it's 5570 :)

    Take care,
    Anand
    Reply
  • Silver47 - Friday, January 28, 2011 - link

    "3, 4, 7"

    "5 sir!"

    "5!"
    Reply
  • strikeback03 - Friday, January 28, 2011 - link

    I thought the number of the counting shall be three? Reply
  • jnmfox - Thursday, January 27, 2011 - link

    Would this be powerful enough for HDTV playback with Sage TV?

    "For HDTV Playback: 3Ghz processor or higher or a slower processor in combination with a video card utilizing DXVA support and using a decoder which supports DXVA"
    http://www.sagetv.com/requirements.html?sageSub=tv

    TiA!
    Reply
  • QChronoD - Thursday, January 27, 2011 - link

    I realize that this review was for an ITX board, but what would you guys say are the odds of this chip (or something with equal performance) shipping in the next few months in a thin and light laptop?

    Also could you do a quick test of minecraft? The ION 3D you reviewed the other week was passable and it seems like this is faster, but you never know...
    Reply
  • djfourmoney - Friday, January 28, 2011 - link

    Several were shown with the announcement of the HD6990 - http://www.hardwarezone.com/features/view/131493

    Reply
  • Stuka87 - Thursday, January 27, 2011 - link

    I am happy that the Brazos has turned out to be as good as I hoped it was going to be.

    And its awesome to see AMD hand it to Intel in something :)
    Reply
  • sprockkets - Thursday, January 27, 2011 - link

    obvious troll, is a dumbass Reply
  • zodiacfml - Thursday, January 27, 2011 - link

    Way back, AMD not making an Atom is a painful decision, I think they had Brazos on plans.
    Anyways, what a very nice piece of kit. I have a dual core atom desktop that has unusable 1080 video playback.
    This is the HTPC to get, simple, small, low power, and cheap.

    Regarding power supply, I have a question of design.
    If notebooks and netbooks can get by a DC power adapter, can mini-itx boards use one especially now that many replacement power adapters are available. This is a problem with my atom system since it uses a standard 300w power supply which is inefficient and huge (standard case).
    Reply
  • cyrusfox - Friday, January 28, 2011 - link

    About PSU, yes many low power designs are out there. Just check out PicoPSU
    http://www.mini-box.com/s.nl/sc.8/category.13/.f

    Some premium boards come with the power supply onboard. I hope to see a brazos board come along with the same feature. I would imagine that the cheap 20 pin pico PSU will work fine on these new brazos boards as the draw is so low (will that work Anand, a 20 pin connecter in place of the 24pin connector on this board?) Great thing about these PSU is they are silent(no fans) and very efficient( I have heard 96%). But that isn't counting the conversion lost from AC to DC. You need an AC adapter to go with these supplies, but the AC adapter does not need to output the same amount of watts as the picoPSU. Definitely a solution you should look into.
    Reply
  • zodiacfml - Sunday, January 30, 2011 - link

    cyrusfox, thanks for the link! Reply
  • msroadkill612 - Friday, February 04, 2011 - link

    yeah - me too - have been looking for this for ages

    am also curious about 20 pin thing - my guess is u may have to pass on a pciE card. - any one know?
    Reply
  • Metaluna - Monday, January 31, 2011 - link

    There's also the Antec ISK 100 case that has a Pico-like PSU built in:

    http://www.newegg.com/Product/Product.aspx?Item=N8...
    Reply
  • strikeback03 - Friday, January 28, 2011 - link

    In my carputer the one problem I found was that while a PicoPSU might overall put out enough power to run a board, individual rails might have issues. I was using an Atom330 board and the PSU slowly died, had plenty of 12V power available but findings on the forums indicated that the board overdrew the 5V rail. Reply
  • msroadkill612 - Friday, February 04, 2011 - link

    thanks for the heads up Reply
  • nlr_2000 - Thursday, January 27, 2011 - link

    Brazos does "very again well" against Atom on absolute performance, die size and price.

    Thank you for the review
    Reply
  • DanaG - Friday, January 28, 2011 - link

    One thing I'm wondering: how does Brazos compare to Atom and ARM (such as QNAP TS-219P+) on power usage and network performance? From reviews of ARM devices, it looks like they hit a speed limit that may be caused by the SoC devices themselves.

    I do know that Realtek Gigabit Ethernet sucks, so it'd be worthwhile to benchmark with an Intel NIC and a Broadcom NIC, as well.
    Reply
  • Aone - Friday, January 28, 2011 - link

    I reckon Anand Lal Shimpi should know that neither Atom D510 nor Atom 330 supports EIST or Idle Power State. If he does compare the power consumtion he, at least, must mention it.

    And i'm not sure that Atom platform supports DDR3-1333 as Brazos. Usually Atom sticks with DDR2-667.
    Unfortunately, mem config is absent from description of test configs.
    Reply
  • DanNeely - Friday, January 28, 2011 - link

    DDR3-800 based atoms launched a few months back; the performance difference vs their DDR2 siblings was negligible. Reply
  • krish123 - Friday, January 28, 2011 - link

    Hi Anand,
    You are comparing the gaming performance of Core I5 2500K / higher clocked graphics with E350, but u should also show the power consumption/price difference on those parts. .. I feel ., kind of biased review, when you are trying to compare only the positive side of intel parts and not the negative side

    I will wait for your review on Llano and bulldozer

    Regards,
    Kicha
    Reply
  • 7Enigma - Friday, January 28, 2011 - link

    Did you even read the text or just look at the pretty graphs?

    The i5 is in there as a bracketing agent so that those of us with desktop systems can get a feel for the performance difference to these Atom/APU products. There was never a direct comparison drawn unless you ignore all the text and just stare at the pictures.
    Reply
  • OneArmedScissorB - Friday, January 28, 2011 - link

    Did you even read the pretty graphs? They were very nearly neck and neck in those comparisons. Why omit the most significant differentiation between the two?

    Sandy Bridge and Bobcat as HTPC platforms are more comparable than anything else. They are both at least "good enough" for general use, have similar IGP and video acceleration capabilities, and are both readily available in mini-ITX.

    And yet, there's a 3.1 GHz Athlon II X2 with an 890GX board in the power results, likely full size ATX, in place of every single last one of the numerous mini-ITX platforms they could have just copied and pasted into the chart from past articles. Wtfbbq?!?

    That was not just an oversight, but a very significant mistake that detracts from the usefulness of the entire article.
    Reply
  • bjacobson - Saturday, January 29, 2011 - link

    hm, that seems a bit of an overstatement. I don't really see how the 2500's are at all in the same market as this. Reply
  • code65536 - Friday, January 28, 2011 - link

    I wonder, does Bobcat support hardware-assisted virtualization? Reply
  • silverblue - Friday, January 28, 2011 - link

    I read that it does. Reply
  • macs - Friday, January 28, 2011 - link

    Can someone test the power consumption of an i3 2100 on itx mobo? I suppose it will be quite close to brazos, at idle at least Reply
  • flyck - Friday, January 28, 2011 - link

    at idle the SB cpu uses 30% more power (or 4W) (2400S), during playback the difference between the cpu will also be around 1W.
    the chip consumption is however not known.
    Reply
  • duploxxx - Friday, January 28, 2011 - link

    Nice review, always a good review from Anandtech, thx

    For now brazos could use only another cpu speedbump to totally destroy all atom based solutions and even more get into the ULV regions. A wider APU platform offer (different clockspeeds) would be better though.

    To crush next gen Atom I think AMD knows what to do, update like 69xx series on uarch for gpu, add higher speed or more core in the new 28nm package all combined with turbo modes by the end of this year and Intel will never have a chance unless they bring a new uarch.
    Reply
  • bjacobson - Saturday, January 29, 2011 - link

    meh, they can just compete with the ULV Pentiums.
    To me that's what Brazos is competing with performance wise-- of course it stomps Atom...so what's next? ULV chips...and it has a much harder time with them.
    Reply
  • beginner99 - Friday, January 28, 2011 - link

    ...as mentioned before I do not see the benefit of playing MW 2 at 20 fps on a 1024x768 resolution. Who will do that? probably no one because it is still unplayable.
    I mean this thing is nice and better than Atom (assuming same price) but I think the GPU is a waste of die space. Why not beef up the decode engine? I mean this thing will be used mainly for media stuff and not for gaming. I mean not being able to play 1080p youtube perfectly is already fail for a nettop/htpc because that is a very, very likely usage scenario (1080p tv or screen).
    More fixed function hardware fro media would be better maybe even something like QuickSync so you could actually trans-code on your htpc. Or said otherwise functionality you get with these broadcom cards, which quite a few atom system have.

    I would still 100% choose this over Atom.
    Reply
  • nitrousoxide - Friday, January 28, 2011 - link

    Then it won't be an APU :)

    And AMD doesn't have smaller GPU...HD5400 Cedar is the smallest available design AMD has lol

    While it can't run Modern Wardare 2, I'm quite satisfied with the fact that it can run L4D2, Warcraft 3, Starcraft 2 and I can even play Command and Conquer Red Alert 3 tuned to "medium".
    Reply
  • sebanab - Monday, January 31, 2011 - link

    I also think the 3D functions of the chip might as well not be there at all.
    But they should come in handy for Aero effects in Win7 and why not , Win8.
    Reply
  • Shadowmaster625 - Monday, January 31, 2011 - link

    1080p youtube is just too slow to be practical for anything other than watching trailers for the really big movies that come once a year, like Avatar. Most people's internet connections simply cannot keep up to stream realtime 1080p. And even if you did have the bandwidth, that dont mean that youtube is going to have it.

    720p is much more reasonable, though even then comcast 20mbps + youtube have trouble keeping that pipe filled most of the time. Most of the time I go with 480p just because less wait is more important than the added quality.
    Reply
  • krumme - Friday, January 28, 2011 - link

    I like the review, epecially the new HD bm and the recommandations. Looking at the hd bm themselves one had to wonder if 7200rpm was not enough for the bobcat? but i guess there is room for personal interpretation. I think Anand really likes his ssd more than i do :)

    I still feel the emphasis on multithreaded and heavy workloads is a little to much, and the same for the emphasis on the gaming side. But ofcourse there must be some readers that will use this for for gaming.

    I hope AMD get this bobcat on lower leaking gf 32nm process, so we can have even better battery life.
    Reply
  • bjacobson - Saturday, January 29, 2011 - link

    You know I've found my 160GB 5400RPM drive to be plenty fast for my Atom. I doubt a 7200RPM would be too terribly much of a bottleneck. Might impact battery life though compared with SSD on a laptop.

    I found the multi-threading very useful-- it showed us that even in perfect-scalable applications (Cinebench), the e350 still stomps a dual core dual threaded Atom. That's the conclusive evidence I was looking for that this is a better chip.
    Reply
  • macs - Friday, January 28, 2011 - link

    My thought:

    - AMD E-350 itx board: 100-130 $

    - Intel Sandy Bridge Pentium G620T (rumored release date 2/27/11, dual core, 35w TDP) 70$ paired with a H67 ITX board like the Foxconn one (75$) TOT:145$

    I suspect that we will have similar power cosumption (probably Idle is even better for Intel) but the performance are way better in Sandy Bridge and totally worth the 10/50$ price difference.

    Am I wrong? (sorry for my poor English)
    Reply
  • beginner99 - Friday, January 28, 2011 - link

    nope. had the exact same thought. SB also offers quicksync which could be very useful on a HTPC.
    Plus if you factor in total system cost especially if you use a ssd and a nice rather expensive HTPC case, the difference is not that big anymore.
    Power consumption on desktop is also not a big issue unless you run it as a server 24/7.
    Reply
  • duploxxx - Friday, January 28, 2011 - link

    perhaps before you praise the quicksync that much you might want to check at which cpu it is actually supported, yes yes here is our beloved intel again cutting features for certain lower end parts :D

    http://www.anandtech.com/show/4083/the-sandy-bridg...

    sure cpu of G620 will be better then e350 but GPU won't and the E350 is fully passive and probably most will be right at the 100$ mark while there is only 1 board with that kind of price and the 70$ is bulk 1000 pieces price......

    but sure you can always select the SB in a month or 2 i.s.o. the brazos, just a matter of positioning, my goal is to replace my atom netbook and that will be done very fast.

    never ever will this be same power consumption.... a 35W rated cpu against a 18W get real
    Reply
  • sebanab - Monday, January 31, 2011 - link

    Llano should take care of our confused friends here. Reply
  • silverblue - Monday, January 31, 2011 - link

    It might, but we're still talking a memory bandwidth limitation along with four tweaked Stars cores. For the moment, we don't know how the latter will affect things, but the former certainly presents an issue especially when you think about enabling AA. It'll be a big boost over Brazos, but then again, we expected that. Perhaps it'll have enough grunt to play most modern games with medium and higher details at a good frame rate, but I can't help but feel sceptical. Reply
  • mosu - Sunday, January 30, 2011 - link

    totally wrong Reply
  • silverblue - Monday, January 31, 2011 - link

    Come on... a two word answer isn't an answer. :) Reply
  • Shadowmaster625 - Monday, January 31, 2011 - link

    If intel charges $70 for budget SB 60 days after launch, I'll eat my socks. That goes against everything they stand for. $100, bare minimum. Reply
  • rs2 - Friday, January 28, 2011 - link

    I'm wondering about the suitability of one of these systems for use as a low-volume, low power consumption web server. Obviously performance would not be stellar, but I'm wondering if it would be reasonably acceptable for a low-traffic, non-mission-critical server (for more intense loads I have a quad-core box that I can use, but the power bill associated with running it 24/7 in true server fashion makes me cringe).

    It seems like the E-350 is about as fast as a 3.6 GHz Pentium 4 from way back when. If so then it seems like it would make a passable server (I'm just thinking basic stuff, like Apache httpd, PHP, and MySQL...maybe Tomcat at most), no?
    Reply
  • jjcrandall - Friday, January 28, 2011 - link

    supermicro makes a atom based 1u server, so i imagine that the amd platform does a much better job than that. Reply
  • lordmetroid - Friday, January 28, 2011 - link

    How hot are are these new processors running? Are they possible to passively cool? Reply
  • 7Enigma - Friday, January 28, 2011 - link

    Last line of the second page:

    "ASUS is working on a passively cooled E-350 motherboard which I should have within the next week."
    Reply
  • silverblue - Friday, January 28, 2011 - link

    Aha! I KNEW it!

    I knew somebody would replace the fan in no time at all... :) And losing that fan's power is even better for a system that doesn't drink too much in the first place.
    Reply
  • nitrousoxide - Friday, January 28, 2011 - link

    A fanless HTPC with decent performance...that's what people are expecting for years :)

    However I doubt if it is entirely silent as you still have to use case fans and noisy HDDs :)
    Reply
  • silverblue - Friday, January 28, 2011 - link

    If it's no louder than my Dreamcast, I can certainly live with that. :P Reply
  • cybersax2 - Friday, January 28, 2011 - link

    Nice review!

    I'm a bit surprised that the E350 isn't compared to an Atom D525 + ION2 combo (like the ZOTAC ZBOX ID41 uses).
    Reply
  • silverblue - Friday, January 28, 2011 - link

    He trolls every AMD launch article on here and Toms. Don't worry; unless he decides to actually contribute, he won't be at either place for long. Reply
  • ash9 - Friday, January 28, 2011 - link

    Would be interesting to run a few optimized programs (OpenCL) and benchmark the differences. Hidden value - can it get any better?

    asH
    Reply
  • haplo602 - Friday, January 28, 2011 - link

    Hi Anand

    Seems nobody does answer the following questions:

    1. Is the iGPU disabled automaticaly when a discreete one is plugged in ?
    2. If not, does the system support Hybrid CF (or will it ) ?
    3. Can the iGPU be used as a GPGPU device when a discreet GPU is plugged in and/or the iGPU is not driving a display ?

    thanks ...
    Reply
  • sebanab - Monday, January 31, 2011 - link

    In my opinion the hole discrete GPU side of Zacate should be ignored completely.
    At the moment it does not bing any considerable advantages to add a discrete GPU, and if it will very few people would consider doing it.

    But these questions will nonetheless be very interesting in the Llano review!
    Reply
  • silverblue - Monday, January 31, 2011 - link

    You definitely have a point. In most tests, it really doesn't look as if adding a discrete card would make any real difference, be it a 5570 or a 5870. The Bobcat cores don't look to be capable of handling enough data so perhaps the 6310 was the sweet spot for performance without the whole thing being too CPU-limited.

    Now if they added two more cores, perhaps we'd see a better GPU on-die, but the situation could still be the same. Then there's the matter of the 4x port.

    I suspect the same will go for Llano; AMD knows that Phenom and Athlon II are CPU limited until the highest resolutions for most modern titles so there's little point giving them an overpowered GPU that is underfed. Putting a GPU on Bulldozer could be too expensive for the moment, but it's unlikely that Bulldozer would have issues feeding a relatively powerful graphics card.
    Reply
  • Matt310 - Friday, January 28, 2011 - link

    At home in my desktop machine I have a 4-lane PCI-x hardware RAID card with four 1.5TB drives configured in RAID 5 for speed hand fault tolerance. For a while now I've been wondering how the i/o performance of a good RAID card would fair if it was moved to a low-power (e.g. Atom- or Brazos-based) mini-ITX NAS system. I've been googling for a review of this setup but not been able to find anything conclusive.

    Think a lot of people would appreciate if you guys could give this a try and report on your findings!

    Thanks Anand!
    Reply
  • msroadkill612 - Friday, February 04, 2011 - link

    Well you should know - dont the raid cards have their own smarts & put little work on the cpu?

    My guess is it would be a great unit for a dedicated server. In the old days, many dedicated servers kept their 486 pcS for a decade as they ran fine.

    a post above linked to a comparison of running pciE graphics cards at 1,4,8,16 lanes, & the difference between 4 x &16 by was slight. I imagine its the same with raid cards.
    Reply
  • jjcrandall - Friday, January 28, 2011 - link

    Anand, I didn't see what hdd you used for the tests. It would be very interesting to see how this platform performed in some of the benchmarks with a ssd. I know pcmark & vantage productivity benches use hdd test's to a certain degree.

    This really shows either how poor netburst was as an architecture, or how good existing sub 10w designs are.
    Reply
  • Shadowmaster625 - Friday, January 28, 2011 - link

    It seems like the E-350 was built for overclocking. Especially on the desktop or with any cooling solution that has an extra 10 Watts of thermal headroom. Sooner or later someone is going to release a mini-ITX board that can OC an E-350 up to 2.5GHz and beyond. Right? It's not like the new intel cpu where you cannot overclock it, right? Reply
  • bjacobson - Saturday, January 29, 2011 - link

    yeah I definitely want to overclock it!
    In a laptop when you've got an 8-10 hour battery life, if you know you're only going to need the laptop for 4 hours today, why not just crank up the speed? Great having that option.
    Reply
  • Arnulf - Friday, January 28, 2011 - link

    "Unfortunately in modern titles that's not always enough to have a playable experience, but with older games you should be able to do more with Brazos than you ever could with Atom or even ION for that matter. The CPU/GPU balance in the E-350 is good enough that I feel like Llano could make for a pretty decent value gaming machine."

    Given that Zacate/Brazos performs at the level of 5+ years old midrange desktop CPU with midrange GPU and considering the quoted statement above it would be very interesting to see some actual performance data taken in titles that Zacate/Brazos *can* run at bearable FPS rates.

    You could look up your performance charts from 5 years ago, dig out the old titles and rerun the benchmarks with those titles on Zacate/Brazos. This should give people better idea what they can expect from this generation of APUs. Sub-20 FPS rates in modern titles at rather low resolution and with all details at minimum isn't exactly what people have their sights on when deciding whether to shell out money for their new toy. Comparison with other crappy graphics that is on par with 5 year old solutions may give AMD's platform some edge over competition from Intel's camp, but it utterly meaningless as nobody is planning to use it in such setting.

    Myself, I'd *love* to see some factual data from titles such as CoD2, DiRT, Doom3, etc. - the games you were referring to in the paragraph quoted above and all of which you already have reference benchmarks of (in order to have something to compare Zacate/Brazos against, without having t rerun all the benchmarks with old software titles). Such a benchmark with usable performance and enough eye candy will help put Zacate/Brazos' use in gaming setting into perspective, very much unlike those ~20 FPS tests you did above.

    These merely indicate that integrated graphics of today are still useless for gaming.
    Reply
  • bjacobson - Saturday, January 29, 2011 - link

    Hm that's a good point. Likely I'd be playing a copy of UT2k4, perhaps quake 3 online, or some older online RTS. How do these fare at those games? Reply
  • cjs150 - Friday, January 28, 2011 - link

    Nice review.

    There is a passively cooled E-350 board coming out. The review convinces me that this will be the right choice for the HTPC for the main room.

    Why: TV already connected to an AV receiver so an E-350 system can use the HDMI connection to the AV receiver. E350 more than powerful for BluRay, fine for some light web surfing, ripping music. Maybe stick a TV tuner in as well and that is the main room sorted.

    I have no intention of playing modern games on it so lack of power is not a problem. I have a bigger PC in study to rip movies if it takes too long on the E350.

    But I also have an Atom based system. Again passively cooled. I use it as a server for the home network. It works just fine. It was cheap, ultra easy to build and just works. The fact that Atom is rubbish does not matter because as long as it can shift data round the network I need nothing more.

    So every CPU has a purpose, just with some we have to look hard
    Reply
  • BernardP - Friday, January 28, 2011 - link

    AMD might score technical successes with its new lineup (Bobcat, llano, Bulldozer), but it will have trouble marketing them if only meaningless C Series, E Series, A series (Llano) and whatever Series (Bulldozer) are used.

    last time I looked, a C350 and E350 were Mercedes-Benz models, while CSeries is an upcoming Bombardier jet:

    http://en.wikipedia.org/wiki/Bombardier_CSeries

    In addition the numbers seem meaningless. Even intel's Core naming scheme is clearer than AMD's.

    AMD simply needs 3 distinct marketing names that the general public will understand for its 3 new product lines code names: Bobcat - Llano - Bulldozer. "Fusion" and "Vision" just don't cut it anymore, as they have already been used for years by AMD.

    One can easily see that there is a problem is the naming scheme when OEMs have to use the Zacate and Brazos code names to try to describe the product, like this example:

    http://www.asus.com/product.aspx?P_ID=9BmKhMwWCwqy...
    Reply
  • nitrousoxide - Friday, January 28, 2011 - link

    Yeah...the naming scheme ASUS and ACER use makes people mad when choosing laptops :) Reply
  • StardogChampion - Friday, January 28, 2011 - link

    A low-powered mini-ITX board with 4-5 SATA III ports, maybe a eSATA (Hudson FCH supports 6 x SATA) port and PCI-e slot for adding a SATA controller card (even if it's x4 it's still better than PCI) for $100 has me thinking home server. I don't see a board out there that comes close to these features at this price. If you want a D510 board with 6 x SATA II you have to shell out $200. Take one of these E-350 mini-itx boards, put it in a Lian-Li Q08 case with a bunch of 2TB drives, Amahi/WHS/etc. and you've really got something. It's a waste of GPU but how do you beat it?

    And, for building friends/family/neighbors desktops to replace the big outdated tower sitting on the floor that they just use for surfing, email, MS Office: once this comes out I'll never build another big tower again.
    Reply
  • fr500 - Friday, January 28, 2011 - link

    Using mainstream games for this reviews make little sense. It would be better to show it with some good indie games at 720p, say super meat boy (won't run well on an i3 530's IGP or an 9300-ITX IGP for instance it needs good steady FPS to be playable).

    Maybe Braid, Trine, Torchlight.

    I don't see many people using this for MW2.
    Reply
  • ProDigit - Friday, January 28, 2011 - link

    I would have preferred to see the E-350 compared to an Atom N550,especially with power consumption.
    I believe the N550 beats the crap out of the E-350 (power-wise). After all, I guess AMD wants to put this apu in netbooks, no?
    Reply
  • LeftSide - Friday, January 28, 2011 - link

    Does anybody know if bobcat is going to support ecc ram? I want to build a low power server with ecc ram. Reply
  • msroadkill612 - Friday, February 04, 2011 - link

    I have read it doesnt Reply
  • AmdInside - Friday, January 28, 2011 - link

    Will it run Angry Birds? Reply
  • kenyee - Friday, January 28, 2011 - link

    Wish you touched more on that...that's listed as a weakness, but nothing about whether anything this year will be able to decode 3D bluray movies... Reply
  • bjacobson - Saturday, January 29, 2011 - link

    what are people using OpenCL at this point these days for anyways? Reply
  • spiked_mistborn - Saturday, January 29, 2011 - link

    I agree with you geekfool. Where are the OpenCL or DirectCompute benchmarks (especially OpenCL since that is cross-platform and seems to be building steam)? This article does not feel complete without some type of GPU acceleration test since that is one idea that AMD has been pushing since we first started hearing about fusion. The idea was that the low performance of the cpu core on compute intensive workloads could be compensated for by shifting highly parallel workloads to the SPs. If I remember correctly Intel recently released OpenCL support for their CPUs, and there are also upcoming ARM based SOCs that will have OpenCL support.

    While OpenCL may not matter to everybody today, I think that in a couple of years any devices that are released that don't support it will be skydiving without a parachute.
    Reply
  • Klimax - Saturday, January 29, 2011 - link

    Just few things:
    What board with D510 was used and was it with latest BIOS?

    I have noticed that after new BIOS is loaded on intel boards I get about ~20% increase in write and ~10% in read performance on SSD. Along with few percent on graphics benches. (Tested between October and January)

    Could somebody here test it? (I used BLKD510MO as cheapest D510 board with 1GB DDR2 and Win7 x64 and cheap Kingston SSD)

    This is somewhat missing IMHO.

    Thanks.
    Reply
  • themw6 - Saturday, January 29, 2011 - link

    Hi Anand, as many of us are interested in using zacate in a NAS - could you run a quick truecrypt benchmark? I'm especially interested in AES-256 performance. Additionally, network throughput would be great, too ;)

    Something else: you measured 24 watts idle power consumption. Could you provide more details on your test platform? Others had engineering samples with as little as 11 watts at idle. I'm guessing you used some 850W-rated PSU? ;) Anyway, thanks for the review!
    Reply
  • MrSpadge - Saturday, January 29, 2011 - link

    With such a low clock speed and such low power draw and a desktop platform, which can easily provide power but needs some more performance.. you'd think AMD would have put a proper Turbo mode in there. 1 thread 2.0 GHz, 2 threads 1.6 GHz - easy.

    MrS
    Reply
  • rashire - Saturday, January 29, 2011 - link

    First Thanks to Anand for doing a HTPC review for this chip, I've been waiting on this since your initial comments in Nov.

    But I'm kind of disappointed after reading the review. While it definitely looks like a better alternative to atom+ion i was just expecting a bit more horse power from it.

    I was hoping for a product i was certain would have the power to do downloads, archive extraction and other CPU intensive processes in the background while doing 1080p playback or running an emulator in the foreground. And from the review i just don't see enough power to be sure my media and data-storage processes without potentially maxing out the CPU under certain combinations.

    Looking forward to next gen discrete solution to replace my mid-tower HTPC.
    Reply
  • GullLars - Sunday, January 30, 2011 - link

    I'd like to see a clock for clock comparison of Bobcat, Athlon II, Phenom II, Atom, Core, Nehalem, and Sandy Bridge.
    From what i can see in these graphs, Bobcat performs about 60-90% of Athlon II clock for clock.
    I look forward to seeing overclocking of bobcat :)
    Reply
  • Malih - Sunday, January 30, 2011 - link

    Zacate (board) ~$100
    WiFi ~$25
    small Case ~$75
    PSU ~$50
    4GB RAM ~$50
    widescreen LCD ~$125
    storage ~$100

    around $500 for a good enough complete system, but of course you can always get cheaper case, PSU, storage, LCD and 2GB RAM
    Reply
  • THE_DOM!NAT0R - Monday, January 31, 2011 - link

    The mini-ITX boards based on Brazos are quite good as it seems. They surpass Atom for HTPC's and have more features like more sata ports. I would like to see a comparison with about 5 mini-itx Brazos boards by Anand. Also, if you have a power supply that is rated for max efficiency at about 15-30 W, you can accurately compare the power figures between Brazos boards and Atom. I've seen some reviews and they use power supplies that are rated at about 300W; their lowest reasonable efficiency is 10% load, that isn't designed for 15<P<30. Reply
  • MrMilli - Monday, January 31, 2011 - link

    I would really like to see a comparison against the Athlon II 250u. Reply
  • jamesgor13579 - Monday, January 31, 2011 - link

    Until it can replace my ION/Linux/XBMC system, I am not interested. AMD has been doing well on the Linux front recently, but I don't know about their video decode. Reply
  • Metaluna - Monday, January 31, 2011 - link

    I'm also interested in how these boards perform in Linux (or even if the new GPU is supported at all in current mainstream distros like Ubuntu 10.10). I'm interested in building a small nettop, mostly for web browsing but possibly HTPC use at some point. Reply
  • Jeffk464 - Tuesday, February 01, 2011 - link

    "The good news is 1080i60 works just fine.

    Brazos’ GPU also fails all of our deinterlacing tests,"

    The second sentence follows the first one and seems to directly contradict it. I was planning on buying one of these for a htpc and the ability to play 1080i60 as well as 1080p24 are absolutely critical. 1080i60 is one of the main standards of broadcast(over the air TV) and 1080p24 is the standard for blue ray. So my question then becomes since the author seems to have contradicted himself can brazos handle the interlaced 1080i60 content? This article should have cleared things up not made it more confusing.
    Reply
  • Belard - Tuesday, February 01, 2011 - link

    It would fail at bluRay 3D.

    the 1080i is a bit unimportant since flatscreen TVs don't do interlace.
    Reply
  • doctorpink - Tuesday, February 01, 2011 - link

    if i remember correctly, nano runs 1.8ghz and is much faster but more power hungry...

    are there netbooks with nano anyway?
    Reply
  • grego3d - Tuesday, February 01, 2011 - link

    I'm sold on this Sandy Bridge / Atom killer. Well, maybe not Sandy Bridge killer yet, but i'm sure the second generation fusion processors will be. Now I need help finding a great little mini-itx case. I'd love to build a mini pc the size of my Wii. I have found a few small cases, but where on earth are the slot feed DVD/BD players. I always start my daily reading right here @ Anandtech.com so please save me some time and help us all out by rounding up a case review for this new Fusion platform. Go AMD Fusion - Boo Intel (and your 1$Billion oops!) Reply
  • Belard - Tuesday, February 01, 2011 - link

    Sure is such excitement over such a tiny chip.

    It also reminds me why one of the P4s Dells I use at work is sooooooooooo Slooooooooooooooow.

    I'll admit I'm using my very old computer (AMD X2 3800) as an HTPC somewhat, since it wasn't worth selling when I upgraded. But its a 90watt CPU... so saving power is the big thing, eh?
    Reply
  • lovansoft - Tuesday, February 01, 2011 - link

    I've got an Acer Aspire 5517 with an AMD Athlon tk-42 processor and integrated HD3200 video. It's also 1.6 Ghz so I wanted to showcase a relative clock for clock comparison. It is a 20W chip with 1meg cache built on a 65nm process and no VT. It's not an Athlon II, just an Athlon 64 x2, I believe.

    On Cinebench R10:
    Single Thread: E-350=1174, TK-42=1340 :-: 1174/1340 = 87.6%
    2 Threads: E-350=2251, TK-42=2373 :-: 2251/2373 = 94.4%
    (1174+2251=3425)/(1340+2373=3713) = 92.2%
    Scaling seems better on the new chips than on the older ones.

    At idle with the screen off the laptop pulls about 18 watts. In Cinebench on a single thread it pulls about 30 watts, with 2 threads it pulls about 33 wats. Opening the screen to run the LCD at full brightness adds about 9 watts at any time.

    I ran these tests with a Kill-A-Watt meter. It's not quite an exact comparison, but is pretty close. But to see that they kept performance close, added graphics, and still managed to shave 10% off TDP it's pretty dang impressive.
    Reply
  • lovansoft - Tuesday, February 01, 2011 - link

    Actually, I just let my laptop sit idle for a while. Now, idle power usage dips down to 11 watts with the lid closed and generally stays switching between 11-13 watts. Hmmm, the power usage on these new chips aren't quite as I would expect unless it's a platform thing. This article shows that the new chips pull 9W at full load under Cinebench. My testing shows I ramp from about 12W up to 33W which is a 21W increase by taxing the TK-42, right inline with the 20W spec giving my rounding of numbers. All other parts being equal and I only ramp up 9W instead of 21W then my peak should be about 12W less, or about 21W total instead of 33W total. That would be a significant gain. Interesting that this article has the new platform at 32.2W with the same workload. That's about 50% higher than my rough estimations. Is it because it's a desktop board and not a laptop design? Reply
  • sebanab - Wednesday, February 02, 2011 - link

    Thanks for the performance comparison. I really helps putting the Zacate into perspective.
    On the power consumption comparison:
    Ofcourse the desktop board system will consume more than a laptop with same specs.
    Your laptop consumes less because of different PSU , less USBs , less components in general (PCIex) and so on.
    So in order to make a correct comparison , wait for a HP DM1z review for example.
    Reply
  • lovansoft - Wednesday, February 02, 2011 - link

    I figured as much, but 50% seems high. But then again, it really is only a few watts... Lower efficiency PSU, a couple more chips to provide some extra ports. A couple of watts here and there do add up I suppose. And to think it provides more than 90% of my current performance into only 2/3 the power. I'd think if they can up it to 2Ghz it'd be about the same as mine without tapping the power too much. From the speculation I've read, it makes it seem that the revision coming in a year should grow the performance by quite a bit without really increasing power, about what you'd expect from a die shrink.

    You know, with this architecture, it'd be nice if a board could be made that would have multiples of these chips for server use. From my experience in SMB, I rarely find servers being CPU bound. Usually if they are, then there is some runaway process that needs to be tamed. Maybe this current generation isn't quite fast enough, but with a process shrink and some speed adjustments, getting a few of these on a board would make a very low energy server. But it'd only be feasible if there were something they could use that built in graphics portion for. Otherwise it'd be a waste.

    Oh, and Cinnebench R10 on an AthlonXP 3000+ (2Ghz) = 1438.
    Single Thread: E-350=1174, XP3000=1438 :-: 1174/1438 = 81.6%
    1.6/2=80% Seems to be about the same IPC as the AthlonXP line.
    I don't have power numbers for it, though.
    Reply
  • torkemada - Wednesday, February 02, 2011 - link

    Is E350IA-E45 board HDMI 1.4? Info on the net is confusing. Reply
  • Hrel - Thursday, February 03, 2011 - link

    that VIA chip is pretty impressive. Reply
  • msroadkill612 - Friday, February 04, 2011 - link

    I thought I read somewhere that the GPU shared the L2 cache with the cpu Reply
  • ElRey - Tuesday, February 08, 2011 - link

    Add me to the list of folks wanting to know if this is powerful enough to use as a DVR / HTPC build (ie. not just streaming).

    Is the partial help for encoding in the E-350 good enough? I know the USB tuner I have has pretty high requirements. Maybe there is a better tuner that could go in the PCI-E slot?

    It would be cool to see a HTPC shootout with this, Sandy Bridge i3, and the Athlon II mentioned in the article.
    Reply
  • msroadkill612 - Tuesday, February 08, 2011 - link

    I saw one mobo which had 1 x pcie, 1 pci, and 2 of the little pcie slots, so you do have other options.

    personally i would only consider a a dual tuner card (take care - they also refer to analogue/digital as dual tuner).

    The nature of free to air is that the good shows are always shown at the same time.
    Reply
  • silverblue - Monday, February 21, 2011 - link

    http://www.asus.com/product.aspx?P_ID=qSoDxhM5mAk1...

    £110 inc. VAT in the UK and it's not even passively cooled, nor does it have WiFi.

    The Deluxe model offers DTS Surround Sensation and a couple more USB 3.0 ports, plus comes in Mini ITX as opposed to uATX. Both seem to offer performance as well as power saving settings, and 5 SATAIII ports. I'd expect the Deluxe to be very difficult to find, however.
    Reply
  • misterg - Friday, March 11, 2011 - link

    Quick question: I am considering to use an e-350 brazos as a small and very power efficient system that runs windows 7 in the metering cupboard. Display keyboard mouse will not be needed, other than when installing the system.

    The OS (win7) run from a USB stick
    Some storage in 2.5" disks is added but spun down when not accessed
    PicoPSU or other very efficient PSU
    Low voltage / eco RAM

    What is a realistic power envelope for this when at idle? The 24 watt does not sound really great to me, I'm hoping to be able to reach well below 10 watts..
    Reply
  • FDIV-Bug - Monday, May 09, 2011 - link

    Hey guys!

    I have the same MSI E350IA-E45 board with Windows 7 SP1 x64 installed, but a cant log cpu usage.
    I already tried Core Temp 99.8 (log only one core), AMD Overdrive 3.2.3.0457 (it doesn't even open) and AMD AMD System Monitor Version 1.0.5 (just doesn't log correctly) .

    If you have any suggestions, I'm all ears
    Reply

Log in

Don't have an account? Sign up now