The Notebook Review forums were hosted by TechTarget, who shut down them down on January 31, 2022. This static read-only archive was pulled by NBR forum users between January 20 and January 31, 2022, in an effort to make sure that the valuable technical information that had been posted on the forums is preserved. For current discussions, many NBR forum users moved over to NotebookTalk.net after the shutdown.
Problems? See this thread at archive.org.
← Previous page

    A10 cpu performance

    Discussion in 'Hardware Components and Aftermarket Upgrades' started by Soul0Reaper, Nov 28, 2012.

  1. tilleroftheearth

    tilleroftheearth Wisdom listens quietly...

    Reputations:
    5,398
    Messages:
    12,692
    Likes Received:
    2,717
    Trophy Points:
    631
    HTWingNut, thanks.

    Hopefully this will put this topic to rest now? (I don't have any hope... ;) ).

    Even in a gaming scenario, a GPU is an add-on to the CPU+RAM combo/platform that does the WORK of the system. If there is a platform that is capable of more WORK, then every component included in that system from the sound card, network cards, video cards and yes, even the storage subsystem (whether they are HDD's or SSD's or Hybrid's) will benefit from the platform that can do more WORK.

    In some metrics doubling the raw Compute performance will only give a small increase in performance (if at all) of the components mentioned above - but either way: it won't be less. In some (as your tests have proven so well...) a 20% difference in REAL WORLD USE can be seen and easily appreciated.

    This is huge. It's like buying tomorrow's performance today (and for only ~$300 too - according to the numbers thrown around in this thread...). To be used and enjoyed every single day for the next XXX days of ownership. To me, that's priceless.

    While I have some clients that value their $$ more than their time - I also have more clients that see the benefit of spending 20-30% more now and enjoying a level of computing (continuously, for the next half decade or more...) that far exceeds buying one or two more systems in that same time frame (when the setup, testing and verification steps are taken into account).

    While I can appreciate the notion 'buy what you need, not what you want', with computers we NEED as much (performance) as possible - ALWAYS. We will never need less. Not even Joe's grandma who Skype's with her great granddaughter once a month.

    Buy within a budget (I believe in those), buy without getting in debt, buy to get work/gaming done 'now'.

    But don't buy a computer with blinders on. The CPU+RAM is what makes a computer do work (period). RAM (maxed out...) allows the Processor to get out of it's own way (as much as it can...).

    Although up to 20% doesn't seem like much improvement for over 50% more 'power' (actually; clock speed) - I would be interested to know what the MINIMUM FPS rates were for each.

    Also; was this run with 16GB RAM for all tests? LOL... I would bet that this would make a huge difference too (with everyone's bleating recommendation that 4GB RAM is ENOUGH!) :)
     
  2. Karamazovmm

    Karamazovmm Overthinking? Always!

    Reputations:
    2,365
    Messages:
    9,422
    Likes Received:
    200
    Trophy Points:
    231
    I dont think that you can read whats in the parenthesis, 10fps. And that was arranged when I used a clevo number, thus with enduro issues at the time that the number was generated.

    Anyway as I said, I dont care about what you imagine, I care about numbers, show me those.

    It is a bottleneck.
     
  3. miro_gt

    miro_gt Notebook Deity

    Reputations:
    433
    Messages:
    1,748
    Likes Received:
    4
    Trophy Points:
    56
    and I think you have quite an issue with your bottleneck. That, besides your ignorant posts about what you care and what you dont.

    the big difference in regards with this thread between you and me is that I think the combination of A10 + 7970 is fine, and you think it's a bottleneck thus is not fine.

    - will the top i7 right now get out more of what 7970 has to offer - probably. But in the same matter two years from now another new i7 will get even more out of the 7970, so in this case will you consider the current i7 a bottleneck ? ... or you want me to come back with numbers ... LOL.

    "bottleneck" is a very general term, as it could be a tiny bottleneck, or it could be a huge bottleneck. That's why I asked you what do you understand by that, but I see for you bottleneck is just bottleneck and that's it. Well OK, it's your opinion. Nonetheless I can probably show you tiny graphics performance difference only by changing RAM modules in the computer and I can see that this for you will be a bottleneck.

    mehh ...

    on top of that computer performance is highly software dependent, i.e. some programs will run super fast while others will bog down on the same hardware, which is due to how those were coded. Is that mean that those that run slower would imply that there is hardware bottleneck?

    anyways we'll see what the GX60 owners have to say, at least they are those that spent the money for the product. I wont be one of them as 15" laptop is too big for me.
     
  4. Karamazovmm

    Karamazovmm Overthinking? Always!

    Reputations:
    2,365
    Messages:
    9,422
    Likes Received:
    200
    Trophy Points:
    231
    I actually just posted the number difference between an enduro troubled 7970m is actually 10fps more than the gx60, I really cant understand whats your point, because you dont have one, you dont have any numers to corroborate your lack of thesis, or idea or whatever, and you just keeping back at something that I never said, which in your mind basically goes that I said that the gx60 is a bad buy, which I never said that in all the threads that I talked about this model.

    I said, as per the OP request that I do have numbers (which you just didnt bother to even look, and you should), and that they do bottleneck the 7970m, which they do, and really its not even fair to compare an i7 clocked at 2ghz because the arch is just better than what trinity is.

    And really going for the software thing right now, is just totally uncalled for, you know we are just talking about gaming performance, you have to accept that we have some awfully coded games and thats that. There is no point in arguing, "you guys should have 8 threads going on all the time", its not going to happen, just go with it.

    basically what you are actually trying to argue, is just a failure all the way through, btw have you actually saw a number on the gx60 lounge aside the awful scores from meaker? no you havent, Im still trying to see if the drivers are going to make a difference, which apparently arent.

    Still, just dont give me that crap of some games are actually cpu bottlenecked, because i posted at least 7 games right there, and you havent actually proved anything, just that you dont know what you are talking about. Which you dont.

    Im really a patient man, but you just crossed the line.
     
  5. Atom Ant

    Atom Ant Hello, here I go again

    Reputations:
    1,340
    Messages:
    1,497
    Likes Received:
    272
    Trophy Points:
    101
    All right guys, let's see some official tests what CPU performance matters for gaming in general;

    Shogun 2
    [​IMG]
    As you can see CPU performance ain't affect anything


    Battlefield 3
    [​IMG]
    CPU performance ain't affect anything


    Skyrim
    [​IMG]
    CPU performance ain't affect anything


    Dirt Shodown
    [​IMG]
    Core counts make some difference, but core speed won't matters


    The funny thing you try to compare quad core A10 performance with twin core i5 based on CPU benchmarks, but otherwise you do not have idea how is A10 performing under games. Than you make consequences A10 bottleneck ha ha ha... :D

    The only real MSI GX60 tests we have are here;
    Shogun 2;
    [​IMG]

    And Sleepingdogs;
    [​IMG]

    And if you want CPU bottleneck, run games in 800X600 resolution near low details and no matters what kind of CPU you have you will be bottlenecked, I guarantee:thumbsup:.
     
  6. Karamazovmm

    Karamazovmm Overthinking? Always!

    Reputations:
    2,365
    Messages:
    9,422
    Likes Received:
    200
    Trophy Points:
    231
    you know I have explained already how that shogun2 benchie is wrong in so many ways. You see the FX line is just about 20% faster than trinity for desktops, still it doesnt even get close to what intel provides on a high clocked cpu, really thats just garbage right there.

    The strange thing is that you dont recognize the benchies on the reviews that were done on the gx60, still defending things with a clear amd game. When a sli 680m gives you lesser fps than a single 7970m, there is really a bias right there. And yes Im talking about sleeping dogs.

    cpu performance actually affects gaming, your bias towards amd is actually well known. otherwise we wouldnt be having this argument, the argument would be is it worth a 30-50% increase on performance if you pay 300 bucks more? which we never engage.

    And besides why are you comparing between the intel cpus? shouldnt you just compare to the trinity ones? even desktop ones would still bottleneck a card that is actually slower than a 7970m, I cant see your point, I only see blindness.

    As I said I have opteron cpus to do some very multithreaded work (actually 4 right now, just bought a new mobo), but trinity isnt about performance, the FX line or the server line is.
     
  7. HTWingNut

    HTWingNut Potato

    Reputations:
    21,580
    Messages:
    35,370
    Likes Received:
    9,877
    Trophy Points:
    931
    That BF3 bench is total bunk. It's likely the single player intro and no multiplayer. I have plenty of evidence showing FPS drops drastically with either a dual core or slower CPU in BF3 multiplayer, like 30-50%. Skyrim is universally known to be strongly GPU bottlenecked. I can't speak for Shogun 2 because I never played it nor looked at it.

    Like I stated before, anyone with an Intel CPU can just download Throttlestop, and run some benchmarks or play games yourself and record FPS. It's clear as a sunny day the impact that a slower CPU has.

    The other thing that bothers me is this notion of running 1080p Ultra settings, yadda. How long do you think you'll be able to run Ultra settings? Current gen games are already borderline playable with everything maxed out. You can force it to a GPU limit, but to what end? So you can play at 25FPS? Or just reduce detail a little to get 40FPS?

    Yes ideally it would be nice to have the actual AMD gaming results, but we don't. So I also feel it is a fair assessment to use Intel results. If CPU benches are roughly equal it is a fair assessment to make that the AMD will perform similarly. A10 is roughly equal to i5-3210m. And I highly doubt that AMD has some supercompute instruction for games that improves performance by 50% or more. That would be simply unprecedented.

    Edit: Just ran Dirt 3 with CPU at 2.0GHz and 3.1GHz and results are clear: 66.8 FPS vs 83.2 FPS using ultra preset at 1080p.

    Downloading Sleeping Dogs extreme texture pack now then will do the same.

    Edit: Sleeping dogs:

    Extreme settings:
    2.0GHz 29.0 FPS
    3.1GHz 30.0 FPS

    Normal Settings:
    2.0GHz 67.4 FPS
    3.1GHz 88.7 FPS

    At Extreme settings CPU does barely make a dent, but it's barely playable framerates. Run at Normal you get a 30% FPS improvement due to CPU and at a playable framerate.
     
  8. Atom Ant

    Atom Ant Hello, here I go again

    Reputations:
    1,340
    Messages:
    1,497
    Likes Received:
    272
    Trophy Points:
    101
    Ok, we know how A10 perform in CPU synthetic benches, video encoding, but how do you know if the same thing is true for games?
    For example there was Dirt showdown above in my post, where you could see only core counts make difference. So where would you put the Quad core A10 on that chart? Closer to i5 or i7? Because tell the truth I have no idea, and doubt if you know without testing...

    Changing CPU frequency by throttlestop, maybe not the same as a factory lower frequency, but fully operational CPU. Especially seeing the 4 benchmarks (that I posted above) where CPU frequency ain't do any difference.

    " How long do you think you'll be able to run Ultra settings?"

    As long as GPU enough for that, than need higher GPU ;).
     
  9. baii

    baii Sone

    Reputations:
    1,420
    Messages:
    3,925
    Likes Received:
    201
    Trophy Points:
    131
    Well, lowering clock speed is not exactly a fair test consider most game are 2 or 4 thread optimized.
     
  10. HTWingNut

    HTWingNut Potato

    Reputations:
    21,580
    Messages:
    35,370
    Likes Received:
    9,877
    Trophy Points:
    931
    Why isn't it a fair assessment? its just showing that slower CPU affects GPU performance which was started previously that it didn't.

    Sent from my YP-G70 using Tapatalk 2
     
  11. baii

    baii Sone

    Reputations:
    1,420
    Messages:
    3,925
    Likes Received:
    201
    Trophy Points:
    131
    The comparison was A10 (which is said to be close to a i5) vs i7 quad. Depending on how many threads a game utilize, a i5 may not be slower then i7 at all. Capping clock speed is not a fair way to mimics architecture difference.
     
  12. Charles P. Jefferies

    Charles P. Jefferies Lead Moderator Super Moderator

    Reputations:
    22,339
    Messages:
    36,639
    Likes Received:
    5,075
    Trophy Points:
    931
    I reviewed a notebook with an AMD A6 dual-core a while back and it was fine for general usage. I included some benchmark numbers in the review:
    HP ENVY Sleekbook 6z: Performance
     
  13. Atom Ant

    Atom Ant Hello, here I go again

    Reputations:
    1,340
    Messages:
    1,497
    Likes Received:
    272
    Trophy Points:
    101
    So why do not you buy an MSI GX60 to play with it and test? You already have the Intel + 680M for making competition and I'm sure could be interesting test for many people.
     
  14. HTWingNut

    HTWingNut Potato

    Reputations:
    21,580
    Messages:
    35,370
    Likes Received:
    9,877
    Trophy Points:
    931
    Send me the laptop or the money and I'll be happy to.
     
  15. HTWingNut

    HTWingNut Potato

    Reputations:
    21,580
    Messages:
    35,370
    Likes Received:
    9,877
    Trophy Points:
    931
    No, but the point is that if CPU benchmarks are approximately equal, it is safe to say that the gaming performance is roughly equal. Tell me what you think is so different about the Trinity architecture that could make up for a 30-50% improvement in FPS?
     
  16. baii

    baii Sone

    Reputations:
    1,420
    Messages:
    3,925
    Likes Received:
    201
    Trophy Points:
    131
    The point I trying to make is that i7 power come from the fact that it is 4c/8t, a i7 quad running 2c/4t(say software restriction) with same clock speed will perform the same as a i5. The proper way to get a i7 to run like a i5 is to disable 2 cores, not reduce the clock speed.
    multithread benchmark does not translate in console ports and badly optimize engines.
     
  17. HTWingNut

    HTWingNut Potato

    Reputations:
    21,580
    Messages:
    35,370
    Likes Received:
    9,877
    Trophy Points:
    931
    I wasn't trying to get it to perform like an i5. I was just showing the relationship of clock speed to GPU performance, period, nothing more, whether 1 core, 2 cores, 4 cores or 8 cores! If an Intel quad core CPU @ 2GHz in in every way faster than an AMD A10 in every benchmark whether single threaded or a billion, and it has an impact on GPU performance, then it can be concluded that the A10 will also hinder performance. It's not a hard relationship to understand.

    I'll be happy to throw in my i5-3360m CPU in my Sager and produce some benchmark results at 2GHz and 3GHz as well. But results will be the same, if not worse performance because of only 2 cores. And Hytherthreading, except in very unique circumstances, only offers a 5-10% improvement in general performance AT BEST. If I could turn off hyperthreading I would. The A10 is not even a true quad core even either.
     
  18. Link4

    Link4 Notebook Deity

    Reputations:
    551
    Messages:
    709
    Likes Received:
    168
    Trophy Points:
    56
    So much nonsense in this thread, we know the A10 won't be as fast as an i7 Quad. Lets not forget that this is a Notebook comparison, if you want the best possible performance, you buy a desktop workstation. Don't compare things that come in different sizes and weigh much more, these are laptops not desktops. Always take into account power efficiency/battery life and heat when it comes to notebooks. Just don't run things in low resolution and then make a big deal of a few more fps when the native resolution of the display is 1080P. Don't put the CPU intensive settings to max and keep the GPU intensive settings to minimum just to get at least a sizable difference that won't even matter in real use cases. And please don't say something is way faster just because it performs 30% better when both cases get over 60fps, it just doesn't matter, your screen doesn't support anything above 60Hz. And good luck seeing a difference when minimum fps is over 30, human brain can't decode that much data. And don't say you have fast reflexes, because human reaction time for vision (190ms) is even slower than that for sound (160ms) (ever wondered why they use gunshot to start printing events), so by the time you see a frame and react to it about 5 other frames have already passed.
     
  19. miro_gt

    miro_gt Notebook Deity

    Reputations:
    433
    Messages:
    1,748
    Likes Received:
    4
    Trophy Points:
    56
    looks like everybody is pretty happy with the GX60, including actual owners, only you went trolling at page 15-16 there ... just like you do here

    but then what do I know ... at this point enough to not waste time writing 3 pages reply of what you'll apparently not understand anyways, as it seems after all said by now. Consider this as your next avatar:
    http://www.attunity.com/sites/default/files/bottleneck-2.png


    - sorry OP for the way this thread turned out, by all means get the GX60, you'll be happy with it.
     
  20. tilleroftheearth

    tilleroftheearth Wisdom listens quietly...

    Reputations:
    5,398
    Messages:
    12,692
    Likes Received:
    2,717
    Trophy Points:
    631
    Link4 you are so wrong...

    Average response time doesn't mean squat when talking about individual response metrics. If anything over 60Hz means so little - why do 120Hz LED monitors look so good?

    If you can't tell what the point of this thread is... why bother responding with nonsense of your own?
     
  21. Fat Dragon

    Fat Dragon Just this guy, you know?

    Reputations:
    1,736
    Messages:
    2,110
    Likes Received:
    305
    Trophy Points:
    101
    You're saying people are happy with their machines. He's saying the APU bottlenecks the 7970m performance. Those are not mutually-exclusive statements; the 7970m's power is enough to make most people very happy even when it's gimped by a slower processor. Karamazovmm has admitted many times that it's still a fast machine and most owners are thrilled with the performance. He's accepted your point. If you can accept his point that the system is indeed bottlenecked by the APU, I think we can all move on. This isn't a radical assertion - almost any system with more than one component is going to be bottlenecked in some way. It's unusual for the processor to be a bottleneck on the GPU and not the other way around, but it is in this case and so be it.

    Summary: It's fast. It would be faster with an i7. It would be more expensive with an i7. fin
     
  22. Link4

    Link4 Notebook Deity

    Reputations:
    551
    Messages:
    709
    Likes Received:
    168
    Trophy Points:
    56
    You don't get my point, even though 120Hz monitor may look smoother because a repeated frame on it is shorter than on a 60Hz monitor, you can only get those on expensive stereoscopic 3D notebooks, and since most screens have a 60Hz refresh rate those extra frames go to waste.
     
  23. daviepubman

    daviepubman Notebook Enthusiast

    Reputations:
    0
    Messages:
    16
    Likes Received:
    0
    Trophy Points:
    5
    In games and bechmarks noticeably so. In Windows Media Player and Excel, not ;)
     
  24. LamaNr2

    LamaNr2 Newbie

    Reputations:
    0
    Messages:
    1
    Likes Received:
    0
    Trophy Points:
    5
    Folks,
    would u go for i7 3630 + 670mx? It's 300$ more than gx60, or it's not worth it?
    I am a huge fan of hearts of iron and europa universalis type of games, and I'm worried a10 4600m might be to low, since I, believe, that type of games depend on cpu mostly. Or I'm just plain wrong.
     
  25. Karamazovmm

    Karamazovmm Overthinking? Always!

    Reputations:
    2,365
    Messages:
    9,422
    Likes Received:
    200
    Trophy Points:
    231
    well I can play EUIII perfectly on a mbp 2011. HOI is another mather so many performance hits because its coded so poorly. And for 300 more you are really entering the territory of the 7970m or 680m + i7.

    have you checked the gt60 barebones with the 680m? its actually at 1540

    and depend if you are a real strategy fan really look elsewhere
     
  26. nipsen

    nipsen Notebook Ditty

    Reputations:
    694
    Messages:
    1,686
    Likes Received:
    131
    Trophy Points:
    81
    Yes.

    No, wait.. of course, what I mean to say is.. "Oh, no, of course you might be /happy/ with your intel i7/670m combo, I'm not going to dispute that. The feeling of happiness is completely subjective. And there's no doubt in my mind that people say you need an underclocked i7 quad-core to run a game that runs without a hitch on a lower range dual-core. And for all I know, they may be right. But because the happiness factor is so important, you should certainly consider buying the combo that technically speaking has a lower range of graphics power, and that costs more. Because processing power, I hear, is really the only thing that really matters in a game. Of that there can be no doubt, and now I will waffle on about it for several paragraphs, while pulling in evidence that I will invariably interpret to fit my argument, regardless of what it actually shows".
     
  27. Link4

    Link4 Notebook Deity

    Reputations:
    551
    Messages:
    709
    Likes Received:
    168
    Trophy Points:
    56
  28. Karamazovmm

    Karamazovmm Overthinking? Always!

    Reputations:
    2,365
    Messages:
    9,422
    Likes Received:
    200
    Trophy Points:
    231
    thats very interesting! thanks for the link!
     
  29. Silverfern

    Silverfern Notebook Deity

    Reputations:
    96
    Messages:
    955
    Likes Received:
    1
    Trophy Points:
    31
    i am not sure about this, but i think the A10 mobile chip performs at similar level as the intel T9500 from 2008?
     
  30. nipsen

    nipsen Notebook Ditty

    Reputations:
    694
    Messages:
    1,686
    Likes Received:
    131
    Trophy Points:
    81
    ..honestly, this is going to entertain me for months.

    See, it's extremely simple. The games that have cpu-bound threads - and have significant performance differences when running on the a10 2.4Ghz and the i7's 3.1-3.3Ghz - are few and far between. Very likely none of the games in the test run at Anandtech actually are in that category.

    Instead you're going to see framerate drops like that when the bus-speed is lower than max (the a10 dynamically clocks up and down - not really a good scheme for gaming on a power-supply). So if you set the a10 locked to the highest "p-state" when running the laptop, you would very likely not see those gaps. There's bound to be a higher max fps on lower detail, that's a given. Conversely, you would be very likely to actually get higher lower fps on the a10/7970m combo in crossfire in some cases where the game is gpu-bound. Which.. it will be when turning the detail and AA up.

    And any other game that isn't so sensitive to clock-speeds, and agrees more with multicore setups - and you're not going to notice anything. Like the test shows - 0% difference. But that the other gaps can be smoothed over by just setting the processor to the highest p-state -- it's never going to be discovered, I guarantee you.

    Mystical results, it's going to be, and AT are going to find a new category of games that have miraculously been optimized to run on ATI-drivers on AMD processors.

    Or maybe we'll be treated to something even more fantastic? Only time will tell! :)
     
← Previous page