AMD FX-8150 Black Edition Processor Launch Review @ HH

Discussion in 'Reviews & Articles Discussion' started by craig5320, Oct 12, 2011.

Thread Status:
Not open for further replies.
  1. Stuart_Davidson

    Stuart_Davidson Well-Known Member

    Joined:
    Nov 20, 2002
    Messages:
    5,843
    Likes Received:
    188
    Trophy Points:
    73
    Closing the thread temporarily so I can respond to some of the points up to here rather than end up with crossed conversations. Thread will be reopened shortly (within the hour) after I post.
     
  2. Stuart_Davidson

    Stuart_Davidson Well-Known Member

    Joined:
    Nov 20, 2002
    Messages:
    5,843
    Likes Received:
    188
    Trophy Points:
    73
    OK, i'll pick a few of the more serious claims out of the thread for response.

    Firstly regarding bias, pleasing PR people etc. That has never been the case on this site there are various reasons for this but the most simple is that you please one company and by doing that automatically alienate another. So the site would gain nothing through doing this. Additionally claims of bias, pandering to reps are thrown at every site when someone doesnt like (or understand) what they see but I guarantee that the majority of publications have had run-ins with pretty much every major manufacturer but very few have any bias. I can remember unhappy calls from AMD, Intel, NVIDIA, ATI... the list goes on and on. If we (or I) were biased then that wouldn't be the case.

    What I can say is that on the occasions where complaints have come in from PR because they are unhappy with what we say (its their job to be that way) that come the next launch we often (always) hear "yeah, we gave you a hard time but you guys were right that last product wasn't as great as we hoped. This new one is awesome though".

    On the awarding of products. We review products which are sent to us, the majority of the time they are good (what sort of idiotic PR person sends out truly bad products????) but when something isnt good, we say so. This also ties in a little with the accusation of pleasing sponsors. Yes there is a list above of products which received awards (many sites are the same) but I think we can put the argument to bed with a single link:

    http://www.hardwareheaven.com/reviews/1082/pg7/razer-starcraft-2-peripherals-review-conclusion.html

    Razer have adverts running on the site at the moment(depending on your region), and have in the past. Do you think they were pleased with review scores as low as 4/10? No. Do you think we care? No. Do they respect our work and still send us products? Yes.

    Its not just Razer either, there are plenty of articles I could list which prove my point... we were one of the first sites to come out and tell the truth about Lucid tech ( http://www.hardwareheaven.com/articles.php?articleid=147 ) or even on the gaming side where last week Rage got 50/100. You can bet that PR company didn't "like" the score. That said, even when we are positive about a review there are still people who are not happy. Last week I had a great email chain with someone who wasnt happy about an 8/10 when compared to other sites ratings. Not happy with 8/10... madness. Did it cause us to change anything? Nope. Did we get sampled a new product from the same place after this yes.

    I'm sure editorial seems very exciting and full of bias, covert activities and the like from the outside looking in but really its quite simple. As long as testing is thorough and articles complete then everyone gets on with their job. Using AMD, the same company we are accused of pleasing in this thread, as an example we have on numerous occasions slated their stereoscopic 3D tech, would we do that if we were biased?

    Speaking of thoroughness, the quality of testing and game selection. Is it any co-incidence that the games on AMD's list of "approved" titles are some of the best and most recent releases? When selecting *real world* tests to perform we always look at what consumers are doing. In this case I went to the top 10 games of the moment, and our review index, to see what is most approprate to a gamer today. That resulted in Deus EX, F1 and Shogun (offering a range of different styles and engines in the DX11 era). Mixed in with that were a selection of AA/AF and detail settings.

    Remember this is a platform review, not a graphics specific review, so 15games with mutliple resolutions isn't relavent. All of the guys on the editorial side had input in this decision though. What we did instead was cover other aspects which are important to a platform/CPU review. Like how well it compares on PCIe speed... or SATA speed. Show me the huge number of other sites who covered this, an important factor to take into account when choosing which CPU to go for.

    I'd also like to point those questioning our practices to sections of a post which I made a while ago:

    "Regarding our benchmark scores vs other sites. I have posted a number of times before that you cannot compare results from site to site, part of the reason is that the sections being tested will be different, part is that the systems will be different and part is that the quality of work on these other sites. Some will be ok but there are a huge number of sites who publish rubbish. Whether it is recycled results from months ago for some cards to timedemos which mean nothing or even just pure guesswork in some cases.

    All of our tests are performed on a clean OS with the latest patches, drivers, game patches, DirectX etc. Everything is real world gameplay and each test is on one of the most demanding sections of gameplay, not just the 1st level of a game. We take a lot of time to find these sections of gameplay so that our results represent a worst case scenario... and therefore they are normally lower results than other sites or end users but they are far more useful."

    Looking through other reviews of the FX-8150 its absolutely clear to me that people are reusing their old performance graphs (with results obtained on drivers with varied perfomance) in comparison with new results. How anyone can trust that is beyond me... there is one I can see where the FX-8150 has results shown along with the original perfomance of the socket 1155 I5-760. That looks nice and thorough on the surface, comparing against processors from a year or more ago, but do you think you would get the exact same performance from that 760 if you ran the same test today with the updated drivers, BIOS, OS patches and application updates? No, you wouldn't.

    And that doesnt mean the older products always go up in perfomance either. New application updates can for example boost the perfomance of newer models and reduce it on older parts as the new features on current CPUs etc are supported by the application.

    That to me makes the results of other sites which draw conclusions about the FX-8150 against older results invalid and useless and I am happy our conclusion, results or awards dont agree with them.

    One other point on testing practices for now, the comment about droping the resolution/detail and seeing what the processors score... unless you actually game at those settings it is pointless. I can buy a 990X CPU, 24GB of DDR3-2000+ and dual 6990's... get hundreds of FPS in pretty much any game at 800x600 but do you think i'm going to game at that on my 2k worth of components? Absolutely not. We test in scenarios, for the real world testing, that are relavent to the hardware at hand. We test each component used in the review (all of the 2600K results were performed within 24hrs of the FX-8150 results) and everything used was up to date (no timedemo's either).

    Finally, once again on hand picking benchmarks... if we did that for this review we would have included the really obscure tests that AMD recommended for video conversion or test only against the 2500k etc. We didnt, we went with what we felt consumers would use, regardless of what AMD wanted.

    Hopefully that answers some of the nonsense claims from this thread and we can get back to talking about the product itself... one which we fully recommend for those playing the latest games... or for those that want to get involved in some overclocking fun. So yes, it does deserve a "recommended award". It's not a bad CPU, its not the CPU a lot of people wanted but the outrage elsewhere is more a result of (incorrect) expecations being disappointed than anything else. It certainly could do with being *a little* cheaper but in the long run it will serve a lot of consumers well.
     
    blibbax and YnHDaRkMaN like this.
  3. Takaharu

    Takaharu Unus offa, unus iuguolo

    Joined:
    Aug 31, 2009
    Messages:
    3,386
    Likes Received:
    401
    Trophy Points:
    108
    I’m going to have my say in this. I’m not going to comment on the rating or the review per se, just the findings of it.
    Right, first of all the CPU performance was shown using SiSoft Sandra and Cinebench. The FX-8150 fell significantly short of the 2600k. 12 points doesn’t sound a lot but bear in mind that a CPU with similar performance will have only a few points difference and similarly with Cinebench. With both tests, the FX fell notably short in performance compared to the 2600k.
    Secondly, the media encoder benchmarks. There’s 10 seconds between the FX and 2600k, which, as any PC enthusiast will tell you, this is huge. For a PC, 10 seconds is forever. In real life, this is hardly noticeable; it’s not noticeable at all if you’ve just gone from a slower PC to a faster PC. With the dBpoweramp benchmark there’s even more of a gap. We can quite happily see a gap here.
    Sod the CPU usage. If video playback makes that much of a difference to you when choosing a CPU there’s something wrong with the way you choose your processors.
    3DMark 11. Now here’s where we really get into the interesting bits for the gamers. The FX has rivalled the 2600k quite happily here and the 2600k costs around £50 more. Good times for the budget.
    The Sandybridge-beating continues with the Deus Ex benchmark. The FX gets a noteable lead on the 2600k with the framerates. AMD’s gaming lead continues when moving onto the F1 2011 benchmark. The framerate is so close you couldn’t give a damn but on a benchmark it looks good. The same goes for Shogun 2.
    So what have we established up to this point? When gaming using an AMD graphics card the FX-8150 pulls slightly ahead of the 2600k. When doing anything other than playing games the 2600k beats the FX-8150 by a reasonable amount. The PCMark 7 results clarify this, further backed by the benchmarks that follow; when I said anything, I meant anything.

    Who should buy this? Gamers looking for a decent-spec PC who won’t do a lot other than play games.
    Who shouldn’t buy this? Performance enthusiasts, people looking for a cheap workstation or people that want a high-performance all-round machine.

    Last but not least I’d also like to state that I happily increased Veridian’s rep for the post above (well, at least I will do once I can) and I really must insist on stating the blindingly obvious to eblg73: the components that are listed with reviews that score 8/10 or higher are decent components. Strange as it may seem, decent components get decent review scores. In a scale of 1-10 you’re not going to get a lot of difference in reviews’ scores. If HH were to use scores based on 100, which they choose not to, then you’d see more of a difference.
    Right, that’s me done before I lose track and start flaming.
     
    Calliers likes this.
  4. blibbax

    blibbax nahm8

    Joined:
    Sep 6, 2008
    Messages:
    6,210
    Likes Received:
    203
    Trophy Points:
    73
    Incidentally, the AMD CPU's lead in graphics bottlenecked scenarios suggests a platform advantage, rather than that the AMD CPU is better for gaming in itself. A few other reviews have suggested this iirc, even with Nvidia cards. When the bottleneck shifts to the CPU, the Intel platforms tend to pull ahead.

    I'll also give V3 +1 rep for his post above when I can.
     
    synthesis204 likes this.
  5. eblg73

    eblg73 New Member

    Joined:
    Oct 12, 2011
    Messages:
    10
    Likes Received:
    0
    Trophy Points:
    0
    Question : So you will continue to use them insane detail settings in the same games benchmarks for newer CPU tests ? So fully maxing the GPU out so the CPU has got close to zero influence. Seems logic to you, doesn't to me as this proves close to nothing of the CPU capabilities. Looking at other CPU reviews here there were hardly any games involved. And now just that particular selection arises :confused:

    To me it still seems the games that were chosen and the detail levels set were quite clearly picked from the AMD FX reviewers guide. If you can proove me wrong plz do so... I'm not an Intel fan boy I just find it hard to believe that you would select them real world settings after a meeting with your collegue reviewers. :hmm: Correcter game scores, yeah right...

    I think we agree to disagree...
     
    Last edited: Oct 13, 2011
  6. george__

    george__ New Member

    Joined:
    Sep 22, 2011
    Messages:
    14
    Likes Received:
    0
    Trophy Points:
    0
    The review on AnandTech made me think the CPU was getting its butt kicked, your review not so much. Interesting...
     
    Last edited: Oct 13, 2011
  7. Stuart_Davidson

    Stuart_Davidson Well-Known Member

    Joined:
    Nov 20, 2002
    Messages:
    5,843
    Likes Received:
    188
    Trophy Points:
    73
    The tests in future will depend entirely on what gamers are doing and what the CPU is targeted at. If the 2700K is released next week then yes... probably the same tests. If it is released in November, probably not as BF3 (as an example) would be a more relavent test than Deus.

    Also, if the CPU had close to zero effect there would be no score fluctuation. There was which meant that it (or at least the platform) had an impact.

    In terms of the number of games chosen. Depends on a few factors. The target market, the time available, the number of comparison products used. We might for example decide to test 5 comparison CPUs and a new model. That reduces the time available for additonal tests so 1-2 games might suffice. Or equally if something is a budget CPU it might not have a target market for gamers so one game or just 3Dmark would be enough for the consumers who are interested in it. (In this particular case i would also say that the fact we were looking at a platform, including a gfx card, also played a small part in the decision to add an extra game)

    You dont seem to grasp the concept that we tailor our reviews to the products though. Its not a set in stone test suite as tests which apply to one product dont necessarily have any use on another. We're not going to waste peoples time with meaningless tests when we could write about something worthwhile (one of the benefits of not relying on re-using old results over an over).

    An interesting comment... so you stated earlier that you had 5days with a Bulldozer CPU. And now you have access to the reviewers guide also? Says a lot.

    Lets just be clear though... the settings we tested at are not exactly as the AMD sample benches. Looking through the guide they offer. Shogun 2 suggests 4aa/8af (we used 0/0 having already used AA/AF in the other games. Who knows what map they used, or it may have been the timedemo which IS NOT representative of actual gameplay). Deus EX the settings do match one suggestion from 4-5, but thats just coincidence because it happens to be the highest settings you can set the game too (a common setting). F1 2011 isn't even listed as an option in the reviewer guide.

    So of three tests one matches and two don't... hardly "picking" things from the AMD press material.

    As for you finding it hard to believe we would choose those settings after meeting with other reviewers (for everyone else thats the meeting AMD have with press worldwide where they go through slides about the new product). Now its just getting silly... clearly you are clutching at straws, if anything you know that AMD focused mainly on BattleField 3 as a recommended game. Where's that in our review?
     
  8. eblg73

    eblg73 New Member

    Joined:
    Oct 12, 2011
    Messages:
    10
    Likes Received:
    0
    Trophy Points:
    0
    Pretty hard to demo a beta online shooter... no idea how AMD did it, to me it sounds practically impossible to get similar results. Plus they closed the beta on sunday :(

    Gonna give 2011 a spin, you used the in game benchmark ? I don't grasp the fact you only tested IGP on the SB and on BD you pull out these games... Also not run 3DMark 11 at the stock performance level, , but go straight extreme... beats me...
     
    Last edited: Oct 13, 2011
  9. Stuart_Davidson

    Stuart_Davidson Well-Known Member

    Joined:
    Nov 20, 2002
    Messages:
    5,843
    Likes Received:
    188
    Trophy Points:
    73
    Its ok, you just go ahead and gloss over the fact you have just been proven wrong multiple times....

    As for F1, no of course it isn't the inbuilt benchmark.
     
  10. eblg73

    eblg73 New Member

    Joined:
    Oct 12, 2011
    Messages:
    10
    Likes Received:
    0
    Trophy Points:
    0
    On certain levels yes you proved me wrong (just verified the detail settings and indeed some are not as printed on my sheets, sorry for that) At first glance overlooking your test settings looked pretty familiar. ... it remains awkward,... but that's your call how you run your benchmarks... and for the rating it remains... well .... you will never be able to persuade me on that part :)
     
    Last edited: Oct 13, 2011
  11. Stuart_Davidson

    Stuart_Davidson Well-Known Member

    Joined:
    Nov 20, 2002
    Messages:
    5,843
    Likes Received:
    188
    Trophy Points:
    73
    I'd say more than certain levels... but hey, we shall have to agree to disagree...

    I'm sure everyone is looking forward to the link you are about to provide to your own testing?
     
  12. blibbax

    blibbax nahm8

    Joined:
    Sep 6, 2008
    Messages:
    6,210
    Likes Received:
    203
    Trophy Points:
    73
    I love proofreading!
     
  13. MaxP779

    MaxP779 New Member

    Joined:
    Dec 20, 2009
    Messages:
    3
    Likes Received:
    0
    Trophy Points:
    0
    Should tell you something... people go to the trouble of signing up to call the reviewer on his BS.

    From the conclusion:

    "There are two areas where the FX-8150 excells though, those are gaming and overclocking"

    What?! Its absolutely terrible at gaming!! The benchmarks done here are all GPU limited that is bulldozer would be destroyed if proper CPU taxing gaming benches were run, at lower resolution and lower settings. Ill bet none of those CPUs even went over 50% load during the gaming benchmarks, all the load was on the GPU. Proves nothing.
     
  14. Stuart_Davidson

    Stuart_Davidson Well-Known Member

    Joined:
    Nov 20, 2002
    Messages:
    5,843
    Likes Received:
    188
    Trophy Points:
    73
    Tell me Max, what settings do you game at?

    EDIT 2: Your point about GPU limits was already addressed earlier and is irrelevant.
     
  15. mad0314

    mad0314 New Member

    Joined:
    Oct 13, 2011
    Messages:
    2
    Likes Received:
    0
    Trophy Points:
    0
    I joined to post in this thread, but because you respond with rational though instead of the mindless word vomit other spew.

    I have seen several people insist that the FX 81X0 is a good CPU for gaming, and one of the sites that are regularly linked is this one.

    I have the same problem with the gaming section of the article as stated before, and I know you addressed it already but let me try to approach it from a different view. You say that this CPU excells at gaming. That would imply that it does good compared to other processors. But you can not tell the performance of a CPU in a game when the CPU is not allowed to reach its maximum performance. That would be like trying to see which of 2 cars is faster, but you can only go the speed limit. I understand that you do the benchmark based on "real world" settings, but you can't make that statement from that test.

    I feel it would add a lot more value to your articles if you added a lower resolution run of the same game to test the performance of the CPU (just because you run a test at "real world" settings doesn't mean you can't run the test at more than 1 setting, and it adds a lot of value to the article IMO). No, people will mostly not be playing at these lower resolutions, but I think it is still valuable data. I mean, if you have 2 CPUs - A and B - and they both get the same FPS at the highest graphics settings, lets say 40, but when you take out the GPU bottleneck CPU A gets 90 FPS and CPU B gets 45 FPS, would you still say it "excells at gaming?" No, it will hardly be relevant for much longer as more demanding games come out. These tests are to measure the power of the CPU in these games, and when you only include a GPU bottlenecked test, people take it the wrong way.

    I would especially love to see these tests done since you stated you use a clean OS with the latest drivers, BIOS, etc. I know what you mean about cross site tests not being comparable, as I saw at least 1 site using an older BIOS (forgot which version exactly but I'm pretty sure it was 1 version older).
     
  16. Stuart_Davidson

    Stuart_Davidson Well-Known Member

    Joined:
    Nov 20, 2002
    Messages:
    5,843
    Likes Received:
    188
    Trophy Points:
    73
    In a way I get where you are coming from and in the past we actually used to do the whole lower res, settings test in CPU reviews but the problem still stands about that not reflecting what gamers actually see in gaming. There is also an issue that often you begin running into other limitations which skew the results.

    We could for example use Left 4 Dead, Portal 2, or any source engine game because they are popular and run at 800x600 with min detail. The CPUs give us FPS through the roof but get to the 300fps cap that is on the engine.

    We are always going to go with a real world snapshot of what todays performance is, i find it far more valuable to the user than anything else. Even if there is the chance that a couple of years down the line there could be a variation in gaming performance (not too likely as engines dont advance that much, sadly).
     
  17. mad0314

    mad0314 New Member

    Joined:
    Oct 13, 2011
    Messages:
    2
    Likes Received:
    0
    Trophy Points:
    0
    Would you consider stating what the CPU usage was during those tests at least? It can be an indication of headroom on the CPU.
     
  18. Stuart_Davidson

    Stuart_Davidson Well-Known Member

    Joined:
    Nov 20, 2002
    Messages:
    5,843
    Likes Received:
    188
    Trophy Points:
    73
    That might be something we add in the future, yes.
     
  19. eblg73

    eblg73 New Member

    Joined:
    Oct 12, 2011
    Messages:
    10
    Likes Received:
    0
    Trophy Points:
    0
    Think if you snoop around eg here at eg nordichardware they test a lower res too with medium detail and then a similar high detail/res test you guys ran. The latter maxing out the used ATI 6970. In the second graphs scores are all alike across the CPUs.(with the exception of a few dual cores) Coz it's pure GPU/platform bound. CPU has got very little influence in the equation. Bench Alien vs Predator in DX11, no cpu will give you a benefit as the output scores will be all quite similar. My point remains : As a reader you can't take much conclusions if you only list the latter slide as it's the GPU that is on its knees. It might be real world scenario, but it will not lead to a point to say : upgrade as this thing rocks in games. Tells hardly anything specific about the cpu, more about the GPU and the whole platform around it.

    I'll try F2011 on the SB, GT and FX with the GTX580. Keep you posted
     
    Last edited: Oct 13, 2011
  20. Trusteft

    Trusteft HH's Asteroids' Dominator

    Joined:
    Nov 2, 2004
    Messages:
    23,875
    Likes Received:
    3,648
    Trophy Points:
    153
    I have zero interest in a CPU right now, or to get into this discussion, but I will ask this. As a gamer and hypothetical potential buyer of a CPU for gaming, let's say I have a system and CPU even older than the one I have now. What exactly would a test/review/benchmark of a game run with 2 different CPUs, at 800x600 at 270fps vs 295 provide me? That the one at 295 is faster and I can get an extra 25 fps running game X at 800x600.... What does this mean to me? Does it mean the CPU with the lower frame rate is too slow? Not fast enough? That the faster one is superb? How do I know that neither of them is not fast enough? I know I am not going to play anything at 800x600 that is demanding enough to matter. So, I will need a game and resolution at high enough/modern enough to tell me something useful without having to go back to more and more reviews and benchmarks to try to GUESS how they compare to the old ones (both CPUs and games) and if either of them is what I am looking for.

    I rather have more modern games and at resolutions I am going to use. Then see which of the two (or more) CPUs provide the best solution for me.

    I don't know, it's just makes sense to me.
     
Thread Status:
Not open for further replies.

Share This Page

visited