FX series CPUs and Gaming in 2015 and beyond

[cowboy44mag]

I'm posting this and asking the question why do so many "experts" think that FX CPUs can't game?

I have seen many threads of people asking for expert advice and not really getting expert advice so I am posting this to basically ask why. I have seen people who already have a good AMD AM3+ motherboard asking for a good upgrade path and being told to go to i3 Intel. I read that and think what in the world are they thinking??

Lets look at a few facts about PC gaming. Most PC games are developed for console and ported from console to PC as consoles sell vastly more copies of said game than PC does. So for the past 8 years or so games were made for the Xbox 360 and ran on a maximum of 3 cores, however the vast majority of the games made in the last 8 years ran on 2 cores. That made the iCore the legend that it is in the gaming community, as Intel computers have very powerful core per core performance. If Intel has a weakness it is only apparent in heavily multi threaded applications where the i3 and even the i5 struggle a bit.

Now we have wholly new console systems that utilize AMD hardware. The PS4 (most powerful of the consoles) has 8 jaguar cores running at a max of only 2.0Ghz. Developers aren't going to able to just write and code as usual as trying to code for 2 or even 3 cores isn't going to cut it. With the weaker cores at low speeds they are going to have to make highly multi threaded games that spread the workload to as many of the cores as is possible.

In practice we already have a game we can look at that is a true prelude into what video games of the future are going to look like. I am referring to Shadow of Mordor. Minimum requirements Intel i5 quad core or AMD Phenom II 965 quad core Recommended Intel i7 quad core (with 8 threads) or AMD 8320 8 core. Notice that i3 isn't even listed, it can't run the game well enough to be recommended. Yet there are still "experts" recommending people to go to i3 systems for gaming... Really, really bad advice for future gaming.

As games become more and more optimized for the new consoles they are going to be more and more multi threaded. In benchmarks that will test the entire CPU, such as Cinebench R15 the FX series fairs much better than in benchmarks that will only test one or two cores that have dominated the benchmark world for a long time now. In the old standard of benchmarking, where one or two cores are tested even an i3 can out benchmark an FX 8370, but that particular horse has been beat to death for far too long. In Cinebench R15 my FX 8370 @ 4.5Ghz scores 712-725 reported scores of i5 4690k @ 4.5Ghz are in the 690-700 range. So an older socket FX-8370 is outperforming the much newer, more expensive "superior" i5 4690K when more than one or two cores is benchmarked.

Given that games are becoming more and more multi threaded, given that in multi threaded benchmarks the FX 8370 can outperform the i5 4690K, at a better price, why do so many "experts" say that the FX line can't game? Paired with my Sapphire R9 290 there isn't a game out there that my FX 8370 can't play at ultra settings, newer games at really nice FPS. These first round games developed for the new consoles aren't even optimized yet either. Shadow of Mordor never uses more than 6 cores on my FX 8370 and never pushes any core past 60% utilization. Once games can utilize 6-8 cores at 80, 90, 100% utilization things like bottlenecking won't be an issue anymore for the FX line. And how will even the i5 4690K with 4 cores and 4 threads fair in a game optimized for 6, 7, 8 cores and 6-8 threads? I'm sure it will still run the game, but it won't be outperforming a processor with 8 cores 8 threads.

Shadow of Mordor is the writting on wall for future gaming. i3 computers won't game much longer, i5 compters are going to be the minimum required and therefore may not be able to get ultra settings anymore, and i7 computers are going to be the new best gaming system, closely followed by FX 8 core systems (until Broadwell 8+ thread and next gen AMD is released late 2015, 2016).

I think its past time that anyone recommends an i3 as a gaming rig. the much vaunted i5 was the gaming rig of the last generation, this generation its going to be CPUs that can handle 8 threads or more, so your looking at i7 being the very best, for now, with AMD FX 8320+ being the next best choice. Notice I'm not an AMD fanboy, I'm saying that the i7 is going to be the best processor for future games, I'm just pointing out that games are becoming more and more multi threaded and i5s are already being listed as the minimum required. We are going to have at least 6 or 7 more years of game development for the 8 core PS4 before its replaced by the next consoles. In that time we are going to see games utilizing 6+ threads to their full potential.

With that said FX 8370 is a better recommendation than any i3, and will more than likely be a better recommendation than i5. Best budget gamers would have to start with FX 6300.

 

[synphul]

I think intel's aware of the gaming market. The old intel may have been a bunch of stuffy suits but that's been changing for years ever since they embraced the enthusiast market. They've been making and marketing 'k' series cpu's specially catered to the overclockers, have been participating in overclocking competitions for quite awhile. I'm sure they're aware that just about every list focusing on the 'best gaming cpu's' at various price points are flushed with their products.

Amd apu's are more powerful than intel's cpu with igpu's which makes sense for a small compact system like a game station. It's true that amd includes more cores for gaming consoles but unlike pc's for instance the xbone runs two os's side by side in parallel. Something pc's don't typically do, so essentially in an 8 core xbone you've got two quad core os's running side by side. Not one os using all 8 cores. We've also seen that more cores doesn't always mean better computing since many 8 core amd's struggle to keep up with 4 core intel chips performing the same tasks. Usually amd and lower heat don't typically go hand in hand but similar to over radding a custom cooling system it may make sense to run multiple slower cores to keep overall heat under control in such confined spaces. Similar to running 8 fans at low speed to achieve the same cooling as 2 or 4 fans at full speed but at quieter noise levels.

I have a feeling amd being so cheap has a lot to do with it as well. Consoles strive to keep their prices down and intel would probably cost more unless they decided to sell their chips at a loss similar to what they've done in the mobile handheld market. It's hard to have your cake and eat it too, performance costs money. Despite apu's doing a fine job in consoles, compared to desktops with dedicated separate cpu's and discrete gpu's even similar apu's to what's at the heart of xbones and ps4's can't keep up. There's always going to be a difference between the two pieces of hardware, pc's and consoles. They fill different roles and if a pc were to be streamlined to do one task the way consoles are (like a bitcoin miner is), then it would no longer be a pc but a console.

As powerful as pc's are, they do different tasks and take on different roles. A bitcoin miner can be had for the same price or less than a high end gaming machine and yet process bitcoins multitudes faster being streamline to do that one task. Consoles are no different, so much more can be achieved with less hardware. It's true ms is an amd partner, they're also an intel partner. Just because intel isn't powering the consoles doesn't mean they've let the gaming industry slip through their fingers, on the pc side of things intel is dominating the gaming market.

Here's a couple links concerning intel and gaming. Just pointing out they're far from ignoring gaming.

https://software.intel.com/en-us/blogs/2013/03/28/pc-gaming-is-big-and-getting-bigger-good-time-to-be-a-developer

http://gamepolitics.com/2015/03/05/intel-promises-help-foster-more-diversity-video-games-industry#.VTrIAyHBzRY

 

[junkeymonkey]

intel is too greedy to let a market share slip by so easily

you can see in my last post article links how amd had good things going for them and got in a mess some how and things slipped away


just to add back in the good old days there [amd ] chips had a better user rating per review compaired to todays chips with near the amount of rreviews

old stuff
97% 5 egg with review titles like -- ''awsome!''
http://www.newegg.com/Product/Product.aspx?Item=N82E16819103377

96%
http://www.newegg.com/Product/Product.aspx?Item=N82E16819103379

new stuff
84%
http://www.newegg.com/Product/Product.aspx?Item=N82E16819113286

80%
http://www.newegg.com/Product/Product.aspx?Item=N82E16819113285&SortField=0&SummaryType=0&PageSize=10&SelectedRating=-1&VideoOnlyMark=False&IsFeedbackTab=true#scrollFullInfo


back then amd had good stuff and I was glad to run them now its just ''good enough for gaming ''' what ??

 

[cowboy44mag]

http://www.techpowerup.com/210960/amd-bets-on-directx-12-for-not-just-gpus-but-also-its-cpus.html

All the nay-sayers can say what they like, and everyone is entitled to their opinion, however I'm not the only one who is looking towards the future. The biggest problem with the mass core FX processors were they were too far ahead of their time. Its the same problem AMD had with 64bit arch. By the time anyone cared to develop software for the hardware AMD's reputation took a blow and everyone else caught up. Wouldn't it be something if the "ancient" four year old Piledriver arch could put AMD right back in the gaming world because of a simple API update long overdue.

Although DX12 could be huge for Piledriver and repair AMD's reputation a little, I'm looking forward to Zen. AMD has learnt not to go too far ahead of themselves and the software restraints this time. AMD are very good with making APUs and multicore beasts, an APU multicore beast with greatly improved IPC (which is what most are thinking Zen is) will go a long way to help them. To be realistic even if Zen is a huge success it won't bring AMD totally out of the hole they are in by itself. They need Zen to be a huge success and they need the next several generations after to be just as great.

 

[junkeymonkey]

the fx was a mistake not ahead of its tine ??

''Ultimately AMD’s focus on new “growth areas” isn’t the culprit. What has hurt AMD is a big bet on a Bulldozer architecture — in which two CPU integer cores share a floating-point unit and other components — that simply didn’t work out''

then-CEO Rory Read said at a Deutsche Bank event.

did not work out , not ahead of its time

maybe zen may do something ,but until you got that retail ready to buy chip in hand it really doing nothing kinda like there new r9 300 cards that are to be so good but theres yet to be a real retail ready to buy card to prove all that ??

as I said AMD is a expert at saying what they will have for you next year, but sure has a hard time putting there money where there mouth is

 

[cowboy44mag]

the fx was a mistake not ahead of its tine ??

''Ultimately AMD’s focus on new “growth areas” isn’t the culprit. What has hurt AMD is a big bet on a Bulldozer architecture — in which two CPU integer cores share a floating-point unit and other components — that simply didn’t work out''

then-CEO Rory Read said at a Deutsche Bank event.

did not work out , not ahead of its time

maybe zen may do something ,but until you got that retail ready to buy chip in hand it really doing nothing kinda like there new r9 300 cards that are to be so good but theres yet to be a real retail ready to buy card to prove all that ??

as I said AMD is a expert at saying what they will have for you next year, but sure has a hard time putting there money where there mouth is

Your making the same mistake a lot of others are quoting. Bulldozer was a mistake - a failure. Bulldozer is not Piledriver. Piledriver corrected most of Bulldozer's short comings. Yes, the arch is not perfect, but it is a lot more powerful than most want to give it credit for. With DX 12 finally supporting multiple cores communicating with the GPU the FX 8 core processors will no longer be bottlenecked, and their true potential can finally be realized.

As far as the R9 300 series, I wouldn't be a bit surprised if AMD was holding the launch date until at least after Windows 10 launches. It would be a very prime time to do so. DX 12 has boosted the R9 290X to a level past the 980 in benchmarks, nearly topping the Titan X (with the newest drivers from AMD- now Nvidia will probably release new drivers of their own soon as well). If DX 12 can boost the 290X like that then the 300 series would benefit much more. I think they are holding the launch date so the benchmarking of the 300 series is done on DX 12. The lower benchmark scores the 300 series would get on DX 11 could only hurt AMD, not help them. With Windows 10 launching so soon it would be foolish of AMD to let the 300 series be judged by DX 11 standards.

http://www.gamespot.com/articles/video-thousands-of-units-battle-in-rts-ashes-of-th/1100-6425784/

Whereas games in the past gave the FX processors problems with bottlenecking the GPU and could tear if there were more than 50, 60 characters on screen at the same time (mainly do to the limitations of DX 11 and horrible console port coding) here we see just what DX 12 is going to do for AMD FX gaming. Now that is being done with just an FX 8350- the FX 8370 and 9000 series would run it even better. Thousands of characters at 4K resolution 8350 and dual R9 290X. Several months ago most who are putting AMD down here would have said that is impossible - Its impossible for 8350 to support dual R9 290X at 4K resolution - Its impossible to have thousands of units battling at 4K. That video was also shot using Mantle, by all accounts DX 12 is going to be even better than Mantle. That is what DX 12 and Windows 10 is going to do for the FX processors. That is what DX 12 and Windows 10 is going to do for gaming.

 

[synphul]

Amd has had some good theories, best laid plans - both with 64bit architecture and affordable multicore cpu's. It just hasn't panned out or been in sync with the rest of the industry as if they're off in left field rather than on target. Behind it's time or ahead of it's time matters little if the end result is lacking performance. The same things that hurt bulldozer are still plaguing piledriver since they attempted to put a bandaid on a bad situation. Continuing on a poorly designed architecture rather than ditching it for something better. The continue 'module' design isn't doing them any favors. Intel already figured out long ago that silicon has physical limitations in terms of frequency and that pushing it 'faster' isn't near as important as refining the instructions per clock. Amd's still stuck on the old concept that you can just keep throwing more cores and higher speeds at it to solve the issue.

When intel made the mistake on it's p3's by deepening the pipeline with deeper branch prediction, it should have offered more performance in theory. When it got put into real world scenarios outside of a perfect synthetic bubble and showed how badly such deep branch prediction could hurt performance, they pulled back, regrouped and looked to improve other areas. Essentially what amd had they been faced with the same problem is try to deepen the branch prediction even further. The 'throw more at it' gameplan just isn't getting it done so hopefully zen has ditched the previous follies.

The days of trying to trump your competition with gimmicks like 'we have higher ghz' or 'we have double the cores' (so it MUST be double the power, naturally) are over. Results are what matter. It may not sound like much to the average consumer if a new company came out with some off beat single clustered module that ran at only 2ghz, but so long as it proves itself and does the job faster than the competition with competitive power consumption and manageable temps it really doesn't matter how they get it done. Just that they're getting it done.

 

[cowboy44mag]

Amd has had some good theories, best laid plans - both with 64bit architecture and affordable multicore cpu's. It just hasn't panned out or been in sync with the rest of the industry as if they're off in left field rather than on target. Behind it's time or ahead of it's time matters little if the end result is lacking performance. The same things that hurt bulldozer are still plaguing piledriver since they attempted to put a bandaid on a bad situation. Continuing on a poorly designed architecture rather than ditching it for something better. The continue 'module' design isn't doing them any favors. Intel already figured out long ago that silicon has physical limitations in terms of frequency and that pushing it 'faster' isn't near as important as refining the instructions per clock. Amd's still stuck on the old concept that you can just keep throwing more cores and higher speeds at it to solve the issue.

When intel made the mistake on it's p3's by deepening the pipeline with deeper branch prediction, it should have offered more performance in theory. When it got put into real world scenarios outside of a perfect synthetic bubble and showed how badly such deep branch prediction could hurt performance, they pulled back, regrouped and looked to improve other areas. Essentially what amd had they been faced with the same problem is try to deepen the branch prediction even further. The 'throw more at it' gameplan just isn't getting it done so hopefully zen has ditched the previous follies.

The days of trying to trump your competition with gimmicks like 'we have higher ghz' or 'we have double the cores' (so it MUST be double the power, naturally) are over. Results are what matter. It may not sound like much to the average consumer if a new company came out with some off beat single clustered module that ran at only 2ghz, but so long as it proves itself and does the job faster than the competition with competitive power consumption and manageable temps it really doesn't matter how they get it done. Just that they're getting it done.

Don't get me wrong, I agree with a lot of what you are saying. While I don't believe that Piledriver is a failure and I do believe that with DX 12 it will give that "struggling" arch a boost of new life (albeit it at the end of its life). If DX 12 can boost the current 8 core Piledriver processors to between i5 and i7 in gaming it would at least give AMD a chance to liquidate some if not most of their remaining stock. A lot of budget minded gamers will be happy as well to have a great bang for your buck gamer that should last at least as long as the life span of the current console systems.

Pushing higher clock speed is still relevant somewhat, however AMD is working on something totally different with Zen. While I would be surprised if Zen still didn't overclock well I think it will focus more of power consumption, IPC, and fielding something that directly challenge the i7 line. If in the process they are still able to add "moar cores" - if those cores have much better IPC so be it.

 

[junkeymonkey]

''As far as the R9 300 series, I wouldn't be a bit surprised if AMD was holding the launch date until at least after Windows 10 launches''

I do find what you said funny -- were holding back our hot new item until - until what ?? were near bankrupt and we don't need fresh revenue flow ??

while our competitors reap a greater market share that can hurt us even more ?? no they ran there mouth and now cant produce

''Your making the same mistake a lot of others are quoting. Bulldozer was a mistake - a failure. Bulldozer is not Piledriver''

so AMD's CEO is making the same mistake as everyone else as well ?? piledriver is a Bulldozer architecture remember ?? [ In 2010 AMD revealed that the 2nd generation was scheduled for 2012; AMD referred to this generation as Enhanced Bulldozer. This later generation of Bulldozer core was codenamed Piledriver. ]

bulldozer- pushed over all amd built up--- piledriver --beat them into the dirt -- steamroller -flattened them out -- excavator digs the grave then undertaker , well you know what that guy does [LOL] the real AMD road map

I think you need to research AMD some more

 

[cowboy44mag]

''As far as the R9 300 series, I wouldn't be a bit surprised if AMD was holding the launch date until at least after Windows 10 launches''

I do find what you said funny -- were holding back our hot new item until - until what ?? were near bankrupt and we don't need fresh revenue flow ??

while our competitors reap a greater market share that can hurt us even more ?? no they ran there mouth and now cant produce

''Your making the same mistake a lot of others are quoting. Bulldozer was a mistake - a failure. Bulldozer is not Piledriver''

so AMD's CEO is making the same mistake as everyone else as well ?? piledriver is a Bulldozer architecture remember ?? [ In 2010 AMD revealed that the 2nd generation was scheduled for 2012; AMD referred to this generation as Enhanced Bulldozer. This later generation of Bulldozer core was codenamed Piledriver. ]

bulldozer- pushed over all amd built up--- piledriver --beat them into the dirt -- steamroller -flattened them out -- excavator digs the grave then undertaker , well you know what that guy does [LOL] the real AMD road map

I think you need to research AMD some more

Funny, every time a doom and gloom Intel fan says AMD is doomed, they aren't. I don't want this thread to turn into yet another AMD is doomed and will be bankrupt in months. There are a ton of those going back three, four years, yet AMD is still here. AMD isn't going anywhere anytime real soon. I can't understand why anyone would want them too. If AMD were to go under the cost of Intel systems would skyrocket, that wouldn't be good for anyone. Competition drives innovation and price.

Yes, they (AMD) need to do better financially. AMD's biggest challenge isn't innovation, and it isn't the technology its marketing. Marketing pressure is what forced AMD's hand to bring a Bulldozer to market before it was ready. They can't afford a bad review of the 300 series, that is exactly why they are holding the release of the 300 series. The potential performance gains from DX11 to DX12 is exactly why they are delaying it. AMD can't afford another product getting a bad or underwhelming initial review. By holding off until Windows 10 launches the 300 series will be benchmarked using Windows 10 / DX 12 and will get much better reviews. If AMD can improve on its marketing it would be in much better condition, and you can't do that without good reviews. Nvidia doesn't hold much of an overall performance edge over AMD (like Intel they hold a good edge in efficiency), the AMD cards are priced much better, yet they only hold 25% of the market. Marketing needs to be improved.

As far as Piledriver being exactly the same the Bulldozer, that is like saying an Excursion is the same truck as an F350. They are built on the same framework after all, heck even have the same power plants, so I guess they are the same exact thing. Now try hooking a 30' fifth wheel trailer to that Excursion... oops what do you know they aren't the same. Piledriver uses the same arch, but it is a much more perfected version of the arch. Piledriver performance wise (especially when all cores are utilized) isn't that far behind iCore (realistically around the performance of Sandy Bridge), and competes well with i5 however efficiency wise they realized they could never get the power consumption to the level iCore (Haswell) is on. They have to do a die shrink and change arch for efficiency, and more of a performance boost, which is why AMD started development of Zen. Bulldozer was far from ready to launch, and that most definitely is no one's fault but AMD (I won't make any excuses for them there). Hopefully they fully learnt their lesson about rushing something to the market before its totally ready. Zen will show everyone if AMD learnt anything from its mistakes with Bulldozer's launch.

 

[NightAntilli]

Just leaving this here, not directly FX, but it shows something...

http://www.expreview.com/39963.html

 

[junkeymonkey]

never said I was a intel fan boy , but did say I had built amd exclusive up until fx AM3+ platform and felt let down after time and money spent trying to get something satisfactory out of it .. then cut my losses and went intel for my first and only time ever and cant say it was a wrong move .. not to say its the greatest [like why cant intel give 2 full x16 slots as amd can with out the need of a plx chip ? ] little things like that , but overall more satisfied.. 2 years ago if you asked me to build intel I would of said no thanks ... that's all

amd just fallen into a lot of talk but no action and everything is next year that seems not to materialize/postponed/or falls short of the promises. that's just where there at today

 

[cowboy44mag]

never said I was a intel fan boy , but did say I had built amd exclusive up until fx AM3+ platform and felt let down after time and money spent trying to get something satisfactory out of it .. then cut my losses and went intel for my first and only time ever and cant say it was a wrong move .. not to say its the greatest [like why cant intel give 2 full x16 slots as amd can with out the need of a plx chip ? ] little things like that , but overall more satisfied.. 2 years ago if you asked me to build intel I would of said no thanks ... that's all

amd just fallen into a lot of talk but no action and everything is next year that seems not to materialize/postponed/or falls short of the promises. that's just where there at today

Don't get me wrong, I've never said you were right or wrong to switch. I remember spending what seemed like weeks learning how to correctly configure Skyrim's ini file to make it run "right". Getting Ultra settings at 1080p with no dramatic drop in FPS I was happy, however with an Intel I probably wouldn't have had to do that to begin with. There is no doubt AT ALL that with the last generation of games Intel had the lead in performance over AMD, that's not to say that upper end FX processors can't run those titles on Ultra with the right GPU and a little "tinkering".

Looking forward to Windows 10 and DX 12 gaming is where the excitement is. To summarize a good posting on another thread- The current API DX 11 can utilize a few cores however it has to rely upon a single primary thread which can become a bottleneck. The "well threaded" games of the DX 11 generation would then use other cores for physics, game engine, ect. DX 12 / Mantle / Vulkan (Mantle II) will allow the game to use circa 6 cores. If the developer is good they can use multiple threads on other parts of the game as well. There is however a practical limit to this though. It has been theorized that 4 cores will be the bare minimum for games moving forward with 8 being the new "sweet spot" with few practical returns above 8. That is not to say that smaller developers won't still do things the way things have been with one primary core being used, however the AAA titles being produced by large studios will want their titles to shine. Those titles will be pushing up to 8 cores, each one capable of direct communication with the GPU via DX 12 / Mantle / Vulkan and that will fall directly into the FX 8 core lineups strongest suite multi threading across all 8 physical cores. The console systems will be relegated to 6 cores, however everyone already knows that gaming on a PC yields much higher "eye candy" than the consoles.

 

[junkeymonkey]

but AMD is not true 8 core its 4 core + 4 integral core or you can assume 4 core + 4 physical hyper threating .. so intel i7 has hyper threading that's way faster and way more efficient ??? plus the sad fact the i5 can still out preform that so called amd 8 core anyway [ 32 0f 42 tests] for what ever this site is worth

http://www.cpu-world.com/Compare/446/AMD_FX-Series_FX-8350_vs_Intel_Core_i5_i5-4670K.html

that core thing from amd is just a 1/2 truth but unlike intel who nickel and dimes you per each benefit so amd could of charged you for a chip that just 4 core and extra for the one that gives hyper threading but with amd its all hyper threading because they can only compete that way

say AMD offered there chips with out that thaty would be such a dog and set further back

intel hyper threading
http://en.wikipedia.org/wiki/Hyper-threading

amd Clustered Integer Core

http://en.wikipedia.org/wiki/Bulldozer_(microarchitecture)

http://en.wikipedia.org/wiki/Piledriver_(microarchitecture)

so overall in the end you may as well go i7 or xeon and get the more efficient and faster core -- you get what you pay for [and some times less ]

also again in that piledriver link you can see another AMD talked something up and never produced it [Komodo platform] just like the 1090 chipset boards that never made it and so on and so on .. with all the let downs and dropped ideas , delays ,postponed its hard to get any hopes up with anything amd has to say or do these days , its all about what they will have next year that may never come .. as they say '' if you hold your breath too long you will turn blue in the face '' just take a breath of fresh air and move on


just to add what about there good Opteron support ?? all gone with am3+ as well ?? use to get support for most all chips for the socket but with am3+ that went in to the can as well -- ya they seen am3+ was a mistake and quit while they were a head

http://www.servethehome.com/amd-opteron-3200-series/

 

[cowboy44mag]

but AMD is not true 8 core its 4 core + 4 integral core or you can assume 4 core + 4 physical hyper threating .. so intel i7 has hyper threading that's way faster and way more efficient ??? plus the sad fact the i5 can still out preform that so called amd 8 core anyway [ 32 0f 42 tests] for what ever this site is worth

http://www.cpu-world.com/Compare/446/AMD_FX-Series_FX-8350_vs_Intel_Core_i5_i5-4670K.html

that core thing from amd is just a 1/2 truth but unlike intel who nickel and dimes you per each benefit so amd could of charged you for a chip that just 4 core and extra for the one that gives hyper threading but with amd its all hyper threading because they can only compete that way

say AMD offered there chips with out that thaty would be such a dog and set further back

intel hyper threading
http://en.wikipedia.org/wiki/Hyper-threading

amd Clustered Integer Core

http://en.wikipedia.org/wiki/Bulldozer_(microarchitecture)

http://en.wikipedia.org/wiki/Piledriver_(microarchitecture)

so overall in the end you may as well go i7 or xeon and get the more efficient and faster core -- you get what you pay for [and some times less ]

also again in that piledriver link you can see another AMD talked something up and never produced it [Komodo platform] just like the 1090 chipset boards that never made it and so on and so on .. with all the let downs and dropped ideas , delays ,postponed its hard to get any hopes up with anything amd has to say or do these days , its all about what they will have next year that may never come .. as they say '' if you hold your breath too long you will turn blue in the face '' just take a breath of fresh air and move on


just to add what about there good Opteron support ?? all gone with am3+ as well ?? use to get support for most all chips for the socket but with am3+ that went in to the can as well -- ya they seen am3+ was a mistake and quit while they were a head

http://www.servethehome.com/amd-opteron-3200-series/

I can see where your coming from, AMD has over hyped things in the past that have not made the big "splash" they were supposed to. There is no denying that. I'm fairly certain that DX 12 won't be one of those things however as it is Microsoft who actually produced that. It can be claimed to be largely based on Mantle, but DX 12 will have much better support and use by developers. If Microsoft themselves weren't as certain DX 12 will provide large boosts and it was just AMD claiming everything I might be a little more leery, however Microsoft is also putting their reputation on the line. The great thing with DX 12 is we at least won't have to wait long to see what the results will be for FX, i5, and i7. I'm simply predicting (looking at all the available facts and information) that with DX 12 support games developed for 8 threads will run really well on higher end FX. No, I don't think anyone truly believes that the i7 will be trumped (there is no way) but I do fully expect the higher end FX to benchmark somewhere above i5 and below i7.

 

[cowboy44mag]

Just leaving this here, not directly FX, but it shows something...

http://www.expreview.com/39963.html

Again, not exactly FX but thought you might be interested in this. Zen is shaping up to be something special😀

http://www.planet3dnow.de/vbulletin/attachment.php?attachmentid=32376&d=1430143447

 

[synphul]

It does look like amd is further helped by dx12 but it also appears to be a comparison with amd's apu vs the i3 igpu. The fact the integrated gpu on the amd apu is stronger than intel hd 4400 is nothing new at all. Can't really make out all of what that has to say since it's in chinese or something. The translation is very rough (thanks google lol). Maybe a boost to the mobile market or those strictly running apu's but still won't compare to upper end cpu's with discrete gpu's which are what many gamers run. It also wouldn't be fair to expect discrete graphics performance out of an apu/igpu. Just comparing the sheer size of physical hardware, one occupies a small portion of a cpu chip the other has an entire board dedicated to it's graphics processing power.

 

[NightAntilli]

Zen is looking very traditional, which means single core performance will shoot up like a rocket, at the cost of multi-threading.

 

[cowboy44mag]

Zen is looking very traditional, which means single core performance will shoot up like a rocket, at the cost of multi-threading.

That's exactly why its looking special. Its a relatively simple design, very reminiscent of the beloved Phenom II. AMD is already very good at multi-threading, and I would be very surprised if these processors didn't offer at least 8 cores of performance. I'm glad to see rather than going to 16 or 32 cores or something totally unwarranted they are instead focusing on core per core performance. That has been their biggest "weakness" in current years. Going to be interesting to see how far they can close that gap.

On a side note, Zen isn't a "bad" name, but I think they would have better marketing success calling it Phenom III. There are a lot of die hard AMD fans who would go nuts based just on the name alone, not to mention a huge boost in single core performance.

At any rate, its looking like FX will have a very worthy successor😀.

 

[synphul]

They may be going with "zen" to break away from previous brands and make sure it's differentiated and not compared to other prior cpu's. Just like intel moved away from the older celeron/pentium lineup with the new core series. That way people will tend to look at it as it's own entity, not just a revived phenom, athlon or yet another reincarnation of bulldozer/piledriver.

 

[junkeymonkey]

that's what there saying here as well
http://www.techpowerup.com/212112/amd-zen-cpu-core-block-diagram-surfaces.html

then it like said there falling back to the phemon style ?? that's going backwards right ?? but I guess it did work and going Athlon for there apu's as well ..

thing is what has intel really got to counter it? again this is next year speculation, and intel is not feeling a real threat and not really having to use any of there technology due to lack luster competition

figure intel has come out with 2 or 3 generations and skylake in the wings and staying on that tick tock road map not having to do any kind of super slam cause someone [amd] is not hot on there heals ... ho hum

so by the time amd releases this chip where is intel going to be sitting with there next generation release with something up there sleeve ? 2016/17 is a long way off

 

[Clariska]

To be honest CPU's only play a little role in gaming...I would give an explanation.My Friend and i did a test.I am running amd 4170 and his running an I7 4790k.We used his screen card as my lil 6570 wouldnt have done much -_-.We did test based on games up until 2014.When it came to low reso The intel were Killing the AMD by far.But coming up to 1920x1080 the Intel only had a 20-25% in performance gain.

Now here's my argument.He paid 12k when his cpu got released here.I paid 1.4k....In America i see the difference as the amd/intel cpu's are alot closer in pricing.There's other stuff on intel i also don't like.I think amd/ati drivers are alot better than Intels
.South Africa's the best and worst place if you like your amd as 90% people here runs on intel and when you mention amd you will get flamed for amd(over heat)Well my cpu idles on 29C 43C in gaming and stress test push it up to 54C on air cooling(maybe the 6 120mm fans has something to do with it 😀)

Junkeymonkey does make some good points.Well there's a reason why intel is so far ahead as i use to work directly with intel.They did feel pressure around 2008 so they used a cheesy strat that caught AMD offguard(not going into detail about this)Its just hard for me to believe that amd didnt recover from this yet.I do think the gaps gonna close during 2016/2017 as the person that put intel on a back foot during 2000 era are back with amd.Btw Junky intel didnt really license 64bit out of amd.What happened was amd gave intel the diagram for 64bit and said here's the tech see what you can do with it.

PS;All these people always throwing around websites please start doing hand on testing as no 2 pieces of hardware are the same.For instance my friend also had a 4170 cpu and he's struggling to reach 4.6ghz.This cpu of mine i had up to 4.8ghz and can most prob easily reach 5ghz if i really want to.The review on hardware also tells you so much.Running my amd for 3 years and shes mostly on.Never really get switched off.Didnt give me any problems yet...My brothers I3 that gets switched off most of the time failed so much already and having so much problems with heat its scary....

 

[cowboy44mag]

To be honest CPU's only play a little role in gaming...I would give an explanation.My Friend and i did a test.I am running amd 4170 and his running an I7 4790k.We used his screen card as my lil 6570 wouldnt have done much -_-.We did test based on games up until 2014.When it came to low reso The intel were Killing the AMD by far.But coming up to 1920x1080 the Intel only had a 20-25% in performance gain.

Now here's my argument.He paid 12k when his cpu got released here.I paid 1.4k....In America i see the difference as the amd/intel cpu's are alot closer in pricing.There's other stuff on intel i also don't like.I think amd/ati drivers are alot better than Intels
.South Africa's the best and worst place if you like your amd as 90% people here runs on intel and when you mention amd you will get flamed for amd(over heat)Well my cpu idles on 29C 43C in gaming and stress test push it up to 54C on air cooling(maybe the 6 120mm fans has something to do with it 😀)

Junkeymonkey does make some good points.Well there's a reason why intel is so far ahead as i use to work directly with intel.They did feel pressure around 2008 so they used a cheesy strat that caught AMD offguard(not going into detail about this)Its just hard for me to believe that amd didnt recover from this yet.I do think the gaps gonna close during 2016/2017 as the person that put intel on a back foot during 2000 era are back with amd.Btw Junky intel didnt really license 64bit out of amd.What happened was amd gave intel the diagram for 64bit and said here's the tech see what you can do with it.

PS;All these people always throwing around websites please start doing hand on testing as no 2 pieces of hardware are the same.For instance my friend also had a 4170 cpu and he's struggling to reach 4.6ghz.This cpu of mine i had up to 4.8ghz and can most prob easily reach 5ghz if i really want to.The review on hardware also tells you so much.Running my amd for 3 years and shes mostly on.Never really get switched off.Didnt give me any problems yet...My brothers I3 that gets switched off most of the time failed so much already and having so much problems with heat its scary....

The role of the GPU in video gaming can not be underplayed, but if the cpu bottlenecks it communication with the GPU then you have problems. Most of the problems that the current generation FX processors have had with gaming is when due to DX 11 the primary thread gets overloaded and causes throttling. In poorly coded games you could have a Titan X but if the primary thread gets overloaded your gaming experience is going to suck.

That is why DX 12 is such a big deal. Mantle has already proven that when you have multiple cores communicating with the GPU and not just one primary one then you don't experience bottlenecking. DX 12 is more "mature" and will have much wider support than Mantle but does the same thing. Instead of having only one core do all the work of communicating with the GPU and getting overloaded, now with DX 12 6 to 8 cores can communicate with the GPU. Will this place the higher line FX processors between the levels of the i5 and i7 in gaming? To be fair, I'm predicting it will, but only when Windows 10 and DX 12 launch and we have games developed for DX12 and can actually benchmark everything will we know for sure. At least by the end of the year we should have some real answers. One thing is for certain though DX 12 will breath new life into the FX series in gaming. If DX 12 boosts the FX series to a level just above the i5 then the wait for Zen (probably Q2 2016) won't feel that long. It would be good for AMD to move some of its existing FX processors in mid/high level gaming computers in the meantime.

 

[cowboy44mag]

that's what there saying here as well
http://www.techpowerup.com/212112/amd-zen-cpu-core-block-diagram-surfaces.html

then it like said there falling back to the phemon style ?? that's going backwards right ?? but I guess it did work and going Athlon for there apu's as well ..

thing is what has intel really got to counter it? again this is next year speculation, and intel is not feeling a real threat and not really having to use any of there technology due to lack luster competition

figure intel has come out with 2 or 3 generations and skylake in the wings and staying on that tick tock road map not having to do any kind of super slam cause someone [amd] is not hot on there heals ... ho hum

so by the time amd releases this chip where is intel going to be sitting with there next generation release with something up there sleeve ? 2016/17 is a long way off

First off, as much as I like the idea of a "Phenom III" I do agree with synphul that they don't want anyone jumping to the conclusion that Zen is a re brand of Phenom.

I really think that Zen is a new generation of a Phenom style process "on steriods". The few slides released so far don't show much, but from what most can tell it will be 14nm, and has the possibility of boosting single core performance over Piledriver by 40-60%, which would be huge for AMD. The Zen performance line Bristol Ridge is expected to be 8 core. AMD is making a bold move to catch up to Intel by going from 32nm to 14nm in one move. The interesting thing is Intel has already seen that die shrinks alone are having more of a diminished return, and I question how much more performance they can squeeze out of the Core arch. It looks like when Bristol Ridge launches it will be around the performance level of of Broadwell. Intel may retake the raw performance crown with Skylake and the die shrink to Cannonlake, however AMD is good at undercutting cost to make it more competitive. The exciting thing is IF Bristol Ridge in its first generation can trade blows with Broadwell AMD has a lot of breathing room on a new arch where Intel's Core arch may be reaching the extreme limit of diminished return and they may have to switch to a whole new arch before too much longer.

 

[junkeymonkey]

don't seem to be hurting intel any .. so that still puts amd behind the 8 ball

http://www.pcworld.com/article/2465211/intel-microsoft-promise-directx-12-could-halve-pc-graphics-power-draw.html

another read for you

''If OpenGL is capable of delivering all these improvements, why does Mantle exist at all — and why haven’t developers taken advantage of them already?''

why ?? cause it goes against Microsoft and bill cant license it and collect money from it and keep control over developers , and they cant have that ..[its not proprietary Microsoft]

http://www.extremetech.com/gaming/179010-who-needs-directx-amd-nvidia-and-intel-team-up-demonstrate-ultra-low-overhead-opengl


be sure to read some comments at the bottom like this ...

''one of the many many reasons i have stuck with windows Xp is because i use 3d modelling software that utilizes OpenGL, From Windows Vista onwards Opengl is Redirected through DirectX and Performance is Dire ''

 

[cowboy44mag]

don't seem to be hurting intel any .. so that still puts amd behind the 8 ball

http://www.pcworld.com/article/2465211/intel-microsoft-promise-directx-12-could-halve-pc-graphics-power-draw.html

another read for you

''If OpenGL is capable of delivering all these improvements, why does Mantle exist at all — and why haven’t developers taken advantage of them already?''

why ?? cause it goes against Microsoft and bill cant license it and collect money from it and keep control over developers , and they cant have that ..[its not proprietary Microsoft]

http://www.extremetech.com/gaming/179010-who-needs-directx-amd-nvidia-and-intel-team-up-demonstrate-ultra-low-overhead-opengl


be sure to read some comments at the bottom like this ...

''one of the many many reasons i have stuck with windows Xp is because i use 3d modelling software that utilizes OpenGL, From Windows Vista onwards Opengl is Redirected through DirectX and Performance is Dire ''

Honestly, no matter how well Zen does AMD is still going to be behind the 8 ball. Intel is too large for them not to be. Even if Zen comes out and is on even ground with or slightly better than Broadwell in performance Intel still has the upper hand. Intel can push out Skylake and Cannonlake earlier and AMD simply can't release generation updates at the same rate that Intel can. That and no matter how close the performance may be, Broadwell will still have power efficiency over a first generation Zen. Even if Intel would decide to go to a whole new arch, with the resources they have they could do so in the same time frame it takes them to to release a next generation Core arch. Intel simply has too much talent and money for AMD to hold an edge on them for long. An AMD line that is very competitive with upper end Intel however would be great.

As far as OpenGL goes, its in the same boat as Mantle and Vulkan for the most part. OpenGL could be the very best API ever created, but if the developers don't use it, it really doesn't matter. I think we all know that DX 12 is going to be highly utilized and Vulkan, OpenGL will be used here and there but nothing big enough to really matter. The simple fact that DX 12 is Mantle like in its operation should really give AMD a boost in its current generation FX processors. I simply don't think too many developers are going to do anything much with OpenGL, or Mantle / Vulkan.