AMD CPU speculation... and expert conjecture

Page 288 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Status
Not open for further replies.


Even so, there's a big difference. Usually it's much much harder to make a game for PC for that exact reason. For Consoles, it's literally make it and you're done. Usually the reason that console publishers (like polyphony with Gran Turismo), is because the game that they have in mind, will probably only work of a specified amount of hardware, even without that, there's the factor that when you are making a game like per say, Battlefield, you need to make sure that the settings can get as much performance to quality as possible on as many pc's as possible, which is harder than it seems.

Even so the game may be awesome for console, it's simply that the people aren't good enough to do PC (or they may just choose not to do PC for other reasons). Plus, if the game is console only, you can always get an emulator. So far, you can really get any system you want, get the game, and done (except the PS, you need the actual drive on your PC for that).
 
I'm sorry but I have to say this.

I was sitting on the couch in my office and my eye wandered to the bookshelf, and something caught my eye. A magical EVGA Nvidia 9400 GT was just sitting there. I was so amazed that when I searched it up I found it to be far worse than my current Radeon 5570. 🙁 However it does have an SLI slot, so there's a chance I can use it for my secondary display. :3
 


http://www.youtube.com/watch?feature=player_detailpage&v=2QkyfGJgcwQ
 


1) That 30% refers to total CPUs: Athlons + Semprons + FX + Opterons + ...

And the FX series includes quad, hexa, and octo cores. The FX-6000 and 8000 series joined must be about a 5% of the market.

2) Look to the professional market. What CPUs is AMD promoting for their Firepro? Intel Xeon? Why? Because whereas AMD has enabled Gen 3 support in Firepro series, the Opteron 6300 series only have Gen 2 support. Therefore, unless you want to bottleneck your GPGPU, you have to pair it with an Intel Xeon. This is that Apple has made with Mac Pro: Intel Xeon + AMD Firepro.

Look at Cray as well. Titan was Opteron 6300 + Nvidia K20X. Now Cray has abandoned AMD for Intel Xeons because Cray required Gen 3 support. The corporate vide president and general manager of AMD's Server Business Unit reply to Cray's move to Intel was:

We will have PCI Express Gen 3 as well. The switch by Cray to PCI Express Gen 3 is expected on that side of the business, because PCI Gen 3 is where all the discrete GPU accelerators will be connected to, so we have Gen 3 in the works as well. There is nothing preventing putting an AMD chip into a Gen 3 based system.
Click to expand...

No what I agree with such strategies, like giving Gen 3 support to the your GPGPU but then not giving it to your own CPUs, obligating customers to buy a chip from another brand, merely explaining to you that AMD has already made similar moves.
 
Are the Steamroller APUs still coming this year? I'm holding off on upgrading my laptop until the next gen comes out.

I haven't heard any rumors about the IGPU in the Steamroller APUs - is it possible AMD is holding back the release to respond to Intel Iris Pro? For example, do you think AMD might follow Intel in adding a large 4th level cache on the package?
 


iGPU in Kaveri is supposed to rival the HD 7750 in flagship parts for DT, not sure what laptop iGPU will look like, but I would expect it will outclass anything from Intel, even the Iris Pro stuff.

Kaveri will be released Q4 this year, with likely availability being late this year to early 2014.
 
Toms new article is an interesting read. Definately shows one of the weaknesses of AMD not focusing on all aspects of the CPU design.

image003.png
image009.png


I Haven't seen toms do an external memory latency test before, most of the time its cache latency ... ugly side of AMD, very ugly.

one interesting point here is people arguing that faster memory = bad latency ... 1600 CL7 scored worse than 2133 cl 10 pretty much across the board on both Intel and AMD on the latency test.

 
Yeah AMD's IMC needs some serious work. I would like to know if they attempted to OC the NB to 2.4~2.6Ghz and redo the test, just to see what kind of improvement was available.

Memory timings are values measured in clock ticks. 1600 CL7 is 1,600,000,000 cycles upon which it takes 7 of them to respond. 2133 CL10 is 2,133,000,000 cycles with 10 cycles response time. So 7/1,600,000,000 vs 10/2,133,000,000 for how long it takes to respond. They end up nearly identical to each other.
 


Desktop Kaveri (Steamrolloer APU) is coming this year. Mobile Kaveri is waited for next year (~ Feb/March). The iGPU in the top desktop Kaveri is waited to be a HD 7750 DDR3 but with core clock upgraded to 900MHz. However, there are rumours that Kaveri comes with next gen Volcanic Islands graphics. So far as we know, there is no L3/L4 caches in Kaveri.
 


The people who argues faster memory = bad latency must believe that CL numbers are times, when they are clock ticks relative to the frequency of the memory module.

The cycle time for DDR3-1600 is 1.25 ns. The cycle time for DDR3-2133 is 0.94 ns.
7 cycles of DDR3-1600 are 8.75 ns whereas 10 cycles of DDR3-2133 are 9.38 ns. Therefore those 1600 CL7 modules have slightly better latency than 2133 CL10 modules. However 2133 CL9 modules will have better latency (8.44 ns) than any of those.

Finally, a 2133 CL7 module has much less latency (6.56 ns) than a 1600 CL7 module (8.75 ns).
 
Interesting reading, seems to confirm everything that I have said in this thread:

http://www.pcper.com/reviews/Editorial/AMDs-Processor-Shift-Future-Really-Fusion/Future-Still-Fusion

No Steamroller for AM3+, only legacy support for 8000/9000 series.

AMD will continue selling 8000/9000 series before releasing top high-performance APUs. No 4M APU for 2014, 3M APU still possible in late 2014 (but unlikely).

Warsaw will be substituted by Excavator version of Berlin.

HyperTransport obsolete.

2/4 cores for mainstream desktop/laptop. 6/8 cores for enthusiast desktops. 4/8/12/16 cores for servers.

Full transition to APUs by about 2015.

Some more predictions/thoughts that the article doesn't consider:

ARM APU for 2016

Move to ARM (from x86) for 2018-2020.
 


Just like every game released in the past 4-5 years? BF4 is going to perform about the same as BF3. Which is to say: GPU bound in singleplayer, with somewhat better scalability in multiplayer.
 


No one is going to move exclusively to ARM, simply due to that thing called Windows.
 


I think you're seriously underestimating the new Frostbite engine...it will probably kill dual cores in multiplayer, and they will likely be able to run single player albeit with a relatively poor experience compared to what it could be with a quad or better.
 


Dice has already stated that BF4 runs about the same as BF3.
 
i have a couple of questions about the jaguar and radeon cores in the new console socs:
when new nodes become ready, will sony and microsoft make the socs on new nodes e.g. 20nm/16nm/14nm?
with node shrinks, will they be able to clock the cores higher or the performance improvements will be more subtle or none at all?

i ask this after reading microsoft's publicity on squeezing out minor clock boosts on their final xbone(R) soc.
 


Umm ... you realize he's just stating what everyone else has been stating the last few months. It's like guy A talks to guy B who talks to guy C who then *informs* guy A about the "news" and guy A goes off and writes an article about it.

We all know that FM2+ will be the first SR chips in a 2M configuration, that isn't being debated. The question is about the mainstream performance sector @180~200 USD. A four core APU isn't going to cut it, neither is the current 8350 or the golder chip 9 series. They will need to release a SR module based desktop chip to satisfy this significant market segment, that module doesn't need a GCN based GPU on it at all. So the fact that they can't fit a GCN GPU on the same die as 3~4M is about as relevant as them not being able to fit a Southern Islands GPU on a die with 3~4M. Of course I'm expecting AM3+ to be phased out in favor of FM2+ as there is no need to maintain two different socket types.

Basically and AMD version of the below linked Intel 4+1 die configuration
http://wccftech.com/haswell-die-configurations-intel-ivy-bridge-revealed/

Just enough of a "iGPU / whatever" unit to function for HSA but no where near what's necessary for full blown graphics rendering. And with AMD's modular design they can do this much easier then Intel can.
 


Their consoles, consoles don't receive upgrades to their core specifications as it really f*cks with developers. So don't expect any upgrades to the CPU speed. At most there maybe be a XBONE / PS4 "slim" which includes die shrunk components that are cooler / less power hungry, but those components will behave exactly how the original ones do. Consoles can get an ~80~150% performance boost due to the developers being able to code for bare metal, screwing with component specifications mid-cycle toss's that out the window.
 


Very old news with the same old (probably wrong) assumptions. "ITS WHAT EVERYBODY IS SAYING"

* NO more new uarch AM3+ upgrades is pretty clear, since even Warsaw will have a **NEW COMPATIBLE** socket (old G34 is EOL to)... and Warsaw is still PD.

* Hypertransport is not obsolete, even if all adapters will be PCIe, every CPU and APU since this far is a Hypertransport switch for the "internal" Xbar, and HSA IOMMU standard is based on HTT internals. Matter of fact Piledriver was supposed to debut this Xbar/HTT at 4Ghz version with better power management (AND HTT HAS MUCH BETTER POWER MANAGEMENT POSSIBILITIES THAN CRAP PCIe), which seems didn't happen. Perhaps new version of PD Warsaw, and PD FX can have it, even for AM3+, only this will make an upgrade worthed.

Hypertransport perhaps makes the same or more sense in GPGPUS than in CPUs... a MCM of GPUs is what is needed to crush Nvidia definitely without redemption and their monster chip approaches, it would be better in all aspects, and it will be single "chip", single socket SKUs. In this light i'm still hoping for a HTX+PCIe combo slots, though that probably end of 2014 or 2015.

* Everything APU, as in a chip with a GPU is simply NOT the best option for server... server needs GFLOPS not graphics... so a "middle" ground must be found. Else is the same of AMD abandoning server altogether... a very strange and penalizing move, considering that there is an Open Platform definition for server.
 


That is what is about thick skulls, HSA is not about iGPU not even GPGPU... though that is what everybody in HSA foundation is concentrating now by agreement... but there will be plenty of HSA systems with DSP and other "accelerators" and without any GPU in the vicinity.

 
Intel Dual Core has been obsolete in BF3 multiplayer for a while. No one benchmarks it but if you hunt around on some forums, i3 owners will constantly complain about performance in large maps. If they think their chip is good for BF3 multiplayer, they will upload videos to youtube showing frame rates on small maps.

Intel Dual without HT is barely good for single player, useless for multiplayer.

Intel Dual with HT is kind of playable in multiplayer and works fine in single.

I personally do not trust pcper that much. They are the ones who conveniently started raising a ruckus over frame times using Nvidia provided tools at a time when AMD was about to release 7990, which would have easily taken the performance crown from Nvidia for a single card.

The fallacious logic in pcper's article is that they are assuming that because we are seeing enthusiast motherboards for FM2+ that that means we WONT be seeing a refresh on the enthusiast platform.

It is the equivalent of seeing a new platform of Intel CPUs, like LGA1155, not seeing LGA2011 motherboards updated, and then going "OMG LGA2011 IS DEAD!!!"

If you want my opinion, AMD will continue to release higher end CPU only parts, but they will focus on APUs. AMD needs to coerce people into the APU market so people will develop HSA applications as there will be a base of users to run the software. It is like how Intel is trying to coerce people into LGA2011 platform by offering 4820k for cheaper than 4770k, and even going as far as to deceitfully try and brand Ivy Bridge quad with HT as superior to Haswell quad with HT by implying that a higher model number makes a better chip.

Not a lot of people are going for it though, because they keep screaming "MUH PCIe 3.0!!!"

The whole article is poorly written. One of the main points is that "Intel is focusing on 1155/1156/etc instead of 1366/2011/etc and AMD is doing the same with their line up". Their conclusion, from that line of logic, is somehow that means that AMD is going to abandon the market served by AM3+/LGA2011/1366/etc. The problem is that Intel still serves that market. In fact, they just announced new CPUs for it a while ago.

Josh wrote that article with the logic of a middle school child. It's appalling to me to even read it.

Not to mention I implied earlier that AMD would be going APU only and I got reamed on this forum for it. There's a big difference between shifting focus to something and completely ignoring everything else for the sake of focusing on something.

AMD is making a big APU push, but that doesn't mean they're going to completely forget about >=6 core CPUs. They are a lower priority right now and rightfully so. Not to mention AMD having enthusiast FM boards is a good thing to help them shake their "value only" stigma they have, which I feel is even stronger with their APUs, as APUs right now seem to be a "well you can't afford an AM3+ or Intel quad so just get an APU instead" High end FM2 boards would at least get people to consider it as something more than a HTPC build or "too poor to go full rig with dGPU" build.
 
Status
Not open for further replies.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom