AMD CPU speculation... and expert conjecture

Page 362 - 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.


That's because APU's bottleneck at the memory bus not at the core clock. There is a community of folks who like to OC APU's and they have found that OCing the GPU clock doesn't do much for performance. What you want to OC instead is the CPU_NB and the memory itself. Those two will net you a much larger increase in performance the bumping up the iGPU clock.
 


I think this is the 2012 slide where AMD said that Kaveri will be the first Teraflop apu
 
It is still possible that the iGP, AMD have found how to scale it higher without the need for core speed, similarly on the IPC side, a 3.6ghz CPU is probably faster than the 4.2ghz current chips by a substantial margin. Improving the architecture, notably the prediction misses, cache latency the IMC allows the CPU to push the iGP further, also seems as though they are able to bypass bandwidth limitations somehow as 1080 Medium is very impressive, it is possible that AMD are delivering 7750/770/250X performance on chip which is itself a massive microengineering success.

As to how it performs, wait and see, maybe you will be very surprised, anyways I am waiting for BF4's mantle update to see how AMD hardware peforms, I had many issues with Intel hardware, notably had to upgrade the i3 which spikes badly, the i5 is better but again a i7 exhibits issues of spikes and I can somehow seeing this related to HT. BF is neutral and can allocate to resources, the problem I think lies in that the i3 and i7 have half the threads dedicated to HT and HT cannot keep up with the loads resulting in spikes and lag. perfoverlay.drawgraph 1 is the best way to monitor spikes, i5 is much better than both the i3 and i7, similarly the FX 6/8 parts tend to hang with the GPU showing no bottlenecking and thus a i5 or FX 6/8 is probably going to be my recommendation for CPU's for BF4 until, if EA fix HT.
 

jaguar is designed that way. the igpus in the consoles aren't exactly complex high end design, and tsmc has much better track record than glofo. although i didn't think it'd be this impressive.
 


you want to claim I am wrong yet you only offer your opinion. APU cores are weak, always have been.

CPU_001.png


the lowly fx-4100 is 46% faster than the A10 5700 and clocked slower. The fx-4320 is 58% faster and the fx-8350 and the I5 3470 is 73% faster. good luck catching that i5.

This is the same architecture. Id call that a weak ass core any day.

CPU_01.png


again, 4320 = 40% faster, 8350 = 56% faster, I5 3470 = 73% faster.

CPU_01.png


4320 - 52% , 8320 - 83%,I5 3470 - 89%

CPU_03.png


4320 - 33%, 8320 - 50%, i5 3470 - 62%

... there is plenty more where these came from

now tell me using these RL results how removing the L3 cache in favor of making the cpu weaker and adding an IGP is such a great thing to ever happen to the computing industry?

How is kaveri going to catch the I5? 8350? ... heck even the 4320 ...

Id say the a10 5700 is a pretty good spot to start with Kaveri figures since they both clock at 3.7 ghz. As you put it "in ordinary cpu workloads" shown above, its starting at negative 40% already just to get to the 4320.

Kaveri is not going to be the answer your hoping it will be. It will not catch the i5 3470 very often (if ever,) it may here and there, but that will take some luck "in ordinary cpu workloads"

Tell me just exactly how I am "just plain wrong"?
 


I could give my opinion on why there is no Steamroller FX or similar chip (Phenom or whatever brand you chose), only 'Athlon' version of Kaveri (i.e. APU with the iGPU disabled), but I think that AMD can do it better:

Lisa Su, Senior VP & GM of Global Business Units at AMD, delivered the opening keynote and the message was clear: AMD is positioning its Accelerated Processing Units (APUs) -- which combine traditional multi-core CPUs and a discrete multi-core graphics processing unit on a single chip -- to dominate the market from smartphones to servers.
Click to expand...
 


Both are average values

http://techreport.com/news/25636/amd-kaveri-hitting-desktops-on-january-14-with-mantle-support



That is from Lisa Su 2012 talk. This is from this week talk

73727.jpg


Finally reduced to 856 GFLOP.
 


I already said this to you. Also you continue using essentially the same logic than in your previous attack to AMD ARM line.

Here you pick very old benchmarks (not optimized for AMD architecture, one of them is a Nvidia sponsored game) and an old 5700 Trinity APU and you claim this is how Kaveri will perform.

You miss the estimation of Kaveri CPU performance in the BSN* article, you miss the leaked benchmarks comparing Kaveri to Bulldozer and Piledriver FX, and you miss the BF4 benchmark given by AMD during October talk, where a Richland APU got the 98% of the performance of a FX-6350 and the 96% of the performance of the FX-8350 (the three using a R9 280X and playing @ 1080p ultra). I think you continue missing the subsequent discussion on multiplayer BF4. And you miss that AMD is in the consoles now, with a CPU based in jaguar cores.

Can you compare the performance of the PS4 CPU to Kaveri CPU? I can.

Maybe you don't still understand this, but game developers will be offloading the consoles CPUs and running the heavy computations on the consoles GPUs. That is why both consoles have GPGPU abilities and HSA support.

You also miss that MANTLE aims to liberate some CPU bottlenecks that exist in current gaming technology. This is from Oxide talk at APU13:

Mantle Unleashed: How Mantle changes the fundamentals of what is possible on a PC. Over the last 5 years, GPUs have become so fast that it has become increasingly difficult for the CPU to utilize them. Developers expend considerable effort reducing CPU overhead and often are forced to make compromises to fully utilize the GPU. This talk will discuss real-world results on how Mantle enables game engines to fully and efficiently utilize all the cores on the CPU, and how it’s efficient architecture can eliminate the problems of being CPU bound once and for all.
Click to expand...

 
Sorry Juan but I fail to see how 5700 is old compared to 4350.... The game doesn't matter when comparing apples to apples. 5700 vs 4300, 5700 lose sometimes by 50% add 30%/50% (Kaveri) to that and you're still behind or equal to.

"Richland APU got the 98% of the performance of a FX-6350 and the 96% of the performance of the FX-8350" This I would argue is a un-optimized part of the game... As it does not use the cores of the 6350 and 8350.
 
How does this even remotely make sense to you?

The company that sells a 200w+ stock CPU is too afraid to push their desktop APU past 100w?

http://semiaccurate.com/2012/11/20/overclocking-amds-fx-8350/

Thomas got his FX 8350 to run at 4.5ghz with 1.45v on the stock FX cooler.

Please tell me again why 100w limit needs to exist?
Click to expand...

Because APU's were designed first and foremost for mobile. I'm guessing the APU's can't handle the load.

Steamroller APU will have a brick wall of frequency just like Intel CPUs do when it comes to OCing, I guarantee it.
Click to expand...

Intel runs into a brick wall because they eventually run into a giant GPU bottleneck. Remember our previous discussion on Draw Calls? There is a point where the CPU is going fast enough where you aren't going to improve the latency involved in Draw Calls any farther. At that point, the GPU is the primary system bottleneck. So the fact Intel eventually hits a brick wall is evidence they are essentially maxing out for that given program, and the GPU config, even dual-GPU, is the primary bottleneck. Meanwhile, AMD is still showing linear FPS gains, even at 4.5GHz, indicating they haven't maxed out the CPU yet.

In simpler terms: Intel has basically maxed out its current architecture. Hence the lack of performance for each new release.
 


Old APU compared to Kaveri. Old games are not optimized for APUs, because they were absent in previous consoles.

Let us assume that the benchmark did not use all the cores, as you say. Assume that it only uses 4 cores, the benchmark still shows a 4-core Richland APU (without L3 cache) competing with 4 cores of the FX chips (L3 cache). Substitute the FX-6350 and 8350 with a FX 4350 and the result is the same. The FX-4350 will not be 50% faster than the FX-6350 ~ APU.

You have to approach to the gaming performance of Kaveri by combining all the factors: (i) games are now being developed for balanced APUs, (ii) APUs run faster under W8.1, (iii) APUs run faster with last drivers, (iv) Steamroller improvements over Piledriver, (v) HSA improvements (including hUMA contribution to memory bandwidth), and (vi) MANTLE.

Add (i--vi) together and Kaveri will be a good APU for mainstream gaming. Enthusiasts can purchase a 9000 series FX or an i7-4770k and a pair of high-end cards.
 
Juan we do agree that richland and piledriver is basicly the same, expect for powermanagement and the like, right? what does the richland APU have that the fx doesn't? And specifically what is it that makes the APU be the only one recieving performance increases:

The six factors, Except for the piledriver -> steamroller/HSA (Won't help in singlethreaded), doesn't everything else apply to fx and older apu too..?
 
You guys are fighting over this when you're missing something obvious.

APU will be strong in games where Intel was usually strong (low threads) and FX will be strong in other fields.

You are making the mistake of thinking that a game can scale to many cores all the time if it is capable of scaling to many cores or that if it can't, it never will.

The second one is true but the first one is not. I recall FC3 running at 15fps on my 5ghx FX 8350 and 1.3ghz 7970 and glancing over at my second monitor and seeing one core fully loaded while the others do nothing. I also recall playing the same game and watching total CPU usage go past 75%, meaning great core scaling.

I can guarantee you that the slide with APU right behind FX 8350 that was part of AMD's presentation was a very undemanding part of the game, think like that part where it's all scripted.

Put APU and FX 8350 in a 64 person multiplayer map and I don't think 2m/4c SR would keep up with FX 8350 even with 30% total improvement in speed.

It is really easy to cherry pick benchmarking scenarios to make a point with hardware. It should be obvious from the fact that there's much inconsistency between BF4 and BF3 benchmarks. Do you think the ones with FX 8350 on top are lying while the ones with i3 doing better than FX 8350 are the only true ones?

No, they are both true, the reviewer is either inadvertently or intentionally picking a part of the game to benchmark which shows certain strengths and weaknesses.

Yes, you will find benchmarks in games where the game engine only uses 4 or less threads and SR APU will be doing very well. And then you will see games where FX 8350 walks away.

But if you are going to say that SR APU is equal to FX 8350, then it must be equal to it all the time. If you are going to say they are not the same, then they can be same at some times but there must be a time when the FX 8350 walks away and the APU can't touch it.

What bothers me is that the benchmarks noob posted are not using more than 4 threads. Meaning that it's NOT going to give FX an advantage in those benchmarks yet FX does have an advantage.

Find a benchmark where 3930k and greater pull ahead of Intel Quads and you will see APU can not keep up.

This just bothers me to no end. AMD gets consoles and gets game developers to optimize for many cores, and then turns around and releases nothing higher than quad core.

It's AMD squandering their software and hardware advantage right out of the gate, which is no surprise to me at all. I expected them to be ready with new 8 cores and shiny benchmarks in highly multi-threaded parts of multi-threaded games with an SR FX CPU with 290x or even TWO 290x in crossfire and instead AMD is here talking about how Mantle will make their $130 APU get close to 60fps and how the $130 APU can run BF4 on medium at 1080p at 31fps.
 


Alright!
 


The APU lacks L3 cache. I don't know if there are other differences which AMD has not disclosed. What I know is what follows.

The BF4 benchmark given at the October talk to OEMs shows a Richland APU performing as well as FX-6350 and FX-8350. The APU gives a 98% and 96% of the performance of each FX respectively.

The 'old' argument that FX is 30--50% faster because has L3 cache doesn't apply here.

The argument that Suez map is single player and not well-multithreaded doesn't mean anything, because a FX-4350 will not be 30--50% faster than the FX-6350 and FX-8350.

The argument that the optimizations for APUs also work for FX is not reflected in the BF4 benchmark. If the performance of the APU increases up to match the FX-6350 and the FX-8350. Why has not increased the performance of the FX chips? The FX chip run also W8.1, the same drivers and used the same 2133MHz memory and HDD than the APU.
 
Interesting data about MANTLE disclosed at APU 13.

http://translate.googleusercontent.com/translate_c?depth=2&hl=es&ie=UTF8&prev=_t&rurl=translate.google.com&sl=fr&tl=en&u=http://www.hardware.fr/news/13445/apu13-amd-mantle-premiers-details.html&usg=ALkJrhgnY37aJJ61I7lorqV5I0j-_HjSSw

A part that caught my attention was that about how AMD would pair a APU with a dGPU. Previously in this tread two options were discussed. In one the iGPU was devoted entirely to compute and the dGPU to graphics. In the second option the iGPU and the dGPU would work in tandem doing compute or graphics or both. This is from the talk:

The application also acquires the ability to take control of the multi-GPU and to decide where to run each command issued. Why AMD has provided in Mantle access CrossFire compositing engine, data transfer between GPU etc.. Will allow multi-GPU modes that go beyond the AFR and adapt better example to use GPU computing in games or asymmetric multi-GPU systems as is the case for APU combined with a GPU. For example it is possible to imagine the GPU load based rendering and APU handle the post processing.
Click to expand...
 
But Juan that is a single case out of many, okay so in the BF4 it won't be 50%, what about the million other applications (Which I btw doesn't think is right either, 10-25% is more likely).

"It's not reflected" in a canned benchmark by AMD who is trying to sell APU's. Would it not be odd if AMD choose a benchmark where the richland would perform poorly? That would be terrible marketing. And before you go claim that "this is how it's going to be" , you have to have more of the picture. Meaning more than one benchmark.

If we look at the 5800k vs 4300 in gaming, which is what most of us is interested in:
http://www.anandtech.com/bench/product/675?vs=700 (Yes anandtech, both are AMD chips so there shouldn't be a problem right?)
4300 - 5800k - 7850k?(Added 30%)
118 - 97 - 126 Dragon Age
64 - 51 - 66 Dawn of War
82 - 73 - 95 WoW
44 - 41 - 53 Starcraft

Not a horrible result but I doubt it is ~i5 territory.
 


get over yourself. Very old benchmarks? I specifically looked for games released in 2013 ... ermago ... thats soo friggin old.

a10 5700 = piledriver without l3 cache
4320 = piledriver fx
8350 = piledriver fx

whats soo old about the a10 5700? all that richland brought was higher clock speeds over trinity, kaveri brought lower clock speeds. you can't compare the 4.2 ghz richland to the 3.7 ghz kaveri and say its going to be 30% faster on top of being slower clock. 3.7 ghz piledriver vs 3.7 ghz kaveri is a fair comparison.

If anything id describe AMD's APU as a gpu with an integrated cpu while Intel is making cpus with an integrated gpu.

The reason your stuck on this "BF4 ERMAGO BENCHMARK" is because it doesn't stress the cpu AT ALL. Its a gpu bound benchmark. Thats why it was chosen, its a tactic called marketing.

As for the BSN article, its was pulled from your website.
 
Status
Not open for further replies.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom