AMD CPU speculation... and expert conjecture

Page 357 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Status
Not open for further replies.


bf3_2560_1600.gif

crysis3_2560_1600.gif

metro_lastlight_2560_1600.gif
 


When I gave the link to the October talk by the first time

http://www.tomshardware.co.uk/forum/352312-28-steamroller-speculation-expert-conjecture/page-179#11836419

I mentioned explicitly that Kaveri APUs will be named A10-7x00. Evidently the new A10-6790K is Richland... and however it is very close to performance of FX-6350 (98%) and FX-8350 (96%).

As said then, that slide #13 seems to indicate that AMD plans are to replace the entire FX line by APUs.

If some dick or jane ask me anything with courtesy and good manners, I will help; if by the contrary he or she has bad manners and writes insults, then I simply laugh.
 


Do you read the posts that you reply? No comparison of Kaveri APU, with Bulldozer and Piledriver FX in your link.

http://www.chinadiy.com.cn/html/21/n-11921.html

That is the chinese site comparing Kaveri APU, with Bulldozer and Piledriver FX :pt1cable:



You missed that multiplayer was discussed before in this thread. I also gave my estimation of the performance of the new A10-6790K in multiplayer :lol:

Try again.



SARCASM

Kaveri will be 0.006% faster and then AMD will rethink their plans and will release a new FX Bulldozer chip with 21 cores at 6.66GHz, with turbo of 6.67GHz

END SARCASM
 


1. I don't play at 1600p, but thanks
2. I play games that are advanced on Nvidia, but thanks for the information, ill be sure to tell me friends to get the 290x. (>.>)

As a side note, any news on the 780ti's real performance?
 


they got their information from a similar site to yours, wich doesn't even explain how they came up with a 1.8 ghz bd/pd fx cpu score (probably with some flawed math calculations)

Its even on the chineese web page as a disclaimer.

Concrete results we would not describe, because foreign media gives us a report containing bulldozer, Piledriver, and rollers (Kaveri APU) III processor performance comparison, reference of exposure results today.

original:
http://amdfx.blogspot.com/2013/10/amd-kaveri-benchmarks-more-steamrolling.html

but something interesting from the chineese site

But better than glad too early, after all, it was only a project template, its performance is how we have to wait after Kaveri APU is available, we will know, and desktop version of the FX series processors until at least the middle of next year to meet us.

wow ... your refrence even mentions fx cpus. seems you're the only one set agianst it.

As for your BF4 mp complaint, I gave you another single player game option to discuss. why did you ignore it?

While (Juanrga = wrong) {printf("JUANRGA IS NEVER WRONG")};

Is kaveri going to improve 40% to catch the 4320 or 60% to catch the 8350 in that other option? Why would marketing choose BF4 (single player) over Batman?
 
I think that AMD replacing the fx line with apus is great, If you think about all of the LGA 1155/1150 i7 processors are APUs, now IF amd can make an apu that can compete in this market segment, as the fx 8xxx series was originally intended for, then that would be great. In order to do so they need an apu with either 8 cores or 4 cores and 8 threads. I don't think that this is too far out. In fact I have thought ever sence I joined this thread that the steamroller for the enthusiaest market would be a A-10 on steroids.
 

okay. if you are so confident, why not post them here, in this thread? why are you so reluctant to posting the calculations and the results?
your current actions speak contrary of your confidence. they actually show the lack of confidence.

i have obtained some approximations from the slide #13, but i still need your full data for verification and comparison. i also need your initial calculations and results. please post them. only you can do it.

i re-assure you i am not posting insults. i even edited out as much sarcasm as possible so that you don't get the wrong impression. please do not confuse accusations(based on your actions) as insults. you may feel insulted but i cannot help you with that. imho, your actions and constant avoiding only solidifies my suspicions. after you prove me wrong, i'll withdraw my accusations and apologize - simple. i've already explained why i cannot find the information you keep refusing to post. you're the only one who can track your own posts.


edit:
summarizing the chinese site's and planet3dnow.de's info(speculations):
the 2M sample used in the benchmarks have :
base clock 1.8 ghz
lowest clock 1.2 ghz
and turbo clockrate 2.3 ghz.
it had 2 modules, 4 cores.
2x 2MB L2 cache.
a0 stepping.
may correspond to a 35w kaveri apu, possibly aimed at notebooks/laptops.
igpu clockrate may be 500 mhz.
the benchmark used was BOINC.

the a1 stepping sample in the chinese site's link may have increased lowest clockrate of 1.4 ghz.
curiously, the first sample's test bench had 32 GB ram, the second sample's had 16 GB ram. but the 2nd one's performance was shown improved. no mention of system memory data rate or model.

additionally, those links say that floating point calculation capability may have been hampered by lack of L3 cache. is there any crediblity to this? afaik games also use fp capabilities in a cpu, and if such cpu lacks L3 cache, won't it's performance be less than ones with L3 cache ((same cpu core uarch) as long as the game is actually taxing the cpu)? 😗

 


The title would suggest otherwise.

More performance with More Cores
 



The Crossfire performance is similar at other resolutions. E.g.

bf3_1600_900.gif

bf3_5760_1080.gif


The 780 Ti

perfrel.gif

 


In your head that must be true. Reality is very different :lol:

Those are scores measured using ESs. All them are clocked at 1.8GHz. This was explained before, but you don't read.



:rofl:



'original' is only in your head again. That site repost from others.



Yes a site reporting leaked benchmarks comparing Kaveri to BD/PD FX mentions FX. WOW! that is truly impressive.

And of course I am not against FX. That is again in your head. However, you seems very reluctant to the idea that AMD must be replacing the entire FX line with APUs. Otherwise you would not be doing all this show.



The answer is very simple. The point of this discussion about the slide #13 was explained three or four times before. A new explanation will not change anything.
 


Not only it is great the idea of AMD replacing FX line with APUs(*), but matches with what AMD is doing in servers where the Opteron 3330 CPUs (4 and 8 PD cores) are replaced by 4 SR core Berlin APU (*).

I would like to see 8-core APU, but there is none in the near horizon.

(*) And the CPU variant.
 


That is funny because I am the only who has posted both absolute values of FP and percentages. Also I already explained you that intermediate calculations depend of the media used to analyze the slide.



Or you can spend your precious time searching the post where I wrote the numbers in a 2x3 'table' instead asking others to search for you.



You insulted others calling them liars.



No. The lack of L3 means nothing. It is not affecting the integer results. The floating point results must be related to SR using a new simplified (streamlined) FPU, although AMD said that would not affect performance. Other ES of Kaveri don't show that performance regression of the FPU. Therefore, maybe there had a problem with that test or a problem with that ES unit. We don't know now.



Yes. The FX-6350 is ~2% faster and the FX-8350 is ~4% faster in that slide.
 
Good and bad news (if this is legit)

http://www.bitsandchips.it/9-hardware/3614-kaveri-notevoli-miglioramenti-per-la-cache-l1-ed-l2-non-per-l-imc-ddr3

The bad is that IMC is not significantly improved. This means that the leaked benchmarks showing a 2x memory bandwidth with 1600MHz were probably measuring aggregate bandwidth in dual graphics config.

The good is that the L2 cache is a 20% faster. Up to now all the known details were that the L1 cache was improved, but the L2 cache had the same speed and only receive a power consumption improvement.
 


An 8 core APU is probably not possible at 28nm right now, but AMD still could have a hyper threading technology for the enthusiast APUs.
I think that the new branding for the FX line will be as follows:

4xxx series: Athlon
6xxx series: A-10
8xxx series: Phenom
A-4 will still remain low power
A-8 will be between Athlon and A-10

Remember the Phenom IV rumor, hence the Phenom Branding

Amd recycled the Athlon name for the low end, so I would only wake sense to have the Phenom back as the high end.

I still think that there will be an 8 core or 4 with some kind of hyper threading for the A-12 or Phenom line.

Remember Slide 13 said MORE performance with MORE cores.


 

did you post them? could you re-post them, please? i'll address your 'explanation' after you post all of your findings.

i did search, and could not find any. i could say you never posted them but i know about tom's forum search function, it tends to skip relevant results some times.
and i did not ask anyone else to search, i asked You. You, who posted them in the first place, and in possession of the method of calculations as well as the results. please post those.

don't make this about others. i accused you of lying.
and.. to make this crystal clear, i didn't call anyone else liar. at least, to my knowledge, not without reason. like i said, don't make this about others. oh, you still haven't posted the data. i hope you stop avoiding.

see, you post only the end results, not the calculations or intermediate results. those would be really helpful.

edit:
i was searching the site (that juanrga linked) for more kaveri info and came across their a10 7800k table. it's a summary for their findings(rumors):

APU [A10-6800K] [A10-7800K (speculated model number)]
Production Process [32nm PD-SOI] [BULK 28nm Gate First]
Core [Richland] [Steamroller B]
Frequency (Turbo) [4.1 GHz (4.4 GHz)] [2.9 GHz (3.2 GHz)]
iGPU [VLIW4] [GCN 2.0]
Stream Processor [384] [512]
Frequency iGPU [ 844 MHz ] [1 - 1.2 GHz]

from what i know about substrates, the choice of silicon and clockrates don't seem very encouraging..... hopefully the real deal would be better. i mean.. intel's cpus are made on bulk and they can hit up to 3.9 ghz turbo at stock. it'll also depend on how improved kaveri's (in turn steamroller's) turbo controller and in-built thermal sensors turn out to be. especially in laptops where cooling would be restricted. the igpu clockrate looks good. imo, the igpu looks very powerful even from the speculated specs. 512 gcn 2.0 cores at 1 ghz base clock is.... monsterous for an igpu. as long as amd feeds this monster well, it'll deliver.

one more thing, hawaii is not gcn 2.0. kaveri is gearing up to jump ahead of amd's existing gpu lineup by a generation. may be temporarily though, since 20nm gpus are very likely to be gcn 2.0.
 
Hopefully there's as many 'k' models as trinity/richland to sort out clock speed if its low, was the leaked photo not a 3.5ghz though? Would this nor suggest 3.9-4ghz turbo?
 


You don't need K model to overclock with AMD APU, you can still OC via the bus. A K model just makes it easier to do. It isn't locked down like Intel CPUs are.
 


They could do it, just have to weaken the iGPU side. The reason AMD APU's are so powerful graphics wise is that 50% of the die is the iGPU. This is also the reason AMD's APU's make no sense in systems with moderately powerful dGPU's. Your using a crippled CPU that you paid full price for. It's also the reason Intel makes a desktop model with an incredibly weak iGPU, if any at all. Minimize the amount of die space wasted on a feature that won't be used.
 


Why make all of Steamroller APUs ?

Why not make one or two plain and simple CPUs for Steamroller, replace the GPU with more CPU power ?
 

Erm, what was I thinking/typing 🙂

But the point still holds - if the 6C/8C FX chips show no differentiation in FPS, that simply suggests the scenario is not scaling beyond 4 threads. You have to ask why AMD would pick such a poor example to demonstrate more performance with more cores, when the extra cores are not even being used. I guess some marketing guy put the PPT together in a hurry.
 


That is what I meant, by saying Phenom, I meant a normal CPU, like the new Athlons.

I agree and 8 core apu is possible with a smaller iGPU, maybe the iGPU is only 25% of the die instead of 50% for a weaker iGPU.

AMD also could also introduce a hyper threading like technology in their enthusiast class chips.

What ever enthusiast steamroller chip is released tomorrow, you can be sure that it will not carry FX branding, due to what AMD released to OEMs. I can also guarantee you that there is a new enthusiast chipset, because my friend from ASUS, whom is a head tech told me that they are working on a new ROG board with a new chipset, that replaces 990fx.
 
What Steamroller should be :

A cpu as good as an i5 2500k.
A cpu as good as an i5 3570k

An APU with weak GPU
An APU with Medium GPU
An APU with High end GPU


That's all we need from AMD.
 


An 8 core APU in a single die seems difficult. It would have a minimum size of about 480 mm2 @ 28nm and that is a big die. But maybe a MCM version by combining two 4 core dies could be made. The problem here is that would be probably too expensive considering the demand.

I think there are two realistic scenarios: Either (i) AMD replaces the entire FX line with APUs or (ii) replaces the FX-4000 with APUs and releases a refresh of the FX-6000/8000/9000 series using the same PD modules used in Warsaw CPU.

If Kaveri is a 20% faster than Piledriver. I think the option (ii) rules until 2015 when Carrizo (excavator APU) is released and then the FX line is abandoned definitively.

If Kaveri is a 45% faster than Piledriver. I think the option (i) rules. The chinese site that I mentioned above says that Kaveri will compete with Haswell i7-4770k.
 
Status
Not open for further replies.