• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Some Trinity A10-5800K numbers

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Frakk

Member
Joined
Dec 3, 2011
Location
UK
http://news.softpedia.com/news/AMD-...sted-Provides-50-Faster-Graphics-259848.shtml

Those impressive GPU improvements are very welcome AMD, so i'm not just picking on the deflating data buried in the numbers.....

As you can see the CPU performance is up 12.8%, now that's not to be sniffed at, that is until you analyze those numbers.

Llano A8-3850 runs @ 3.2ghz (Turbo)

Trinity A10-5800K @ 4.2ghz (Turbo)

That's a 31% clock increase over Llano

Take the 12.8% away from the 31% gives you 18.2%

Not taking into account clock scaling droop, 18.2% is how far Trinity is behind Llano (per core - per clock)

Llano is an Athlon II, which is a Phenom II minus L3

I don't have the exact numbers in my head but Bulldozer is about 20% - 25% behind Phenom II.

I probably don't need to finish what i'm saying here.
 
Last edited:
Yeah, bummer. I'd definitely buy that sort of thing for a laptop (I have an E350 APU in a netbook and I really like it) but not for a desktop. It's an interesting direction - they just need to keep CPU performance up to be competitive.
 
Piledriver is going for higher clock speeds at the same TDP.

So lets assume worst case scenario, (clock + core performance is 0% over BD)

The i5 2500K is 44% faster at stock per core than BD

Realistically assuming PD will be clocked 20% higher @ 4.2Ghz it would put PD 25% behind the 2500K per core, and 7% ahead of Thuban.

In multi threaded it would top the i7 2600K by about 5%

Now, IF PD Overclocks to 5Ghz on something like a CM Hyper 212 Evo, and has a down clocked FX-83## part costing less than the 3570K (as BD currently has) then that would be ok, not great but ok.

If this thing turns out to be an overclocking monster then that 'would' be great!

I hope it will have 5 - 10% clock for clock / core for core performance increase over BD and overclock to 5Ghz on a reasonable priced cooler, yet still costing less than a 3570K

I hope that is not to much to ask.
 

Attachments

  • 41694.png
    41694.png
    28.7 KB · Views: 204
  • 41695.png
    41695.png
    29.9 KB · Views: 203
Last edited:
Nice job on the GPU side... but thats about it. IPC needs to come up to get enthusiasts on the AMD side Im afraid.
 
Nice job on the GPU side... but thats about it. IPC needs to come up to get enthusiasts on the AMD side Im afraid.

ED, i'm quite worried about this... On the Trinity side there great, IPC performance is not that important, what matters there is [cost / efficiency / IGP performance]

For the market APU's are aimed at they are superb.

It is the enthusiasts market that looks on very thin ice, once the price of them settles it will no doubt cost less than Intel, it will also overclock pretty well no doubt.

But if those performance numbers pan out they way they look to AMD will have an impossible task trying keep the customers they still have, let alone to get some of its lost customers back on board.

Yes they absolutely must get the Per clock performance up, even just a bit along side the large clock speed increase.

If not they may have to dip out of the enthusiasts market all together, leaving Intel with no competition at all.
 
Sorry, that was more in reference to desktop PD... and really didnt have a place here. My bad. :)
 
'steamroller'. Is this what everyone is hearing? = If you mean are we sort of hearing the same rumor...the answer is yes and makes sense overall.
 
It appears that AMD is not worried about performance anymore on the CPU side but rather putting the emphasis on the GPU side:
http://www.decryptedtech.com/index....n-every-laptop-on-the-planet-today&Itemid=139

I heard that they plan to ramp up performance again with 'steamroller'. Is this what everyone is hearing?

Apparently yes, Steamroller is set to be the next 'real' performance jump for AMD.

The problem is the last time AMD where competitive in performance is when they launched the chip you see in my sig, that was against the i7 9## series which shortly followed.

Thuban remains AMD's most powerful chip, since then Intel have successfully leveled up by a significant chunk in the form of Sandy Bridge.
Meanwhile AMD have gone backwards by about the same amount Intel went forward, by the time they get back to a where they were with Thuban (steamroller) Intel will have stepped it up again.....

Reading what you posted there i have to wonder if AMD have completely lost the plot?

CPU performance is not important, people do not care as there moving to Cloud Computing?
That line of thinking explains why there just bought out a server maker.

I'm sitting here thinking,- are they just looking for excuses to hide the fact that there struggling financially to keep up, they can't afford to compete with Intel.

Or should Rory Read loose his job?

Why did they throw good money at bloody servers?

BTW,- AMD.... PC Game sales this year are at the highest of what they have ever been!

Get a grip, the enthusiasts market is going to be going strong for some years yet.
 
Last edited:
Apparently yes, Steamroller is set to be the next 'real' performance jump for AMD.
For the past couple of years, it seems like every time an AMD product gets released the general response is "don't worry, <next core architecture> will totally fix this problem!". Kinda silly.

I see what AMD is going for with the integrated GPU market; I like the APU idea. I think they're also banking on much wider support for GPU accelerated apps in the future. If everyone really jumps on board with OpenCL optimization the performance difference would be much smaller, or possibly even in AMD's favor. That's a big gamble, though, and unfortunately it means that AMD is looking really weak in the short term.
 
For the past couple of years, it seems like every time an AMD product gets released the general response is "don't worry, <next core architecture> will totally fix this problem!". Kinda silly.

I see what AMD is going for with the integrated GPU market; I like the APU idea. I think they're also banking on much wider support for GPU accelerated apps in the future. If everyone really jumps on board with OpenCL optimization the performance difference would be much smaller, or possibly even in AMD's favor. That's a big gamble, though, and unfortunately it means that AMD is looking really weak in the short term.

Oh yes i get it, manufacturers are looking for ways to build xbox 720 / PS4 ect... into the back of slimline TV's. (As one example)

For that the GPU / CPU needs to be tiny, cool and very efficient. which is exactly what PD based APU's are, especially the 17.5w and 25w ones, they are quite impressive frankly, and Intel on that front are already lagging behind falling even more behind...

Add to that AMD convinced Adobe to take advantage of the powerful compute advantage APU's have with CS6

More such things will no doubt follow, thus rendering IPC performance there increasingly pointless.

BUT, IPC performance matters in so many area which can not be side steeped by IGP compute performance its not going to have a big enough impact.
And AMD's bet the enthusiasts market is already on its last legs is vastly misplaced, in fact it continues to grow, that will stop and will one day be consigned to the history books, but not for a long while yet.
 
I typed into g00gle: computing + with + graphics + processors. Many interesting hits. I had been just glancing at the phenomenom of using graphics processors for real computing power. Below is link that goes to g00gle page I went to. However that is not a direct link to the site where that information was located. What we as focused home users maybe missing is the big picture. Again that link is full of high-powered links for study of what is on the horizon and AMD is fully in the mix.


The straight path of continuing to lash together ever-increasing numbers of microprocessors has hit a wall. Titan, whose design relies heavily on the use of many thousands of graphics processors — the kind of chips originally designed for gaming. Titan has the potential to be twice as fast and three times as energy efficient as the fastest supercomputer in the world today, Japan’s K computer.
 
Not that it matters too much but I was just curious with PD's improvements how much performance is being left off the table by not utilizing the scheduler improvements that come with Windows 8. If anyone here reviews this cpu down the line make sure to add win 7 (without hotfixes and with them) and win 8 consumer preview benchmarks. Win 8 might be at RC by then.
 
The scheduler thing was bunked I thought. People tested BD on W8 and found little to no gains.
 
The scheduler thing was bunked I thought. People tested BD on W8 and found little to no gains.

I could be wrong but it was my understanding that as Windows 8 develops, they would be collaborating with AMD to further enhance the scheduler.
 
I could be wrong but it was my understanding that as Windows 8 develops, they would be collaborating with AMD to further enhance the scheduler.

Right..... Win 8 was but a mear shell with BD tested on it, Win 8 is still not much more than Win 7 in a new suit.

MS are apparently including some optimisations into Win 8 for both BD and ARM CPU's

Do I think it will make any difference? if any, little.
 
I could be wrong but it was my understanding that as Windows 8 develops, they would be collaborating with AMD to further enhance the scheduler.
Keep chasing that proverbial carrot... I hope for AMD's sake it improves things by leaps and bounds howeer. :thup:
 
Back