• Notice regarding image hosts and attachments. See here.

  • Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Some news on Piledriver

humbug said:
I don't include APU's in a market they would pull out of, as you rightly point out they are good chips.
At least for the low cost and mobile market, and they will do well there.

The problem the higher end x86 based CPU's have is that people use Intel Ivy Bridge as a yard stick, people tend to discount anything that does not measure up to it in single or low threaded x86 as pointless and worthless.

So unless there is a significant catchup to Intel in single threaded performance the AMD bashing will continue, AMD have to show the Piledriver chips can be comparable to Intel so as to take away the ammunition AMD bashers throughout the Internet have.

Consider this, how often do you see people pulling up single or low threaded app's comparing i3's to FX-8150's to say "oh look at how much better Intel are than AMD, even the cheapest Intel chip beats the most expensive AMD chip"

I have even seen that reasoning used to justify an i3 over an FX-8150 in raw performance for all games including highly threaded games like BF3, the sad thing is almost every time that is done people buy into it without question and then flush their money down the toilet on a locked 2 core chip.
AMD have no chance is such a climate, even with app's changing reviewers such as Anand seemingly insist on using old out of date low threaded app's to review CPU's with, ask yourself why they still use Photoshop CS4 and not CS6 when Photoshop CS6 is whats now and is a win for AMD over Intel.
Click to expand...


It is a shame because realistically there isn't going to be significant catchup in single threaded performance, and agressively persuing it is a waste of their more limited resources.

TBH I think the mainstream cpu market is the low cost desktop/mobile. Whatever is left of the desktop market, 75% (home and business) would be better served by an AMD APU.The lack of need for a discrete gpu means smaller form factors without a compromise on graphical capability, almost nobody wants a hulking tower.
As has been covered already, it's also a low cost all in one that could (not amazingly) run a recent game with adequate graphical settings (for now). In terms of gaming the trend is toward better threaded games and it's not long before they will be Intel/AMD agnostic, as you say we see it now on BF3. That also means you wouldnt want to reccomend lower than a 2 module amd apu. With gaming in mind, GPU wise, larger cases could house a card at the later convenience of the owner much as they do now anyway.
 
^^^ for now i would recommend a 2 module Piledrive or Bulldozer or a 4 core Deneb over an i3, simply because they do better in multi threaded app's and they can be overclocked to catch up with the i3 in low or 2 threaded app's / games. if not a little more.

Moving up the 3570K is the only thing to recommend, even if a decent socket 2011 Z77 Motherboard will cost you upwards of £140.

If the 4 Module FX Piledriver turns out to be a high clocking 8 core Thuban, personally i would be happy with that as i do think low threaded games are on the way out, as are low threaded app's.

Having said that i would have to see how it deals with Nvidia's Physx, even Intel overclocked to high heaven struggle with it in Borderlands 2 simply because Nvidia don't want it to run well on the CPU.

Now, Bulldozer does significantly better than my Thuban in 3DMark 11 Physx, which would suggest it's properly threaded, how that translates into Borderlands 2 or other Physx games i have no idea.
I might actually have to buy the game and run tests on it.
 
Last edited:
I agree they'l have a longevity that will better the i3.

I kind of hope Physx fails in the end, and nvidia may be feeding it a slow death with the way they mess about with it. Maybe that's just the bitternes talking :p , my last nvidia was the ol' 8800gt. By chance all amd since then, although doesn't seem im missing out.

If Physx struggles on a clocked i5, it does sound like it may single threaded. A simple way to hamper it. No doubt nvidia have a contingency patch they can whip out to take care of that if any need/threat arises. /conspiracy truth.
 
AMD, doing less with more since bulldozer, lets hope they price it right from the off this time.. single threaded still looks dung.
 
AAAAAAAAAAAAAND wait for the next attempt... AGAIN :p

Please AMD! Pull something amazing off! I really feel like going AMD with my next build for some reason... But at this rate i'll be waiting 4 years before they release a chip better than mine. :rolleyes:

Still, its more fun to cheer on the team that is falling way behind :D
 
Last edited:
acme15 said:
AAAAAAAAAAAAAND wait for the next attempt... AGAIN :p

Please AMD! Pull something amazing off! I really feel like going AMD with my next build for some reason... But at this rate i'll be waiting 4 years before they release a chip better than mine. :rolleyes:

Still, its more fun to cheer on the team that is falling way behind :D
Click to expand...

hhmmmm.... dare i say it, Wait for Steamroller, that's when the real improvements happen.

You may call me an idiot :D

Good Grief, why is it so hard to want something different? Get a grip AMD
 
I've been reading on.... Those benchies seem downright awful. I'm faster on a 4 core, 8 thread i7 950 @ 4.2ghz

8.5sec on piles diver vs 6.049 on i7 950

as for superpi, i'm sub 10sec vs 20 secs. Think I'll be passing on this as a potential alternative - I'll be faster sticking with 2 year old kit!
 
I believe the issue with Superpi is that it uses legacy x87 code which is far from efficient on the Bulldozer arch. A lot of the die space is dedicated to new ISA instructions which are much better when a program takes them into account.
 
Tattysnuc said:
I've been reading on.... Those benchies seem downright awful. I'm faster on a 4 core, 8 thread i7 950 @ 4.2ghz

8.5sec on piles diver vs 6.049 on i7 950

as for superpi, i'm sub 10sec vs 20 secs. Think I'll be passing on this as a potential alternative - I'll be faster sticking with 2 year old kit!
Click to expand...

I think there's more to it than that to be honest. I doubt your i7 cost ~£160 did it?

AMD can't win the performance crown but they can do something with price/performance (maybe :) )
 
Tattysnuc said:
I've been reading on.... Those benchies seem downright awful. I'm faster on a 4 core, 8 thread i7 950 @ 4.2ghz

8.5sec on piles diver vs 6.049 on i7 950

as for superpi, i'm sub 10sec vs 20 secs. Think I'll be passing on this as a potential alternative - I'll be faster sticking with 2 year old kit!
Click to expand...

Bulldozer and Piledriver have no hardware to decode the anicent x87 instructions since it is a new core. It is all done in software. The reason Intel cores have it is a holdover from the Core days. If Intel brings out a brand new core then expect not to see it included, as the current socket 1155 CPUs use cores descended found in the original Core CPUs. The Phenom II used cores which can be traced back to the K7 or K8 IIRC.

TBH,only SuperPi and PhysX use it. PhysX does not run well on even Intel CPUs,when compared to running it on the graphics card, and it seems Nvidia sort of done it on purpose:

http://techreport.com/news/19216/physx-hobbled-on-the-cpu-by-x87-code

It is basically of no use to anyone.


Now, David Kanter at RealWorld Technologies has added a new twist to the story by analyzing the execution of several PhysX games using Intel's VTune profiling tool. Kanter discovered that when GPU acceleration is disabled and PhysX calculations are being handled by the CPU, the vast majority of the code being executed uses x87 floating-point math instructions rather than SSE. Here's Kanter's summation of the problem with that fact:

x87 has been deprecated for many years now, with Intel and AMD recommending the much faster SSE instructions for the last 5 years. On modern CPUs, code using SSE instructions can easily run 1.5-2X faster than similar code using x87. By using x87, PhysX diminishes the performance of CPUs, calling into question the real benefits of PhysX on a GPU.

Kanter notes that there's no technical reason not to use SSE on the PC—no need for additional mathematical precision, no justifiable requirement for x87 backward compatibility among remotely modern CPUs, no apparent technical barrier whatsoever. In fact, as he points out, Nvidia has PhysX layers that run on game consoles using the PowerPC's AltiVec instructions, which are very similar to SSE. Kanter even expects using SSE would ease development: "In the case of PhysX on the CPU, there are no significant extra costs (and frankly supporting SSE is easier than x87 anyway)."

So even single-threaded PhysX code could be roughly twice as fast as it is with very little extra effort.

Between the lack of multithreading and the predominance of x87 instructions, the PC version of Nvidia's PhysX middleware would seem to be, at best, extremely poorly optimized, and at worst, made slow through willful neglect. Nvidia, of course, is free to engage in such neglect, but there are consequences to be paid for doing so. Here's how Kanter sums it up:

The bottom line is that Nvidia is free to hobble PhysX on the CPU by using single threaded x87 code if they wish. That choice, however, does not benefit developers or consumers though, and casts substantial doubts on the purported performance advantages of running PhysX on a GPU, rather than a CPU.

Indeed. The PhysX logo is intended as a selling point for games taking full advantage of Nvidia hardware, but it now may take on a stronger meaning: intentionally slow on everything else.
Click to expand...
 
Last edited:
I might regret getting my bulldozer in the next few days then again i might not. It will be interesting to see how it compares to the bulldozer and how it overclocks.
 
You must log in or register to reply here.
Back
Top Bottom