AMD Piledriver rumours ... and expert conjecture

[Reynod]

We have had several requests for a sticky on AMD's yet to be released Piledriver architecture ... so here it is.

I want to make a few things clear though.

Post a question relevant to the topic, or information about the topic, or it will be deleted.

Post any negative personal comments about another user ... and they will be deleted.

Post flame baiting comments about the blue, red and green team and they will be deleted.

Enjoy ...

 

[newb88]

I am excited about Pile Driver but how long can AM3+ mobos last? If Pile Driver is the last generation, i will rather gamble and wait for Steam Roller-based APUs...

 

[gamerk316]

Almost guaranteed a win against Bulldozer. The questions is will the 8350 close the gap with the I7 2600k/2700k/3770k? We appear to know that the 8350 will have a base clock of 4 Ghz (great for the AMD Marketing Department) and a turbo clock of 4.2 Ghz.

How well can it OC I wonder?

 

[esrever]

overclocking the FX-8150 just makes it an 1100T equal, nothing more.
and certainly doesn't factor in the performance comparison vs Intel.
and those clocks are not daily usage clocks like 'verbalizer' stated.
just for show with tons of nitro cooling.

irrelevant in the real-world.

Not really.

 

[cgner]

On May 29 AMD stock was $6.46 and today it's $3.88, which is a decline of about 40% - pretty close to that 50% in a bit less than a quarter (by 2 days).

AMD had to sell more senior notes at 7.5% interest to pay off the senior notes now due at something like 6% interest, which is akin to using a higher-interest credit card to paying off a lower-interest credit card, which is generally a pretty bad financial practice at least for consumers. People who do that are generally one paycheck away from bankruptcy..

40% is nothing. Rest is just formalities that happen every day.

 

[esrever]

Almost guaranteed a win against Bulldozer. The questions is will the 8350 close the gap with the I7 2600k/2700k/3770k? We appear to know that the 8350 will have a base clock of 4 Ghz (great for the AMD Marketing Department) and a turbo clock of 4.2 Ghz.

It will close the gap but not really by enough to be a large factor. Sandy -> Ivy was very small, Bulldozer->Piledriver would be guaranteed to be that much of an improvement from just the clock increase. Depending where you look, the FX8350 should be as powerful or more powerful as a i7 3770k in things like video conversion and rendering while it will be slower than an i3 at gaming most likely.

 

[gamerk316]

And now, we find FPS is a horrible benchmark to use for gaming:

http://techreport.com/articles.x/21516/1
http://techreport.com/articles.x/23246

Granted, I knew some of this, but amazing to see and track the per-frame latencies. Second of the two is most interesting, as you can easily pick out the higher latencies on AMD's chips...

Remember how we always wondered why the PII X4 pulls ahead of both the X6 and FX? Now we know:

Looks to me like the FX CPUs have an Amdahl's Law problem. Even though they have a relatively large amount of cores for their given product segments, their per-thread performance is fairly weak. The Bulldozer architecture combines relatively low instruction throughput per cycle with clock speeds that aren't as high as AMD probably anticipated. That adds up to modest per-thread performance—and even with lots of cores on hand, the execution speed of a single thread can limit an application's throughput.

*whistles*

But its REALLY an interesting read. Kinda makes you re-think about how valid FPS is as a benchmarking tool...Its also worth noting that no processors can spit out a "steady" 60 FPS [16.7ms] rate...

 

[gamerk316]

^^ Well, my point is basically to show that while the AVERAGE performance over a second looks flat, frames are hardly comming out at a steady rate. Within that second, there are frames spat out early, frames that are delayed, hence the famous "microstutter" that you see in dual-GPU configs. The first article even notes NVIDIA has a ton of HW specifically to battle this.

So yeah, frame latencies, on both the GPU and CPU side, are probably the gaming benchmark I've been looking for. Does an overall better job then low resolution testing does of showing how much extra oomph one CPU has over another.

Proposing Toms does its own investigation into this area.

EDIT:

The A8-3850 and i5-3570k come out as the two big winners in this area.

 

[jdwii]

^^ Well, my point is basically to show that while the AVERAGE performance over a second looks flat, frames are hardly comming out at a steady rate. Within that second, there are frames spat out early, frames that are delayed, hence the famous "microstutter" that you see in dual-GPU configs. The first article even notes NVIDIA has a ton of HW specifically to battle this.

So yeah, frame latencies, on both the GPU and CPU side, are probably the gaming benchmark I've been looking for. Does an overall better job then low resolution testing does of showing how much extra oomph one CPU has over another.

Proposing Toms does its own investigation into this area.

EDIT:

http://techreport.com/r.x/cpu-gaming-2012/value.gif

The A8-3850 and i5-3570k come out as the two big winners in this area.

Their is a lot of blue up their and the green is on the bottom.

 

[gamerk316]

^^ But at the same time, Intel really doesn't have a ~$150 option, which AMD does have plenty of.

 

[viridiancrystal]

^^ Well, my point is basically to show that while the AVERAGE performance over a second looks flat, frames are hardly comming out at a steady rate. Within that second, there are frames spat out early, frames that are delayed, hence the famous "microstutter" that you see in dual-GPU configs. The first article even notes NVIDIA has a ton of HW specifically to battle this.

So yeah, frame latencies, on both the GPU and CPU side, are probably the gaming benchmark I've been looking for. Does an overall better job then low resolution testing does of showing how much extra oomph one CPU has over another.

Proposing Toms does its own investigation into this area.

EDIT:

http://techreport.com/r.x/cpu-gaming-2012/value.gif

The A8-3850 and i5-3570k come out as the two big winners in this area.

Isn't that part of what Vsync is for? Though, once you dip below the same fps as your monitor is, Vsync is a liability.

Almost guaranteed a win against Bulldozer. The questions is will the 8350 close the gap with the I7 2600k/2700k/3770k? We appear to know that the 8350 will have a base clock of 4 Ghz (great for the AMD Marketing Department) and a turbo clock of 4.2 Ghz.

In some heavily threaded tasks, I think the 8 core PD will be better than the 4c/8t i7's. BD is already close, and PD is quite a step. In single threaded and games, PD still falls short of SB/IB, but better than PII, if I had to guess.

 

[jkhoward]

don't just speak on it, prove it then.....
let me know when your ready to realize the truth as well
😉

without proof then your words carry no weight.
are you on a fanboi kick too?

I do believe that you are the Intel fanboi for bagging on everyone in this forum about being a AMD fanboi. I highly doubt that the AMD Piledriver will come close to the new Ivy Bridge CPU myself. But this is a forum to discuss what we would like the Piledriver to be, is it not?

 

[jkhoward]

Isn't that part of what Vsync is for? Though, once you dip below the same fps as your monitor is, Vsync is a liability.

In some heavily threaded tasks, I think the 8 core PD will be better than the 4c/8t i7's. BD is already close, and PD is quite a step. In single threaded and games, PD still falls short of SB/IB, but better than PII, if I had to guess.

When I was running benchmarks between the 8-core Bulldozer and Intel i5-2500 the AMD system scored 10K on a multi-threaded application and the Intel scored around 8K. Single threaded performance is where it is at right now. Until the day AIM and Notepad will use all eight core... 😱 ... Intel has AMD beat.

 

[-Fran-]

Isn't that part of what Vsync is for? Though, once you dip below the same fps as your monitor is, Vsync is a liability.

Yes and no, vsync is supposed to "lock" the frames output to your monitor's reported frequency, but it doesn't state "how". That's why Lucid and nVidia have implemented very different approaches to do so. Also, each game should implement it to fit their own input polling, because doing it from the driver will sometimes cause way too much pain.

Unless I'm remembering it wrong (gamerk can correct me here), using vsync is not the same across the board for all games.

In some heavily threaded tasks, I think the 8 core PD will be better than the 4c/8t i7's. BD is already close, and PD is quite a step. In single threaded and games, PD still falls short of SB/IB, but better than PII, if I had to guess.

Even if PD is equal clock per clock to PhII, if it OCs well enough it can be a worthy "upgrade"... I think, haha.

Cheers!

 

[Dangi]

Hey guys I have a question for you

Now that Piledriver "officially" launch is in october, there will be new chipset ??

As far as I know 990X/99FX doesn't have PCI 3.0 capability, yet no card fully uses the bandwich of PCI 2

 

[cgner]

Even if PD is equal clock per clock to PhII, if it OCs well enough it can be a worthy "upgrade"... I think, haha.

Cheers!

Overclock is a moot point. A good overclock usually adds less than 10% increase in actual performance. Archie makes all the difference.

 

[jkhoward]

not what you said earlier.
please get your lie right next time.

LOL.

 

[fazers_on_stun]

Is the "Steamroller and later being discontinued" rumor the one that says that there won't be an AMD CPU for the performance segment after Piledriver?

Yep - from OBR's website which was discussed a couple pages ago. IIRC he was the guy who leaked internal AMD docs showing BD would miss the target performance by a big margin, around this time last year.

 

[-Fran-]

Overclock is a moot point. A good overclock usually adds less than 10% increase in actual performance. Archie makes all the difference.

Not quite. Usually, a 500Mhz clock increase can do a lot for min frame rates. Also, it depends on the "default" clock speed from the CPU how good of a bargain it is. And notice the i5 k are the most recommended CPUs because of just that, they're highly OCable.

With my 965, is from 3.4 to 3.97Ghz, that's a tad less than a 600Mhz increase (with all power savings activated). Or a little more of 1/6th of the default's clock. Considering at 4Ghz it performs close to a 3.1Ghz i5 (2.8 being the default and 3.1 the quad bump), PD will need a lot higher clocks to even dream to get close to a OCed 2500k, let alone a 2600k for multi.

This doesn't paint the whole picture, but helps me demonstrate my point:

I wish I had some sort of sample data for a game, but Cine actually reflects most of what happens in the games world (at least in most benchies) from what I've noticed.

Cheers!

 

[cgner]

Not quite. Usually, a 500Mhz clock increase can do a lot for min frame rates. Also, it depends on the "default" clock speed from the CPU how good of a bargain it is. And notice the i5 k are the most recommended CPUs because of just that, they're highly OCable.

With my 965, is from 3.4 to 3.97Ghz, that's a tad less than a 600Mhz increase (with all power savings activated). Or a little more of 1/6th of the default's clock. Considering at 4Ghz it performs close to a 3.1Ghz i5 (2.8 being the default and 3.1 the quad bump), PD will need a lot higher clocks to even dream to get close to a OCed 2500k, let alone a 2600k for multi.

This doesn't paint the whole picture, but helps me demonstrate my point:

http://dl.dropbox.com/u/8247566/cine_3.98ghz_TuneUP_64bits.jpg
http://dl.dropbox.com/u/8247566/cine_i5-2300_64bits.jpg

I wish I had some sort of sample data for a game, but Cine actually reflects most of what happens in the games world (at least in most benchies) from what I've noticed.

Cheers!

Yes, thats kinda my point.

 

[fazers_on_stun]

Well, I think the lowest stock value ever was $1.82, in November 2008. To lose roughly 40% in a bit less than a quarter is not a good thing, but the stock market certainly isn't the place wherein you'll find definite proofs of a company being healthy or unhealthy -- it's just a single part of the equation. Intel stock falls short of being a performer, and Microsoft, for instance, has been a disappointment for almost 10 years; but still, Microsoft arguably has its strongest portfolio ever and many record results, despite the challenges ahead. AMD isn't in a Microsoft-like position, that's for sure; but it's doing a lot better in all departments than it did in the very recent past.

The senior notes part seems quite worrisome, but I don't understand it well enough to comment any further.

Still, an other outlook is that the new management is shaking things up and seems to bring a lot of fresh air into the headquarters. If many people who were deemed key to the company have left, many big names -- including the CEO -- have come from a much more comfortable position in the industry and accepted the challenge.

Thought it was a buck sixty-three, but not important . The worrisome part would be the low guidance AMD provided last quarter, plus the fact that so far no PD (shades of BD a year ago), plus Intel threatening with Haswell.

The senior notes show AMD is doing what it has to to continue in business - not a good thing but at least they are not declaring Chapt. 11 or anything. People (more likely corporations) buying these are a bit more secure than ordinary stockholders, because the notes are "senior" to any other equity such as stock certificates, so the note-holders would get paid first in event of a bankruptcy and liquidation (unlikely IMO at this point).

If there should be a global crisis (and my personal opinion would be Israel going at it with Iran, in the next couple months, > 50% probability, and shutting down a large part of global oil shipments for some time), then all bets are off of course. There could be some fairly long-lasting consequences and weaker companies dying off.

 

[fazers_on_stun]

40% is nothing. Rest is just formalities that happen every day.

Well you asked "Are the stocks losing 50% of value in a quarter?" so that's about what I did 😀.

And if you go by stock value since March 26th ($8.24), it's more than 50% - 53% to be precise, at todays close of $3.85. That would be 5 months BTW.

And no, having to sell higher-interest senior notes to meet previous notes, is not common at all. IIRC the original notes, due this year, were convertible to ordinary stock if the stock price hit $20 a share. Unlikely at this point.

BTW, all this is 'been here, done that': http://www.tomshardware.com/forum/264129-28-loss-worse-expected-stock-slumps-after-hours-trading

 

[esrever]

don't just speak on it, prove it then.....
let me know when your ready to realize the truth as well
😉

without proof then your words carry no weight.
are you on a fanboi kick too?

I responded with as much evidence as your post, why be mad at me?

http://www.hardwarecanucks.com/forum/hardware-canucks-reviews/47155-amd-bulldozer-fx-8150-processor-review-21.html
here is FX8150 vs 1100t at stock, its pretty even. Taking out all the synthetic tests, the 8150 wins vs the 1100t.

Now I say this back to you:
without proof then your words carry no weight.
are you on a fanboi kick too?

 

[cgner]

Well you asked "Are the stocks losing 50% of value in a quarter?" so that's about what I did 😀.

And if you go by stock value since March 26th ($8.24), it's more than 50% - 53% to be precise, at todays close of $3.85. That would be 5 months BTW.

And no, having to sell higher-interest senior notes to meet previous notes, is not common at all. IIRC the original notes, due this year, were convertible to ordinary stock if the stock price hit $20 a share. Unlikely at this point.

BTW, all this is 'been here, done that': http://www.tomshardware.com/forum/264129-28-loss-worse-expected-stock-slumps-after-hours-trading

40% isnt 50%. 5 months is not a quarter. Deal with it.

 

[-Fran-]

Yes, thats kinda my point.

Well, my point is not about Intel vs AMD; that's the real moot point for me now... Kinda got strayed away there, but i wanted to show that OC does play a factor; a big one done right; just need where to measure it to really notice it.

Regarding PD, I'm just saying: Ok, it seems to clock a tad better (according to what Trinity showed) than BD and it seems to have narrowed (hopefully, REMOVED) the clock per clock disadvantage with the PhII. So, that brute-force speed increase is a welcome one, for PD will be the last AM3+ CPU from AMD (according to official wording, so far).

The FX83xx doesn't have to overcome the i5 2500k (nor the bigger siblings) to fit a lot of folks tastes here, but just overcome PhII steadily at games AND "heavy threading" tasks to make it a good CPU for a (let's be clear here) dinosaur platform. All that with no need for a nuclear reactor to power it, off course. That's not so idiotic to ask, right?

Cheers! 😛

 

[viridiancrystal]

Well, my point is not about Intel vs AMD; that's the real moot point for me now... Kinda got strayed away there, but i wanted to show that OC does play a factor; a big one done right; just need where to measure it to really notice it.

Regarding PD, I'm just saying: Ok, it seems to clock a tad better (according to what Trinity showed) than BD and it seems to have narrowed (hopefully, REMOVED) the clock per clock disadvantage with the PhII. So, that brute-force speed increase is a welcome one, for PD will be the last AM3+ CPU from AMD (according to official wording, so far).

The FX83xx doesn't have to overcome the i5 2500k (nor the bigger siblings) to fit a lot of folks tastes here, but just overcome PhII steadily at games AND "heavy threading" tasks to make it a good CPU for a (let's be clear here) dinosaur platform. All that with no need for a nuclear reactor to power it, off course. That's not so idiotic to ask, right?

Cheers! 😛

Not at all. If BD was more power efficient than it is, it would have been accepted more easily. I thought about overclocking my 8120, but I didn't like the idea of having all that heat in my room. Funny though, I bought a nice big air cooler only not to put it to good use yet. I hope PD is a good overclocker.