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 ...

 

[vishalaestro]

i think this time the fx piledriver processor's will match atleast the performance of an intel core i5 2400 because the bulldozer processor was a delayed release ..if we compare an fx 8150 with the 1st generation core i5 processors we will get almost the same performance from both of the i5's and fx 81xx cpu's .. since the ivy bridge is only 5 % increase in per clock performance compared to sandy bridge ..it's a great chance for amd to come back into the game..and also fx piledriver offers 15-20% of performance increase over bulldozer processors i think the game is on..what do you think guys..

 

[sarinaide]

The purists will say no and make arguements that AMD's IPC's are years behind Intel, we will just have to wait for the first synthetics to come out to better say where AMD are. If trinity is anything to go by then it is promising, as to whether its a match or better, don't really want to say that, I think it will match intel on price/performance but perhaps not on hardware itself.

 

[jdwii]

i think this time the fx piledriver processor's will match atleast the performance of an intel core i5 2400 because the bulldozer processor was a delayed release ..if we compare an fx 8150 with the 1st generation core i5 processors we will get almost the same performance from both of the i5's and fx 81xx cpu's .. since the ivy bridge is only 5 % increase in per clock performance compared to sandy bridge ..it's a great chance for amd to come back into the game..and also fx piledriver offers 15-20% of performance increase over bulldozer processors i think the game is on..what do you think guys..

Under multithreaded apps the 8 core PD will probably be competitive with a 4 Core with HT But when it comes to gaming i'm sure core to core(Not clock to clock) it will only be competitive to a Pentium IVY/Sandy core give or take a couple percents. Saying that Amd needs to price the 8 core Piledriver at the I5 or below and make sure street prices are about Equal to a I5.(220-240$)

I'm pretty sure that's how AMD will price their products after all their CEO says their done with the "high-end" which means I7 not I5(mid-range) Pricing. Amd can't price their processors at a high price while giving their users less just because they say their out of the high-end.

Scares me seeing some Trinity laptops come up with a 850$ price tag! Way overpriced they needs to be priced at 650$ and below. Even 600-650$ is pushing it.

 

[amdfangirl]

Marketecture... AMD trinity is clocked higher than Intel, is a "quad core" and has Radeon graphics. To the average consumer the specs make it look faster.

 

[sarinaide]

Market myopia is used for selling things, this isn't new....how much IPC do you really need to play a game, considering its the biggest market for performance systems. At the end of the day a chip is a chip.


I don't really think any AMD user is expecting it to beat intel in the synthetic stakes, but overall it is looking like a step in the right direction.

 

[vishalaestro]

Marketecture... AMD trinity is clocked higher than Intel, is a "quad core" and has Radeon graphics. To the average consumer the specs make it look faster.

even i we see the benchmarks of amd trinity apu's in tomshardware site they offer atleast 10-15% increase from previous gen apu's..that's encouraging

 

[gamerk316]

I always wonder if most 'software' issues are actually slightly busted hardware.

Yep. I've worked on Fault Detection systems for the past few years, and over time, you see the oddest failure modes possible, and they produce the oddest result set. Detecting faulty hardware is NOT easy, especially if the rest of the system is currently running and using said hardware when it finally goes.

Point being, for all you SW Engineers out there: When was the last time you bothered to check the return of a standard C 'printf' statement to see if the output actually succeeded? Its a failure mode so unlikely, you don't even bother to code for it. But if the GPU takes that exact moment to decide to die, your wonderfully coded application breaks in a totally unreproducable way. But at the end of the day, its unreasonable to try and handle every possible exception under the sun in SW.

 

[-Fran-]

Yep. I've worked on Fault Detection systems for the past few years, and over time, you see the oddest failure modes possible, and they produce the oddest result set. Detecting faulty hardware is NOT easy, especially if the rest of the system is currently running and using said hardware when it finally goes.

Point being, for all you SW Engineers out there: When was the last time you bothered to check the return of a standard C 'printf' statement to see if the output actually succeeded? Its a failure mode so unlikely, you don't even bother to code for it. But if the GPU takes that exact moment to decide to die, your wonderfully coded application breaks in a totally unreproducable way. But at the end of the day, its unreasonable to try and handle every possible exception under the sun in SW.

Since University times, HAHAHA.

I'm now a Java kid, so I don't have to deal with that. xD!

And regarding PD... It looks good so far, so I expect to see a jump in performance for the avg consumer that gets a PD based computer. I wonder if it will suffice for me though... A PhII at 4Ghz is still a very competent CPU.

Cheers!

 

[lilcinw]

If I am able to snag an FX-8350 for less than $250 I will be all over it. Otherwise I will have to drop to an 8320.

Regardless it looks like it will be a good boost over my FX-4100.

 

[gamerk316]

Since University times, HAHAHA.

I'm now a Java kid, so I don't have to deal with that. xD!

And regarding PD... It looks good so far, so I expect to see a jump in performance for the avg consumer that gets a PD based computer. I wonder if it will suffice for me though... A PhII at 4Ghz is still a very competent CPU.

Cheers!

Same problems exist in Java; forcing try/catch blocks doesn't solve that particular problem unless you have a generic "catch (exception)" catch statement. I too often see SW engineers code for the expected errors, but NOT the unexpected ones. Then the first time you get a rounding error in a FP division, the application takes a right hand turn, crashes with a 0xC0000005 (STATUS_ACCESS_VIOLATION) and vanishes.

At the very least, apps should be coded so if they crash, they disable all functions, take you back to teh menu, give you the oppertunity to save your data, and give you valid support information. [/rant]

 

[truegenius]

Marketecture... AMD trinity is clocked higher than Intel, is a "quad core" and has Radeon graphics. To the average consumer the specs make it look faster.

sandy/ivy bridge cpu is also acting as a marketing term 😗
even if its a celeron cpu in comparison with phenom 2 x4 :lol:

 

[truegenius]

If I am able to snag an FX-8350 for less than $250 I will be all over it. Otherwise I will have to drop to an 8320.

Regardless it looks like it will be a good boost over my FX-4100.

$250 before release
looks like it is cheap or pd is cheap
quickly test it and kindly tell us something about ( i will keep it secret till release 😗 )

 

[de5_Roy]

if i compare pd launch to bd launch... top pd cpu might be priced around $260-270 msrp, sold at $280~ since pd promises 10-15% performance improvement over bd. amd's marketing will also publicize people (read - sucker) into ghz and core advertisements and what not. don't expect any less for the flagship lineup from amd.
availability will determine retail price so my assumption could easily be wrong.
pd will get into sub $200 range within a year - that'd be the best time to buy one. am3+ owners would be the ones to benefit the most from pd price cuts.

 

[viridiancrystal]

So where does everyone think PD falls based on Competition? IPC and overall.
Lets say BD is 1 PII is 4 SB is 9 and IB is 10.

personally I think for IPC it will fall around 5, and for overall it will fall around 8.5.

 

[lilcinw]

@truegenius I am talking post-release day.

@De5_roy I am hoping MicroCenter has some decent prices (and stock) shortly after release.

 

[de5_Roy]

microcenter's didn't lower bd prices for a long time iirc. cheap cpu+motherboard combo deals started popping up recently after amd declared the price cuts.
i'd blame glofo's capability of producing bd cpus for bd's initial pricing. this time it could be different, since glofo has more mature 32nm node.

 

[lilcinw]

I got a good deal two months after release, $50 discount on MB+CPU. They didn't do anything with the 8xxx for a long time but as you point out GF didn't help availability.

Hopefully things will be better this time around.

 

[-Fran-]

Same problems exist in Java; forcing try/catch blocks doesn't solve that particular problem unless you have a generic "catch (exception)" catch statement. I too often see SW engineers code for the expected errors, but NOT the unexpected ones. Then the first time you get a rounding error in a FP division, the application takes a right hand turn, crashes with a 0xC0000005 (STATUS_ACCESS_VIOLATION) and vanishes.

At the very least, apps should be coded so if they crash, they disable all functions, take you back to teh menu, give you the oppertunity to save your data, and give you valid support information. [/rant]

I use "finally" to always check everything after a catch sentence. It's a must for a pro-Java developer 😛

Also, not quite agree with the generic exception catching. It's unhealthy to mask exceptions when you don't know what's coming. It's always better to include the pertinent catch to apply a pertinent fix, so once something wrong happens, you just have the trace-ability for it and don't have to search in the sea of generic catches (which most crappy devs use, sorry). This is why so many C programs have horrible memory leaks and never fix them (although there is Valgrind >_>). When you try to catch a generic problem, I'm sure you won't cover all the possible outcomes there could be and will be masking the real problem most of the time. I hate that when I see code... And I double hate: try{ /*blah*/ } catch(Exception e){}. Not even a God Damn log of the exception; FFFUUUUU! D:

----

I'm willing to pay up to USD$300 for PD, if it's really worth that much. For USD$300 it MUST beat consistently the i5 2500k, which is the best CPU right now for most scenarios and get on even terms with the i7 2600K.

If PD gets close to the PhII at 4Ghz with stock clocks (which could be around them; 3.8 and turbo to 4.2 are the most possible clocks), I'll get it at USD$200 tops, not a cent more for it's flagship.

If nothing of the above happens, I'll get an i5 rig 😛

Cheers!

 

[de5_Roy]

one thing amd has reiterated over and over again that pd is incremental improvement over bd. so an overclocked phenom owner or bd owner might not have enough incentive to get a pd rig unless it is really cheap. i assume steamroller is going to be the one that is able to tempt amd owners to upgrade with its performance figures. i think most of the pd performance improvements might come in the form of better power use (compared to bd) because of the improved architecture, rcm tech and the way windows 8 is rumored to be more multithreading and power friendly than win 7. and i've read many many amd owners dismiss power usage/efficiency as a 'non issue for "real gamers" or "enthusiasts"'.

 

[-Fran-]

one thing amd has reiterated over and over again that pd is incremental improvement over bd. so an overclocked phenom owner or bd owner might not have enough incentive to get a pd rig unless it is really cheap. i assume steamroller is going to be the one that is able to tempt amd owners to upgrade with its performance figures. i think most of the pd performance improvements might come in the form of better power use (compared to bd) because of the improved architecture, rcm tech and the way windows 8 is rumored to be more multithreading and power friendly than win 7. and i've read many many amd owners dismiss power usage/efficiency as a 'non issue for "real gamers" or "enthusiasts"'.

I run the PhII at 4Ghz with CnC and C1E activated. It IDLEs to 800Mhz and 0.9v when not in use. So, PD would have to beat that as well, which I'm sure it will.

Now... 4Ghz at 1.48v is a lot of power to be honest; if PD is really really efficient it could be worth higher than USD$200 while being a sidegrade in performance. Benchies will tell 😛

Cheers!

 

[jdwii]

So where does everyone think PD falls based on Competition? IPC and overall.
Lets say BD is 1 PII is 4 SB is 9 and IB is 10.

personally I think for IPC it will fall around 5, and for overall it will fall around 8.5.

I would have to say it depends on the Instruction set but on average 3.5 and at times 4.5-5. But the Clock speed well be higher and if Ivy is a 10 and sandy is a 9(i would say 9.5) then Piledriver will be a 7 overall.


I would also like to say its not fair to rate this off a scale of overall because if were talking about gaming PD will be a 7 compared to the I5 which is a 10.
Now when it comes with video encoding(which i care a lot about) the PD will walk all over the I5 at least under good encoding software the only software i like is hand brake and i tried the stuff that cost 100$! LOL

For plan gaming i don't even need to max out every game and if i did my 6950 would cause a bottleneck not my x6.

 

[gamerk316]

I would have to say it depends on the Instruction set but on average 3.5 and at times 4.5-5. But the Clock speed well be higher and if Ivy is a 10 and sandy is a 9(i would say 9.5) then Piledriver will be a 7 overall.


I would also like to say its not fair to rate this off a scale of overall because if were talking about gaming PD will be a 7 compared to the I5 which is a 10.
Now when it comes with video encoding(which i care a lot about) the PD will walk all over the I5 at least under good encoding software the only software i like is hand brake and i tried the stuff that cost 100$! LOL

For plan gaming i don't even need to max out every game and if i did my 6950 would cause a bottleneck not my x6.

With the exception of a handful of games that care about memory bandwidth, or use very graphically "weak" engines (source), GPU's are the limiting factor. In games, PD and BD are going to be identical. Mark it down.

 

[esrever]

Scares me seeing some Trinity laptops come up with a 850$ price tag! Way overpriced they needs to be priced at 650$ and below. Even 600-650$ is pushing it.

It seems all trinity laptops so far are $100 overpriced compared to the llano offerings. Im hoping to see some more laptops come out for more competition soon. Hopefully this comes before september.

 

[jdwii]

With the exception of a handful of games that care about memory bandwidth, or use very graphically "weak" engines (source), GPU's are the limiting factor. In games, PD and BD are going to be identical. Mark it down.

Your completely wrong, Games benefit from clock speed and IPC both of which improved on PD. Now will it be better then the Phenom when it comes to gaming Yes i think so not by much but by 5-10%. I don't even understand why you would say this of course its not going to be "identical". Will it beat sandy no but 90% of the time its your GPU that's the bottleneck like with my Phenom/ I can even come close to maxing out skyrim. Overclock Piledriver to 4.6Ghz and you will game perfect maybe not at 120FPS but you will game better on PD then you will from a 360/PS3/Wii U?/

People at least try to look up stuff before saying statements like this.

 

[viridiancrystal]

With the exception of a handful of games that care about memory bandwidth, or use very graphically "weak" engines (source), GPU's are the limiting factor. In games, PD and BD are going to be identical. Mark it down.

I've said that before about BD too. Even is graphically "weak" engines (I hate wording it like that, makes it sound like source doesn't look good) once you hit a cpu bottleneck, your at or over a playable frame rate anyway.

Also, the debate earlier on this thread about games not scaling over 4 cores is perfectly false. More than 4 cores will be necessary in the future.